Citation
Cognitive Factors Contributing to Verbal Memory Performance in Clinical Pediatric Populations

Material Information

Title:
Cognitive Factors Contributing to Verbal Memory Performance in Clinical Pediatric Populations
Creator:
Jordan, Lizabeth L
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (5 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology
Clinical and Health Psychology
Committee Chair:
HEATON,SHELLEY C
Committee Co-Chair:
GEFFKEN,GARY R
Committee Members:
BAUER,RUSSELL M
ELDER,JENNIFER HARRISON
Graduation Date:
8/9/2014

Subjects

Subjects / Keywords:
Academic achievement ( jstor )
Attention deficit hyperactivity disorder ( jstor )
Childhood ( jstor )
Control groups ( jstor )
Memory ( jstor )
Memory encoding ( jstor )
Memory retrieval ( jstor )
Orthopedics ( jstor )
Pediatrics ( jstor )
Physical trauma ( jstor )
Clinical and Health Psychology -- Dissertations, Academic -- UF
cognition -- memory -- pediatric
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Psychology thesis, Ph.D.

Notes

Abstract:
Memory problems and other cognitive dysfunction are commonly reported within clinical pediatric populations, including children with medical and/or psychological diagnoses. However, these memory problems are poorly characterized and are often described as heterogeneous. In part, the heterogeneity of memory problems within pediatric populations are likely related to variations in underlying neuroanatomical dysfunction (e.g. seizurefoci associated with epilepsy, injury loci resulting from traumatic brain injury). In addition, heterogeneity of memory difficulties may be related to varied abilities in other cognitive domains (e.g. attention difficulties associated with attention deficit hyperactivity disorder, processing speed impairments after traumatic brain injury). The overarching goal ofthis dissertation was to better understand the memory performance profile and its underlying contributing cognitive components of clinical pediatric populations. Neuropsychological data and medical information was extracted from the records of clinical pediatric patients who were evaluated in the Shands Hospital / University of Florida Pediatric Neuropsychology Clinic or the University of Florida's Pediatric Neuro psychology Research Laboratory. Data was analyzed using univariate and multivariate statistics. Overall, results revealed that the verbal memory difficulties associated with a variety of clinical pediatric populations seems to be stemming from a weakness in verbal encoding processes. Verbal retention and retrieval skills appeared similar to those of healthy normative data and clinical control samples. Furthermore,the relatively weak verbal encoding performance of clinical pediatric patients was significantly related to performance on separate measures of verbal knowledge/reasoning. Although the clinical pediatric populations did not necessarily demonstrate a clinically significant verbal encoding impairment,this dissertation study suggests that clinical pediatric populations have a verbal memory profile characterized by a relative weakness in the verbal encoding stage of memory processes. Future research should replicate these findings in independent samples of clinical pediatric populations and with other measures of verbal (and visual) memory. If these findings are successfully replicated, then memory treatment and intervention programs should focus on there habilitation of encoding abilities as well the development of verbal knowledge skills. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: HEATON,SHELLEY C.
Local:
Co-adviser: GEFFKEN,GARY R.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-08-31
Statement of Responsibility:
by Lizabeth L Jordan.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
8/31/2016
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

Journal of the International Neuropsychological Society (2009), 15 , 613 – 617 . Copyright © 2009 INS. Published by Cambridge University Press. Printed in the USA. doi:10.1017/S1355617709090651 613 INTRODUCTION Clinicians appreciate that healthy people can obtain some low scores on a test battery. This psychometric principle has been supported in the research literature when considering performance across batteries of cognitive tests (Axelrod & Wall, 2007 ; Binder et al., 2009 ; Crawford et al., 2007 ; Heaton et al., 1991 , 2004 ; Iverson & Brooks, in press ; Iverson et al., 2008a , 2008b ; Schretlen et al., 2008 ). The presence of low memory scores in healthy people (Brooks et al., 2007 , 2008 ; Palmer et al., 1998 ) and the potential impact this has on identifying subtle or prodromal memory problems (de Rotrou et al., 2005 ) are of special relevance to neuropsychologists. Palmer et al. ( 1998) illustrated that 39.4% of healthy older adults obtained at least one score at or below 1.3 standard deviations ( SD s) and 12.9% obtained at least one score at or below 2.0 SD s. In studies of two large standardization samples, Brooks et al. ( 2007 , 2008 ) found that approximately one half of healthy older adults with below-average intelligence would meet the psychometric criterion for mild cognitive impairment (MCI; Petersen et al., 1999 ). Descriptive studies that highlight the likelihood of an isolated low memory score provide useful information on the potential to misdiagnose memory problems in older adults. For example, de Rotrou et al. ( 2005) reported that 48% of older adults, who were identied as having MCI at baseline based on the presence of a low memory score, had normal cognitive functioning at a 1-year follow-up. The need to understand normal variability across a battery of neuropsychological measures, and thus the presence of some low scores in healthy people, should not be limited to adults and older adults. To our knowledge, information on the base rates of low scores on memory batteries in healthy children does not exist, even though performance on these batteries is used as the foundation for clinical inferences relating to memory problems. The purpose of this descriptive study was to illustrate that the principles of multivariate test BRIEF COMMUNICATION Healthy children and adolescents obtain some low scores across a battery of memory tests BRIAN L. BROOKS , 1 , 2 GRANT L. IVERSON , 3 , 4 ELISABETH M.S. SHERMAN , 1 , 5 , 6 and JAMES A. HOLDNACK 7 1 Neurosciences Program , Alberta Children’s Hospital , Calgary , Alberta , Canada 2 Department of Pediatrics , University of Calgary , Calgary , Alberta , Canada 3 Department of Psychiatry , University of British Columbia , Vancouver , British Columbia , Canada 4 Research Department , British Columbia Mental Health & Addiction Services , Coquitlam , British Columbia , Canada 5 Department of Pediatrics , University of Calgary , Calgary , Alberta , Canada 6 Department of Clinical Neuroscience , University of Calgary , Calgary , Alberta , Canada 7 Clinical Content Development , Pearson Assessment , San Antonio , Texas (Received December 6 , 2008 ; Final Revision February 18 , 2009 ; Accepted February 18 , 2009 ) Abstract Obtaining some low memory scores across a battery of tests is common. The purpose of this study was to examine the prevalence of low scores on the Children’s Memory Scale (CMS ). Participants were 1000 children and adolescents between 5 and 16 years of age from the CMS standardization sample. Consistent with research on other batteries, having some low memory scores is common in healthy children and adolescents. The prevalence of low memory scores also increases with lower intelligence. Clinicians should be cautious when interpreting isolated low memory scores as sole evidence of memory impairment. Knowing the prevalence of low scores as a supplement to clinical judgment should reduce the likelihood of misdiagnosing memory problems. ( JINS , 2009, 15 , 613– 617 .) Keywords : Children , Pediatric , Memory , Psychometric , Base rate , Misdiagnosis Correspondence and reprint requests to: Brian L. Brooks, Neurosciences Program, Alberta Children’s Hospital, 2888 Shaganappi Trail NW, Calgary, Alberta, Canada T3B 6A8. E-mail: brian.brooks@albertahealthservices.ca

PAGE 2

B.L. Brooks et al. 614 interpretation, which have been demonstrated in adults and older adults, are applicable to children and adolescents. METHODS Participants Participants for the present study included 1000 healthy children and adolescents between 5 and 16 years of age (mean = 9.7, SD = 3.2) from the Children’s Memory Scale (CMS; Cohen, 1997 ) standardization sample. An equal number of boys and girls were in each age group. Ethnicity included 68.4% White, 16.1% African American, 11.6% Hispanic, and 3.9% as “other.” The sample was strati ed by level of parent education (i.e., eighth grade or less, 9–11 years, high school graduate or equivalent, 1–3 years of technical school or college, and 4 or more years of college). The standardization sample was recruited from 149 sites from the western, north central, northeastern, and southern regions of the United States. Children were excluded from the standardization sample if they were reading below their grade level, had repeated a grade, were receiving special education, were previously diagnosed with a neurological disorder, or had sustained an injury that would have put them at risk for having memory problems (Cohen, 1997 ). The treatment of participants and the collection of data were done in compliance with the Helsinki Declaration. Use of the archival CMS data was approved by the University of Calgary research ethics board. A subsample of the CMS standardization group ( n = 209) was administered the Wechsler Intelligence Scale for Children (Third Version) (WISC-III; Wechsler, 1991 ) as part of a linking study. The WISC-III linking sample ranged in age from 6 to 16 years (mean = 10.1, SD = 2.8), was 48.8% male, 77.0% White (11.0% African American, 9.6% Hispanic, 2.4% other), and had a mean WISC-III Full Scale Intelligence Quotient (FSIQ) of 105.3 ( SD = 14.4, range = 70– 146). Inclusion of the WISC-III linking data allows for strati cation of the CMS data by level of intelligence (i.e., WISC-III FSIQ). Although the subsample with WISCIII FSIQ data had more people identi ed as Caucasian compared to the entire sample, 2 (1) = 6.12, p = .01, there were no statistically signi cant differences in age ( p = .08), sex ( p = .75), or performance on the CMS index scores ( p values = .32–.71) for the WISC-III subsample compared to the total standardization sample. Measures There are six primary subtests on the CMS (i.e., Stories, Word Pairs, Dot Locations, Faces, Numbers, and Sequences). These six subtests contribute to eight age-adjusted index scores, six of which were included in the base rate of low scores analyses: Learning, Verbal Immediate, Verbal Delayed, Verbal Delayed Recognition, Visual Immediate, and Visual Delayed. The General Memory index and the Attention/Concentration index were not included in the analyses. Analyses Analyses involved examining performance on all six index scores, simultaneously . The cutoffs used for analyses of the CMS data included <16th percentile (<1 SD ; index < 85), <10th percentile (index < 81), 5th percentile (index 76), and 2nd percentile (<2 SD s; index < 70). The prevalence of low CMS index scores was examined for the total sample (5–16 years; N = 1000) and for the three levels of intellectual abilities: below average (FSIQ = 70–89; n = 30), average (FSIQ = 90– 109; n = 93), and above average (FSIQ = 110+; n = 86). RESULTS The base rates of low CMS index scores in children and adolescents are presented in Figures 1 and 2 . Using the <1 SD cutoff score ( Figure 1 ), one or more low index scores was found in 37.6% of the total sample and three or more low scores were found in 10.6% of the sample. One or more index scores <10th percentile was found in 30.2% (data not shown), one or more index scores 5th percentile was found in 22.4% ( Figure 2 ), and one or more index scores <2 SD s was found in 12.4% of healthy children (data not shown). There were no substantial differences in the prevalence of low scores across the age bands. For example, the prevalence of one or more index scores 16th percentile included 47.1% of 5to 8-year-olds, 43.2% of 9to 13-year-olds, and 47.5% of 13to 16-year-olds. These slight, but not substantial, variations were present across different cutoff scores and for different numbers of low scores. Although age had limited impact on the base rates, intellectual functioning had considerable in uence on these base rates. Compared to children with above-average intelligence, those with below-average intelligence were 7.1 times more likely, 95% con dence interval for odds ratio = 2.9–17.6; 2 (1) = 19.79, p < .001, to have one or more scores <1 SD ( Figure 1 ). When considering 5th percentile ( Figure 2 ), 33.3% of children and adolescents with below-average intelligence had one or more low index scores compared to 3.5% with above-average intelligence. In other words, children with below-average intelligence were 13.8 times more likely than those with above-average intelligence to have one or more index scores 5th percentile, 95% con dence interval for odds ratio = 3.7–50.86; 2 (1) = 19.91, p < .001. When considering a more stringent cutoff, such as <2 SD s (data not shown), 20.0% of children and adolescents with below-average intelligence had one or more low index scores compared to 2.0% with above-average intelligence, 2 (1) = 10.82, p < .001; odds ratio = 10.5 (95% con dence interval = 2.2–48.2). DISCUSSION It is common for healthy people to have some low scores across a battery of tests (Axelrod & Wall, 2007 ; Binder et al., 2009 ; Brooks et al., 2007 , 2008 ; Crawford et al., 2007 ;

PAGE 3

Low scores on the CMS 615 Heaton et al., 1991 , 2004 ; Iv erson & Brooks, in press ; Iv erson et al., 2008a , 2008b ; Palmer et al., 1998 ; Schretlen et al., 2008 ). As a result, accumulating literature suggests that the psychometric interpretation of a test battery should include a multivariate approach. In other words, test scores should be interpreted simultaneously using empirical data because examining individual test scores can lead to overinterpretation of one or more isolated low scores (Brooks et al., 2007 , 2008 ; de Rotrou et al., 2005 ; Palmer et al., 1998 ). To our knowledge, this is the rst study that examines and presents the prevalence of low memory scores in a sample of healthy children and adolescents. The results of this descriptive study clearly demonstrate that having at least one low memory score is common in many healthy children and adolescents. However, it would be considered uncommon (i.e., prevalence rate of approximately 10% or less) to have four or more index scores <1 SD . There were not any substantial differences in the prevalence of low CMS scores across the age groups. In other words, prevalence rates were fairly consistent from 5 to 16 years of age. This is likely the result of using age-adjusted standard scores. As the cutoff for identifying cognitive problems becomes more stringent, the number of low scores below those cutoffs considered uncommon also declines. For 4.5 60.0 30.0 23.3 6.7 2.2 17.4 3.5 0.0 0.0 10.620.037.6 11.8 18.3 36.60 10 20 30 40 50 60 1 or more 2 or more3 or more 4 or moreNumber of low scoresCumulative Percent Total Sample Below Average Average Above Average Fig. 1. Prevalence of low CMS index scores (<1 SD ) by level of intelligence. Total sample N = 1000. Intelligence is based on WISC-III FSIQ scores and includes below average (FSIQ = 70–89; n = 30), average (FSIQ = 90–109; n = 93), and above average (FSIQ = 110+; n = 86). Analyses involved examining all index scores simultaneously . Analyses included six index scores (Learning, Verbal Immediate, Visual Immediate, Verbal Delayed, Verbal Delayed Recognition, and Visual Delayed). The Attention/Concentration and General Memory indexes were not included in the analyses. For index scores, <16th percentile or <1 SD is equal to an index <85. Standardization data are from the Children’s Memory Scale. Copyright © 1997 by NCS Pearson, Inc. Used with permission. All rights reserved. 9.8 4.3 33.3 20.0 6.7 6.7 8.6 2.2 3.5 0.0 1.4 0.00.0 22.4 21.50 10 20 30 40 50 60 1 or more 2 or more3 or more 4 or moreNumber of low scoresCumulative Percent Total Sample Below Average Average Above Average 1.2 Fig. 2. Prevalence of low CMS index scores ( 5th percentile) by level of intelligence. Total sample N = 1000. Intelligence is based on WISC-III FSIQ scores and includes below average (FSIQ = 70–89; n = 30), average (FSIQ = 90–109; n = 93), and above average (FSIQ = 110+; n = 86). Analyses involved examining all index scores simultaneously . Analy ses included six index scores (Learning, Verbal Immediate, Visual Immediate, Verbal Delayed, Verbal Delayed Recognition, and Visual Delayed). The Attention/Concentration and General Memory indexes were not included in the analyses. For index scores, 5th percentile is equal to an index 76. Standardization data are from the Children’s Memory Scale. Copyright © 1997 by NCS Pearson, Inc. Used with permission. All rights reserved.

PAGE 4

B.L. Brooks et al. 616 example, having one or more low index score is found in 37.6% when considering the 1 SD cutoff but is found in 22.4% when using the 5th percentile cutoff. Being able to determine the number of low CMS scores that would be considered uncommon is a strength of these analyses. Interpretation of memory performance should be done in the context of overall level of intellectual functioning (Iverson & Brooks, in press ). A person who is lower functioning cognitively will have more low scores, and be at greater risk for misdiagnosis of memory problems (i.e., false positives), than a person who is higher functioning (and at greater risk of having a missed diagnosis, i.e., false negatives). The number of low index scores (<1 SD ) that would be considered uncommon (i.e., approximately 10% prevalence in the sample) was four or more for those with below-average intelligence, three or more in children with average intelligence, and two or more for those with above-average intelligence. Evidently, some caution should be exercised when interpreting test scores by level of intelligence in children because illness, injury, and/or developmental disorder may produce both intellectual and memory impairment in previously healthy children. Consequently, both low intelligence quotient (IQ) and a high prevalence low memory scores may be very clinically signi cant, depending on the presenting problem. Importantly, although multiple low memory scores may not be diagnostic of a speci c disorder, the presence of such low scores, particularly in the context of normal intelligence, may indicate a relative weakness in mnemonic functions that is typical of some childhood developmental (e.g., memory for faces in developmental disorders; Williams et al., 2005 ) or neurological conditions (e.g., temporal lobe epilepsy). The existing literature on the prevalence of isolated low memory scores across a battery of tests has focused on older adults (Brooks et al., 2007 , 2008 ; de Rotrou et al., 2005 ; Palmer et al., 1998 ), particularly in the context of identifying memory changes consistent with prodromal dementia or MCI. The results of the present study, which considered similar analyses of base rates of low memory scores but in a pediatric standardization sample, were fairly similar (albeit slightly higher) to the results presented by Brooks et al. ( 2007) for the Neuropsychological Assessment Battery (NAB; Stern & White, 2003 ) Memory module and by Brooks et al. ( 2008) for the Wechsler Memory Scale–Third Edition (WMS-III; Wechsler, 1997 ). The similarities across the CMS, NAB Memory module, and WMS-III studies are also maintained when considering performance strati ed by level of intelligence. The consistency in ndings across different memory batteries with very different standardization samples suggests that the presence of low memory scores is not an artifact of any particular battery and is not attributable to a particular age group. Iverson and Brooks ( in press ) suggested that clinicians and researchers should be familiar with the following ve psychometric principles when interpreting multiple test scores. As suggested, low test scores (1) are common across all batteries, (2) depend on where cutoff scores are set, (3) depend on the number of tests administered, (4) vary by demographic characteristics of the examinee, and (5) vary by level of intelligence. The descriptive analyses presented in this article clearly illustrate principles 1, 2, and 5. It will be important for future research to examine how different demographic characteristics impact the prevalence of low scores and the interpretation of performance on any test battery. Although it is important for clinicians to understand that low scores are common and to consider these principles, it can be challenging to consider this information readily in everyday clinical practice unless easy-to-use interpretive tables and/or gures are readily available. There are a few limitations to this study that warrant a brief discussion. First, like other studies involving a standardization sample, memory problems were not screened for a priori in the normative group. Although inclusion in the standardization was contingent on not having medical, neurological, or psychiatric conditions that could negatively impact memory performance, it is possible that some children and adolescents with memory problems were included. However, if a small proportion of the CMS standardization sample did have primary memory problems atypical of most healthy children, it is likely that this proportion would be quite small and unlikely to account for a large percentage of the prevalence rates presented in Figures 1 and 2 . Second, the sample of children and adolescents who were part of the WISC-III linking study was small in size, consisted of relatively higher functioning youth, and the number of children with below-average intelligence was relatively small compared to the other intelligence groups. The inclusion of the WISC-III (Wechsler, 1991 ) as the measure of intelligence warrants some discussion. Since the standardization and publication of the CMS in 1997, a newer version of the Wechsler Intelligence Scale for Children (WISC) has been published [e.g., Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV); Wechsler, 2003 ]. Although the correlation between the WISC-III and the WISC-IV FSIQ scores is high ( r = .87; Wechsler, 2003 ), the WISC-IV yielded slightly lower scores compared to the WISC-III. However, given that the CMS sample was given the WISC-III approximately 6 years after it was normed and published in 1991, and it has been 6 years since the WISC-IV was normed and published (i.e., 2003), children’s WISC-IV FSIQ scores today might be similar to the WISC-III FSIQ scores from the CMS standardization sample. Despite these similarities in time from publication of the respective WISC version, clinicians and researchers should use caution when interpreting the prevalence of low scores based on level of intelligence that is derived for a measure of intelligence other than the WISC-III. Users should also exercise some caution when interpreting the prevalence of low CMS scores for those with intelligence scores that are close to the cutoff for the different IQ ranges. It might be important to consider the base rates in the obtained classi cation as well as the base rates in the neighboring classi cation when drawing any conclusions on performance.

PAGE 5

Low scores on the CMS 617 Knowing the prevalence of low scores in healthy children and adolescents is designed to supplement other psychometric interpretive methods (e.g., discrepancies between indexes) and clinical decision-making. The results of this study suggest that some caution is needed when interpreting isolated low CMS subtest scores as sole evidence of memory impairment. The goal is to use this information to reduce the likelihood of misdiagnosing memory problems. The lower prevalence of low scores in healthy children of above-average intelligence also holds potential in reducing the chances of a missed diagnosis of memory problems in children and adolescents who are higher functioning. ACKNOWLEDGMENTS Portions of this data were presented at the 37th annual meeting of the International Neuropsychological Society, February 2009. The information in this manuscript and the manuscript itself has never been published either electronically or in print. The authors thank the Psychological Corporation (Pearson Assessment) for use of the data presented in Figures 1 and 2 . Standardization data are from the Children’s Memory Scale. Copyright © 1997 by NCS Pearson, Inc. Used with permission. All rights reserved. The authors also thank the editors and reviewers for their helpful comments on earlier versions of this manuscript. Drs. B.L.B., G.L.I., and E.M.S.S. have no known, perceived, or actual con ict of interest with this study. Dr. J.A.H. is a senior research director with Pearson Assessment, which is the publisher of the Children’s Memory Scale . REFERENCES Axelrod , B.N. & Wall , J.R. ( 2007 ). Expectancy of impaired neuropsychological test scores in a non-clinical sample . International Journal of Neuroscience , 117 ( 11 ), 1591 – 1602 . Binder , LM. , Iverson , GL. , Brooks , BL. (2009 ). To err is human: “abnormal” neuropsychological scores and variability are common in healthy adults . Archives of Clinical Neuropsychology , 24 ( 1 ), 31 – 46 . Brooks , B.L. , Iverson , G.L. , Holdnack , J.A. , & Feldman , H.H. ( 2008 ). Potential for misclassi cation of mild cognitive impairment: A study of memory scores on the Wechsler Memory ScaleIII in healthy older adults . Journal of the International Neuropsychological Society , 14 ( 3 ), 463 – 478 . Brooks , B.L. , Iverson , G.L. , & White , T. ( 2007 ). Substantial risk of “Accidental MCI” in healthy older adults: Base rates of low memory scores in neuropsychological assessment . Journal of the International Neuropsychological Society , 13 ( 3 ), 490 – 500 . Cohen , M.J. ( 1997 ). Children’s Memory Scale, manual . San Antonio, TX : The Psychological Corporation . Crawford , J.R. , Garthwaite , P.H. , & Gault , C.B. ( 2007 ). Estimating the percentage of the population with abnormally low scores (or abnormally large score differences) on standardized neuropsychological test batteries: A generic method with applications . Neuropsychology , 21 ( 4 ), 419 – 430 . Heaton , R.K. , Grant , I. , & Matthews , C.G. ( 1991 ). Comprehensive norms for an extended Halstead-Reitan Battery: Demographic corrections, research ndings, and clinical applications . Odessa, FL : Psychological Assessment Resources, Inc . Heaton , R.K. , Miller , S.W. , Taylor , M.J. , & Grant , I. ( 2004 ). Revised comprehensive norms for an expanded Halstead-Reitan Battery: Demographically adjusted neuropsychological norms for African American and Caucasian adults professional manual . Lutz, FL : Psychological Assessment Resources . Iverson , G.L. & Brooks , B.L. (in press ). Improving accuracy for identifying cognitive impairment . In M.R. Schoenberg & J.G. Scott (Eds.), The black book of neuropsychology: A syndromebased approach . Iverson , G.L. , Brooks , B.L. , & Holdnack , J.A. ( 2008 a). Misdiagnosis of cognitive impairment in forensic neuropsychology . In R.L. Heilbronner (Ed.), Neuropsychology in the courtroom: Expert analysis of reports and testimony , (pp. 243 – 266 ). New York : Guilford Press . Iverson , G.L. , Brooks , B.L. , White , T. , & Stern , R.A. ( 2008 b). Neuropsychological Assessment Battery (NAB): Introduction and advanced interpretation . In A.M. Horton, Jr. & D. Wedding (Eds.), The neuropsychology handbook , (pp. 279 – 343 ). New York : Springer Publishing, Inc . Palmer , B.W. , Boone , K.B. , Lesser , I.M. , & Wohl , M.A. ( 1998 ). Base rates of “impaired” neuropsychological test performance among healthy older adults . Archives of Clinical Neuropsychology , 13 ( 6 ), 503 – 511 . Petersen , R.C. , Smith , G.E. , Waring , S.C. , Ivnik , R.J. , Tangalos , E.G. , & Kokmen , E. ( 1999 ). Mild cognitive impairment: Clinical characterization and outcome . Archives of Neurology , 56 ( 3 ), 303 – 308 . de Rotrou , J. , Wenisch , E. , Chausson , C. , Dray , F. , Faucounau , V. , & Rigaud , A.S. ( 2005 ). Accidental MCI in healthy subjects: A prospective longitudinal study . European Journal of Neurology , 12 ( 11 ), 879 – 885 . Schretlen , D.J. , Testa , S.M. , Winicki , J.M. , Pearlson , G.D. , & Gordon , B. ( 2008 ). Frequency and bases of abnormal performance by healthy adults on neuropsychological testing . Journal of the International Neuropsychological Society , 14 ( 3 ), 436 – 445 . Stern , R.A. & White , T. ( 2003 ). Neuropsychological assessment battery . Lutz, FL : Psychological Assessment Resources . Wechsler , D. ( 1991 ). Wechsler Intelligence Scale for Children— Third edition . San Antonio, TX : The Psychological Corporation . Wechsler , D. ( 1997 ). Wechsler Memory Scale—Third edition . San Antonio, TX : The Psychological Corporation . Wechsler , D. ( 2003 ). Wechsler Intelligence Scale for Children—Fourth edition . San Antonio, TX : The Psychological Corporation . Williams , D.L. , Goldstein , G. , & Minshew , N.J. ( 2005 ). Impaired memory for faces and social scenes in autism: Clinical implications of memory dysfunction . Archives of Clinical Neuropsychology , 20 ( 1 ), 1 – 15 .



PAGE 1

1 COGNITIVE FACTORS CONTRIBUTING TO VERBAL MEMORY PER FORMANCE IN CLINICAL PEDIATRI C POPULATIONS By LIZABETH L. JORDAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULMILLMENT OF THE R EQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 201 4 Lizabeth Jordan

PAGE 3

3 To the pediatric patients and their families who have contributed to this dissertation

PAGE 4

4 ACKNOWLEDGMENTS I thank my family, friends, faculty, and colleagues who have greatly supported me throughout graduate school and this dissertation project. This dissertation study and written document could not have been completed without their support.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 BACKGROUND ...................................................................................................... 13 Introduction ............................................................................................................. 13 Three Verbal Mem ory Processes ........................................................................... 14 The Encoding Process ..................................................................................... 15 The Retention Process ..................................................................................... 16 The Re trieval Process ...................................................................................... 17 Summary of the Three Memory Processes ...................................................... 18 Neural Correlates of Verbal Memory Processes ..................................................... 18 Prefrontal Cortex .............................................................................................. 19 Neocortex, Medial Temporal Lobe, Diencephalon, and Basal Forebrain .......... 20 Summary of the Neural Correlates of Verbal Memory ...................................... 21 Other Cognitive Factors Underlying the Verbal Memory Process ........................... 22 Verbal Knowledge ............................................................................................ 22 Focused & Sustained Attention ........................................................................ 23 Processing Speed ............................................................................................ 24 Working Memory / Executive Functioning ........................................................ 25 Summary of Cognitive Contributors .................................................................. 26 Development of Verbal Memory ............................................................................. 27 Development of Verbal Encoding & Retrieval Strategies .................................. 28 Developing Neuroanatomy ............................................................................... 29 Development of Other Cognitive Factors .......................................................... 30 Summary of Verbal Memory Development ....................................................... 31 Verbal Memory Problems in Clinical Pediatric Populations ..................................... 31 Attention Deficit Hyperactivity Disorder ............................................................ 32 Pervasive Developmental Disorders ................................................................ 34 Epilepsy ............................................................................................................ 35 Traumatic Brain Injury ...................................................................................... 37 Childhood Cancer ............................................................................................. 39 Summary of V erbal Memory Problems in Clinical Pediatric Populations .......... 40 Summary of Background Information ...................................................................... 41 2 OVERVIEW AND STATEMENT OF THE PROBLEM ............................................. 42

PAGE 6

6 Aim 1 ....................................................................................................................... 45 Specific Aim 1a ................................................................................................. 45 Specific Aim 1b ................................................................................................. 45 Aim 2 ....................................................................................................................... 45 Specific Aim 2a ................................................................................................. 46 Specific Aim 2b ................................................................................................. 46 Specific Aim 2c ................................................................................................. 46 Specific Aim 2d ................................................................................................. 47 Aim 3 ....................................................................................................................... 47 Specific Aim 3a ................................................................................................. 47 Specific Aim 3b ................................................................................................. 48 Specific Aim 3c ................................................................................................. 48 Specific Aim 3d ................................................................................................. 48 3 METHODS .............................................................................................................. 49 Participants ............................................................................................................. 49 Procedure ............................................................................................................... 51 Measurement of Verbal Memory ............................................................................. 52 The Children’s Memory Scale (CMS) ............................................................... 52 Measurement of Other Cognitive Domains & Motor Abilities .................................. 53 Measures of Verbal Knowledge ........................................................................ 53 Measures of Focused & Sustained Attention .................................................... 55 Measures of Processing Speed ........................................................................ 55 Measures of Working Memory/Executive Functioning ...................................... 56 Statistical Analyses ................................................................................................. 57 Aim 1: Understanding Delayed Verbal Memory Performance in a Large, Mixed Clinical Pediatric Sample .................................................................... 58 Aim 2: Understanding the Verbal Memory Profile of an ADHD Sample ........... 59 Aim 3: Understanding the Verbal Memory Profile of a TBI Sample .................. 63 Sample Size Considerations ................................................................................... 63 Aim 1: Understanding Delayed Verbal Memory Performance in a Large, Mixed Clinical Pediatric Sample .................................................................... 63 Aim 2: Understanding the Verbal Memory Profile of an ADHD Sample ........... 63 Aim 3: Understanding the Verbal Memory Profile of a TBI Sample .................. 64 4 RESULTS ............................................................................................................... 66 Aim 1 Results .......................................................................................................... 66 Exploratory Factor Analysis .............................................................................. 66 Structural Equation Model ................................................................................ 67 Aim 2 Results .......................................................................................................... 69 Aim 2 Participant Sample ................................................................................. 69 Memory Profile of ADHD Sample ..................................................................... 70 Relationship between Memory Abilities and Cognitive Variables in ADHD ...... 70 Aim 3 Results .......................................................................................................... 71 Aim 3 Particip ant Sample ................................................................................. 71

PAGE 7

7 Memory Profile of Pediatric TBI Sample ........................................................... 72 Relationship between Memory Abilities and Cognitive Variables in Pediatric TBI ................................................................................................................ 73 5 DISCUSSION ......................................................................................................... 95 Cognitive Predictors of Verbal Memory in a Mixed Clinical Sample ........................ 95 Distinct Cognitive Domains Identified in Clinical Pediatric Populations ............ 95 Verbal Knowledge Predicted Delayed Verbal Memory Performance ............... 97 No Significant Contribution from Other Cognitive Domains to Verbal Memory ....................................................................................................... 100 Memory Profile and Contributing Factors in Pediatric A DHD ................................ 103 Relative Weakness in Verbal Encoding Associated with Pediatric ADHD ...... 103 Factors Associated with Verbal Encoding Problems in Pediatric ADHD, Mood, and Learning Disorders .................................................................... 106 Memory Profile and Contributing Factors in Pediatric TBI .................................... 108 Verbal Encoding Weakness Uniquely Associated with Pediatric TBI ............. 108 Cognitive Factors Contributing to Verbal Encoding Weakness Associated with Pediatric TBI ........................................................................................ 110 Overall Conclusions .............................................................................................. 113 Comments on Methodology .................................................................................. 116 Implications & Limitations of an Archival Clinical Study .................................. 116 Utility of Encoding, Retention, and Retrieval Measures .................................. 118 Memory Measure ........................................................................................... 122 Discussion Summary ............................................................................................ 123 APPENDIX A DETAILED DESCRIPTION OF CMS VERBAL MEMORY MEASURES ............... 128 Stories Subtests .................................................................................................... 128 Word Pairs Subtests ............................................................................................. 128 CMS Standardized Scores .................................................................................... 129 B SOURCES OF DATA ............................................................................................ 134 LIST OF RE FERENCES ............................................................................................. 136 BIOGRAPHICAL SKETCH .......................................................................................... 144

PAGE 8

8 LIST OF TABLES Table page 3 1 Summary of CMS outcome variables ................................................................. 65 3 2 Measures of other cognitive domains ................................................................. 65 4 1 Demographic information of Aim 1’s heterogeneous clinical sample ( N = 234) .. 74 4 2 Additional i nformation of Aim 1’s heterogeneous clinical sample ( N = 234) ....... 75 4 3 Cognitive performance based scores of Aim 1’s heterogeneous clinical sample ( N = 234) ................................................................................................ 75 4 4 Verbal memory performancebased scores of Aim 1’s heterogeneous clinical sample ( N = 234) ................................................................................................ 76 4 5 Spearman rho correlations between CMS delayed verbal memory index and other cognitive measures in Aim 1’s heterogeneous clinical sample. ................. 77 4 6 Pattern matrix of rotated factor loadings from exploratory factor analysis .......... 78 4 7 Structured matrix of rotated factor loadings from exploratory factor analysis ..... 78 4 8 Goodness of fit statistics ..................................................................................... 78 4 9 Structural equation modeling (SEM) results ....................................................... 79 4 10 Demographic information of Aim 2’s ADHD and clinical control sample ............. 80 4 11 Additional characterization of Aim 2’s ADHD and clinical control sample ........... 81 4 12 ADHD group mean performance on verbal memory measures (N = 99) ............ 82 4 13 ADHD group mean performance on cognitive measures (N = 99) ..................... 82 4 14 Clinical control group mean performance on verbal memory measures (N=23) ................................................................................................................ 83 4 15 Clinical control group mean performance on cognitive measures (N=23) .......... 83 4 16 Results of regression analyses in Aim 2’s combined group of ADHD and clinical control patients ....................................................................................... 84 4 17 Demographic information on Aim 3’s TBI and orthopedic injury control groups ................................................................................................................ 85

PAGE 9

9 4 18 Additional characterization of Aim 3’s TBI and orthopedic injury control groups ................................................................................................................ 86 4 19 TBI group mean performance on verbal memory measures (N = 41) ................ 87 4 20 TBI group mean performance on cognitive measures (N = 41) .......................... 87 4 21 Orthopedic injury control group mean performance on verbal memory m easures (N=19) ................................................................................................ 88 4 22 Orthopedic injury control group mean performance on cognitive measures (N=19) ................................................................................................................ 88 5 1 Correl ations between measures of verbal memory & other cognitive domains in Aim 1’s clinical pediatric sample ................................................................... 125 5 2 Post hoc mediation model results for Aim 1 ..................................................... 126 5 3 Collinearity diagnostic statistics of the measured cognitive variables ............... 126 A 1 Normative descriptive statistics of the verbal delayed contrast score ............... 133 B 1 Breakdown of data sources .............................................................................. 135

PAGE 10

10 LIST OF FIGURES Figure Page 4 1 Disorders represented in Aim 1 .......................................................................... 89 4 2 Hypothesized structural equation model.. ........................................................... 90 4 3 Final structural equation model. .......................................................................... 91 4 4 Disorders represented in Aim 2 samples ............................................................ 92 4 5 Verbal memory performance profile of the ADHD and clinical control groups.. .. 93 4 6 Distribution of TBI severity in Aim 3’s TBI sample .............................................. 93 4 7 Disorders repres ented in Aim 3’s samples ......................................................... 94 4 8 Verbal memory performance profiles of Aim 3’s TBI and orthopedic injury control groups.. ................................................................................................... 94

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COGNITIVE FACTORS CONTRIBUTING TO VERBAL MEMORY PERFORMANCE IN CLINICAL PEDIATRIC POPULATIONS By Lizabeth L. Jordan August 2014 Chair: Shelley C. Heaton Major: Psychology Memory problems and other cognitive dysfunction are commonly reported within clinical pediatric populations , including children with medical and/or psychological diagnoses . However, these “memory problems” are poorly characterized and are often described as heterogeneous. In part, the heterogeneity of memory problems within pediatric populations are likely related to variations in underlying neuroanatomical dysfunction (e.g. seizure foci associated with epilepsy, injury loci resulting from traumatic brain injury). In addition, heterogeneity of memory difficulties may be related to varied abilities in other cognitive domains (e.g. attention difficulties associated with attention deficit hyperactivity disorder, processing speed impairments after traumatic b rain injury ). T he overarching goal of this dissertation was to better understand the memory performance profile and its underlying contributing cognitive components of clinical pediatric populations. Neuropsychological data and medical information was extracted from the records of clinical pediatric patients who were evaluated in the Shands Hospital / University of Florida Pediatric Neuropsychology Clinic or the University of Florida’s Pediatric Neuropsychology Research Laboratory. Data was analyzed using univariate and

PAGE 12

12 multiva riate statistics. Overall, results revealed that the verbal memory difficulties associated with a variety of clinical pediatric populations seems to be stemming from a weakness in verbal encoding processes. Verbal retention and retrieval skills appeared similar to those of healthy normative data and clinical control samples. Furthermore, the relatively weak verbal encoding performance of clinical pediatric patients was signific antly related to performance on separate measures of verbal knowledge/reasoning. Although the clinical pediatric populations did not necessarily demonstrate a clinically significant verbal encoding impairment, this dissertation study suggest s that clinical pediatric populations have a verbal memory profile characterized by a relative weakness in the verbal encoding stage of memory processes. Future research should replicate these findings in independent samples of clinical pediatric populations and with other measures of verbal (and visual) memory . If these findings are successfully replicated, then memory treatment and intervention programs should focus on the rehabilitation of encoding abilities as well the development of verbal knowledge skills.

PAGE 13

13 CHAPTER 1 BACKGROUND Introduction Memory is a complex cognitive system that can be conceptualized in a number of ways. The following background sections review the current understanding of the human declarative verbal memory processes. Declarative memory (i.e. explicit memory) refers to information that is remembered with conscious awareness and includes episodic and semantic memory. Verbal memory refers to memory for verbal information (e.g. facts, words, stories, etc.) as opposed to nonverbal infor mation (e.g. faces, pictures, etc.). The following background review focuses on declarative verbal memory processes for reasons related to several facts: First, verbal memory is easier to examine than nonverbal memory because verbal processes (e.g. strategies) can be more overtly observed. Second, performance on nonverbal memory tests is often heavily dependent upon verbal processes. For example, individuals often use verbal strategies to remember nonverbal information (Paivio, 1991). Third, in contrast to nonverbal memory, verbal memory abilities are progressively developing throughout childhood and adolescence (Dehn, 2010). Taken together, these facts indicate that declarative verbal memory involves complex processes. Compared to nonverbal memory, ver bal memory processes are more likely to be 1) easily examined, 2) influenced by other cognitive abilities, and 3 ) disrupted by abnormal development or childhood injury. Therefore, the following background sections primarily focus discussion on declarative, verbal memory. However, distinctions between nonverbal and verbal memory are highlighted when relevant.

PAGE 14

14 This background chapter discusses declarative verbal memory within an informationprocessing model comprised of three processes: encoding, retention, and retrieval. Research on neuroanatomic correlates as well as underlying cognitive factors supporting the human declarative verbal memory process is described. In addition, information on how the verbal memory process normally develops in healthy children is reviewed. Finally, this background chapter reports the current understanding about the verbal memory process in clinical pediatric populations. Overall, the following sections review the current literature on the cognitive model, underlying neuroanatomy, and development of human declarative memory. In addition, a brief overview is provided on memory problems experienced by pediatric clinical populations. Three Verbal Memory Processes In cognitive psychology, memory is often discussed within an inf ormationprocessing framework. Early memory research (i.e. Atkinson and Shiff r in, 1971; Waugh & Norman, 1965) suggests that the verbal memory process involves a flow of information through memory “stores.” First, perceived information enters a temporary , primary store. Next, information transfers to a long term, secondary store from where information can later be accessed. Over the past few decades, the exact nature of the primary and secondary store has been debated and reconceptualized (Baddeley, 2000). Today, researchers generally agree that verbal information is transferred and utilized between stores through an integrative threestage process: 1. Encoding of perceived verbal information from the short term store into the long term store, 2. Retent ion and consolidation of verbal information within the long term store, and 3. Retrieval of retained verbal information from the long term store via the short term store. The

PAGE 15

15 following background sections provide an overview of the encoding, retention, and retrieval processes involved in the transfer of verbal information between stores. The Encoding Process The verbal memory process begins with the encoding process, when information is initially acquired and processed for long term storage (Wagner, 20 02). Some verbal information can be encoded incidentally with little cognitive demand or effort. For example, we typically process information into the long term store unconsciously at a constant level without any particular intention (Moscovitch 1994; D ehn, 2010). However, the amount of verbal information later remembered after incidental encoding is minimal compared to the amount of information remembered after intentional encoding (Brown & Craik, 2000). In fact, use of encoding strategies is unique t o intentional encoding and contributes significantly to the amount of information later remembered. The successful encoding of verbal declarative information depends upon intentional encoding strategies. For example, simple repetition (also known as rehearsal) of verbal information improves encoding (Brown & Craik, 2000; Wilson, 2009). In addition, active elaboration (i.e. processing of semantic characteristics) of verbal information improves encoding beyond simple repetition (Wilson, 2009). In general, research indicates that elaborative, meaning based encoding significantly improves subsequent memory compared to processing of only superficial characteristics (e.g. size, shape, sound, etc.; Craik & Lockhardt, 1972). Interestingly, research suggests that even nonverbal information is best encoded with semantic, verbal based strategies (e.g. verbal descriptions; Paivio, 1991). In addition to elaborative rehearsal, verbal encoding may be improved with mental reorganization of

PAGE 16

16 information (Brown & Craik, 2000). For example, grouping (i.e. “chunking” or “clustering”) information to be remembered into units seems to improve verbal encoding (Miller, 1956). Taken together, encoding of verbal memory can vary widely in terms of strategies used (e.g. repetitio n, semantic elaboration, chunking, etc.). However, memory problems may occur when patients have difficulty engaging in the encodings strategies typically used to improve verbal memory abilities (Bauer, 2003). The Retention Process After the initial encodi ng process, verbal information can either by forgotten or retained in the long term, secondary store. Successful retention of verbal information is dependent on the initial consolidation of verbal information from an unstable temporary store into the more stable long term store (Bauer, 2003; Dehn, 2010). Consolidation is a critical early stage of retention that begins immediately after encoding and continues over a period of time spanning days, weeks, or even months (Dehn, 2010). Research suggests that rapid forgetting of verbal information from the long term store (e.g. amnesic memory performance) is often reflective of poor consolidation ( Zola Morgan & Squire, 1993). Furthermore, forgetting of verbal information from the long term store may appear tem porally graded, such that information more recently consolidated is more easily forgotten than information more remotely consolidated (Bauer, 2003; ZolaMorgan & Squire, 1993). Researchers have proposed that long term consolidation (and retention) of verbal information is highly dependent on encoding and retrieval processes. That is, verbal information that is repeatedly reencoded and retrieved is better consolidated and retained than verbal information that is not repeatedly activated (Brown & Craik, 2000; Dehn, 2010). Moreover, repeated verbal encoding and retrieval distributed over

PAGE 17

17 intervals that gradually lengthen usually leads to even better consolidation (Brown & Craik, 2000; Wilson, 2009). Taken together, rapid forgetting of verbal memories likely reflect s some disruption in the initial consolidation or repeated encoding and retrieval of verbal information during the long term retention process. The Retrieval Process When verbal information is successfully encoded and retained in the long term st ore, it can be accessed through the retrieval process of verbal memory. The retrieval of verbal information can occur with conscious awareness of knowledge (i.e. “declarative” or “explicit” memory) or without (i.e. “nondeclarative” or “implicit” memory) ( as reviewed in Moscovitch, 1994). It is hypothesized that nondeclarative memory retrieval occurs in an automatic fashion with little cognitive demand. In contrast, declarative memory retrieval requires deliberate, effortful strategies. As previously men tioned, this background chapter focuses primarily on verbal declarative memory. Retrieval of verbal declarative memory requires at least two conditions: 1. The information is stored in long term memory, and 2. The information is accessible by means of the present cues (Brown & Craik, 2000; Rugg et al., 2002). In general, retrieval of verbal declarative memory benefits from additional cues (e.g. category or phonemic cues during cued recall or options presented during recognition; Brown & Craik, 2000; Wils on, 2009). Consequently, recognition tests are normally more sensitive than recall tests for determining how much verbal information was actually stored in long term memory. Furthermore, a large discrepancy between recognition and recall performance (i.e. clinically significant worse recall than recognition) may indicate verbal memory retrieval problems (Lezak, 2004; Verfaelle & Keane, 2002).

PAGE 18

18 Summary of the Three Memory Processes Researchers describe memory within an informationprocessing model that des cribes flow of information through three stages of memory: 1. Encoding, 2. Retention, and 3. Retrieval. Importantly, these three verbal memory processes are interdependent, and memory problems cannot necessarily be localized to or explained by only one of these memory processes. For example, retention of memory into long term storage is dependent on the degree and amount of repeated encoding; retrieval of memory is dependent on successful long term retention as well as appropriate cues that are consistent with initial encoding (Paller, 2002). Nonetheless, the verbal memory processes of encoding, retention, and retrieval are a useful heuristic for characterizing variations in memory problems (Bauer, 2003). The following sections of this background chapter refer to the verbal encoding, retention, and retrieval processes in order to provide a conceptual framework for discussing verbal memory. Neural Correlates of Verbal Memory Processes The neural correlates of memory are not completely understood. Arguabl y many brain areas are involved, which suggests a complex neural network underlying verbal memory. As technology used to study the brain improves, research indicates that the verbal memory processes involve two discrete brain structures: the prefrontal cortex (PFC) and medial temporal lobe (MTL). In the previous sections, we discussed a cognitive model where three verbal memory processes (i.e. encoding, retention, and retrieval) are interdependent with each other. Lesion and neuroimaging research additionally indicate that verbal encoding, retention, and retrieval processes are supported by interdependent brain function (Nyberg, 2002). The following sections

PAGE 19

19 describe the contributions of the PFC, MTL, and cross cortical networks to verbal memory processes . Prefrontal Cortex The prefrontal cortex (PFC) is associated with control of attention, working memory, inhibition, and executive function across various cognitive tasks (GoldmanRakic, 1996; Mayes, 2002; Wagner, 2002; Stuss & Benson, 1984). Within verbal memory, PFC controlled processes seem to play an important role in both strategic encoding and retrieval memory processes. In regards to verbal encoding, activation of the PFC is associated with the successful manipulation of information to be remember ed during elaborative and clustering strategies (Mayes, 2002; Moscovitch, 1994; Wagner, 2002). In comparison, patients with frontal lobe lesions demonstrate difficulties with elaborative verbal encoding strategies. These verbal encoding deficits are red uced when information to be remembered is presented in a semantically organized structure (Kopelman & Stanhope, 2002). In the latter condition, researchers hypothesize that the verbal memory improves because demand for frontal lobe functioning is reduced. Similar to the role in encoding, the PFC is critical for various controlled aspects of retrieval. In particular, the PFC has been associated with monitoring of controlled memory search, evaluation of accuracy, and reconstruction of memory traces from long term storage (Mayes, 2002; Paller, 2002; Wagner, 2002). In fact, patients with frontal lobe lesions demonstrate confabulation, sourcemonitoring errors as well as disorganized verbal memory searches during retrieval tasks (Kopelman & Stanhope, 2002; Paller, 2002). Furthermore, the relationship between PFC activity and verbal

PAGE 20

20 retrieval appears mediated by the availability of retrieval cues such that increased PFC activity is reported whe n retrieval cues are reduced ( Wagner, 2002). It is also importan t to note that normal functioning of the PFC is supported by a number of white matter connections within the frontal lobes and between distinct supporting structures (e.g. to basal ganglia and cerebellum). Consequently, brain damage to other structures wi thin these cortical networks or to the white matter pathways may disrupt PFC functioning. Disruption anywhere within the PFC network may cause impairments of encoding and retrieval processes, even though discrete lesions may not be present within the PFC. Neocortex, Medial Temporal Lobe, Diencephalon, and Basal Forebrain While the PFC is commonly associated with strategic encoding and retrieval processes, the neocortex, medial temporal lobe (MTL), and associated structures are most often associated with long term retention and retrieval of verbal information (Moscovitch, 1994; Shimamura, 2002). To elaborate, individual components of a complex verbal memory are stored in various neocortical areas (Paller, 2002). Collectively, the neocortical brain regions store the complete verbal memory (Paller, 2002). During consolidation, the hippocampus and surrounding MTL structures (e.g. parahippocampal, perirhinal, entorhinal, and subiculum) coordinate activity between the neocortical areas (Mayes, 2002; Moscovitch, 1994; Paller, 2002; Nyberg, 2002). The hippocampus mediates verbal memory storage for a considerable time before the memory is completely consolidated and stable in a neocortical long term store (Shimamura, 2002; Zola & Morgan, 1993). Then, during the retrieval process, the hippocampus coordinates the re activation of the same neocortical areas (Brown & Craik, 2000; Mayes, 2002; McDermott & Buckner, 2002; Moscovitch, 1994; Nyberg,

PAGE 21

21 2002; Shimamura, 2002). Of note, lesion studies have demonstrated that the hippocampus is not necessary for all forms of verbal memory retrieval to occur (e.g. some recognition judgments rely solely on entorhinal cortex). However, the hippocampus does appear to significantly facilitate the recall of memorized verbal informat ion (Broadbent, 2002; Mayes, 2002; McDermott & Buckner, 2002; Shimamura, 2002). Research suggests that the diencephalon and the basal forebrain mediate the establishment of long term verbal declarative memory in the neocortex (Paller, 2002; Zola Morgan & Squire, 1992). (The diencephalon includes the mammillary nuclei, mammillothalamic tract, and anterior nucleus of the thalamus; the basal forebrain includes the medial septal nucleus and diagonal band of Brach; Paller, 2002; ZolaMorgan & Squire, 1992). The verbal memory performance of patients with diencephalon or basal forebrain damage is notable for poor source monitoring, temporal order judgments, and confabulation (Paller, 2002; ZolaMorgan & Squire, 1992). These memory errors indicate a disruption during the processing (e.g. encoding or retrieval) of verbal information in relation to prior knowledge. In other words, diencephalon and basal forebrain functioning may be critical for attending, inhibiting, reconstructing, or associating new verbal memory in relation to prior knowledge (Paller, 2002; ZolaMorgan & Squire, 1992). These brain areas likely process verbal information in relation to prior knowledge as it transfers between the MTL, PFC, and neocortical areas (ZolaMorgan & Squire, 1993). Sum mary of the Neural Correlates of Verbal Memory Multiple brain areas support the verbal memory processes. The role of each described region does not exclusively occur within a single verbal memory process

PAGE 22

22 (Nyberg, 2002). For example, the PFC appears impor tant for controlled verbal encoding and retrieval processes. In addition, the MTL (and neocortical areas) is required for verbal retention as well as some retrieval processes. Thus, disruption of the verbal memory process may result from dysfunction of t he PFC, MTL, or cross cortical networks and white matter. Other Cognitive Factors Underlying the Verbal Memory Process Some verbal memory processes appear to occur independently from other cognitive factors. For example, initial encoding of verbal memory can be automatic and unintentional with minimal cognitive demands (Wilson, 2009). However, variations in declarative verbal memory are common and likely relate to normal variation in other cognitive domains (Brooks, 2009). More specifically, researchers have suggested that verbal knowledge, focused and sustained attention, processing speed, and working memory/executive functioning contribute to verbal memory performance of healthy individuals (Ornstein et al., 2006). The following sections discuss the research that support s a model where other cognitive factors likely contribute to verbal memory abilities. Verbal Knowledge In general, there exists a relationship between learning/memory abilities and intellectual functioning, such that low intellectual functioning is associated with poor memory skills (Brooks, 2009). Research suggests that the causeand effect nature of this relationship is likely bi directional, such that memory influences knowledge and at the sametime knowledge impacts memory (Baddeley, 2000). Here, we specifically discuss how knowledge influences the memory processes. Although prior knowledge could negatively interfere with the accuracy of memory retrieval through proactive

PAGE 23

23 interference, research more frequently indicates that increased knowledge and familiarity improves initial encoding abilities (Baddeley, 2000; Brown & Craik, 2000; Hitch, 2006; Ornstein et al., 2006). For example, some researchers suggest that faces are easily well recognized because all healthy humans have “expertise” in the visual perception of faces (Brown & Craik, 2000). In comparison to nonverbal knowledge, verbal knowledge is more consistently related to general verbal (and nonverbal) memory performance (Cohen, 1997). Theoretically, verbal knowledge and familiarity provides structure and content for rich elaboration and inter item associations between prior and current knowledge (Dehn, 2010). Taken together, increased verbal knowledge likely contributes to the verbal memory performance by improving verbal encoding strategies. Current research reports that semantic knowledge is represented in distributed cortical areas. In particular, the anterior temporal lobe and parietal lobe have been associated with the storage of verbal information (Martin & Chao, 2001; Paller, 2002) . Focused & Sustained Attention On any cognitive task, focused/selective and sustained attention improves performance (Posner & Peterson, 1990). Specifically within memory, verbal encoding processes are improved when an individual pays attention to the in formation to be remembered (Dehn, 2010; Moscovitch, 1994; Ornsetin et al., 2005; Wilson, 2009). Beyond attention to salient emotional and physical cues, purposefully directed selective and sustained attention can remarkably improve verbal encoding (Ornstein et al., 2006). For example, focused and sustained attention is required for even passive rehearsal of verbal information (Baddeley, 2000; Brown & Craik, 2000). Moreover, attention to verbal information further determines the degree at which informat ion will be processed

PAGE 24

24 during elaborative encoding (Craik & Lockhardt, 1972). In addition, selective retrieval of verbal information requires focused and sustained attention (Brown & Craik, 2000). Taken together, research indicates that focused and sustained attention contribute to verbal memory encoding and retrieval. In regards to neuroanatomy, attention is supported by a network of brain regions; however, the parietal lobes and anterior cingulate within the frontal lobes appear to be especially important for focused attention (GoldmanRakic, 1996; Posner & Peterson, 1990; Stuss & Benson, 1984). For example, individuals with frontal lobe lesions often demonstrate impairments of directed and sustained attention across modalities, including visual, audito ry, and kinesthetic domains (Stuss & Benson, 1984). These findings indicate that similar brain regions (i.e. the PFC) supports selective and sustained attention as well as strategic verbal encoding and retrieval processes. Processing Speed Processing spe ed refers to an individual’s efficiency to complete a cognitive task pro. Research indicates that individuals with more efficient processing speed (as measured by their naming fluency) are able to additionally remember more information than their sameaged peers (Diamond, 2006; Schneider & Pressley, 1989). To explain this relationship, Baddeley suggests encoding is timedependent with more elaborative processing requiring more time to process (Baddeley, 2000; Brown & Craik, 2000). For example, integrati on of new information with prior knowledge can extend the duration of the encoding process (Ornstein et al., 2006). Furthermore, efficient retrieval of verbal memory requires adequate retrieval speed and fluency (Dehn, 2010). Overall, increased processing efficiency allows for more strategic encoding and retrieval of verbal information (Dehn, 2010; Lezak, 2004; Ornstein et al., 2006).

PAGE 25

25 Unlike other cognitive domains, processing speed is not necessarily localized to a single brain area; rather, processing speed is reliant on white matter connections (i.e. myelinated axons of neurons) between brain regions. Research indicates that processing speed during encoding, consolidation, and retrieval depends on the degree of integration and interconnectedness betw een the neocortical areas storing the long term verbal memory (Dehn, 2010). Therefore, the integrity of connections between cortical areas including the hippocampus contributes to verbal memory performance. Working Memory / Executive Functioning Baddeley (2000) well describes a working memory store that exists separately from, but is highly related to a long term memory store. In brief, he defines working memory as a timelimited short term store where information is actively processed and manipulated be fore entering the long term store. Researchers often describe Baddeley’s concept of working memory within a larger, concept of “executive functions,” which include any controlled, higher level processes such as inhibition, cognitive flexibility and set sh ifting (Diamond, 2006). Hitch (2006) reports that separating the neuropsychological domains of working memory and executive functions is problematic because the constructs overlap in terms of conceptual framework as well as their ability to be measured in cognitive testing; consequently, the terminology of working memory and executive functioning is used here interchangeably. Executive functioning is theoretically the most important of the many neuropsychological abilities. Baddeley (2000) hypothesizes that all verbal information must pass through the short term, working memory store before entering the long term store. Moreover, he suggests that elaborative processing and reorganization of verbal information during encoding occurs in the working memor y store (Baddeley, 2000;

PAGE 26

26 Brown & Craik, 2000). In addition, researchers suggest that verbal retrieval processes require working memory/executive functioning to control, monitor, coordinate, evaluate, and revise search processes (Dehn, 2010). Research supports these hypotheses and has demonstrated that healthy verbal memory performance correlates significantly with measures of executive functioning (Cohen, 1997; De Alwis et al . , 2009; Wilson, 2009). In addition, Baddeley and Wilson (1988) described a Dysexecutive type of amnesia, wherein primary executive impairments can result in clinically significant verbal encoding and retrieval problems. Additionally, c ognitive rehabilitation studies report that treatment programs targeting the improvement of executi ve functioning additionally improves verbal encoding (Wilson, 2009). As with focused and sustained attention, executive functioning is associated with activity in the frontal lobes (Stuss & Benson, 1984; GoldmanRakic, 1996; Diamond, 2006). Consequently, PFC activity during strategic verbal encoding and retrieval processes may be related to executive demands during these verbal memory processes. Overall, research evidence highly suggests that executive functioning contributes to verbal memory processes. Summary of Cognitive Contributors Normal variation of verbal memory performance appears to be partially explained by variations of other cognitive abilities, including verbal knowledge, focused and sustained attention, processing speed, and working memory /executive functioning. In particular, elaborative verbal encoding strategies may be mediated by verbal knowledge, focused and sustained attention, processing speed, and working memory/executive functions. Additionally, effortful and systematic verbal retrieval searches are mediated by similar cognitive factors. Taken together, several other cognitive factors seem to contribute to the verbal memory processes.

PAGE 27

27 Development of Verbal Memory Memory is reportedly a cognitive ability present very early in hum an development. For example, implicit recognition of nonverbal information (e.g. faces, sounds) is exhibited as early as the first two days of life -significantly earlier than the development of other cognitive skills (Baddeley, 2000; Dehn, 2010). In a ddition, delayedmatchedto sample studies indicate that infants as young as 2021 months are able to retain and explicitly retrieve nonverbal information when encoding demands are minimized (Diamond, 2006). This evidence indicates that early in development, humans have the ability to successfully encode, retain, and retrieve memories (de Haan et al., 2006). Nonetheless, the phenomenon of “childhood amnesia” indicates that older children and adults remember very little of the complex, episodic information experienced before the age of 4 years old (Dehn, 2010). Researchers have concluded that nonverbal and simple semantic memory abilities (e.g. knowledge of places, faces, names, etc.) develop before complex, verbal memory abilities (e.g. memory of complete events or episodes) (Ornstein et al., 2006; Dehn, 2010;Hann et al., 2006). Furthermore, declarative memory for complex, verbal information significantly improves throughout childhood and adolescence. In general, research suggests that improvements in chil dren’s long term verbal memory performance are primarily explained by improved encoding and retrieval strategies, while retention abilities remain relatively stable (Dehn, 2010). In addition to maturing neuroanatomy that underlies verbal encoding and ret rieval processes, children learn to become more strategic and develop more complex, effective, and consistent verbal encoding and retrieval strategies with age (Ornstein et al., 2006). Agerelated improvements in other cognitive domains additionally seem t o support improvements in

PAGE 28

28 verbal encoding and retrieval abilities. Although previous research on this topic is limited, the following sections review the current literature about the developing neuroanatomy, strategies, and agerelated cognitive factors t hat underlie typical human development of encoding and retrieval abilities. Development of Verbal Encoding & Retrieval Strategies Throughout childhood and adolescence, children become more strategic in their encoding and retrieval abilities. Simple strategies for encoding (and retrieval) appear early in childhood. For example, children as young as 18months use naming, pointing, and “peaking” (i.e. looking back at stimuli) behaviors when asked to remember nonverbal information, such as the spatial locat ion of items (DeLoache e al., 1985; Ornstein et al., 2006; Schneider & Pressley, 1989). Notably, young children appear to primarily encode information based on visual characteristics (Baddeley, 2000). Around the ages of 7 and 10, children begin to adopt verbal encoding strategies (Baddeley, 2000). Initial verbal encoding strategies rely on simple rote repetition (Ornstein et al., 2006). With age, children’s encoding processes become more active, elaborative, and semantically based (Dehn, 2010; Ornstein et al., 2006). For example, older children demonstrate organized, semantic clustering of verbal information during encoding (Dehn, 2010; Ornstein et al., 2006). Notably, young children will use elaborative encoding strategies when they are directly instructed to. In addition, their memory will benefit from instruction to use elaborative encoding strategies. However, young children exhibit “utilization deficiency .” That is, young children do not spontaneously and reliably use elaborative encoding until later in development despite having the ability and capacity to use elaborative encoding processes (Dehn, 2010; Ornstein et al., 2006; Schneider & Pressley, 1989). Similarly,

PAGE 29

29 children 5 years old and younger will only utilize retrieval cues if they are overtly provided (Schneider & Pressley, 1989). In contrast, older children will use encoding strategies spontaneously and consistently across multiple settings (Dehn, 2010; Ornstein et al., 2006). Additionally, older children will use verbal retrieval cues spontaneously and systematically in order to improve their recall abilities (Dehn, 2010; Schneider & Pressley, 1989). Taken together, developmental research suggests that older children and adolescents begin to use more efficient verbal encoding and retrieval strategies as their metamemory (i.e. knowledge about their own memory abilities) improves (Dehn, 2010). As children mature, they are more likely to monitor their own verbal memory abilities and use appropriate encoding and retrieval strategies. Developing Neuroanatomy Development of verbal memory processes is in part dependent on the development of the neuroanatomy which supports encoding, retention, and retrieval processes. Throughout childhood and adolescence, brain regions and connections are continuously developing. In regards to the MTL, research indicates that at birth the hippocampus is fairly well developed; however, its structure and function co ntinues to significantly mature during the first few years of life (Dehn, 2010; de Hann et al., 2006; Ornstein et al., 2005). Consistent with the development of long term verbal memory abilities, gradual changes in connections between the hippocampus and association areas are experienced during early childhood (Brown & Chiu, 2006; Ornstein et al., 2006). By the age of 6 years old, the structure and function of the MTL is fully developed and as mature as adults’ (Dehn, 2010). In contrast, the PFC develops gradually throughout childhood until young adulthood (Brown & Chiu, 2006). Development of the PFC correlates with improved

PAGE 30

30 focused attention, monitoring, and selection/inhibition of verbal memory (Dehn, 2010; Enns & Trick, 2006). For example, young children with immature frontal lobes exhibit source memory confusion as well as vulnerabilities to distorted memories and false recollections (Dehn, 2010). However, as the PFC matures, older children simultaneously exhibit better verbal memory strategies (disc ussed in more detail below). Furthermore, Sowell and colleagues (2001) reported that frontal lobe maturation was more predictive of verbal delayed memory abilities than MTL development. Overall, the child and adolescent studies are consistent with the a dult research, which similarly report a strong correlation between PFC and MTL activity and verbal memory performance (Brown & Chiu, 2006). Development of Other Cognitive Factors Beyond the development of supporting neuroanatomy, other agerelated cognitiv e factors mediate developmental changes in verbal encoding and retrieval abilities. Dehn (2010) reports that improvements in long term verbal memory performance after the age of 12 years old are attributable to expanding knowledge and other maturing cogni tive domains. For example, the development of verbal knowledge and verbal memory is well correlated across the lifespan ( Schneider & Pressley, 1989) . The correlation between verbal knowledge and memory strengthens with increasing age (Dehn, 2010; Hitch, 2006; Kail 1984; Ornstein et al., 2006). In addition, focused and sustained attention, working memory/executive functions, and processing speed improve rapidly during school aged years and gradually through young adulthood (Diamond, 2006; Schneider & Pres sley, 1989). Development of these cognitive domains likely mediates the development of intentional and strategic verbal encoding and retrieval strategies (Baddeley, 2000; Dehn, 2010; Diamond, 2006; Kail

PAGE 31

31 1984;Ornstein et al., 2006). For example, when speed of processing is equated, children and adults demonstrate comparable verbal memory spans (Schneider & Pressley, 1989). Taken together, research suggests that the development of other cognitive factors (e.g. verbal knowledge, focused and sustained attent ion, processing speed, and working memory/executive functioning) mediates the relationship between increasing age and improved declarative memory abilities. Summary of Verbal Memory Development Overall, the encoding, retention, and retrieval of complex, v erbal memory improves significantly throughout childhood and adolescent. Healthy development of verbal memory abilities has been related to the development of essential neuroanatomy as well as the development of other cognitive factors, such as verbal knowledge, attention, working memory/executive functioning, and processing speed. In general, there are a number of factors which seem to contribute to the healthy development of verbal memory processes. Verbal Memory Problems in Clinical Pediatric Populations Memory problems are frequently reported in pediatric patients with history of neurodevelopment and acquired brain injuries (Dehn, 2010). Pediatric verbal memory problems are heterogeneous and may stem from deficits in the encoding, retention, or retrie val processes (Dehn, 2010). The etiology of verbal memory problems may be related to disruption of the neuroanatomy underlying the memory process (e.g. temporal lobe epilepsy; Felix & Hunter, 2010; Westerveld, 2010). Alternatively, memory problems may oc cur secondary to impairments of some other cognitive factor(s), such as verbal knowledge, attention, etc (Brooks, 2009). A variety of pediatric patients report memory complaints; the following sections review the limited information known about

PAGE 32

32 the memory performance of children and adolescents with history of Attention Deficit Hyperactivity Disorder (ADHD), Pervasive Developmental Disorder (PDD), epilepsy, childhood cancer, and traumatic brain injury (TBI). These pediatric disorders were selected for revi ew because these are common neurodevelopmental and acquired brain injury populations referred to Pediatric Neuropsychology services for evaluation and/or treatment of memory problems. Attention Deficit Hyperactivity Disorder Attention Deficit Hyperactivity Disorder (ADHD) is a complex neurodevelopmental disorder characterized by developmentally excessive inattention, hyperactivity, and impulsivity (APA, 2000). It is considered the most common psychiatric disorder of childhood and affects 5% of school aged children and occurs more frequently in males than females (Goldstein, 2011; Marks et al., 2010). It is suspected that the primary pathology of ADHD symptoms is dysfunction of the dopamine and norepinephrine neurotransmitter systems in the prefrontal cortex and locus coeruleus (Arnsten, 2006). Neuroimaging research further suggests primary frontal lobe pathology (i.e. reduced frontal cortex volumes) as well as diffuse abnormal brain development in regions heavily connected to the frontal lobe (e.g. basal g anglia, cerebellum; Goldstein, 2011; Marks et al., 2010). Approximately 50% of children with ADHD additionally have been diagnosed with Oppositional Defiant Disorder (ODD) or Conduct Disorder (CD; APA, 2000). ). In addition to these disruptive behavior disorders, childhood ADHD is associated with higher rates of mood, anxiety, learning, and communication disorders (APA, 2000). On neuropsychological testing, children and adolescents with ADHD primarily exhibit mild to moderate impairments in attention, working memory, and executive functioning

PAGE 33

33 (Denckla, 1996; Doyle, 2006). In addition, research indicates that children with ADHD experience verbal learning and memory problems. In general, research on the verbal memory performance profile of children with A DHD is limited. However, Denckla (1996) suggests that verbal memory problems in children with ADHD are primarily characterized by encoding and retrieval impairments. More specifically, children with ADHD demonstrate poor attention and suboptimal strategi es during verbal encoding. For example, they do not spontaneously organize information to be remembered (Denckla, 1996). In addition, children with ADHD demonstrate poor verbal retrieval. For example, children with ADHD take longer time to recall the same amount of accurate verbal information as agematched controls (Kourakis et al., 2004). This impairment in retrieval efficiency does not appear to be related to processing speed. In contrast, Denckla (1996) suggests that children with ADHD exhibit inefficient verbal retrieval due to impulsive, disorganized and/or poorly monitored verbal retrieval strategies. For example, children with ADHD often take longer to recall information because they retrieve the same items over and over again (Denckla, 1996) . There has been no research evidence that indicates children with ADHD suffer from retention problems (Denckla, 1996). In fact, children with ADHD demonstrate normal rates of forgetting verbal information (Cahn & Marcotte, 1995) Taken together, curren t literature suggests that children with ADHD suffer from verbal encoding and retrieval problems that are related to poor attention and working memory/executive functioning. However, the current literature is limited to a few number of research papers. T he direct relationship between verbal memory, attention, and working

PAGE 34

34 memory/executive functioning in children with ADHD has not been systematically investigated. Pervasive Developmental Disorders Pervasive Developmental Disorders (PDDs; also known as auti sm spectrum disorders) include a range of neurodevelopmental disorders, including Autism, Asperger’s disorder, Rhett’s disorder, childhood disintegrative disorder, and pervasive developmental disorder, not otherwise specified. It is estimated that 1 in 150 individuals in the USA have some type of PDD, and the incidence of autism may be increasing over time (Wolf & Paterson, 2010). Rhett’s disorder and childhood disintegrative disorder are very rare, and most research on PDDs has focused on Autism and Asperger’s disorder. Although the etiology is not well understood, research reports that compared to healthy controls, individuals with PDDs have significantly increased volume of the frontal, temporal, and parietal lobes (as reviewed in Wolf & Paterson, 2010 and Corbett & Gunther, 2011). In addition, they have decreased neuronal cell size and paucity of long range fibers in a number of brain regions including the ventromedial prefrontal, subgenual prefrontal regions, and amygdala (as reviewed in Wolf & Pater son, 2010 and Corbett & Gunther, 2011). There has been limited research specifically examining temporal lobe dysfunction associated with PDDs. However, evidence indicates that individuals with PDD have dysfunctional amygdalae as well as enlarged and/or a bnormally shaped hippocampi (Dawson et al., 2002; Wolf & Paterson, 2010; Corbett & Gunther, 2011). Diagnosis of Autism and Asperger’s disorder requires impairments in socialization as well as repetitive/restricted behaviors or interests (APA, 2004). Com pared to those diagnosed with Asperger’s disorder, children with Autism additionally exhibit difficulties

PAGE 35

35 with communication (APA, 2004). In terms of their neuropsychological profile, children and adolescents with PDDs exhibit a wide range of cognitive st rengths and weaknesses. The typical profile of a child with autism includes deficits in language, attention, and executive functioning (e.g. problems with mental flexibility, shifting, orienting, disengaging; Corbett et al., 2008; Dawson, 2002; O’hearn et al., 2008; Wolf and Paterson, 2010). Comorbid symptoms of ADHD are common (Corbett et al., 2008; Corbett & Gunther, 2011). In regards to verbal learning and memory, individuals with PDD demonstrate the impaired complex and source memory, but relatively intact single word or pairedword associate memory (Bowler, 2004; Dawson, 2002; Williams, 2006). Additional research reports that verbal memory difficulties associated with PDD are largely reduced with cued recall and recognition tests. Taken together, research suggests that the verbal memory performance of children with PDD suffers from specifically a verbal retrieval deficit (Bowler, 2004). Epilepsy Epilepsy is a neurological condition diagnosed when an individual experiences recurrent seizures accomp anied with abnormal electrical discharge in the brain. It affects 1 2% of school aged children and is the most common childhood neurological disorder (Felix & Hunter, 2010). The etiology of epilepsy can be symptomatic (i.e. from a known cause like brain anomalies, anoxia, infarcts, trauma, etc.), idiopathic (i.e. an unknown cause), or cryptogenic (i.e. probably from a brain abnormality that has not yet been identified; as reviewed in Bennett et al., 2011; Felix & Hunter, 2010; Westerveld, 2010,). Epilepsy syndromes vary in terms of seizure presentation (e.g. myoclonic, tonic clonic, absence, etc) and localization of electrical discharge (i.e. generalized, localized, or undetermined; as reviewed in Felix & Hunter, 2010). Temporal lobe and

PAGE 36

36 frontal lobe epilepsies are the most common types of epilepsy in children (and adults). Absence seizures, which are brief nonconvulsive staring spells, account for 211% of all seizure diagnoses. Within childhood epilepsy populations, overall intellectual and cognitive functioning varies widely from low average to average (Felix & Hunter, 2010; Westerveld, 2010). In general, the neuropsychological profile of children with epilepsy depends on their seizure duration, location, and frequency (Carreno, 2008; Felix & Hunter, 2010; Westerveld, 2010). Increased seizure duration and frequency are both associated with decreased cognitive abilities (Carreno, 2008; Felix & Hunter, 2010). Across epilepsy groups, impairments in attention, working memory, and executive function have been reported (Felix & Hunter, 2010). In particular, studies have found that 30% of children with epilepsy exhibit clinically significant attentional problems (Bennett et al., 2011) Individuals with absence or frontal lobe seizures demonstrate greater sustained attention impairments and executive dysfunction than those with focal seizures not localized to the frontal lobe (Bennett et al., 2011; Motamedi & Meador, 2003; Patrikelis et al., 2009). Temporal lobe epilepsy is often associated with impaired l earning and memory (Motamedi & Meador, 2003; Rzezak et al., 2011; Westerveld, 2010). However, some interesting research suggests that children with absence or frontal epilepsy may perform worse on verbal memory tests than children with temporal lobe epilepsy (Felix & Hunter, 2010). Unlike adults with epilepsy, children with epilepsy do not demonstrate seizure lateralization effects on verbal memory performance (Felix & Hunter, 2010). However, right sided temporal lobe epilepsy patients demonstrate reduced retention of nonverbal information compared to left sided temporal epilepsy patients (Felix & Hunter,

PAGE 37

37 2010). In addition, Borden and colleagues (2007) reported that intellectual functioning mediates memory performance within a sample of children with epilepsy. In addition to seizure effects, research indicates that use of antiepileptic medication also impacts neuropsychological functioning (Bennett et al., 2011; Carreno, 2008; Motamedi & Meador, 2003; Westerveld, 2010). There is some evidence that ef fects of antiepileptic medication have been reduced in more recently developed medications (Westerveld, 2010). However, antiepileptic medication is associated with reduced processing speed, learning/memory, language, and nonverbal processing (Felix & Hunt er, 2010). Traumatic Brain Injury Childhood traumatic brain injury (TBI) is a heterogeneous group that includes mild, moderate, and severe head injuries, including those that result from abuse, falls, motor vehicle, motor vehiclerelated, pedestrian, and bicycle related accidents. In total, TBI affects approximately 180 out of 1000,000 children younger than age 15 per year, and this incidence rate varies by age group and gender (e.g., in late childhood and adolescence, brain injury rates are greater for males than females, Max, 2005). Primary injury, including contusions and diffuse axonal shearing and vascular insults, is caused by direct impact and/or accelerationdeceleration applied to the head. Secondary injury, including hypoxia, brain swelling, and seizures, is additionally likely to occur post injury (Kirkwood et al. 2010). The loci of lesions resulting from TBI can occur in various brain regions depending on the mechanism and severity of the injury; however, the frontal lobes are particularly vul nerable to damage. In addition, axonal shearing can be widespread and diffuse across brain regions.

PAGE 38

38 Given the heterogeneity of brain injuries associated with TBI, the neuropsychological profile of childhood TBI is diverse. Effects in TBI are strongly r elated to injury severity. Mild TBI effects are subtle and typically resolve within a few weeks (Yeates, 2010). In contrast, moderatesevere TBI results in more clinically significant cognitive and behavioral changes that persist beyond 6 months post in jury (Yeates, 2010). The most common cognitive profile of severe childhood TBI includes impairments of processing speed, learning/memory, and functional communication (Kirkwood et al. 2010; Yeates, 2010). Problems with nonverbal abilities, sustained atten tion, and executive functioning are additionally common (Kirkwood et al. 2010; Yeates, 2010). Specifically regarding memory difficulties, research suggests that pediatric TBI patients suffer from both encoding and retrieval problems. For example, Harris ( 1996) reported that pediatric patients with history of severe TBI and impaired verbal recall engage in inefficient, passive rehearsal strategies during encoding. In addition, compared to agematched controls, pediatric severe TBI patients do not benefit f rom repeated learning trials (Harris, 1996). Mottram & Donders (2006) extended these findings and found that although merely repeating information did not improve memory of pediatric TBI patients, slowing down the rate of stimuli presentation during encoding improved subsequent memory. This research suggests that processing speed likely mediates verbal memory encoding in pediatric TBI patients (Mottram & Donders, 2006). Moreover, rehabilitation studies report that pediatric TBI patients are able to effec tively use semantic, elaborative encoding strategies in order to improve their verbal memory (Oberg & Turskstra, 1968). However, they only engage in semantic, ageappropriate

PAGE 39

39 elaborative encoding strategies when explicitly instructed (Oberg & Turkstra, 19 68). That is, pediatric TBI patients do not spontaneously engage in elaborative strategies. In addition to verbal encoding difficulties, research indicates that children with TBI demonstrate retrieval difficulties (Catroppa & Anderson, 2002). More speci fically, severe TBI patients suffer from intrusions and reduced output during verbal recall tests (Yeates et al.,1994). There are limited studies examining the neural correlate disruption underlying memory problems in pediatric TBI. In adult studies, TB I patients exhibit reduced activity of the right PFC during retrieval in a doseresponse relationship to injury severity (McAllister, 2006). In addition, adult TBI patients demonstrate decreased MTL activity during encoding in a doseresponse relationship with injury severity (McAllister et al., 2006) Childhood Cancer Common forms of childhood cancer include leukemia, lymphoma, and solid brain tumors. Leukemia, which is malignant disorder of cells produced in the bone marrow and includes acute lymphoblastic leukemia (ALL) as well as acute myelogenous leukemia (AML), is the most common form of cancer experienced before the age of 20 years old (Krull & Jain, 2010). Lymphomas, which are malignant disorders of the lymph nodes and include Hodgkin lymphoma (HL) and nonHodgkin lymphoma (NHL), constitute the third most common form of pediatric cancer (Krull & Jain, 2010). Brain tumors are the most common childhood solid tumor, and they occur commonly in midline structures (e.g. cerebellum, fourth ventricle, and brainstem) and present with increased intracranial pressure (Rey Casserly, 2010).

PAGE 40

40 When possible, brain tumors are treated by surgical resection. Moreover, children with brain tumors, leukemias, and lymphomas are often treated with chemotherapy or radiation therapy as well as glucocorticoids. Unfortunately, cranial radiation and common chemotherapeutic agents (e.g. methotrexate) target normal brain cells in addition to tumor cells; consequently, these treatments have been associated with white matter abnormalities in the frontal lobe as well as abnormal hippocampal development (Nagel et al., 2004). Use of glucocorticoids (e.g. prednisone) has been shown to cause damage in hippocampus. In general, amount and type of cancer treatment seems to impact neurocognitive outcomes in childhood cancer. For example, high doses of cranial radiation are associated with poor overall intellectual functioning, academic functioning, attention, processing speed, and memory. Methotrexate has been associated with reduced intelligence, attention, processing speed, and executive functions . More specifically, children with ALL treated with chemotherapy exhibit a verbal memory profile suggestive of poor attention and deficits in strategic planning (Hill et al., 1997; Rodgers et al., 1992). Glucocorticoids are associated with poor memory. Overall, the neuropsychological profile of children who have been treated for cancer includes deficits in working memory, attention, and processing speed. Summary of Verbal Memory Problems in Clinical Pediatric Populations Cognitive dysfunction is commonly reported across pediatric clinical populations. Although research often reports that children with neurodevelopmental and acquired brain injuries suffer from “memory problems,” these “memory problems” are poorly characterized. In contrast, more often “memory problems” are described as heterogeneous even within selected populations. In part, the heterogeneity of memory difficulties seems related to heterogeneity of the etiology and the underlying

PAGE 41

41 neuroanatomy dysfunction (e.g. loci of injury in TBI, epilepsy, and cancer). Additionally, the heterogeneity of memory difficulties appears related to heterogeneity in other cognitive domains (e.g. attention and executive dysfunction associated with ADHD, childhood cancer, and PDDs). Summary of Background Information Research in healthy children and adults has long sought to understand the declarative verbal memory process and its underlying cognitive components. A widely accepted cognitive model of declarative verbal memory conceptualizes a system that includes 3 component processes (i.e. encoding, retention, and retrieval). Verbal encoding, retention, and retrieval processes are supported by neural correlates (e.g. PFC, MTL, neocortex) and other cognitive factors (e.g. attention, processing speed, working memory/executive functioning, etc). There is limited research that examines how the verbal memory process, including its underlying cognitive components and neuroanatomic correlates, develops in clinical pediatric populations. Thus far, pediatric resear ch has focused on studying normal development of memory in healthy children and adolescents. Research in healthy children and adolescents suggests that development of successful verbal memory is mediated by the development of underlying neuroanatomy (e.g. , PFC, MTL) as well as other agerelated cognitive abilities, including verbal knowledge, attention, processing speed, and working memory/executive functions (Dehn, 2010; Ornstein et al., 2006). Although memory impairments are common in pediatric clinical populations (i.e. children with ADHD, PDD, epilepsy, TBI, cancer, etc.), very little information is known regarding verbal memory processes affected in pediatric neurodevelopmental disorders and acquired brain injuries.

PAGE 42

42 CHAPTER 2 OVERVIEW AND STATEMENT O F THE PROBLEM The overall goal of this study was to extend previous adult and healthy childhood research findings to pediatric patient populations. Specifically, the study aimed to understand the developing verbal memory process and its’ underlying cognit ive components in clinical pediatric populations with history of neurologic and/or developmental problems. Pediatric patients often experience verbal “memory problems”; however, previous research failed to describe the exact nature of these difficulties o r how they may be occurring. Verbal memory problems can include a variety of impairments within the verbal memory process; for example, difficulties with memory could stem from a purely encoding, retention, or retrieval impairment, or it could stem from a ny combination of impairments (Dehn, 2010). In addition, verbal memory problems may have varying etiologies (e.g. memory difficulties may be a manifestation of primary intellectual, attention, processing speed, or executive dysfunction; Brooks, 2009). Thi s study aimed to better understand the variance of verbal “memory problems” within clinical pediatric populations. First, we examined the relationship between a general measure of delayed verbal memory and various cognitive factors within a large, mixed pediatric clinical sample. More specifically, we developed a model which indicated the relative contributions that cognitive factors have on delayed verbal memory. This model was constructed and tested within a mixed clinical sample so that broad conclusio ns can be made regarding the impact of various contributing factors to delayed verbal memory. Next, we used this general model to more closely examine the verbal memory profiles of a common

PAGE 43

43 childhood neurodevelopmental disorder (i.e. ADHD) and a pediatri c acquired brain injury (i.e. TBI) population. Previous research has well described the attention and working memory/executive impairments associated with ADHD (Denckla, 1996; Doyle, 2006). Although learning and memory problems are reportedly common within ADHD groups, verbal memory problems are not well characterized (Denckla, 1996; Doyle, 2006). Previous research has not clarified whether attention and working memory/executive problems are directly related to the verbal memory impairments reported in A DHD populations. In addition, it is unclear whether verbal memory difficulties experienced by ADHD children are related uniquely to their ADHD diagnosis, or if the verbal memory difficulties are related to common co morbid diagnoses (i.e. disruptive behav ior, mood, anxiety, or learning disorders). Therefore, this study examined the verbal memory profile (i.e. encoding, retention, and retrieval abilities) as well as the role of cognitive contributors to verbal memory performance of children with ADHD. The ADHD verbal memory profile was compared to the verbal memory profile of a Clinical Control group, which included children with disruptive behavior, mood, anxiety, and learning disorders but not ADHD. In addition, we examined the memory of children with TB I because prior research has well described their impairments of attention, working memory/executive functioning, processing speed, and memory (Yeates, 2010); however, verbal memory problems are not well characterized. Similar to the limited information a bout pediatric ADHD populations, it is currently unclear at what stage of the memory process these memory problems are occurring in TBI. Although some limited research suggests that verbal memory impairments are related to injury severity, executive dysfunction, and

PAGE 44

44 poor processing speed, the relative contribution of these and other cognitive factors is unclear. Therefore, we investigated the verbal memory profile (i.e. encoding, retention, and retrieval abilities) as well as the role of contributing fac tors to verbal memory performance of children with TBI. The TBI verbal memory profile was compared to the verbal memory profile of an Orthopedic Injury Control group. In general, a better understanding of the verbal memory process in clinical pediatric populations will allow clinicians to make evidencebased decisions regarding clinical assessment and treatment. Knowledge about impairments within the verbal memory process will inform clinicians with evidencebased recommendations for how to evaluate “memory problems.” For example, this study aimed to provide information about which stages in the verbal memory process (i.e. encoding, retention, retrieval) are most likely to be impaired within a clinical pediatric population. In addition, knowledge about t he underlying cognitive components of the verbal memory process will indicate what other cognitive domains (e.g. verbal knowledge, attention, processing speed, working memory/executive functioning) are likely impacting verbal memory performance and should be additionally evaluated within clinical pediatric populations (Lezak, 2004). Finally, this proposed investigation will inform clinicians about what components of the verbal memory process should be targeted during cognitive rehabilitation efforts to imp rove memory in pediatric populations. For example, treatment recommendations may include a focus on improving a particular stage of the verbal memory process (i.e. encoding, retention, or retrieval) or an underlying cognitive component (e.g. attention, p rocessing speed, etc.; Wilson, 2009). Thus, the overarching goal of this study was to better understand verbal memory processes and underlying contributing cognitive

PAGE 45

45 components in order to provide evidencebased recommendations for assessment and treatmen t of pediatric verbal “memory problems.” Aim 1 The first aim of this study was to examine which cognitive factors contribute to a measure of delayed verbal memory performance (reliant on all 3 memory stages) in a diverse clinical group of children. Speci fic Aim 1a The first step was to use specific neuropsychological measures to construct a 4factor model representing cognitive domains. Hypothesis 1a. We predicted that neuropsychological variables collected from a diverse clinical group of children would load onto four discrete cognitive factors: verbal knowledge, attention, processing speed, and working memory/executive functioning. Specific factor loadings should be consistent with the test development of the neuropsychological measures. Specific Aim 1b Then, we determined which of these four constructed factors predicted performance on a measure of delayed verbal memory. Hypothesis 1b. We hypothesized that each of the cognitive factors (i.e. verbal knowledge, attention, processing speed, and working memory/executive functioning) would significantly predict delayed verbal memory performance. This hypothesis was based on previous literat ure that suggests these cognitive domains contribute to memory abilities in healthy, normally developing children. Due to the anticipated heterogeneous nature of our sample, we did not have any a priori predictions regarding the relative amount of varianc e explained by each significant predictor. Aim 2 The second aim of this study was to examine verbal memory abilities within a sample of children diagnosed with ADHD. More specifically, we sought to characterize the verbal memory performance profile and th e factors (cognitive and/or demographic) that contribute to performance across the three memory stages.

PAGE 46

46 Specific Aim 2a First, we aimed to characterize the verbal memory performance profile of children diagnosed with ADHD relative to a healthy standardization sample. The group mean performance values across the 3 verbal memory stages was determined and characterized in terms of level of impairment (defined as > 1 s.d. below the normative mean). We also examined whether differences in performance across the three verbal memory stages were statistically significant in an ADHD sample, supporting the notion of a meaningful “performance profile”. Hypothesis 2a: Assuming that pediatric ADHD patients have intact mesial temporal functioning, we expected to find deficits in the verbal encoding and retrieval stages, but not in the retention stage. It was further expected that within the ADHD group, performance on measures of verbal encoding and retrieval would be statistically significantly worse than performance on a measure of verbal retention. These predictions were based on the existing literature reporting that learning difficulties are frequently reported in children with ADHD. Specific Aim 2b In order to determine whether the pediatric ADHD performance profi le was unique to this pediatric disorder and not stemming from common comorbid conditions, it was compared to the verbal memory performance profile of a nonADHD pediatric clinical control group. Hypothesis 2b: We expected the ADHD group to perform statis tically significantly worse than the nonADHD clinical control group on measures of verbal encoding and retrieval. This hypothesis was based on previous literature that reports ADHD children demonstrate learning difficulties. In contrast, no significant group difference was expected on a measure of verbal retention. Specific Aim 2c Next, we used the model developed in Aim 1 to determine which cognitive and/or demographic factors contribute to the verbal memory performance profile of children diagnosed w ith ADHD.

PAGE 47

47 Hypothesis 2c: We predicted that within the pediatric ADHD group, attention and working memory/executive abilities would significantly predict verbal encoding and retrieval performance. We did not expect significant group differences on measures of verbal memory retention. These predictions were based on previous literature reporting that attention and working memory/executive functioning problems are very common in childhood ADHD. We did not expect to identify any other cognitive deficits th at could be related to reduced verbal memory abilities. Specific Aim 2d Finally, we aimed to determine if the cognitive and/or demographic factors identified in Aim 2c mediate (as opposed to directly affect) the verbal memory impairments uniquely associated with ADHD (as identified in Aim 2b). Hypothesis 2d: We predicted that the cognitive contributing factors (e.g. attention and working memory/executive functioning) would only partially mediate the effect of ADHD diagnosis on verbal memory performance (e. g. encoding and retrieval impairments). In addition, we expected the direct effects of diagnosis to significantly predict verbal memory performance even in the context of mediating cognitive abilities. Aim 3 The third aim of this study was to examine verbal memory abilities within a sample of children diagnosed with TBI. Similar to Aim 2, we sought to characterize the verbal memory performance profile and the factors (cognitive, demographic) that contribute to performance across the three memory stages. S pecific Aim 3a We aimed to characterize the verbal memory performance profile of children diagnosed with TBI relative to a healthy normative sample. More specifically, the group mean performance values across the 3 verbal memory stages was determined and characterized in terms of level of impairment (defined as > 1 s.d. below the normative mean). We also examined whether differences in performance across the three verbal memory stages are statistically significant in a TBI sample, supporting the notion of a meaningful “performance profile”.

PAGE 48

48 Hypothesis 3a: Assuming that at the group level pediatric TBI patients have intact mesial temporal functioning, we expected to find deficits in the verbal encoding and retrieval stages, but not in the retention stage. It was further expected that within the TBI group, performance on measures of verbal encoding and retrieval would be statistically significantly worse than performance on a measure of verbal retention. These predictions were developed based on the existing l iterature which suggests that TBI is related to encoding and retrieval, but not retention difficulties. Specific Aim 3b In order to determine whether the TBI performance profile was unique to this pediatric disorder and not stemming from non TBI injury fac tors, it was compared to the verbal memory performance profile of an orthopedic injury control group. Hypothesis 3b: We expected the TBI group to perform statistically significantly worse than the nonTBI clinical control group on measures of verbal encodi ng and retrieval. In contrast, no significant group difference was expected on a measure of verbal retention. This prediction was based on previous studies that report verbal memory problems are related to TBI, but not necessarily other bodily injuries. Specific Aim 3c Next, we used the model developed in Aim 1 to determine which cognitive and/or demographic factors contribute to the verbal memory performance profile of children diagnosed with TBI . Hypothesis 3c: We predicted that within the pediatric TB I group, attention, processing speed, and working memory/executive abilities would predict verbal encoding and retrieval performance. We did not have any a priori predictions regarding significant predictors of verbal memory retention. These predictions were based on the existing literature that reports attention, processing speed, and working memory difficulties are common problems after pediatric TBI. Specific Aim 3d Finally, we determined if the cognitive, motor, or demographic factors identified in Aim 3c mediate (as opposed to directly affect) the verbal memory impairments uniquely associated with TBI (as identified in Aim 3b). Hypothesis 3d: We predicted that the cognitive contributing factors (e.g. attention, processing speed, and working memory/ executive functioning) would only partially mediate the effect of TBI diagnosis on verbal memory performance (e.g. encoding and retrieval impairments). That is, the direct effects of diagnosis would significantly predict verbal memory performance even in t he context of mediating cognitive abilities.

PAGE 49

49 CHAPTER 3 METHODS The following sections describe the patient samples, general procedures, neuropsychological measurements, and statistical analyses that were used to examine the verbal memory processes and its contributing factors within a mixed, pediatric sample. Participants The data used for the current study was derived from archival clinic al and research neuropsychological evaluations occurring during a 10 year period (20012011). Participants included pediatric patients who underwent neuropsychological or psychological services through the Shands and University of Florida (UF) Psychology C linic or through research in Dr. Heaton’s Pediatric Neuropsychology Laboratory. Pediatric patients seen through the Psychology Clinic were typically referred by primary caregivers (i.e., parents), physicians, psychologists, or schools for diagnostic clari fication and treatment recommendations for conditions such as pervasive developmental disorders (PDD), ADHD, learning disabilities, communicative disorders, and other cognitive disorders resulting from neurological disease or insult. In addition, patients seen through the Psychology Clinic include pediatric patients with preestablished neurological history, such as TBI, stroke, epilepsy, and cancer. Data utilized from Dr. Heaton’s Pediatric Neuropsychology Laboratory was originally collected under IRB pr otocol 6492001, which aimed to examine the cognitive effects of TBI in comparison to orthopedic injury. To investigate all aims, we only used archival data from children who met the following criteria: (1) between the ages of 6years 0 months and 16years 11months,

PAGE 50

50 and (2) completion of a neuropsychological evaluation including the Children’s Memory Scale (CMS), which was the primary outcome measure of memory abilities in this study. For Aim 1, we did not exclude any patients based on cognitive or diagnostic factors because we wanted to examine the relationship between verbal memory and cognitive abilities across a wide range of neuropsychological functioning and patient populations. For Aim 2, we utilized a subset of the archival clinical data to examine the memory profiles of an ADHD groups and a Clinical Control group. We matched the diagnostic and control group to include patients with comorbid diagnoses of the ADHD and a similar distribution of demographic factors. Patients with any comorbid diag nosis of a medical disorder that impacts the central nervous system (e.g. pediatric patients with history of epilepsy, cancer, and traumatic brain injury) were excluded from Aim 2 analyses. For Aim 3, we utilized a subset of the archival clinical data as well as archival research data to examine the memory profiles of a TBI group and an Orthopedic Injury Control group. Aim 3 specifically required the use of research data because the archival clinic data did not identify any orthopedic injury control pati ents. In addition, the inclusion of research data allowed us to collect more data from TBI patients. More specifically, Aim 3 included a group of children who have sustained a moderatesevere TBI Injury severity in the TBI group was determined according to Glasgow Coma Scale (GCS) ratings at hospital admission (as described in Yeates, 2010). If GCS ratings were unavailable or inappropriate, then reports of loss of consciousness, concussion severity, or posttraumatic amnesia was used to determine injury severity (as described in Hannay et al., 2004). We attempted to match comorbid diagnoses and demographic

PAGE 51

51 factors of the TBI sample to the Orthopedic Injury Control group. In addition, neuropsychological data from TBI and orthopedic injury patients must have been collected within 6 months post injury. Procedure Archival data was collected from neuropsychological and psychological evaluations conducted during the past 10 years (i.e. 2001 – 2011) through various Child Psychology and Pediatric Neuropsychology services within the Shands and UF Psychology Clinic as well as a TBI research study conducted through Dr. Heaton’s Pediatric Neuropsychology Laboratory. Although the archival data from Dr. Heaton’s laboratory was originally collected under IRB approved research protocols, new IRB approval was obtained to extract archival data from clinical evaluations and to combine it with the research data. Data on each participant’s demographic information, medical and psychological history, and neuropsychological f unctioning was extracted from clinical and research charts. Demographic information was used to characterize participant groups as well as to determine norm referenced standardized scores for memory scores unique to this study. Measures of neuropsycholog ical functioning were used to measure encoding, retention, and retrieval of verbal memory as well as verbal knowledge, simple attention, processing speed, and working memory/executive functioning of pediatric patients. Data was analyzed using multivariate and univariate statistical methodology, including exploratory factor analysis, structural equation modeling, repeated measures analyses of variance, and multiple regression analyses. The following sections provide more detailed information about the meas ures and the analyses that were used in order to examine general verbal memory processes and contributing cognitive factors in pediatric clinical samples.

PAGE 52

52 Measurement of Verbal Memory The Children’s Memory Scale (CMS) This study used the two core verbal s ubtests from the Children’s Memory Scale (CMS; Cohen, 1997) to collect verbal learning and memory data. The CMS is a comprehensive measure designed to assess learning and memory processes in children aged 516 years old (Cohen, 1997). The CMS manual repo rts that its’ subtests demonstrate good content and construct validity for measuring verbal memory, as well as good sensitivity for detecting learning and memory dysfunction in pediatric neurodevelopmental and acquired brain injury populations (Cohen, 1997). Although data examining the encoding, retention, and retrieval stages of the memory process can be derived using the CMS, the CMS has not been previously used in this manner to characterize profiles of verbal memory impairments within pediatric sampl es. The core subtests of the CMS include two verbal measures of learning and memory: Stories and Word Pairs. A detailed description of CMS administration and raw data collection is provided in Appendix A . Raw data from the subtests were processed into standardized scores using normative data from the CMS manual and/or as provided for this study by the CMS author and publisher (See Appendix A for detailed methods of data transformation; Cohen, 1997). As previously mentioned, standardization of raw scores provides two advantages for this study: 1. agecorrected scores controls for normal demographic variation between patient and control groups, and 2. scores standardized on the same metric allows for within group comparisons across encoding, retention, and retrieval measures. Table 3 1 summarizes the CMS scores used in order to describe verbal memory profiles. Specifically, we generated four norm referenced scores: 1) delayed verbal

PAGE 53

53 memory abilities (Verb al Delayed Memory), 2) encoding (Verbal Immediate Memory), 3) retention (Verbal Percent Retention), and 4) retrieval (Verbal Recall/Recognition Contrast). These latter three memory scores were selected because of their theoretical correspondence to stages of the memory previous (as described in the Cohen, 1997). Notably, similar measures have been used in adult memory research to estimate encoding, retention, and retrieval abilities (e.g. Lezak, 2004). Measurement of Other Cognitive Domains & Motor Abili ties In addition to measures of verbal memory, this study used performancebased measures of verbal knowledge, focused and sustained attention, processing speed, and working memory / executive functioning. We used measures of these specific cognitive domains because previous literature has reported that development of good verbal knowledge, focused and sustained attention, processing speed, and working memory/executive functioning contribute to development of verbal memory abilities in healthy pediatric populations (Ornstein et al., 2006). The specific performancebased measures used to represent each cognitive domain (summarized below in Table 3 2) were selected based on the convenience of our sample and strong psychometric properties of these measures in normally developing pediatric populations. The following sections describe in detail the subtests from the Wecshler Intelligence Scale for Children – Fourth Edition / Wechsler Abbreviated Scale of Intelligence (WISC IV, Wecshler, 2003a; WASI, Wecshler, 19 99) and the Tests of Everyday Attention for Children (TEA Ch) that were used in this study. Measures of Verbal Knowledge WISC IV / WASI Verbal Comprehension subtests. Verbal knowledge was measured using the three verbal comprehension subtests from the W ISC IV and WASI

PAGE 54

54 (Wecshler, 2003a; Wecshler, 1999). Overall, the WISC/WASI measures demonstrate good reliability and validity; specific individual subtests exhibit good internal consistency, test retest reliability, and construct validity (Wecshler, 2003b) . The three WISC IV verbal comprehension subtests we used are the following: 1. Vocabulary, which examines knowledge of word definitions, 2. Similarities, which examines reasons why two words are similar, and 3. Comprehension, which examines knowledge abo ut social situations. All three subtests are administered orally to the subject; raw subtest scores are standardized into scaled scores using agebased normative information from the WISC IV manual. A composite score, called the Verbal Comprehension Index (VCI), can be additionally calculated in order to summarize patient performance on all three WISC IV verbal subtests. Pediatric patients who underwent neuropsychological evaluation in Dr. Heaton’s Pediatric Neuropsychology Laboratory completed the WASI rather than the WISC IV. The WASI and the WISC IV share two verbal knowledge subtests: Similarities and Vocabulary. Although the individual subtest items are different, the WASI Similarities and Vocabulary subtest are extremely similar to those of the W ISC IV in terms of administration and scoring criteria. In fact, previous literature reports that the difference between scores on the WISC IV verbal subtests and the WASI verbal subtests are negligible (Stancin & Alyward, 2008). The WASI does not have a Comprehension subtest; however, a composite score, called the Verbal Intelligence Quotient (VIQ), can be calculated in order to summarize patient performance on the two WASI verbal subtests. Correlations reported in the WISC IV technical manual indic ate that performance on the WASI VIQ and verbal subtests to the WISC IV VCI and verbal

PAGE 55

55 subtests are comparable Wecshler, D. (2003b). Therefore, this dissertation used scores from both the WISC IV and WASI verbal subtests to provide norm referenced measures of an individual’s verbal knowledge and verbal reasoning abilities. Measures of Focused & Sustained Attention TEA Ch: Sky Search, Score! The Test of Everyday Attention for Children (TEA Ch) was designed to provide a multi dimensional assessment of at tention for children (Manly et al., 1999). The TEA Ch demonstrates good test retest reliability and construct validity. It has been widely used in research of children with ADHD; factor analyses within normative samples indicate that the TEA Ch can be us ed to measure distinct domains of selective, sustained/divided, and controlled attention (Heaton 2001; Manly 2001). In order to measure focused (i.e. selective) and sustained attention, we used scores from the TEA Ch Sky Search and Score! subtests. The S ky Search subtest is a test of selective attention where the child must complete a speeded visual search and circle target shapes imbedded amongst distracters across a large page. The Sky Search score is based on both accuracy of responses and speed of the visual search. The Score! subtest is a test of auditory sustained attention where the child must accurately count a series of audiotaped tones presented over approximately 15 minutes. Raw scores from the Sky Search and Score! were transformed into stan dardized scaled scores based on gender and agebased normative information presented in the TEA Ch manual (Manly et al., 1999). Measures of Processing Speed WISC Processing Speed subtests. Processing speed was measured using two timed subtests from the WISC IV: Digit Symbol Coding and Symbol Search (Wecshler, 2003a). As previously described, the WISC IV subtests have good reliability and

PAGE 56

56 validity indices. The Digit Symbol Coding task requires the child to quickly decode a series of numbers into symbols using a key presented at the top of the page. The Symbol Search task requires the child to quickly examine a set of symbols and decide whether or not a symbol is repeated within each item’s set of symbols. Both processing speed subtests require speeded visual perceptual abilities as well as a motor response; however, the Symbol Search test is less reliant on visual constructional abilities compared to the Coding test. Raw scores from the Digit Symbol Coding and Symbol Search subtests were transformed into agereferenced scaled scores based on normative information provided in the WISC IV manual. Measures of Working Memory/Executive Functioning As described in previous literature, the cognitive domain of “executive functioning” includes working memory, inhibition, and attentional set shifting abilities (Hitch, 2006). In this study, we aimed to distinguish working memory/executive functioning from attention because working memory/executive functioning requires some mental manipulation of information bey ond basic selective and sustained attention. We used the TEA Ch Creature Counting and Digit Span Backwards tests in order to assess executive functioning. TEA Ch Creature Counting. The Creature Counting subtest from the TEA Ch measures children’s abilit y to control attention and accurately switchsets in an efficient manner (Manly et al., 1999). Administration requires the child to accurately count a series of visual stimuli. In addition, the child must be able to switch counting direction (i.e. count u p or down) when prompted. The TEA Ch Creature Counting score is based on both accuracy and speed of task completion. Raw scores from the TEA Ch Creature

PAGE 57

57 Counting subtest were transformed into gender and agebased standardized scaled scores according to the TEA Ch manual. WISC Digit Span Backwards. Working memory / executive functioning was additionally measured using the WISC IV Digit Span Backwards subtest, which is a reliable and valid test (Wecshler, 1991; Wecshler, 2003a). Completion of the Digi t Span Backwards subtest requires individuals to listen to orally presented series of numbers, mentally reorder the numbers, and respond out loud the series of numbers re arranged from lowest to highest. The raw Digit Span Backwards score is the number o f correct responses. Similar to our other measures, this raw score were transformed into an agebased standardized scaled score according to the WISC III/IV manual. Statistical Analyses In order to investigate the three specific aims of this study, we ran three separate sets of analyses. Visual inspection of histograms indicated that the CMS data distributions were adequately normal. Skewness and kurtosis statistics indicated that the distributions of some CMS derived scores were not normally distributed ; however logarithmic transformations, which are typically the most useful data transformations, did not significantly improve normality (Field, 2009). Given that parametric statistics are known to be robust, we conducted parametric analyses on the untransformed data (Schmider, Ziegler, Danay, Beyer, & Bhner, 2010). The main analyses of Aim 1 included an exploratory factor analysis and a structural equation model that estimated the relative contributions of demographic information and other cognitive contributors to delayed verbal memory performance in a mixed pediatric clinical sample. Within the context of the model developed in Aim 1, Aims 2 and 3 were investigated using parallel sets of statistical analyses including

PAGE 58

58 repeated measures analyses of variance (ANOVAs) and multiple regression analyses to examine the verbal memory profiles (i.e. encoding, retention, and retrieval) of: 1) children diagnosed with ADHD compared to clinical controls, and 2) children with history of moderatesevere TBI compar ed to orthopedic injury controls. The following section provides detailed descriptions of the statistical analyses that were used for the three study aims. Aim 1: Understanding Delayed Verbal Memory Performance in a Large, Mixed Clinical Pediatric Sampl e For Aim 1, we conducted two main statistical analyses: an Exploratory Factor Analysis (EFA) and Structural Equation Modeling (SEM). The EFA was used in order to construct a factor model where the measured indicators are neuropsychological measures and t he latent factors represent cognitive domains. Based on the clinical use of the neuropsychological measures, we planned for the EFA to identify a 4factor model. Since we anticipated data collection to include individual cases with some missing data, cas es with missing data were pairwise deleted from the analysis. The EFA was conducted in the most current version of the computer software Statistical Package for the Social Sciences (SPSS 18.0). Next, the cognitive factors identified in the EFA were entered into an SEM in order to investigate the relationship between measures of cognitive abilities and a measure of general verbal memory performance in a large, mixed pediatric sample. SEM is a multivariate statistical method that uses observed and latent variables (i.e. variables that are not directly measured) to examine a specified model (Kline, 2011). In this study, SEM was particularly useful because the analysis allowed for the use of latent factors as predictors.

PAGE 59

59 The initial specified model was developed based on the results of the EFA. Each latent factor and any measured variables that did not load onto a factor in the EFA were predictors in the SEM. The SEM examined the relationship of these predictors with a single observed dependent variable (i.e. the CMS Verbal Delayed Memory Index). Data as examined and transformed as necessary in order to meet multivariate (e.g. normal distribution, homogeneity of variance, sphericity, etc.) and SEM assumptions (i.e. large sample size) (Kline, 2011). All SEM analyses were conducted using the Analysis of Moment Structures (AMOS) software (Arbuckle, 2008). A special form of Maximum Likelihood (ML) estimation for incomplete data (i.e. Data Maximum Likelihood SEM also known as Full Information Maximum Likelihood, FIML) was used to obtain model fit as well as parameter estimations because our data set included instances of missing data (Kline, 2011). M odels were assessed using multiple assessments of fit (e.g.. Chi Square, Akaike Information Criterion (AIC), R oot Mean Square Error of Approximation (RMSEA), etc). More specifically, good fit was determined based on the following criteria: ChiSquare, p <.01; small AIC, RMSEA <.05, etc. (Kline, 2011). In addition, significant parameter estimations (i.e. Beta weig hts) were identified using p<.05 criteria. In addition to examining the fit of the SEM, we examined the correlation matrix of all cognitive and memory variables. The correlation matrix was a useful statistical method to examine relationships between a large number of variables (Kline, 2011). A statistical criterion of p tailed was used to confirm significant correlations between variables. Aim 2: Understanding the Verbal Memory Profile of an ADHD Sample Analysis 2a/b. In order to investigate the verbal memory profiles of our pediatric ADHD and clinicalcontrol group, we examined group means and standard deviations

PAGE 60

60 exhibited on the CMS measures listed in Table 3 1. In addition, a single 2 X 3 repeated measures ANOVA was conducted in order to examine betweenand within group differences on measures of verbal encoding, retention, and retrieval. We used a repeated measures ANOVA because within each individual participant’s scores, measures of verbal encoding, retention, and retrieval are related (Field, 2009). The within group independent variable was memory stage (i.e., encoding, retention, and retrieval), and the betweengroup independent variable was patient group (i.e. ADHD, clinical control). The dependent variable was memory performance as described in Table 3 1. All univariate statistics were conducted using the most current version of SPSS. A moderate statistical criterion (p<.05) was used to identify any significant between and within group effects, whic h were further examined with post hoc tests (e.g. Bonferroni corrected t tests). Effect sizes of the omnibus ANOVA (i.e. partial eta2 index) as well as any post hoc t tests (i.e. d index) were estimated and qualitatively described according to Cohen’s cri teria (Cohen, 1988 as reviewed in Field, 2009). Overall, F ratio statistics, post hoc t values, all probability values, and effect sizes were determined in order to describe significant betweengroup effects (i.e. effect of patient group) and withingroup effects (i.e. effects of memory stage) on verbal memory performance. Analysis 2c. Multiple regression analyses were used to determine which cognitive contributors identified in Aim 1’s model best predict verbal memory performance of children with ADHD. T hree separate stepwiseentry regressions were used to separately examine measures of verbal encoding, retention, and retrieval abilities. The three dependent variables that were examined are the following: Immediate Memory

PAGE 61

61 score (i.e. to estimate encoding abilities), Percent Retention (i.e.to estimate retention abilities), and the Recall/Recognition Contrast score (i.e. to estimate retrieval abilities). Based on the EFA in Aim 1, 5 independent variables were included as potential predictors in the model: one measure of verbal knowledge, one measure of attention, one measure of processing speed, one measure of working memory/executive functioning, and gender. The four cognitive measures were selected based on the results of the exploratory factory analysis from Aim 1. That is, the current regression model used the four measured cognitive variables that best represented four latent cognitive factors identified in Aim 1. For example, the WISC IV provides composite scores for the verbal knowledge (i.e. Verbal Comprehension Index) and processing speed (i.e. Processing Speed Index) subtests. These WISC IV composite scores are good representations of the cognitive factors of verbal knowledge and processing speed indentified in Aim 1’s EFA. Other cognitive domains were represented in the current regression model by the observed measure that loaded best with the latent factor in the EFA. For example, TEA Ch Creature Counting had the highest loading on the Attention factor, so TEA Ch Creature Counting was used in order to represent the cognitive domain of attention. Finally, the gender variable was dummy coded in the regression analysis (Field, 2009). All data was examined and transformed when necessary in order to meet univariate (e.g. normal distribution, homogeneity of variance, etc; Field, 2009) and multiple regression (e.g. multicollinearity, homoscedasticity) assumptions. Cases with missing data in the regression analysis were deleted with pairwise methodology.

PAGE 62

62 F ratio and R square statistics were reported in order to characterize the overall fit of the regression model. In addition, the effect sizes of the overall final model (i.e. f2 index) and included significant predictors (i.e. Beta weights) were reported and qualitatively described according to Cohe n’s criteria (Cohen, 1988 as reviewed in Field, 2009). A moderate statistical criterion (p<.05) were used to characterize the overall fit and to identify any significant independent variables of the final model. Analysis 2d. We planned to use the bootstrapping method to investigate whether significant predictors identified in the results of Aim 2b are mediating the relationship between ADHD status and verbal memory impairments identified in the results of Aim 2a( Hayes & Preac her, 2008). The bootstrapping method is reportedly the best method for examining mediators because it does not require data to be normally distributed and it reduces the possibility of Type 1 errors; in addition, it uses a resampling with replacement method which allows for the analysis of a data set that may include missing data points (Hayes, 2009). We planned to use the SPSS “INDIRECT” macro by Hayes & Preacher to conduct the bootstrapping method (2008). This macro analyzes the effect of a potential mediator (i.e. the “indirect effect”) and reports Pearson correlations as well as the Beta coefficients, t values and probability values for direct and “indirect effects.” For our analyses, we planned examine the direct effect of ADHD status as well as the indirect effect of any significant cognitive contributors identified in Aim 2b on verbal memory abilities (e.g. encoding, retrieval). A moderate statistical criterion (p<.05) was used to identify significant effects.

PAGE 63

63 Aim 3: Understanding the Verbal Memory Profile of a TBI Sample The analyses conducted to examine Aim 3 were the same set of analyses described in the Analyses of Aims 2a/b, 2c, and 2d; however, Aim 3 investigated the memory profile of pediatric TBI patients compared to a control group of orthopedic injury patients. Sample Size Considerations Aim 1: Understanding Delayed Verbal Memory Performance in a Large, Mixed Clinical Pediatric Sample SEM involving 12 independent variables was used to investigate Aim 1. A good guideline for sample size within structural equation modeling requires at least 20 cases per independent variable (Kline, 2011). Therefore, at least 240 cases are needed for adequate statistical power to run our Aim 1 SEM analysis. We assumed that 2 patients were evaluated through the UF & Shands Pediatric Neuropsychology services each week for 40 weeks per year, so we estimated that 800 pediatric patients were evaluated over the past 10 years through the UF & Shands Pediatric Neuropsychology services alone. We expected that about 70% of these patients (N=560) would meet inclusion criteria for our study. Thus, the archived data from the UF & Shands Pediatric Neuropsychology services alone was expected to exceed our power needs. This final sample was expected to include children with a variety of diagnoses, such as ADHD, TBI, PDDs, epilepsy, and childhood cancer. Aim 2: Understanding the Verbal Memory Profile of an ADHD Sample The Aim 2 analyses involved multiple regression analyses (Analysis 2c, d) and a 2 X3 repeated measures ANOVA (Analysis 2a/b). We conducted two separate power analyses to determine the sample size requirements necessary to detect statistically

PAGE 64

64 significant large effect sizes (as defined in Field, 1999) using 1. Regression analyses using 5 independent variables, and 2. A 2 X 3 repeated measures ANOVA. The first power analysis indicated that the regression analysis required a minimum sample size of 67 cases in order to detect a large effect size ( f2 = .35, alpha = .05, power = .95; Field, 1999). In additi on, the repeated measures ANOVA required a minimum sample size of 42 cases in order to detect a large effect size ( f = .40 , alpha = .05, power = .95; Field, 1999). Based on these results, Aim 2 analyses required a minimum of 67 children with ADHD and 42 clinical control patients to detect large effect sizes. As described previously, we expected that there would be a sufficient number of cases to meet our sample needs from the archival clinical data collection. Aim 3: Understanding the Verbal Memory Profi le of a TBI Sample Similar to Aim 2, the Aim 3 analyses involve multiple regression analyses (Analysis 3c, d) and a 2 X3 repeated measures ANOVA (Analysis 3a/b). We conducted two separate power analyses to determine the sample size requirements necessary to detect statistically significant lar ge effect sizes using: 1) Regression analyses using 5 independent variables, and 2) A 2 X 3 repeated measures ANOVA. In contrast to Aim 2, we conducted the Aim 3 power analyses using effect sizes identified in a preliminary study1. We found that the regression analysis required a minimum sample size of 43 cases in order to detect a large effect size ( f2 = .59, alpha = .05, power = .95). In addition, the repeated measures ANOVA required a minimum sample size of 25 cases 1 The prelimary study referred to here was conducted in order to 1) compare the verbal memory profiles (encoding, retention, and retrieval) of TBI patients and orthopedic injury control s, and 2) examine the relationship between other cognitive functions (i.e. verbal knowledge, attention, processing speed, and executive functioning) and verbal memory. Repeated measure ANOVAs revealed that TBI patients exhibited significantly worse verbal encoding ( p <.01, f = . 40 ), but had similar retention and retrieval compared to orthopedic patients. Multiple regression analyses indicated that across groups verbal knowledge and processing speed significantly predicted verbal encoding ( p <.01, f2 = .59 ).

PAGE 65

65 in order to detect a large effect size ( f = . 40, alpha = .05, power = .95). Based on these results, our Aim 3 analyses required a minimum of 43 children with TBI and 25 orthopedic injury control patients to detect large effect sizes. As described previously, we expected that there would be a sufficient number of TBI cases to meet our sample needs (i.e. approximately 160 cases would exceed Aim 1’s minimum sample size) from the UF & Shands Pediatric Neuropsychology services alone. In addition, we planned to acquire further TBI data and all of the orthopedic injury data from research studies conducted through Dr. Heaton’s Pediatric Neuropsychology Laboratory. Therefore, we expected that our Aim 3 sample needs of 43 pediatric moderatesevere TBI pati ents and 25 orthopedic injury control patients would be feasibly met. Table 3 1. Summary of CMS outcome v ariables Theoretical Memory Stage Measured CMS Score Stage 1. Encoding Verbal Immediate (Recall) Memory Stage 2. Retention Verbal Percent Retention Stage 3. Retrieval Verbal Recall/Recognition Contrast Delayed Verbal Memory: reliant on all 3 Stages Verbal Delayed (Recall) Memory Table 3 2. Measures of o ther cognitive d omains Observed Measure Verbal Knowledge Focused & Sustained Attention Processing Speed Working Memory / Executive Functioning WISC IV / WASI Vocabulary X WISC IV / WASI Similarities X WISC IV Comprehension X TEA Ch Sky Search X TEA Ch Score! X WISC IV Digit Symbol Coding X WISC IV Symbol Search X TEA Ch IV Creature Counting X WISC Digit Span Backwards X

PAGE 66

66 CHAPTER 4 RESULTS Aim 1 Results Our review of archival clinical data identified 571 pediatric patients who had undergone neuropsychological evaluations. Of these, 234 pediatric patients (139 males, 95 females) had a CMS Delayed Verbal Memory Index Score and were included Aim 1 analyses. This sample size was very similar to the predicted number of cases ( N =240) required for adequate power to r un the SEM. As shown in Table 4 1 , the average age of the sample was 10 years old, and age ranged from age 5 to 16 years old. The sample included 169 Caucasian (72.22 %), 34 Black (14.53%), 9 Hispanic (3.85%), 3 biracial (1.28%), and 1 Native Hawaiian / Other Pacific Islander (.43%) children. Additionally, 7.69 % of the sample were of unknown race (n= 18) because race was not reported in their archival records. As listed in Table 41 , the sample included individuals with a variety of ne urodevelopmental and acquired brain disorders (See Figure 4 1) . On average, parents of patients included in the heterogeneous sample reported that the patients display clinically significant difficulties with inattention and hyperactivity/impulsivity. In fact, 55.98% (n = 131) of the pediatric patients included in Aim 1 analyses had been previously diagnosed with Attention Deficit Hyperactivity Disorder (ADHD). Of note, this ADHD diagnosis was co morbid with a medical disorder (e.g. epilepsy, cancer, or TBI) in 20% of the ADHD patients (n=26). Exploratory Factor Analysis Detailed information about the sample’s performancebased cognitive abilities is presented in Table 43 . An exploratory factor analysis (EFA) was conducted on the 9 cognitive measures. The Kaiser Meyer Olkin measure verified the sampling adequacy

PAGE 67

67 for the analysis, KMO = .75 (“good” according to Field, 2009). A correlation matrix describing the correlations between the cognitive (and memory) measures is reported in Table 4 5 . As shown, many of the cognitive measures significantly correlated with each other. According to Field’s suggested criteria (2009), the correlations do not indicate any extreme multicollinearity or singularity (i.e. all correlations <.8). Bartlett’s test of sphericity indicated that correlations between the measures were sufficiently large for factor analysis, Chi2 (36) = 260.55, p<.01. Based on the a priori hypothesis that 4 factors would be identified, an initial analysis was run to obtain eigenvalues for a 4factor EFA. Missing data was excluded using pairwise methodology; and an oblique rotation (promax) was utilized to determine the factor matrices. Three components had eignenvalues over Kaiser’s criterion of 1, and in combination explained 62.68% of the variance. The fourth identified factor did not meet Kaiser’s criterion; however, the final sample included in the EFA (N=196) was somewhat small, and only one observed m easure (Digit Span Backwards) loaded onto this factor. In addition, the fourth factor meets Jolliffe’s criterion (eigenvalue over .7), so it was retained in the final analysis. Table s 4 6 and 47 below show the structure and pattern matrix of the factor loadings after oblique rotation. The measures that cluster on the same factors suggest that Factor 1 represents verbal knowledge, Factor 2 represents processing speed, Factor 3 represents attention, and Factor 4 represents working memory. Structural Equat ion Model A structural equation model (SEM) was employed in order to determine which of the factors identified in the above EFA predict performance on a measure of delayed verbal memory. The hypothesized model, which was developed based on the previous E FA, is illustrated in Figure 4 2 where circles represent latent variables (i.e. the

PAGE 68

68 cognitive factors), and rectangles represent measured variables. As shown in the model, all factors were hypothesized to covary with one another. SEM analyses were perf ormed using data from 234 pediatric patients. The data set included missing data; therefore, maximum likelihood estimation, which is robust and imputes missing data, was used to estimate all models. The goodness of fit test statistics, which are displayed below in Table 48 , overall indicated that model was satisfactory. For example, the traditional Chi square test statistic was significant and indicated good fit, Chi2(37) = 73.781, p <.01 Although the root mean square error of approximation (RMSEA) was not less than .05, the RMSEA was satisfactory (i.e. less than .08) according to Bollen and Long (1993) criteria. Due to missing data, Goodness of Fit Index (GFI) and Adjusted Goodness of Fit Index (AGFI) could not be estimated. The unstandardized and s tandardized parameter estimations are reported in Table 4 9 , and the standardized estimates are additionally illustrated in Figure 4 3 . Consistent with the EFA, the SEM revealed significant relationships between factors and their corresponding variables, p <.0 1. Within each factor, the standardized regression estimates were comparable. Finally, the SEM revealed that only the Verbal Knowledge factor predicted delayed verbal memory performance, Beta = .61 p <.01. Factors representing processing speed and attention as well as the measure of working memory did not explain a significant amount of variance within a measure of delayed verbal memory.

PAGE 69

69 Aim 2 Results Aim 2 Participant Sample Our review of archival data identified 99 ADHD patients who had undergone neuropsychological evaluations and did not have any comorbid medical disorder. In addition, 22 patients with clinical diagnoses of learning, mood, and/or other psychological disorder but without any ADHD diagnosis or medical disorder were identified. Of note, only the ADHD sample met the sample size requirements indicated by preliminary power analyses (i.e. ADHD sample included more than 67 patients; however the Clinical Control group included fewer than 4 2 patients). As shown in Table 4 10, 10 years old was the average age in both the ADHD and Clinical Control sample. Both groups included similar distributions of Caucasian and Black individuals; the ADHD group additionally included 1 biracial and Hispanic patient. Within the ADHD sample, 16 patients (16.2%) were diagnosed with ADHD, Predominantly Inattentive Type, 38 (38.4%) with ADHD, Combined Type, and 9 (38.4%) with ADHD, NOS (See Figure 43 ) . Specific subtype of ADHD was unknown for 36 ADHD patients (36.4%) because the subtype was not reported i n the archival records. Less than half of the ADHD group (38.4%) was taking ADHD medication at the time of their neuropsychological evaluations. The ADHD and Clinical Control groups both included individuals with diagnosis of learning disability, mood, o ther psychological disorders, and mental retardation (See Figure 4 4 ) . Group comparisons of parent reported measures indicated that parents report significantly more attention problems and less anxiety in the ADHD sample than the Clinical Control sample. See Table 4 11 for detailed information about the parent report measures.

PAGE 70

70 Memory Profile of ADHD Sample Group mean CMS performances of the ADHD and Clinical Controls groups are shown graphically in Figure 4 5 and listed in Tables 4 12 and 414 . Based on the group mean scores, the ADHD group did not demonstrate any clinically significant verbal memory impairments; the Clinical Control group exhibited a clinically significant impairment on a measure of verbal delayed recognition memory. Mauchly’s test ind icated that there were significant differences between the variances of differences, Chi2 (2) = 14.24, p<.01. Therefore, the condition of sphericity was not met, and GreenhouseGeisser corrections are described here. The 2X3 repeated measures ANOVA did n ot reveal any group differences or interactions; however, the ANOVA identified a significant, medium withingroup effect of verbal memory abilities, F(1.74, 154.86) = 6.72, p<.01, partial eta2 =.07,. Post hoc t tests indicated that within both groups pati ents scored significantly worse on a measure of verbal encoding compared to measures of verbal retention and retrieval, t(99) = 2.59, p<.05, d = .26 (medium effect) and t(110) = 3.82, p <.01, d = .6 (medium to large effect) respectively. There were no differences between verbal retention and retrieval abilities. Relationship between Memory Abilities and Cognitive Variables in ADHD Detailed information about the sample’s cognitive abilities is presented in Tables 4 13 and 415. Based on the finding that the ADHD patients and the Clinical Control patients did not perform significantly different from each other on any of the CMS measures, regression analyses investigating which cognitive factors significantly predicted verbal memory abilities were conducted with the combined ADHD + Clinical Control group mean score. As shown in Table 4 16, WISC IV Verbal Comprehension Index scores significantly predicted the measure of verbal memory encoding, Beta =

PAGE 71

71 .54, f2 = .43 (large effect size). No significant predict ors of verbal memory retention or retrieval were identified. Although the planned analyses included bootstrapping methodology to examine mediation models, mediation analyses (Specific Aim 2d) were not conducted because no group differences were identified on the CMS measures. Aim 3 Results Aim 3 Participant Sample Our review of archival data identified 16 moderate and severe TBI patients had undergone clinical neuropsychological evaluations and did not have any comorbid medical disorder. In addition, 28 TBI and 21 Orthopedic Injury only patients had participated in a previous IRB approved research protocol. Thus, Aim 3 analyzed data from a total of 44 TBI patients and 21 orthopedic injury patients. These sample sizes were very similar to the required s ample sizes that preliminary power analyses recommended. More specifically, the pediatric TBI sample was slightly larger than the recommended sample size of 43, and the Orthopedic Injury sample was slightly smaller than the recommended size of 25 patients . As shown in Table 4 17 , the average age of the TBI patients was 12 years old, and the average age of the Orthopedic Injury Control patients was 11 years old. Both groups included similar distributions of race / ethnicity, except that the TBI group included a greater proportion of Black/AfricanAmerican individuals than the Orthopedic Injury group. Within the TBI sample, 12 patients (27.27%) had history of moderate TBI, and 32 (73.73) had severe TBI (See Figure 4 6 ) . In addition, 11 TBI patients (25.00%) had clinical diagnoses of ADHD, whereas 3 patients (15.79%) in the Orthopedic Injury control group had ADHD. Both groups had similar distributions of learning and mood disorder (See Figure 4 7 ) . There were no TBI or Orthopedic Injury patients with

PAGE 72

72 dia gnosis of other psychiatric disorders or mental retardation. Group comparisons of parent reported measures indicated that parents reported significantly more externalizing behaviors, depressive symptoms, and atypical behaviors in the TBI sample than the O rthopedic Injury Control sample. See Table 4 18 for detailed information about the parent report measures. Memory Profile of Pediatric TBI Sample Group mean CMS performances of the TBI and Orthopedic Injury Controls groups are shown graphically in Figure 4 8 and listed in Tables 4 19 and 4 21. Based on the group mean scores, neither the TBI nor Orthopedic Injury Controls demonstrated any clinically significant verbal memory impairments. Mauchly’s test of sphericity indicated that the samples had good hom ogeneity of variance. The 2X3 repeated measures ANOVA revealed a significant and large withingroup effect of verbal memory abilities, F(2, 90) = 7.25, p<.01, partial eta2 =.14. Post hoc t tests indicated that across both groups, patients scored signific antly better on a measure of verbal memory retrieval compared to a measure of verbal encoding, t(54) = 3.81, p<.01, d = .56 (medium effect size). The measure of verbal retention did not significantly differ from the measures of verbal encoding or retriev al. In addition, the ANOVA reported that on average the TBI group scored lower than the Orthopedic Injury Controls across the verbal memory measures, F(1, 45) = 4.16, p<.05, partial eta2 =.08 (medium effect size). Additional t tests indicated that the TB I group scored specifically lower than the Orthopedic Injury controls on the measure of verbal encoding, t(54) = 3.15, p<.01 (large effect size). There was no significant Group X Verbal Memory Ability interaction.

PAGE 73

73 Relationship between Memory Abilities and Cognitive Variables in Pediatric TBI Detailed information about the sample’s cognitive abilities is presented in Table 4 20 and 4 22. A series of regression analyses indicated that neither gender nor any of the cognitive factors identified in Aim 1 predicted measures of verbal encoding, retention, or retrieval within a pediatric TBI sample. Based on these null findings, bootstrapping methodology was not warranted to examine whether any of the cognitive factors identified in Aim 1 mediate the TBI’s v erbal encoding and retrieval deficits.

PAGE 74

74 Table 4 1. Demographic i nformation of Aim 1’s heterogeneous c linical s ample ( N = 234) Variable N (Valid %) Min. Max. Mean Std Age (in years) 234 5.00 16.92 10.42 2.91 FSIQ 185 40.00 130.00 90.40 14.85 Gender F emale 95 (40.60) 40.60 Male 139 (59.40) 59.40 Handedness Ambidextrous 6 (2.56) Left 20 (8.55) Right 198 (84.62) Race/Ethnicity Biracial 3 (1.28) Black or African American 34 (14.53) Hispanic or Latino 9 (3.85) Native Hawaiian or Other Pacific Islander 1 (.43) Unknown 18 (7.69) White 169 (72.22) Diagnoses TBI 29 (12.39) Epilepsy 22 (9.40) Oncology 23 (9.83) Other Medical Disorder Affecting CNS 26 (11.11) Attention Deficit Hyperactivity Disorder 131 (55.98) Learning Disability 75 (32.05) Mood Disorder (e.g. MDD, GAD) 51 (21.79) Other Psych (e.g. ODD, CD) 48 (20.51) Mental Retardation 11 (4.70)

PAGE 75

75 Table 4 2 . Additional i nformation of Aim 1’s heterogeneous c linical s ample ( N = 234) Variable N (Valid %) Min. Max. Mean Std BASC 2 Scale Externalizing 213 32 99 56.73 13.87 Hyperactivity 213 31 99 59.78 14.51 Aggression 213 37 99 53.10 12.77 Conduct Problems 206 34 99 54.97 13.98 Internalizing Problems 212 31 99 57.34 15.20 Anxiety 213 29 96 54.53 13.07 Depression 212 34 99 57.41 15.21 Somatization 212 35 99 56.00 15.25 Behavioral Symptoms Index 197 33 99 60.57 13.88 Atypicality 213 41 99 60.82 15.69 Withdrawal 213 34 96 55.62 13.65 Attention Problems 213 36 85 62.26 10.51 Adaptive Skills 212 15 63 39.48 10.22 Adaptability 213 18 68 42.82 11.18 Social Skills 213 21 70 44.24 10.21 Leadership 206 19 68 41.54 9.70 Activities of Daily Living 212 14 68 37.95 10.71 Funct.Communication 211 10 66 38.76 11.60 Conners 3 Parent Scale Content: Inattention 154 39 100 73.03 * 14.91 Content: Hyperactivity/Impulsivity 155 39 100 68.25 * 18.02 Content: Learning Problems 155 40 100 73.40 * 16.11 Content: Aggression 155 41 100 62.32 18.67 Content: Peer Relations 155 42 100 64.10 19.26 DSM: Inattentive 153 40 100 71.22 * 14.92 DSM: Hyperactive/Impulsive 153 37 100 66.71 * 17.24 DSM: Conduct Disorder 153 43 100 58.94 16.96 DSM: Oppositional Defiant 153 40 100 62.60 16.65 1. Scores are T Scores ( mean = 50, standard deviation =10) * Asterisk indicates mean scores that meet the clinically significant cutoff acc ording to the Conners 3 manual. Table 4 3 . Cognitive p erformance based s cores of Aim 1’s h eterogeneous c linical s ample ( N = 234) Cognitive Domain Measure 1 N Min. Max. Mean Std N Impaired (Valid %) Verbal Knowledge WISC IV Vocabulary (ss) 195 1 16 8.47 3.07 53 (27.20) WISC IV Similarities (ss) 195 1 17 9.28 3.24 38 (19.49) WISC IV Comprehension (ss) 195 1 19 9.32 3.04 31 (15.90) Processing Speed WISC IV Symbol Search (ss) 200 1 16 8.13 3.14 57 (28.50) WISC IV Coding (ss) 201 1 19 7.32 3.32 83 (41.29) Attention TEA Ch Sky Search Attention Score (ss) 204 1 16 7.97 3.09 58 (28.40) TEA Ch Score! (ss) 209 1 15 7.35 3.60 99 (47.40) TEA Ch Creature Counting Total (ss) 179 1 15 8.13 3.57 66 (36.90) Working Memory WISC IV Digit Span Backward (ss) 136 1 14 7.73 3.01 46 (33.82) 1. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15

PAGE 76

76 Table 4 4. Verbal memory performance based sc ores of Aim 1’s h eterogeneous clinical s ample ( N = 234) Verbal Memory Domain CMS Measure 2 N Min. Max. Mean Std N Impaired (Valid %) Immediate Memory (encoding) Immediate Index (StS) 232 57 125 91.68 16.35 72 (31.03) Stories Immediate Recall (ss) 233 1 20 9.26 3.35 55 (23.61) Word Pairs Immediate Total (ss) 227 1 20 8.29 3.52 73 (32.16) Retention 1 Retention (z) 181 14.99 8.33 0.12 1.81 27 (14.92) Stories Retention (z) 176 29.44 8.33 0.05 2.80 35 (19.89) Word Pairs Retention (z) 170 4.85 8.27 0.22 1.84 34 (19.89) Delayed Recall Delayed Recall Score (StS) 233 50 134 92.58 17.42 75 (32.19) Stories Delayed Recall (ss) 232 1 20 9.04 3.35 51 (21.98) Word Pairs Delayed Recall (ss) 232 1 17 8.60 3.65 72 (31.03) Retrieval 1 Delayed Memory Contrast Score (z) 213 2.96 2.91 0.02 1.17 46 (21.60) Delayed Recognition Delayed Recognition Index Score (StS) 212 50 131 92.88 17.71 62 (29.25) Stores Delayed Recognition (ss) 228 1 20 8.96 3.66 56 (24.56) Word Pairs Delayed Recognition (ss) 223 1 20 8.72 4.00 68 (30.49) 1. Retention scores and the Verbal Recognition vs. Retrieval Contrast Score were derived from the normative sample's mean and standard deviations. See text for more information about the derivation of these scores. 2. (ss) = scaled scores with mean = 10 a nd sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15, and (z) = z score with mean = 0 and sd = 1.

PAGE 77

77 Table 4 5. Spearman r ho correlations between CMS delayed verbal memory index and other co gnitive measures in Aim 1’s heterogeneous clinical s ample. WISC IV/WASI Vocabulary (ss) WISC IV/WASI Similarities (ss) WISC IV Compreh. (ss) WISC IV Coding (ss) WISC IV Symbol Search (ss) TEA Ch Creature Counting Total (ss) TEA Ch Sky Search Attention Score (ss) TEA Ch Score! (ss) WISC IV Digit Span Backward (ss) Delayed Verbal Memory (CMS Verbal Delayed Index) Correl. Coeff. .50** .51** .52** .25** .29** .17* .17* .17* .24** Sig. (2 tailed) 0.00 0.00 0.00 0.00 0.00 0.02 0.02 0.01 0.00 N 194 194 194 200 199 179 204 209 135 WISC IV/WASI Vocabulary (ss) Correl. Coeff. .73** .65** 0.14 .23** .26** 0.14 .15* .23** Sig. (2 tailed) 0.00 0.00 0.05 0.00 0.00 0.06 0.04 0.01 N 195 195 195 194 159 178 182 135 WISC IV/WASI Similarities (ss) Correl. Coeff. .58** 0.11 .27** .19* .17* 0.13 0.16 Sig. (2 tailed) 0.00 0.11 0.00 0.02 0.02 0.09 0.07 N 195 195 194 159 178 182 135 WISC IV Comprehension (ss) Correl. Coeff. .22** .33** .18* 0.11 .22** .30** Sig. (2 tailed) 0.00 0.00 0.02 0.14 0.00 0.00 N 195 194 159 178 182 135 WISC IV Coding (ss) Correl. Coeff. .61** 0.02 .35** .20** 0.16 Sig. (2 tailed) 0.00 0.78 0.00 0.00 0.06 N 200 163 183 187 136 WISC IV Symbol Search (ss) Correl. Coeff. 0.07 .35** .24** .23** Sig. (2 tailed) 0.37 0.00 0.00 0.01 N 162 182 186 136 TEA Ch Creature Counting Total (ss) Correl. Coeff. .20** .17* 0.15 Sig. (2 tailed) 0.01 0.02 0.12 N 177 179 109 TEA Ch Sky Search Attention Score (ss) Correl. Coeff. .19** 0.10 Sig. (2 tailed) 0.01 0.29 N 124 TEACh Score! (ss) Correl. Coeff. 0.12 Sig. (2 tailed) 0.17 N 127 Correlation coefficients presented here are Spearman Rho correlations. *Asterisk indicates significant correlations, p <.05. **Double Asterisks indicate significant correlations, p<.01

PAGE 78

78 Table 4 6. Pattern m atrix of rotated factor l oadings from exploratory factor a nalysis Measure Verbal Knowledge Processing Speed Attention Working Memory WISC IV Similarities 0.93 0.13 WISC IV Vocabulary 0.92 WISC IV Comprehension 0.76 0.13 0.14 WISC IV Symbol Search 0.11 0.78 0.15 WISC IV Coding 0.88 TEA Ch Sky Search 0.43 0.59 0.14 TEA Ch Score! 0.22 0.55 TEA Ch Creature Counting 0.33 0.81 0.16 WISC IV Digit Span Backwards 0.10 0.97 Exploratory Factor Analysis Results: Extraction method was Principal Component Analysis. Rotation Method was Promax with Kaiser Normalization. Rotation converged in 6 iterations. Table 4 7 . Structured m atrix of r otated f actor loadings from e xploratory factor a nalysis Measure Verbal Knowledge Processing Speed Attention Working Memory WISC IV Similarities 0.89 0.16 0.25 0.16 WISC IV Vocabulary 0.91 0.16 0.29 0.27 WISC IV Comprehension 0.83 0.30 0.27 0.40 WISC IV Symbol Search 0.32 0.82 0.26 0.27 WISC IV Coding 0.14 0.86 0.15 TEA Ch Sky Search 0.12 0.53 0.63 TEA Ch Score! 0.21 0.34 0.59 TEA Ch Creature Counting 0.28 0.12 0.79 0.31 WISC IV Digit Span Backwards 0.28 0.20 0.21 0.97 Exploratory Factor Analysis Results: Extraction method was Principal Component Analysis. Rotation Method was Promax with Kaiser Normalization. Rotation converged in 6 iterations. Table 4 8. Goodness of f it s tatistics Chi 2 CMIN/DF NFI RFI RMSEA RMSEA sign. AIC Chi2 Differ. Independence (Null) Model 1 10.41 0.00 0.00 0.20 0.00 129.78 Specified Model 1 73.78 1.99 * 0.87 0.81 0.06 * 0.11 592.55 * A model with “good fit” is indicated by the following measures: CMIN/DF, NFI, and/or RFI is closer to 1.00; RMSEA is lower than .10 and nonsignificant; AIC is smaller.

PAGE 79

79 Table 4 9. Structural equation m odeling (SEM) r esults Dependent Variable Independ. Variable Unstandardized Regression Weights Stand. Regress. Weights Estimate S.E. C.R. P WISC IV Comprehension (ss) Verbal Knowledge 1.00 0.81 WISC IV Vocabulary (ss) Verbal Knowledge 0.92 0.02 48.93 *** 0.79 WISC IV Similarities (ss) Verbal Knowledge 1.00 0.02 51.20 *** 0.81 WISC IV Symbol Search (ss) Processing Speed 1.00 0.77 WISC IV Coding (ss) Processing Speed 0.91 0.02 37.45 *** 0.74 TEA Ch Creature Counting Total (ss) Attention 1.00 0.46 TEA Ch Score! (ss) Attention 0.92 0.04 24.72 *** 0.43 TEA Ch Sky Search Attention Score (ss) Attention 0.99 0.04 25.44 *** 0.46 CMS Delayed Recall Score (StS) Processing Speed 1.07 0.82 1.31 0.19 0.16 CMS Delayed Recall Score (StS) Attention 0.56 1.88 0.30 0.77 0.05 CMS Delayed Recall Score (StS) WISC IV Digit Span Backward (ss) 0.08 0.45 0.17 0.86 0.01 CMS Delayed Recall Score (StS) Verbal Knowledge 4.15 0.64 6.45 *** 0.61

PAGE 80

80 Table 4 10. Demographic information of Aim 2’s ADHD and c linical c ontrol s ample ADHD Group (N = 99) Clinical Control Group (N = 23) t test Variable N (Valid %) Min. Max. Mean Std N (Valid %) Min. Max. Mean Std Age (in years) 99 5.17 16.42 10.06 2.58 22 6.67 14.92 10.63 2.38 ns FSIQ 77 54.00 130.00 90.84 12.64 18 55.00 111.00 85.61 17.74 ns General Ability Index (GAI) 36 67.00 127.00 96.89 12.85 10 47.00 115.00 87.20 23.41 ns Gender Female 37 (37.37) 10 Male 62 (62.63) 13 Race/Ethnicity Biracial 1 (1.01) 0 Black or African American 21 (21.21) 5 (21.74) Hispanic or Latino 1 (1.01) 0 Unknown 9 (9.09) 3 (13.04) White 67 (67.68) 15 (65.22) Handedness Ambidextrous 4 (4.04) 0 Left 4 (4.04) 3 (13.04) Right 88 (88.89) 19 (82.61) Unknown 1 (1.01) 0 Diagnoses / ADHD Information ADHD Inattentive Type 16 (16.2) ADHD Hyperactive/Impulsive Type 0 ADHD Combined Type 38 (38.4) ADHD Not Otherwise Specified 9 (9.1) ADHD Type Not Reported 36 (36.4) Currently prescribed ADHD medication 38 (38.4) Learning Disability 35 (35.35) 15 (65.22) Mood Disorder (e.g. MDD, GAD) 21 (21.21) 10 (43.48) Other Psych (e.g. ODD, CD) 25 (25.25) 4 (17.39) Mental Retardation 1 (1.01) 1 (13.04)

PAGE 81

81 Table 4 11 . Additional c haracterization of Aim 2’s ADHD and c linical c ontr ol s ample ADHD Group (N = 99) Clinical Control Group (N = 23) t test Variable N Min. Max. Mean 1 Std N Min. Max. Mean 1 Std BASC 2 Scale Externalizing 93 37 99 59.18 13.63 22 32 98 56.09 15.10 ns Hyperactivity 93 36 95 62.38 13.99 22 31 99 58.91 15.22 ns Aggression 93 38 99 54.10 12.56 22 37 85 54.00 14.16 ns Conduct Problems 90 34 99 57.96 14.90 22 34 94 54.23 14.22 ns Internalizing Problems 93 31 92 55.56 14.07 22 35 99 62.45 18.24 ns Anxiety 93 29 86 52.92 12.37 22 38 90 59.50 13.63 p=.03 Depression 93 36 99 56.75 14.17 22 37 99 61.91 18.73 ns Somatization 93 35 87 53.47 14.42 22 37 99 59.32 18.66 ns Behavioral Symptoms Index 85 42 99 62.46 12.80 20 33 99 61.70 17.63 ns Atypicality 93 41 99 61.37 15.57 22 41 99 63.23 19.72 ns Withdrawal 93 35 92 55.73 13.10 22 36 96 57.82 17.47 ns Attention Problems 93 36 85 65.83 * 8.38 22 40 85 60.55 10.92 p=.01 Adaptive Skills 92 18 57 37.16 8.73 22 15 60 38.95 11.14 ns Adaptability 93 19 67 41.81 10.19 22 23 65 44.73 12.88 ns Social Skills 93 21 67 42.62 9.17 22 24 57 42.09 9.77 ns Leadership 90 20 61 39.50 8.80 22 21 60 38.68 8.76 ns Activities of Daily Living 93 16 54 35.75 8.47 22 16 59 38.86 11.99 ns Funct.Communication 92 10 61 36.08 10.96 22 10 66 37.64 14.76 ns Conners 3 Parent Scale Content: Inattention 83 43 100 76.24 * 13.36 18 44 94 64.78 15.41 p<.01 Content: Hyperactivity/Impulsivity 82 39 100 71.44 * 16.90 18 40 100 59.67 20.13 p=.01 Content: Learning Problems 83 40 100 76.31 * 15.51 18 47 99 73.78 * 16.21 ns Content: Aggression 83 41 100 63.96 19.32 18 41 100 57.50 18.52 ns Content: Peer Relations 83 42 100 64.69 19.01 18 42 100 63.83 21.35 ns DSM: Inattentive 81 48 100 75.25 * 12.82 18 41 96 63.33 17.22 p<.01 DSM: Hyperactive/Impulsive 81 38 100 69.60 * 16.29 18 40 98 59.83 19.03 p=.028 DSM: Conduct Disorder 81 43 100 61.81 17.89 18 43 100 55.11 18.04 ns DSM: Oppositional Defiant 81 40 100 63.20 16.28 18 40 100 59.78 17.34 ns 1. Scores are T Scores ( mean = 50, standard deviation =10) * Asterisk indicates mean scores that meet the clinically significant cutoff according to the BASC 2 and Conners 3 manual s.

PAGE 82

82 Table 4 12 . ADHD group m ean p erformance on v erbal m emory m easures (N = 99) Verbal Memory Domain Measure 2 N Min. Max. Mean Std N Impaired (Valid %) Immediate Memory (encoding) Immediate Index (StS) 97 60 125 92.47 15.97 27 (27.84) Stories Immediate Recall (ss) 99 1 20 9.16 3.48 27 (27.27) Word Pairs Immediate Total (ss) 96 1 20 8.78 3.23 22 (22.92) Retention 1 Retention (z) 83 14.99 4.37 0.12 2.04 13 (15.66) Stories Retention (z) 82 29.44 4.37 0.34 3.59 17 (20.73) Word Pairs Retention (z) 79 4.85 4.37 0.16 1.66 42 (56.00) Delayed Recall Delayed Recall Score (StS) 97 50 134 93.58 17.14 26 (26.80) Stories Delayed Recall (ss) 98 1 20 8.99 3.50 23 (23.47) Word Pairs Delayed Recall (ss) 97 1 15 8.93 3.48 26 (26.80) Retrieval 1 Delayed Memory Contrast Score (z) 91 2.96 2.91 0.14 1.28 23 (25.27) Delayed Recognition Delayed Recognition Index Score (StS) 91 50 131 94.18 17.25 27 (29.67) Stores Delayed Recognition (ss) 95 1 20 9.21 3.79 26 (27.37 Word Pairs Delayed Recognition (ss) 97 1 20 9.05 3.88 24 (24.74) 1. Retention scores and the Verbal Recognition vs. Retrieval Contrast Score were derived from the normative sample's mean and standard deviations. See text for more information about the derivation of these scores. 2. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15, and (z) = z score with mean = 0 and sd = 1. Table 4 13 . ADHD group m ean p erformance on cognitive m easures (N = 99) Cognitive Domain Measure 1 N Min. Max. Mean Std N Impaired (Valid %) Verbal Knowledge WISC IV Verbal Comprehension Index 78 63 132 93.62 14.45 18 (23.08) WISC IV Vocabulary (ss) 78 2 14 8.45 2.87 19 (24.36) WISC IV Similarities (ss) 78 2 17 9.26 3.06 15 (19.23) WISC IV Comprehension (ss) 78 1 17 9.12 2.92 12 (15.38) Processing Speed WISC IV Processing Speed Index 79 56 123 90.99 13.58 27 (34.18) WISC IV Symbol Search (ss) 79 1 16 8.95 2.74 15.2 (18.99) WISC IV Coding (ss) 79 2 16 7.92 2.98 27 (34.18) Attention TEA Ch Creature Counting Total (ss) 79 1 15 7.19 3.59 39 (49.37) TEA Ch Sky Search Attention Score (ss) 89 1 13 7.60 3.03 30 (33.71) TEA Ch Score! (ss) 91 1 15 7.30 3.65 45 (49.45) Working Memory WISC IV Digit Span Backward (ss) 62 1 14 7.68 2.98 22 (35.48) 1. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15 .

PAGE 83

83 Table 4 14 . Clinical c ontrol group m ean performance on verbal m emory m easures (N=23) Memory Domain Measure 2 N Min. Max. Mean Std N Impaired (Valid %) t test compr. to ADHD 23 12 118 86.39 23.10 7 (30.43) ns Stories Immediate Recall (ss) 23 2 20 9.26 3.89 5 (21.74) ns Word Pairs Immediate Total (ss) 22 3 16 7.95 3.27 9 (40.91) ns Retention 1 Retention (z) 19 1.60 6.60 0.42 1.76 10 (52.63) ns Stories Retention (z) 18 2.25 6.60 0.24 1.93 4 (22.22) ns Word Pairs Retention (z) 18 2.92 3.01 0.31 1.57 10 (55.56) ns Delayed Recall Delayed Recall Score (StS) 23 15 122 87.43 23.02 10 (43.48) ns Stories Delayed Recall (ss) 23 2 20 9.26 3.51 4 (17.39) ns Word Pairs Delayed Recall (ss) 23 1 15 8.35 3.96 10 (43.48) ns Retrieval 1 Delayed Memory Contrast Score (z) 21 1.65 2.08 0.20 1.01 4 (19.05) ns Delayed Recognition Delayed Recognition Index Score (StS) 21 18 118 83.76 * 23.19 8 (38.10) ns Stores Delayed Recognition (ss) 23 2 20 9.17 4.60 8 (34.78) ns Word Pairs Delayed Recognition (ss) 21 2.00 20.00 7.52 4.74 11 (52.38) ns 1. Retention scores and the Verbal Recognition vs. Retrieval Contrast Score were derived from the normative sample's mean and standard deviations. See text for more information about the derivation of these scores. 2. (ss) = scaled scores with mean = 10 a nd sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15, and (z) = z score with mean = 0 and sd = 1. Table 4 15 . Clinical c ontrol g roup mean p erformance on cognitive m easures (N=23) Cognitive Domain Measure 1 N Min. Max. Mean Std N Impaired (Valid %) t test compr. to ADHD Verbal Knowledge WISC IV Verbal Comprehension Index 19 63 124 91.89 16.47 6 (31.58) ns WISC IV Vocabulary (ss) 19 3 16 8.11 3.36 4 (21.05) ns WISC IV Similarities (ss) 19 4 13 8.42 2.95 4 (21.05) ns WISC IV Comprehension (ss) 19 3 16 9.37 3.08 3 (15.79) ns Processing Speed WISC IV Processing Speed Index 19 50 109 83.53 18.25 39 (47.37) ns WISC IV Symbol Search (ss) 19 1 12 6.79 * 3.90 8 (42.11) p=.03 WISC IV Coding (ss) 19 1 12 7.26 3.19 7 (36.84) ns Attention TEA Ch Creature Counting Total (ss) 18 2 15 8.22 3.54 6 (33.33) ns TEA Ch Sky Search Attention Score (ss) 20 1 12 7.75 2.40 5 (25.00) ns TEA Ch Score! (ss) 21 1 14 7.62 4.34 8 (38.10) ns Working Memory WISC IV Digit Span Backward (ss) 13 4 9 6.62 1.71 6 (46.15) ns 1. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15 .

PAGE 84

84 Table 4 1 6 . Results of regression a nalyses in Aim 2’s combined g roup of ADHD and c linical c ontrol p atients Dependent Variable: Verbal Memory Process (CMS Measure) Model Significant Predictor(s) R2 p Cognitive Domain (Measure) B SE B Beta Encoding (Immediate Memory Index) 0.30 <.00 (Constant) 30.80 12.28 Verbal Knowledge (WISC IV VCI) 0.65 0.13 0.54 Retention None identified Retrieval (Delayed Memory Contrast) None identified Table 4 1 6 describes the 3 separate regression analyses conducted to determine which cognitive factors predict verbal memory abilities within a mixed ADHD/Clinical Control group. The multiple regression models were hierarchical, and only significant predictors (p<. 05) were entered into the models in a stepwise fashion,

PAGE 85

85 Table 4 17 . Demographic i nformation on Aim 3’s TBI and orthopedic i njury c ontrol g roups TBI Group (N = 44) Orthop. Injury Control Group (N = 19) t test Variable N (Valid %) Min. Max. Mean Std N (Valid %) Min. Max. Mean Std Age (in years) 43 7 17 12.47 2.97 19 6.50 16.00 11.69 3.12 ns FSIQ 37 40 129 95.03 17.31 19 71.00 119.00 99.16 14.32 ns General Ability Index (GAI) 4 64 96 80.00 13.69 0 Gender Female 21 (47.73) 9 (47.37) Male 23 (52.27) 10 (52.63) Race/Ethnicity Biracial 1 (2.27) 0 Black or African American 8 (18.18) 1 (5.26) Hispanic or Latino 2 (4.55) 2 (10.53) Unknown 2 (4.55) 2 (10.53) White 31 (70.45) 14 (73.68) Handedness Left 3 (6.82) 0 Right 37 (84.09) 17 (89.47) Unknown 4 (9.09) 2 (10.53) Diagnoses Moderate TBI 12 (27.27) Severe TBI 32 (73.73) ADHD 11 (25.00) 3 (15.79) Learning Disability 4 (9.09) 1 (5.26) Mood Disorder (e.g. MDD, GAD) 5 (11.36) 1 (5.26) Other Psych (e.g. ODD, CD) 0 0 Mental Retardation 0 0

PAGE 86

86 Table 4 18 . Additional c haracterization of Aim 3’s TBI and orthopedic injury c ontrol g roups TBI Group (N = 44) Orthop. Injury Control Group (N = 19) t test Variable N Min. Max. Mean Std N Min. Max. Mean Std BASC/BASC 2 Scales Externalizing 32 35 89 54.78 14.28 15 34 74 44.47 9.71 p <.01 Hyperactivity 34 33 92 57.29 16.90 15 33 84 43.33 12.79 p <.01 Aggression 33 35 86 53.09 13.27 15 36 60 44.40 6.96 p <.01 Conduct Problems 34 37 88 51.94 12.27 15 37 74 48.07 9.45 ns Internalizing Problems 33 33 96 55.45 15.01 15 31 78 47.33 13.06 ns Anxiety 34 29 80 50.97 14.16 15 30 79 50.20 14.46 ns Depression 34 34 101 55.85 15.80 15 34 68 44.33 9.24 p <.01 Somatization 34 36 86 55.47 12.43 15 35 76 48.87 10.64 ns Behavioral Symptoms Index 30 30 81 55.17 13.74 15 31 77 47.13 11.76 ns Atypicality 34 36 99 57.35 15.60 15 38 67 47.93 8.34 p <.01 Withdrawal 34 37 76 51.35 9.68 15 37 79 51.00 11.54 ns Attention Problems 33 36 83 58.67 11.57 15 43 84 57.47 12.14 ns Adaptive Skills 32 25 64 42.00 10.31 15 22 64 44.93 10.08 ns Adaptability 21 25 68 45.95 11.33 9 22 58 46.00 10.39 ns Social Skills 33 29 68 44.88 10.95 15 25 71 45.67 11.28 ns Leadership 33 25 63 42.70 9.22 15 23 66 46.73 10.89 ns Activities of Daily Living 12 20 64 38.58 13.17 0 ns Funct.Communication 12 23 65 40.92 10.53 0 ns CPRS R/Conners 3 Parent Scale Content: Inattention 31 40 85 56.65 12.54 18 42 90 59.50 13.71 ns Content: Hyperactivity/Impulsivity 31 43 100 60.42 17.07 18 43 90 53.72 13.75 ns Content: Learning Problems 4 43 75 61.25 13.82 0 Content: Aggression 4 41 87 66.00 * 19.77 0 Content: Peer Relations 4 48 75 64.50 13.08 0 DSM: Inattentive 31 40 82 56.61 12.37 18 44 90 57.83 14.19 ns DSM: Hyperactive/Impulsive 31 43 100 60.68 16.58 18 43 90 53.17 13.17 ns DSM: Conduct Disorder 4 43 65 56.50 9.57 0 DSM: Oppositional Defiant 4 45 79 65.00 * 16.08 0 1. Scores are T Scores ( mean = 50, standard deviation =10).

PAGE 87

87 Table 4 19 . TBI group m ean p erformance on verbal memory m easures (N = 41) Memory Domain Measure N Min. Max. Mean Std N Impaired (Valid %) Immediate Memory (encoding) Immediate Index (StS) 38 51 125 87.71 17.17 14 (36.84) Stories Immediate Recall (ss) 41 3 18 8.41 3.63 12 (29.27) Word Pairs Immediate Total (ss) 41 1 15 7.24 3.84 19 (46.34) Retention 1 Retention (z) 34 2.20 5.58 0.22 1.53 7 (20.59) Stories Retention (z) 31 6.50 7.43 0.07 2.48 7 (22.58) Word Pairs Retention (z) 30 5.77 10.46 0.29 2.52 11 (36.67) Delayed Recall Delayed Recall Score (StS) 38 50 122 90.42 18.09 11 (28.95) Stories Delayed Recall (ss) 41 1 14 8.46 3.93 13 (31.71) Word Pairs Delayed Recall (ss) 41 1 15 7.95 3.39 13 (31.71) Retrieval 1 Delayed Memory Contrast Score (z) 37 2.63 5.45 0.48 1.96 9 (24.32) Delayed Recognition Delayed Recognition Index Score (StS) 37 50 131 95.43 20.16 8 (21.62) Stores Delayed Recognition (ss) 41 1 18 9.05 4.44 12 (29.27) Word Pairs Delayed Recognition (ss) 40 1 12 8.90 4.12 9 (22.50) 1. Retention scores and the Verbal Recognition vs. Retrieval Contrast Score were derived from the normative sample's mean and standard deviations. See text for more information about the derivation of these scores. 2. (ss) = scaled scores with mean = 10 a nd sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15, and (z) = z score with mean = 0 and sd = 1. Table 4 20 . TBI group m ean performance on c ognitive m easures (N = 41) Cognitive Domain Measure N Min. Max. Mean Std N Impaired (Valid %) Verbal Knowledge WISC IV VCI / WASI VIQ 40 31.00 142.00 95.00 21.11 10 (24.39) WISC IV/WASI Vocabulary (ss) 40 1.00 17.00 9.11 3.87 9 (22.50) WISC IV Similarities (ss) 40 1.00 17.50 9.77 3.39 8 (20.00) WISC IV Comprehension (ss) 22 6.00 13.00 9.77 2.27 4 (14.81) Processing Speed WISC IV Processing Speed Index 23 18.00 126.00 83.39 * 21.51 12 (42.86) WISC IV Symbol Search (ss) 22 1.00 12.00 7.59 3.55 8 (28.57) WISC IV Coding (ss) 22 1.00 14.00 7.23 3.19 10 (35.71) Attention TEA Ch Creature Counting Total (ss) 37 3.00 15.00 8.62 3.34 9 (24.32) TEA Ch Sky Search Attention Score (ss) 38 1.00 13.00 7.82 3.49 14 (36.84) TEA Ch Score! (ss) 38 1.00 13.00 6.55 * 3.32 23 (60.53) Working Memory WISC IV Digit Span Backward (ss) 6 5.00 14.00 9.83 3.60 2 (33.33) 1. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15. * Asterisk indicates mean scores that meet the clinically significant cutoff (i.e. 1 standard deviation below the normative mean).

PAGE 88

88 Table 4 21 . Orthopedic injury control group m ean p erformance on verbal memory m easures (N=19) Mem ory Domain Measure N Min. Max. Mean Std N Impaired (Valid %) t test compr to TBI Immediate Memory (encoding) Immediate Index (StS) 18 75 128 103.61 15.63 2 (11.11) p<.01 Stories Immediate Recall (ss) 18 5 15 10.44 2.67 1 (5.56) p = .03 Word Pairs Immediate Total (ss) 18 4 16 10.72 3.64 2 (11.11) p <.01 Retention 1 Retention (z) 18 2.19 2.46 0.02 1.18 3 (16.67) ns Stories Retention (z) 18 0.94 3.94 0.25 1.12 0.00 ns Word Pairs Retention (z) 18 4.23 5.65 0.29 2.09 3 (16.67) ns Delayed Recall Delayed Recall Score (StS) 18 82 134 104.33 12.66 1 (5.56) p<.01 Stories Delayed Recall (ss) 18 4 17 10.00 2.90 3 (16.67) ns Word Pairs Delayed Recall (ss) 18 4 16 11.44 3.38 1 (5.56) p <.01 Retrieval 1 Delayed Memory Contrast Score (z) 18 1.27 5.78 1.33 2.29 2 (11.11) ns Delayed Recognition Delayed Recognition Index Score (StS) 18 57 118 99.22 16.83 3 (16.67) ns Stores Delayed Recognition (ss) 18 4 16 9.94 3.14 3 (16.67) ns Word Pairs Delayed Recognition (ss) 18 2 12 9.78 3.18 3 (16.67) ns 1. Retention scores and the Verbal Recognition vs. Retrieval Contrast Score were derived from the normative sample's mean and standard deviations. See text for more information about the derivation of these scores. 2. (ss) = scaled scores with mean = 10 a nd sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15, and (z) = z score with mean = 0 and sd = 1. Table 4 22 . Orthopedic i njury control group m ean p erformance on c ognitive m easures (N=19) Cognitive Domain Measure N Min. Max. Mean Std N Impaired (Valid %) t test compr to TBI Verbal Knowledge WISC IV VCI / WASI VIQ 19 76.00 125.00 99.37 15.84 5 (26.32 ns WISC IV/WASI Vocabulary (ss) 19 4.00 15.40 9.54 3.88 6 (31.58) ns WISC IV Similarities (ss) 19 5.20 13.90 10.21 2.74 3 (15.79) ns WISC IV Comprehension (ss) 0 NA NA Processing Speed WISC IV Processing Speed Index 1 106.00 106.00 106.00 0 ns WISC IV Symbol Search (ss) 1 10.00 10.00 10.00 0 ns WISC IV Coding (ss) 1 12.00 12.00 12.00 0 ns Attention TEA Ch Creature Counting Total (ss) 17 1.00 14.00 8.82 3.24 2 (11.76) ns TEA Ch Sky Search Attention Score (ss) 17 2.00 12.00 8.76 2.63 3 (17.65) ns TEA Ch Score! (ss) 17 3.00 15.00 9.94 3.86 4 (23.53) p<.01 Working Memory WISC IV Digit Span Backward (ss) 0 NA NA 1. (ss) = scaled scores with mean = 10 and sd = 3, (StS) = Standard Scores with mean = 100 and sd = 15

PAGE 89

89 Figure 41. Disorders r epresented in Aim 1

PAGE 90

90 Figure 4 2. Hypothesized s tructural equation m odel. This depicts the hypothesized Structural Equation Modeling examined in Aim 1. C ircles represent latent variables (i.e. the cognitive factors), and rectangles represent measured variables. Doubleheaded arrows indicate covariance, whereas singleheaded arrows represent one way relationships.

PAGE 91

91 Figure 4 3 . Final s tructural e quation m odel. Figure 43 shows the final SEM with the standardized regresstion estimates added. C ircles represent latent variables (i.e. the cognitive factors), and rectangles represent measured variables. Doubleheaded arrows indicate covariance, whereas singleheaded arrows represent oneway relationships.

PAGE 92

92 Figure 44 . Distribution of ADHD type in Aim 2’s ADHD s ample Figure 44. Disorders represented in Aim 2 s amples

PAGE 93

93 Figure 45 . Verbal memory performance profile of the ADHD and c linical control g roups. Figure 4 5 depicts the group mean verbal subtest scores of the pediatric ADHD and Clinical Control samples. Error bars indicate standard error. The dashed lines provide reference points to compare the group means to normative data. F igure 46. Distribution of TBI s everity in Aim 3’s TBI sample

PAGE 94

94 Figure 47. Disorders r epresented in Aim 3’s s amples Figure 4 8 . Verbal memory performance profiles of Aim 3’s TBI and orthopedic injury control g roups. Figure 4 8 depicts the group mean verbal subtest scores of the pediatric TBI and Orthopedic Injury Control samples. Error bars indicate standard error. The dashed lines provide reference points to compare the group means to normative data.

PAGE 95

95 CHAPTER 5 DISCUSSION The overarching goal of this dissertation was to investigate which cognitive factors contribute to verbal memory performance in clinical pediatric populations. We hypothesized that several different cognitive domains (i.e. verbal knowledge, processing speed, attention, and working memory/executive functioning) were related to verbal memory performance. In addition, we predicted that the relationships between verbal memory and other cognitive domains may vary in specific subsets of pediatric patients (i.e. ADHD and TBI). In general, the results of this study partially supported these hypotheses. The following sections further provide a detailed review of this study’s results in context of our hypotheses, previous literature and additional post hoc analyses. Implications of the study findings are discussed. In addition, considerations about this study’s methods (e.g. limitations of archi ved clinical data collection, the utility of derived memory measures) are reviewed. Cognitive Predictors of Verbal Memory in a Mixed Clinical Sample Distinct Cognitive Domains Identified in Clinical Pediatric Populations The first aim of this dissertation was to determine which cognitive factors contribute to a measure of delayed verbal memory performance in a heterogeneous pediatric sample. In order to examine this aim, we first developed a model that used specific neuropsychological measures to represent distinct cognitive domains. O ur hypothesis that neuropsychological variables collected from a diverse clinical group of children would load onto four discrete cognitive factors was confirmed. The neuropsychological meas ures examined in this study loaded onto four distinct latent factors, which represented the predicted cognitive domains.

PAGE 96

96 The 4 factor model was very consistent with the test development of the WISC IV . For example, the factor loadings of the WISC IV ver bal comprehension (and processing speed) subtests identified in this dissertation were similar or higher to those reported in the healthy standardization sample by the WISC IV Publishers (Wecshler 2003b; Williams, Weiss, & Rolfhus, 2003). The 4 factor mo del was additionally consistent with the test development of the TEA Ch because all of the TEA Ch subtests loaded well together on an Attention factor (Manly et al., 1999) . However, t his finding was somewhat unexpected because the TEA Ch Creature Counting subtest was expected to load with the WISC IV Digit Span Backwards subtest. To elaborate, the TEA Ch publishers reported that the TEA Ch Creature Counting subtest is a good measure of attentional control/switching (Manly et al. , 2001). The cognitive construct of attentional control/switching is conceptually very similar to working memory/executive functioning. Thus, we expected that the TEA Ch Creature Counting subtest score to load with the WISC IV Digit Span Backward subtest, which is reported to be a good measure of working memory (Wecshler, 2003b). The discrepancy between our predictions and the EFA results may be explained by the nature of the TEA Ch Creature Counting score used in this study. More specifically, the TEA Ch Creature Counting score u sed in the current study is based on both accuracy and speed of task completion . Previous literature has reported that this “Timing” score does not correlate well with other traditional measures of working memory/executive functioning (e.g. Stroop, Trails B ; Manly et al., 2001). However, the “Accuracy” score of the TEA Ch Creature Counting subtest, which is not based on any timing requirements, does significantly correlate with other traditional measures of working memory/executive

PAGE 97

97 functioning. Thus, if this study had used the TEA Ch Creature Counting Accuracy score (rather than the timing based score), then the TEA Ch Creature Counting score may have loaded better onto the Working Memory factor (rather than the Attention factor ). In summary , we found that existing neuropsychological tools (i.e. the WISC IV and TEA Ch subtests) can be used to measure distinct cognitive domains (i.e. verbal knowledge, processing speed, attention, and working memory) in a heterogeneous clinical pediatric sample. Given t h is finding, the WISC IV and TEACh measures used in this study should be interpreted to represent their corresponding cognitive domains for the remaining aims of this dissertation. Verbal Knowledge Predicted Delayed Verbal Memory Performance This study fur ther determined which cognitive domains (i.e. verbal knowledge, processing speed, attention, and/or working memory) predicted performance on a measure of delayed verbal memory in a mixed group of pediatric patients . Our hypothesis that each cognitive domain would significantly predict delayed verbal memory performance was only partially supported. We identified a significant relationship between verbal knowledge and delayed verbal memory performance; however, no other cognitive domains significantly predi cted verbal memory performance. The finding that verbal knowledge is significantly related to delayed verbal memory has been previously reported in healthy, normally developing individuals ( Baddeley, 2000; Brown & Craik, 2000; Hitch, 2006; Ornstein et al., 2006). In fact, the CMS manual reported that in the CMS normative data, which was collected from a large group of healthy children, there was a significant correlation between a measure of verbal knowledge (WISC III VCI) and the CMS Verbal Immediate, Del ayed, and

PAGE 98

98 Delayed Recognition Indices (Cohen, 1997) . This is the first known study to report that verbal knowledge significantly predicts verbal memory performance in a large, heterogeneous sample of pediatric patients. In order to better understand the relationship between verbal knowledge and delayed verbal memory in our clinical sample, post hoc analyses were conducted. First, post hoc correlation analyses examined the specific relationships between individual measures of verbal knowledge (i.e. WISC IV Vocabulary, WISC IV Comprehension, WISC IV Similarities) and measures of verbal memory processes (encoding, retention, and retrieval) as well as overall delayed verbal memory performance. As shown in Table 5 1 below, all three measures of verbal knowledge significantly correlated with the measure of verbal memory encoding abilities. One verbal knowledge measure (WISC IV Similarities) significantly correlated with the measure of verbal retrieval. In addition, the measures of verbal encoding and retriev al significantly correlated with the measure of delayed verbal memory; the retention measure did not significantly correlate with the delayed verbal measure. Taken together, these correlations suggested that the indirect effect(s) of verbal knowledge on v erbal encoding and/or retrieval may be mediating the relationship between verbal knowledge and delayed verbal memory performance. Bootstrapping methodology was conducted to investigate the possibility of mediation. As summarized in Table 5 2 , the analyses revealed that the significant relationship between verbal knowledge and delayed verbal memory performance was significantly and partially mediated by the effect of verbal knowledge on verbal encoding abilities. The effect of verbal knowledge on verbal r etrieval abilities was not a significant mediator.

PAGE 99

99 Previous researchers have proposed that the causeand effect relationship between verbal knowledge and verbal memory is bi directional, such that memory influences knowledge and at the sametime knowledge impacts memory (Baddeley, 2000). However, most research supports the notion that increased knowledge directly improves verbal memory abilities (Baddeley, 2000; Brown & Craik, 2000; Hitch, 2006; Ornstein et al., 2006). In fact, literature on healthy, normally developing children reports that there the significantly correlation between verbal knowledge and memory strengthens increases with age (Dehn, 2010; Hitch, 2006; Kail 1984; Ornstein et al., 2006). Moreover, it has been hypothesized that the mechanism of impact appears to be through verbal knowledge’s influence on verbal encoding processes. Dehn (2010) explained that in theory verbal knowledge provides structure for appropriate understanding, rich elaboration, and associations between old and newly learned verbal information. Consistent with the existing literature, the significant post hoc mediation results of this dissertation provide empirical evidence that verbal knowledge influences verbal mem ory performance through its influence on the verbal encoding stage. In summary, this dissertation identified a significant, direct effect of verbal knowledge on verbal memory performance, which appears partially mediated by the indirect effect of verbal knowledge on verbal encoding abilities. Of note, the simple structure of the EFA/SEM analyses suggested that our measure of verbal knowledge (i.e. the latent factor representing of verbal knowledge) was specifically estimating the pediatric sample’s ver bal abilities. That is, the verbal knowledge factor was distinctly identified from other cognitive domains. It is possible that our verbal knowledge factor is additionally estimating general mental functioning (i.e. g ). As proposed historically in

PAGE 100

100 psycho logy, the concept of g is defined as the highest order factor of intelligence which is related to multiple specific cognitive factors, such as verbal knowledge, processing speed, learning/memory (Sattler, 2008). Since this dissertation did not examine any specific measurement of g, the relationship between our measure of verbal knowledge and g is unclear. Future studies should examine the relationship between verbal knowledge and g in clinical pediatric populations because previous literature has indicate d that general intelligence is related to verbal memory (Cohen, 1997 ; Schneider & Pressley, 1989) . No Significant Contribution from Other Cognitive Domains to Verbal Memory Previous literature has reported that processing speed, attention, and working me mory/executive functioning abilities influence verbal memory in healthy children and adults (Cohen 1997; De Alwis et al., 2009; Dehn, 2010; Diamond, 2006; Moscovitch, 1994; Ornsetin et al., 2005; Schneider & Pressley, 1989; Wilson, 2009). In fact, the CMS manual noted that within the CMS normative data, Verbal Immediate, Delayed, and Delayed Recognition Indices significantly correlated with measures of processing speed (WISC III PSI) and executive functioning (Wisconsin Card Sorting Test; Cohen, 1997). How ever, the SEM results of this dissertation did not identify a significant relationship between these cognitive domains and verbal memory performance in a clinical pediatric sample. Based on the SEM results alone, it is unclear why processing speed, attent ion, and working memory did not additionally explain significant variance on a measure of delayed verbal memory in our heterogeneous sample. However, several possible explanations exist. The first possibility is that the latent factor representing verbal knowledge was problematically collinear with our estimates of processing speed, attention, and working

PAGE 101

101 memory; if these cognitive domains/variables were multicollinear, then it would be difficult to assess the individual importance of each cognitive vari able. For example, if the variables were multicollinear, then once the variance in verbal memory is accounted for by verbal knowledge, the other cognitive predictors account for very little variance. In contrast, if the variables were completely uncorrelated, then each cognitive factor (verbal knowledge, attention, processing speed, working memory) would account for different variance in verbal memory. The correlation matrix of all of the cognitive measures us ed in this dissertation (Table 45 ) indicates that the majority of the measures were correlated with each other; however, all of the correlations were less than .80 which provides evidence that the measures were not necessarily multicollinear (Field, 2009). Furthermore, as shown in Table 5 3 , the var iance inflation factor (VIF) and tolerance statistics further indicate that the measures are not problematically collinear (i.e. all VIF statistics are less than 10 and tolerance statistics are greater than .20; Field 2009). Finally, the design of the SEM analysis accounted for collinearity between predicts because all potential predictors of verbal memory were covaried with each other. Thus, the SEM estimated only the unique variance in verbal memory accounted for by each possible predictor (i.e. each cognitive factor/variable), and the null finding that processing speed, attention, and working memory does not significantly predict verbal memory of the heterogeneous pediatric sample is likely valid. Given that the results of the SEM are likely valid, an alternative explanation for the lack of relationship between processing speed, attention, and working memory and verbal memory may be related to the nature of our clinical pediatric sample. To elaborate, the hypothesized relationship between these cogniti ve domains and verbal

PAGE 102

102 memory was based on the existing literature on healthy, normal children. The literature in healthy, normal children reports that processing speed, attention, and working memory all contribute to the development of good memory abiliti es (Brooks, 2009; Ornstein et al. 2006) . The literature on this topic in clinical pediatric populations has been limited . For example, previous research studies have described the relationship between verbal memory performance and individual cognitive domains (e.g. relationship between processing speed and memory has been suggested in TBI studies; Mottram & Donders, 2006); however, this is the first known study to examine the simultaneous relationships between multiple cognitive domains and verbal memory of clinical pediatric populations. In other words , this is the first known study to examine the relative amount of variance in memory explained by separate cognitive domains . As described earlier, the results of our SEM revealed the unique amount of vari ance explained by each cognitive factor. Thus, when verbal knowledge is included in the model, the remaining cognitive factors did not explain a significant amount of variance in memory.1 This is the first known study to report this finding, and future s tudies should also aim to examine the relative contribution of multiple cognitive domains to verbal memory in clinical pediatric populations. 1 Interestingly, an exploratory post hoc analysis was conducted to determine what cognitive factors would explain a significant amount of variance in memory if verbal knowledge was not included in the model of verbal memory performance in the large, clini cal pediatric population. The post hoc multiple regression analysis revealed that when verbal knowledge is not included in the model, a measure of processing speed (WISC IV Symbol Search) becomes a significant predictor of verbal memory, Regression with out any verbal knowledge subtests R2 = .09, F(1, 107) = 10.04, p<.01, B = 1.62, SE (B) = .51, Beta = .29. Although this post hoc analysis is interesting; it should be interpreted with caution due to the primary finding that verbal knowledge does account f or a significant amount of variance in verbal memory, and there is no statistical suggestion to remove verbal knowledge from the model.

PAGE 103

103 Alternatively, our null finding that processing speed, attention, and/or working memory was not significantly related to verbal memory performance of a mixed clinical sample may be explained by the heterogeneous nature of the large sample. The sample included pediatric patients with a variety of neurodevelopmental and acquired brain injury. Based on the sample’s var iety of medical and psychological diagnoses as well as its range of cognitive abilities, the sample likely includes individuals with a range of neural injuries and neuropsychological development. It is possible that cognitive domains of processing speed, attention, and/or working memory only impact verbal memory within particular clinical pediatric populations, and the heterogeneous nature of our sample reduced the ability of our statistical analyses to detect significant relationships. Aims 2 & 3 of this dissertation examined this possibility by further investigating the relationships between cognitive domains and verbal memory performance in specific pediatric neurodevelopmental and acquired brain injury populations. Memory Profile and Contributing Factors in Pediatric ADHD The second aim of this study examined the verbal memory abilities of a specific subsample of children from Aim 1 who were diagnosed with ADHD and did not have any medical history affecting the central nervous system. Relative Weakness in Verbal Encoding Associated with Pediatric ADHD First, we characterized the verbal memory performance profile (i.e. encoding, retention, and retrieval abilities) of children diagnosed with ADHD relative to a healthy standardization sample. Based on previous reports and the assumption that the sample of pediatric ADHD patients had intact mesial temporal functioning, we hypothesized that the ADHD sample would demonstrate deficits in verbal encoding and retrieval, but not in

PAGE 104

104 retention abilities. This hypothesis was only partially supported by our data because the pediatric ADHD sample demonstrated a relative verbal encoding weakness , but no retention or retrieval problems. Although the lower encoding scores were not clinically significant (i.e. mean encoding performance was within 1 standard deviation from the normative mean), the overall memory profile of the ADHD sample suggested that verbal memory difficulties associated with ADHD are likely driven by initial encoding problems; i.e. verbal information that is learned during encoding is retained over time and later retrieved relatively well. This study finding is consistent with previous literature that has reported children with ADHD experience memory encoding problems ( Denckla, 1996). In fact, a re search st udy reported in the CMS manual found that an ADHD sample of children (n = 88) scored significantly lower than an age and gender matched control group on the CMS Verbal Immediate Memory index, but not on any other verbal memory measures (Cohen, 1997). The study reported in the CMS manual was conducted with a group of children with ADHD who did not additionally exhibit co occurring psychopathology such as depression, anxiety, conduct disorder or learning disorder (Cohen, 1997). Interestingly, this c urrent dissertation study identified a verbal memory encoding weakness in a pediatric ADHD sample that did include individuals with comorbid learning and mood disorders. Learning and mood disorders are commonly diagnosed in children with ADHD (APA, 2000) ; therefore, the pediatric ADHD sample studied in this dissertation is likely representative of the pediatric ADHD population. We hypothesized that the verbal memory profile of our ADHD sample was unique to childhood ADHD and the encoding weakness did not stem from common comorbid conditions. In order to examine our

PAGE 105

105 hypothesis, the verbal memory performance profile of our ADHD sample was compared to the verbal memory performance profile of a nonADHD pediatric Clinical Control group. The clinical control group was well matched to the ADHD group and, as expected, only differed on anxiety and ADHD symptoms. However, in contrast to our hypothesis , the ADHD sample did not significantly differ from the Clinical Control group on any measure of verbal memory. In other words, the identified verbal encoding problem was not unique to the ADHD sample. There are several possible reasons why our study may not have identified significant group differences. First, the size of the Clinical Control group did not meet the sample size requirements that preliminary power analyses recommended. Specifically the Clinical Control group was about half the size of the sample needed to conduct adequately powered betweengroup ANOVAs. The small sample size of our Clinical Cont rol group may have limited our analyses ’ ability to detect a group difference between the verbal memory performance of the ADHD and Clinical Control groups. Second, although we did not predict that the ADHD and Clinical Control group would demonstrate simi lar encoding weaknesses on performancebased measures of verbal memory, this finding may be explained by the fact that over 38% of the ADHD patients completed the memory tests while on ADHD medication. It is possible that the medicated ADHD patients were able to perform relatively well in comparison to healthy individuals and the Clinical Control group because their ADHD symptoms (i.e. inattentiveness, distractibility, etc.) were controlled on medication; in addition, the ADHD patients who were not taking any ADHD medication may not have been prescribed

PAGE 106

106 ADHD medication because their symptoms were not significantly impacting their functioning. A third explanation for the similarities between the ADHD and Clinical Control groups’ memory performances may be that the two groups truly experienced similar learning/memory problems. In fact, on average parents of both the ADHD and Clinical Control group members reported that the patients displayed clinically significant learning problems in the home setting. Moreover, the Clinical Control group and the ADHD sample were purposely matched to both include individuals with mood and learning disorders. T he groups’ common encoding weakness may be explained by the finding that the ADHD patients had well controlled ADHD s ymptoms at the time of memory testing, and further related to the finding that both groups had a shared history of learning problems and diagnoses. This hypothesis is consistent with a study reported in the CMS Manual where a sample of learning disabled c hildren (n = 74) scored significantly lower than an age and gender matched control group across the CMS Verbal Immediate, Delayed, and Delayed Recognition Indices. Factors Associated with Verbal Encoding Problems in Pediatric ADHD, Mood, and Learning Disor ders In order to empirically examine what cognitive factors contribute to the ADHD and Clinical Control group’s shared verbal memory profile, we used the model developed in Aim 1 to determine which cognitive and/or demographic factors predict verbal memory enco ding, retention, and retrieval abilities. Based on previous literature and the attention and working memory problems associated with ADHD, we originally predicted that attention and working memory/executive abilities would significantly predict verbal enc oding and retrieval performance within the ADHD sample (Denckla, 1996; Doyle,

PAGE 107

107 2006). Furthermore, we hypothesized that attention and working memory/executive abilities were partially mediating any discrepancies in verbal memory performance between the ADH D and Clinical Control groups. However, given the previous finding that the ADHD sample did not significantly differ from the Clinical Control group on any verbal memory measure, investigation of these two hypotheses were not warranted in this study. Ins tead, we examined the hypothesis that the verbal memory profile of the combined ADHD and Clinical Control group was related to their attention and working memory/executive abilities. In contrast to our hypothesis, regression analyses demonstrated that ver bal knowledge significantly predicted verbal memory encoding abilities; no other cognitive abilities, including attention and working/memory, were related to the groups’ memory profile. Although our hypothesis was not supported, the finding that verbal knowledge was related to verbal encoding abilities of a combined ADHD and Clinical Control group was not surprising given the context of our Aim 1 findings. The results of Aim 1 indicated that verbal comprehension abilities are related to verbal encoding in a large, mixed clinical pediatric sample. The Aim 1 sample was indeed heterogeneous and included a variety of individuals with medical and psychiatric history; however , over 55% of the large sample had an ADHD diagnosis. The Aim 2 sample represented a su bset of indi viduals with ADHD from Aim 1. Given the large overlap of the Aim 1 and Aim 2 samples, it is logical that Aim 1’s SEM results were reproduced in Aim 2’s regression analyses. Future studies should aim to replicate the finding that verbal knowledge is the only cognitive domain related to verbal memory of ADHD patients in an ADHD sample that is independent from this dissertation’s sample. As discussed above in the previous

PAGE 108

108 sections, the exact nature of the relationship between verbal encoding and verbal knowledge in clinical pediatric populations, including ADHD, is not clearly understood. Future research should aim to investigate this relationship. Memory Profile and Contributing Factors in Pediatric TBI The third aim of this study was to descri be the verbal memory performance profile of a pediatric TBI sample and to identify the factors that cont ribute to their verbal memory. Verbal Encoding Weakness Uniquely Associated with Pediatric TBI First, we compared the verbal memory performance profile of children diagnosed with TBI to a healthy normative data from the CMS manual and, separately, an Orthopedic Injury Control sample. Our hypothesis that the TBI sample would demonstrate deficits in the verbal encoding and retrieval stages, but not in the retention stage, was only partially supported by our results. O n the group mean level , a heterogeneous sample of pediatric patients with history of moderate to severe TBI did not demonstrate any clinically significant verbal memory impairments. However, group mean statistics of a TBI sample can be challenging to interpret because the population is very heterogeneous (e.g. injury location, severity, mechanism vary widely ; Yeates, 2010). Therefore, analysis of individual level impairments can be useful in understanding the verbal memory profile of the TBI sample. On the individual level, 2036% of the TBI patients exhibited clinically significant impairments on at least one CMS measure of verbal memory (Table 4 19 ). Th ese rates of individual impairment were similar to the frequencies reported from a small pediatric TBI study (n = 16) in the CMS manual (Cohen, 1997).

PAGE 109

109 The highest rate of individual impairment in this study’s TBI sample was observed in the verbal encoding measure. Consistent with this individual level finding, group mean statistics revealed that the TBI patients displayed a relative weakness on a measure of verbal encoding compar ed to verbal memory retrieval. T hese findings indicate that the verbal memory profiles of pediatric TBI patients are more likely to be marked by verbal encoding problems than verbal retention or retrieval difficulties. Notably, this finding is consistent with the encoding weakness previously identified in the other pediatric sampl es studied in this dissertation (i.e. the ADHD, Clinical Control, and larger, heterogeneous pediatric samples). In order to determine whether this verbal memory encoding weakness was unique to TBI and not stemming from nonTBI injury factors (e.g. preexis ting sample characteristics like a tendency to be inattentive), the verbal memory profile of the TBI sample was compared to the profile of an Orthopedic Injury control group. The TBI and Orthopedic Injury groups were well matched in terms of age, race/ethnicity, and history of learning, mood disorder. The groups differed on parent reported symptoms, such that the TBI group reportedly displayed more externalizing behaviors, depressive symptoms, and atypical behaviors. Given the retrospective nature of thi s study, it was unclear whether these group differences are related to TBI or were premorbid group differences. However, other literature notes that externalizing, depressive, and odd behaviors are common to TBI populations, so it is likely that TBI and Orthopedic Injury control patients are representative of their respective populations ( Max, 2005) . Based on previous research, w e hypothesized that the TBI group would perform significantly worse than the Orthopedic Injury Control group on measures of ver bal

PAGE 110

110 encoding and retrieval. Results of this study confirmed this hypothesis and demonstrated that pediatric patients with history of TBI have an overall reduced verbal memory profile compared to agematched Orthopedic Injury Control patients. Post hoc t tests revealed that the TBI group performed significantly worse than the Orthopedic Injury group on the verbal encoding measure. There were no group differences on verbal retention or retrieval measures. Overall, these findings suggest that pediatric TB I is associated with a unique verbal memory profile, which is distinguished by a relative weakness in verbal encoding. This is the first known study to specifically examine the verbal memory performance profile of a pediatric TBI sample in terms of encoding, retention, and retrieval. The results of this study indicate that pediatric TBI patients experience verbal encoding problems. The encoding problem was identified in comparison to the CMS normative sample and compared to this study’s Orthopedic Injury Control group. This finding is consistent with the existing literature. For example, previous research reported that pediatric TBI patients do not spontaneously engage in active, elaborative encoding strategies, and their learning does not benefit from s imple repetition of information (Harris, 1996; Mottram & Donders, 20 06; Oberg & Turkstra, 1968). Cognitive Factor s Contributing to Verbal Encoding Weakness Associated with Pediatric TBI Similar to our investigation of the memory profile associated with childhood ADHD, we used the model developed in Aim 1 to determine which cognitive and/or demographic factors contribute to the verbal memory performance profile of children diagnosed with TBI . We predicted that within the pediatric TBI group, attention, pro cessing speed, and working memory/executive abilities predicts verbal encoding and

PAGE 111

111 retrieval performance, and these cognitive factors partially mediate the verbal memory impairments uniquely associated with TBI (as identified in Aim 3b). However, the regression analyses indicated that none of cognitive variables significantly predicted verbal memory abi lities. It is possible that these null results occurred because the independent variables entered into the regression model were not neces sarily appropriate within the TBI sample. Each of the independent variables entered into the regression models were selected based on Aim 1’s EFA in order to represent the cognitive domains of verbal knowledge, processing speed, attention, and working mem ory/executive functioning. This method of selecting independent variables was appropriate in Aim 2’s analyses because the ADHD and Clinical Control samples were subgroups from Aim 1’s larger heterogeneous sample. However, Aim 3 examined the verbal memory and cognitive abilities of a TBI sample that overlapped, but was not entirely derived from the Aim 1 sample (See Table 23 in Appendix B for more information) . In fact, the majority of Aim 3 TBI data was derived from a separate research study. Compared t o evaluations in a clinical setting, where all of Aim 1’s data was collected, evaluations in the research study differed because not all research patients completed the WISC IV subtests. Although the WASI Vocabulary and Similarities subtest data can be interpreted similarly to WISC IV Vocabulary and Similarities, research patients were likely to have missing data the WISC IV Processing Speed subtests and the WISC IV Digit Span Backwards subtests. Given the separate sources of data collection, the EFA results may not have necessarily applied to Aim 3’s TBI sample. Consequently, the independent variables entered into the Aim 3 regression models may not have been appropriately selected.

PAGE 112

112 Based on this limitation for interpreting the null findings of Aim 3’s a nalyses, post hoc correlations were conducted. The post hoc correlation examined whether any cognitive variables related to verbal memory abilities. As shown in Table 55 , within the pediatric TBI sample WISC IV Comprehension scores significantly correlated with the measure of verbal encoding, and TEA Ch Sky Search Attention scores correlated with verbal retrieval. A single regression model confirmed that WISC IV Comprehension scores significantly predicted verbal encoding (F(1,16) = 9.72, p<.01, B = 4.70, SE(B) = 1.51, Beta = .62). However, a separate regression model revealed that TEA Ch Sky Search Attention scores did not significantly predict verbal retrieval. These post hoc results suggested that verbal knowledge (as measured by the WISC IV Comprehension) may mediate the relationship between TBI status and reduced verbal encoding abilities. Unfortunately, this study lacked the necessary data (i.e. no WISC IV Comprehension scores for the Orthopedic Injury Control group) to conduct mediation analyses to explore the hypothesis about verbal encoding. Therefore, it is unclear in this study whether verbal knowledge mediates the relationship between TBI and reduc ed verbal encoding abilities. Future studies should aim to determine whether verbal knowledge m ediates the relationship between TBI and reduced verbal encoding abilities. However, based on the other findings of this study’s previous aims, we hypothesize that verbal knowledge does not necessarily mediate the relationship between TBI and reduced verbal encoding. This hypothesis is primarily based on our finding s that verbal knowledge is significantly related to verbal encoding abilities in other clinical pediatric populations. In addition, this relationship has been suggested in previous literature. T he significant relationship

PAGE 113

113 between verbal knowledge and verbal encoding abilities is most likely not unique to patients with history of TBI. F uture studies should aim to better understand the relationship between verbal knowledge and verbal encoding abilities of a pediatric TBI sample. Overall Conclusions The overall goal of this study was to extend previous adult and healthy childhood research findings to pediatric patient populations. Pediatric patients are often referred for Neuropsychological eval uation and treatment of verbal “memory problems;” however, previous research has not described the exact nature of these difficulties or how they may be occurring. This dissertation study found that in clinical pediatric populations commonly referred to a Pediatric Neuropsychology Clinic, verbal memory problems seem to stem from encoding difficulties . Encoding difficulties were identified in the verbal memory performance profiles of a large, heterogeneous sample of pediatric patients with neurodevelopment al and acquired brain injury, a subset of childhood ADHD and Clinical Control patients, as well as in a pediatric TBI sample, which was largely independent from the large, heterogeneous sample of pediatric patients . Verbal memory retention and retrieval abilities appeared relatively intact across the patient samples. T hese finding suggest that pediatric patients referred to a Pediatric Neuropsychology Clinic may be experiencing verbal memory problems that stem from encoding problems, and the information t hat is encoded is stored and retrieved relatively well. The verbal memory profile associated with neurodevelopmental and acquired pediatric disorders (i.e. encoding weakness , but good retention and retrieval) suggests that clinicians should be evaluating t he verbal memory of pediatric patients with an

PAGE 114

114 information processing approach that assesses each stage of the verbal memory process (i.e. encoding, retention, and retrieval). Evaluation of each verbal memory stage can inform clinical recommendations and treatment plans. For example, treatment of verbal encoding problems should specifically aim to teach patients effective learning strategies (e.g. active elaboration of information to be remembered) and/or encourage compensatory strategies during the learning phase (e.g. repeated presentations of information to be remembered; errorless learning presentations of infor mation to be remembered). In addition, this dissertation study found that verbal encoding abilities are related to verbal knowledge, such that better verbal knowledge is associated with better verbal encoding abilities. This finding was demonstrated across three groups: a large, heterogeneous sample of pediatric patients with neurodevelopmental and acquired brain injury, a subset of childhood AD HD and Clinical Control patients, as well as in a pediatric TBI sample which was largely independent from the large, heterogeneous sample. Additionally, this study finding is consistent with other existing literature that reports verbal knowledge is relat ed to verbal memory abilities in healthy, normally developing children. The exact causeand effect nature of this relationship is unclear; however, regression analyses in this study and other previous studies suggest that improving verbal knowledge will i mprove verbal encoding. This hypothesis may have important treatment implications because if cognitive rehabilitation efforts can improve verbal knowledge/abilities, then verbal memory encoding abilities may subsequently also improve. Future studies shoul d aim to identify methods for improving verbal

PAGE 115

115 knowledge/abilities in pediatric patients with memory difficulties and investigate whether treatment of verbal knowledge skills result in additional improvements in verbal memory. The underlying neuroanatomic cause of the verbal memory encoding problems likely varies across clinical pediatric populations depending on the population’s neural dysfunction. However, the prefrontal cortex (PFC) and/or white matter connects with the PFC likely mediate verbal encoding difficulties in clinical pediatric populations. The prefrontal cortex (PFC) appears important for strategic verbal memory encoding processes in healthy individuals, and frontal lobe dysfunction has been associated with encoding difficulties (GoldmanRak ic, 1996; Mayes, 2002; Wagner, 2002; Stuss & Benson, 1984). In addition, white matter connections between the frontal lobes and memory supporting structures (e.g. basal ganglia, cerebellum, and medial temporal lobe) are important for successful encoding; disruption of these white matter connections in clinical pediatric populations could result in encoding processes. Frontal lobe dysfunction and disrupted white matter connections have been commonly reported across a variety of clinical pediatric populations, including childhood ADHD, TBI, Cancer, etc. (Arnsten, 2006; Goldstein, 2011;Marks et al., 2010; McAllister, 2006; Nagel et al., 2004). Therefore, frontal lobe dysfunction and/or disrupted connectivity with the frontal lobes may be underlying the encoding problems displayed by clinical pediatric populations. The verbal memory performance profile of the clinical pediatric populations did not necessarily suggest that neural structures underlying verbal retention and retrieval (i.e. the medial temporal lobe, diencephalon, and basal forebrain) are dysfunctional; however, the strong relationship be tween verbal knowledge (which is theoreticall y

PAGE 116

116 stored in neocortical areas like the parietal and anterior temporal lobe, but retrieved through these same structures) suggests that pediatric patients, especially those with co morbid learning disorders, may be experiencing dysfunction in the medial tem poral lobe, diencephalon, and basal forebrain or the white matter connections between these areas ( Martin & Chao, 2001; Paller, 2002) . In order to better identify the neural dysfunction underlying encoding problems in clinical pediatric populations, futur e imaging studies should be conducted to formally examine the neurocorrelates related to verbal encoding problems of pediatric populations. In general, a prospective imaging study that examines the neurocorrelates of each verbal memory stage (i.e. encoding, retention, and retrieval) would be very useful in understanding the verbal memory performance profile of clinical pediatric populations. Comments on Methodology Implications & Limitations of an Archival Clinical Study This study found a significant rel ationship between verbal knowledge and verbal memory in a heterogeneous sample of pediatric patients who were referred for a neuropsychological evaluation. Given that our data set was large and collected from an archival clinical data review, the demograp hics and neuropsychological performance of the sample were representative of the general pediatric neuropsychology clinic referred population. Thus, the results and clinical implications of this study are ecologically valid and should generalize to clinic ref erred pediatric populations. Although standard of care should encourage neuropsychological evaluation of any pediatric patient with history of neurodevelopmental disorder or acquired brain injury, unfortunately, not all pediatric patients are referred for neuropsychological evaluations. Since this dissertation study, especially Aims 1 & 2, primarily analyzed data from clinic -

PAGE 117

117 referred pediatric patients, the r esults of this study may not generalize to pediatric patients who are not referred to or cannot complete a neuropsychological evaluation. For example, pediatric patients without cognitive complaints or functional difficulties are unlikely to be referred for neuropsychological evaluation. A subset of pediatric patients without complaints or functio nal problems may display a different verbal memory profile than this study’ sample. Although this study did include some patients without any clinically significant performancebased impairments, these patients most likely presented to the Psychology Clinic with reported cognitive complaints and/or functional problems (i.e. the patients were referred for neuropsychological evaluation due to some concern about cognitive difficulties). In addition, this dissertation did not separately examine data from the subset of patients without any performancebased impairments. Thus, the results of this dissertation may not apply to pediatric patients without cognitive comp laints or functional problems. In addition, this study did not include data from patients who w ere unable to complete neuropsychological testing (e.g. patients without sufficient oral language and/or motor abilities to complete). These pediatric patients with severe language and/or motor problems may also display a different verbal memory profile t han this study’s sample. Therefore, the findings of this study should be further examined in a larger, more representative sample of pediatric patients. A future study should plan to collect neuropsychological data from both clinic referred and nonclinic referred children with history of neurodevelopmental disorders and/or acquired brain injury. Replication of this study’s results may clarify evidencebased recommendations for the clinical assessment and treatment of memory problems in a wider range of pediatric patients.

PAGE 118

118 A second limitation of this study’s archival data review was that many of the pediatric patients had missing data. In fact, the review of archival clinical data identified 571 pediatric patients who had undergone neuropsychological evaluations. Of these, less than half ( N = 23 4 ) pediatric patients met inclusion criteria for analysis because the majority of patients were missing the required memory measures. For some analyses, pairwise deletion of cases was required to account for mi ssing data in individual cases. Although an archival data review allowed us to collect a very large amount of ecologically valid data, the amount of missing data was unexpected and may have limited the power of some analyses (e.g. the sample size of the o f the Clinical Control group in Aim 2 was lower than the size power analyses recommended). Future studies should aim to prospectively collect neuropsychological data in order to reduce the amount of missing data and maximize the efficiency of data collect ion. Utility of Encoding, Retention, and Retrieval Measures Bauer (2003) reported that the memory processes of encoding, retention, and retrieval are a useful heuristic for characterizing the nature of memory problems. In addition to examining a general m easure of verbal memory abilities (i.e. the CMS Delayed Verbal Memory Index Score used in the SEM), this dissertation study was designed to examine the verbal memory abilities of pediatric patients in terms of memory processes (encoding, retention, and ret rieval). By examining the individual processes of verbal encoding, retention, and retrieval, this study was better able to characterize the verbal memory profiles associated with clinical pediatric populations. Thus, this dissertation demonstrated that evaluation of a single memory process and/or evaluation of general verbal memory measure (e.g. the Delayed Verbal Memory Index score) may not clearly identify memory deficits and additionally may not reveal enough

PAGE 119

119 information about an individual or group’s verbal memory abilities to guide treatment recommendations. Despite the need to evaluate the distinct verbal memory processes, many neuropsychological tests of memory do not provide distinct measures of encoding, rete ntion, and retrieval. The California V erbal Learning Test for Children (CVLT C; Delis, Kramer, Kaplan, & Ober, 1989), which is the current child version of the California Verbal Learning Test, Second Edition (CVLT II; Delis, Kramer, Kaplan, & Ober, 1989), is one of the few pediatric memory tes ts that provide separate measures of encoding, retention, and retrieval. Specifically, the CVLT C estimates encoding abilities by providing standardized measures of verbal immediate memory, semantic clustering, serial clustering, and the learning curve over repeated learning trials. Retention abilities are estimated through percent retention, which reflects what proportion of information initially learned is stored over time and recalled later. Finally, the CVLT C provides a recall/recognition contrast score to reflect verbal retrieval abilities. Overall, these CVLT C measures of verbal encoding, retention, and retrieval appear to be good methods for evaluating the stages of verbal memory. Unfortunately, the CVLT C measures were not utilized in this dis sertation because it is not as frequently administered in the Shands/UF Pediatric Neuropsychology Clinics as the Children’s Memory Scale (CMS). Historically, the CMS has been the preferred measure of pediatric memory because it comprehensively evaluates v erbal and nonverbal learning and memory. In addition, the CMS includes the Stories subtests to assess oral narrative learning and memory; whereas the CVLT C only evaluates verbal list learning.

PAGE 120

120 The CMS manual provided agecorrected, standardized measures of verbal encoding (i.e. Verbal Memory Index score); however the CMS manual did not clearly provide standardized measures of verbal retention and retrieval. Consequently, we derived agecorrected measures of verbal retention (i.e. the Verbal Percent Retention score) and retrieval (i.e. the Verbal Delayed Contrast score). These derived scores were based on the information that was available in the CMS manual (i.e. agebased group means and standard deviations of percent retention on verbal subtests) and pr ovided by the CMS publishers after a special request (i.e. individual means of the Verbal Delayed Index and Verbal Delayed Recognition Index score). The derived scores were calculated in a manner that was consistent with the CVLT C’s measures (described above) and the CMS manual’s recommendations. Overall, the derived Verbal Percent Retention score and the Verbal Delayed Contrast score were calculated based on good theory and methodology; they are most likely good estimates of verbal retention and retrieval. However, given the derived nature of these scores, our measures of the retention and retrieval should be interpreted with caution. That is, these specific CMS indices have not been previously published or discussed in previous research. As a result, t he psychometrics (i.e. reliability and validity) of the derived retention and retrieval scores are unknown. In particular, the CMS author (Dr. Morris Cohen) and a statistician from the CMS publisher (Dr. Jim Holdnack) expressed concern that the Verbal Del ayed Contrast score may be an unreliable measure of retrieval abilities because one of the scores it is derived from (the Verbal Delayed Recognition Index score) may have a restricted range (L. Jordan, personal communications, 2012). When a contrast score is

PAGE 121

121 derived from the simple difference between scores, then the interpretation of the contrast score is dependent on the distribution of the original s cores that are being compared. Our examination of the normative Delayed Contrast Scores suggests that t he Verbal Delayed Contrast score is normally distributed in healthy children and adolescents (See Table 22 in Appendix A). We hypothesized that the Delayed Contrast Scores appear normally distributed because the restricted range of the recognition score g ets corrected to a normal distribution when it is standardized into the composite Index score. However, the psychometrics and interpretation of derived contrast score may have been improved with regressionbased methodology. Regressionbased methodology has been used to calculate contrast scores in the NEPSY II (Korkman, Kirk, & Kemp, 2007) and the Wechsler Memory Scale – Fourth Edition (WMS IV; Wecshler, 2009). In general, the regressionbased methodology forces one variable to be the control variable and the other to be the dependent variable (e.g. recognition scores as the control and recall scores as the dependent variable). The nonlinear regression technique is a useful method to derive contrast scores because it does not require assumption of normal distributions and accounts for differences in bas e rates across ability ranges. Regressionbased contrast scores are a relatively new concept, and unfortunately, this dissertation study did not have the necessary raw recognition and recall scores to derive a regressionbased Delayed Recognition/Recall contrast score. Future memory s tudies examining the separate verbal memory processes would benefit from using from regressionbased methods to calculate measures of verbal encoding, retention, and retrieval.

PAGE 122

122 Memory Measure This dissertation study estimated verbal memory abilities with the Children’s Memory Scale (CMS). However, there are many other neuropsychological tests that examine pediatric memory abilities. The CMS manual reported that in general the CMS has good reliability with other pediatric memory measures (e.g. WMS III, WR AML, CVLT C); however, future studies should replicate these dissertation results with other verbal memory measures. For example, a study that examines the performance of clinical pediatric populations on the CVLT C would be particularly interesting because the CVLT C’s measures of verbal encoding, retention, and retrieval are published and widely used in clinical practice. In addition, future studies should aim to extend this study by examining visual memory in addition to verbal memory. This study focus ed on the verbal memory profiles because previous literature suggested that in contrast to visual memory, verbal memory processes are more likely to be influenced by other cognitive abilities and disrupted by abnormal development or childhood injury (Dehn, 2010; Paivio, 1991). In addition, verbal encoding, retention, and retrieval were more easily examined with our memory measures than visual memory processes (i.e. a retrieval measure cannot be derived from the CMS nonverbal memory tests). Therefore, thi s study’s findings are only applicable for understanding the verbal memory processes of clinical pediatric populations. In comparison to verbal memory, the examination of visual memory profiles in clinical pediatric populations may yield very different results. In particular we hypothesize that visual memory abilities would be generally intact in clinical pediatric populations. In addition, we would not expect to identify a strong relationship between

PAGE 123

123 verbal knowledge and visual memory. These hypotheses are based on previous research that demonstrates that visual memory is developed very early infancy and is less likely to be disrupted by other cognitive abilities or abnormal brain development/injury ( Dehn, 2010; Paivio, 1991). In fact, the CMS manual r eported that in special group studies of children with neurodevelopment disorders (i.e. learning disorders, ADHD, specific language impairment), CMS Visual Immediate and Delayed Memory Indices were relatively intact compared to the performance of matched c ontrols (Cohen, 1997). Additionally, the CMS manual reported that unlike the CMS verbal memory indices, the CMS visual memory indices do not significantly correlate with measures of executive functioning and language processing. In addition to further st udies of verbal memory, future studies should empirically investigate the nonverbal memory profiles of clinical pediatric populations and aim to identify what cognitive (or demographic) factors contribute to visual memory abilities. Discussion Summary The current set of studies revealed that clinical pediatric populations have a verbal memory profile characterized by a relative weakness in the verbal encoding stage of memory processes. In contrast, their v erbal retention and retrieval skills were similar t o those of healthy normative data and clinical control samples. Although clinical pediatric populations do not necessarily demonstrate a clinically significant verbal encoding impairment at the group level , their verbal encoding weakness appeared most rel ated to verbal knowledge skills . Future studies should aim to replicate these findings in independent samples of clinical pediatric populations and with a variety of memory measures, including both verbal and visual memory tasks . This study’s findings su ggest that clinicians should utilize measures that provide a comprehensive evaluation

PAGE 124

124 of all three pediatric memory stages. If these findings are successfully replicated, it will be important to focus efforts on development of memory treatment/intervention programs that emphasize rehabilitation of encoding abilities as well as foster development of verbal knowledge skills.

PAGE 125

125 Table 5 1 . Correlations between m easur es of verbal m emory & o ther cognitive d omains in Aim 1’s c linical p ediatric s ample Encoding (CMS Verbal Immediate Index) Retention (CMS Verbal Retention Score) Retrieval (CMS Verbal Delayed Contrast Score) WISC IV Vocabu lary (ss) WISC IV Similariti es (ss) WISC IV Compre h.(ss) WISC IV Symbol Search (ss) WISC IV Coding (ss) TEACh Creature Counting Total (ss) TEA Ch Sky Search Attention Score (ss) TEACh Score! (ss) WISC IV Digit Span Backward (ss) Delayed Verbal Memory (CMS Verbal Delayed Index) Correl. Coeff. .77** 0.04 .44** .50** .51** .52** .25** .29** .17* .17* .17* .24** Sig. (2 tailed) 0.00 0.58 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.02 0.01 0.00 N 232 180 212 194 194 194 200 199 179 204 209 135 Encoding (CMS Verbal Immediate Index) Correl. Coeff. 1.00 0.03 .15* .51** .49** .51** .24** .26** .16* 0.08 0.10 0.15 Sig. (2 tailed) 0.70 0.03 0.00 0.00 0.00 0.00 0.00 0.03 0.27 0.14 0.07 N 232 179 211 193 193 193 199 198 178 203 208 134 Retention (CMS Verbal Retention Score) Correl. Coeff. 0.03 1.00 .25** 0.03 0.10 0.01 0.02 0.01 0.16 0.05 0.08 0.12 Sig. (2 tailed) 0.70 0.00 0.71 0.23 0.88 0.83 0.87 0.07 0.52 0.33 0.21 N 179 181 168 150 150 150 154 153 137 159 162 114 Retrieval (CMS Verbal Delayed Contrast Score) Correl. Coeff. .148* .250** 1.00 0.06 .15* 0.08 0.01 0.01 0.10 .16* 0.01 0.03 Sig. (2 tailed) 0.03 0.00 0.46 0.05 0.31 0.88 0.93 0.22 0.02 0.92 0.77 N 211 168 213 176 17 176 182 181 164 186 190 126 *. Correlation is significant at the 0.05 level (2 tailed). **. Correlation is significant at the 0.01 level (2 tailed).

PAGE 126

126 Table 5 2 . Post hoc mediation m odel r esults for Aim 1 Dependent Variable (CMS Verbal Memory Measure) Model Significant Predictor(s) R2 p Cognitive Domain (Measure) B SE B P Encoding (Immediate Memory Index) Verbal Knowledge (WISC IV VCI) a path .61 0.07 .00 Delayed Verbal Memory (Delayed Memory Index) Encoding (Immediate Memory Index) b path .73 .05 .00 .79 .00 Verbal Knowledge (WISC IV VCI) c’ path .23 .05 .00 Verbal Knowledge (WISC IV VCI) c path .71 .07 .00 Table 5 2 describes the effects between the variables entered into the bootstrapping analysis. As shown, Encoding abilities are a significant mediator between the relationship between Verbal Knowledge and Delayed Verbal Memory; More specifically, Encoding abilities partially mediate the relationship. Table 5 3 . Collinearity diagnostic s tatistics of the measured c ognitive v ariables Predictor Tolerance VIF Sky Search Attention Score (Scaled Score) .80 1.25 Score! (Scaled Score) .87 1.15 Creature Counting Total correct (Scaled Score) .87 1.14 Similarities Scaled Score .44 2.28 Vocabulary Scaled Score .38 2.65 Comprehension Scaled Score .48 2.08 Digit Span Backward Span scaled score .85 1.18 Coding/Digit Symbol Scaled Score .63 1.58 Symbol Search Scaled Score .57 1.76 Table 5 4 . Correlations between CMS v erbal memory m easures Retention (Retention Score) Retrieval (Delayed Contrast Score Encoding (Immediate Index) Corr. Coeff. .49** 0.10 Sig. (2 tailed) 0.01 0.57 N 29.00 37.00 Retention (Retention Score) Corr. Coeff. 0.15 Sig. (2 tailed) 0.42 N 29.00 * Asterisk indicates significant correlation, p<.05, ** Double asterisk indicates significant correlation, p <.01

PAGE 127

127 Table 5 5 . Correlations between CMS verbal m emory and o ther c ognitive m easures WISC IV VCI / WASI VIQ WISC IV/WA SI Vocab WISC IV/WASI Similarities WISC IV Comprehens. WISC IV Processing Speed Index WISC IV Coding WISC IV Symbol Search TEACh Creature Counting Total TEACh Sky Search Attention Score TEACh Score! WISC IV Digit Span Backwar d Encoding (Immediat e Index) Corr. Coeff. 0.30 0.19 0.06 .59** 0.06 0.14 0.11 0.22 0.05 0.33 0.87 Sig. (2 tailed) 0.28 0.27 0.74 0.01 0.80 0.58 0.67 0.22 0.77 0.06 0.33 N 15.00 34.00 34.00 18.00 19.00 19.00 19.00 32.00 33.00 33.00 3.00 Retention (Retention Score) Corr. Coeff. 0.30 0.01 0.05 0.19 0.17 0.34 0.19 0.21 0.11 0.13 0.80 Sig. (2 tailed) 0.35 0.95 0.77 0.51 0.54 0.23 0.52 0.30 0.57 0.49 0.20 N 12.00 32.00 32.00 15.00 15.00 14.00 14.00 28.00 29.00 29.00 4.00 Retrieval (Delayed Memory Contrast Score Corr. Coeff. 0.20 0.09 0.02 0.08 0.07 0.29 0.20 0.18 .35* 0.06 0.50 Sig. (2 tailed) 0.48 0.64 0.89 0.76 0.78 0.25 0.43 0.34 0.05 0.76 0.67 N 14.00 33.00 33.00 17.00 18.00 18.00 18.00 31.00 32.00 32.00 3.00 * Asterisk indicates significant correlation, p<.05, ** Double asterisk indicates significant correlation, p <.01

PAGE 128

128 APPENDIX A DETAILED DESCRIPTION OF CMS VERBAL MEMORY MEASURES The following sections describe in detail the administration and scoring of the core CMS Verbal Memory subtests. Descriptions of subtests and scores below are abbreviated from the CMS Manual (Cohen, 1997). Stories Subtests The CMS Stories subtests were designed in order to evaluate a child’s intentional verbal memory of a logical narrative. During the encoding stage of Stories, the child listens to two brief, ageappropriate, and orally administered stories. The child is instructed to listen carefully and is warned about subsequent the subsequent memory tests. Immediately after listening to each story, the child is administered an immediate free recall test (i.e. Stories Immediate Recall subtest). Additionally, the child is administered a delayed freerecall test (i.e. Stories Delayed Recall subtest) 30 minutes after initial encoding. Finally, the child completes a 15item yes no delayed recognition test about each story (i.e. Stories Delayed Recognition subtest). Responses on the Stories free recall and recognition tests are recorded and later coded as either incorrect or correct by the test administrator according to scoring guidelines provided in the CMS manual. In summary, the CMS Stories subtests examines a child’s verbal memory of two orally adm inistered narratives through an immediate and delayed free recall as well as a delayed recognition test. Word Pairs Subtests The CMS Word Pairs subtests examine a child’s intentional learning and memory of verbal paired associates. During the encoding stage of the CMS Word Pairs subtests, the child listens to an orally administered list of 14 word pairs. Immediately

PAGE 129

129 after the list of word pairs is read aloud, the child is provided with one word and asked to recall its word pair from the list. This encodi ng sequence is repeated two additional times with the same list of word pairs rearranged and presented in a different order. In total, the child is presented with a total of three word pair learning trials. Any incorrect responses are corrected during e ach learning trial. The raw CMS Word Pairs Learning score is the total number of correct responses across all three learning trials. Similar to the CMS Stories subtests, the CMS Word Pairs subtests includes recall and recognition memory tests. Immediatel y subsequent to the three learning presentations, the child is administered the Word Pairs Immediate Recall subtest, where the child is asked to freely recall as many of the word pairs as he can remember without any cues. A second, delayed free recall tes t of the word pairs (i.e. the Word Pairs Delayed recall subtest) is similarly administered 2530 minutes later. In addition, delayed memory of the word pairs is examined with the yes/no Word Pairs Delayed Recognition subtest. Responses on the learning t rials as well as the memory tests are recorded and later scored as incorrect or correct by the test administrator. In summary, the CMS Word Pairs subtests examine a child’s verbal paired associate learning and memory through three learning trials, an immediate free recall, a delayed free recall, and a yes/no delayed recognition test. CMS Standardized Scores The CMS manual describes in detail standardized methods for scoring correct and incorrect responses on the CMS subtests (Cohen, 1997]). In brief, correct responses on the learning and free recall tests are given one point; any intrusions or perseverative errors are not scored. Correct responses on the yes/no recognition tests are scored such that both correct “yes” responses (i.e. hits) as well as correct “no” responses (i.e.

PAGE 130

130 true negatives) are counted in the total recognition score. Incorrect responses (i.e. misses and false positives) are not given any points. In summary, raw scores reflecting learning and memory are derived for each CMS verbal subtest. Raw scores from the core CMS subtests can be further processed into standardized scores. The CMS manual provides guidelines on how to derive agereferenced standardized scores in order to compare individual performance to an average performance of other sameaged healthy children. We plan to analyze the following standardized composite scores (mean = 1 00, standard deviation = 15): Verbal Immediate Index, Verbal Delayed Index, and Delayed Recognition Index. The Immediate (and Delayed) Verbal composite scores combine scores from the Stories and Word Pairs immediate (and delayed) recall subtests. The Delayed Recognition composite score combines only scores from the Stories and Word Pairs delayed recognition subtests. These composites scores provide an index to compare overall individual memory performance to a healthy, sameaged group of children. More specifically, the CMS manual reports that the Verbal Immediate Index provides a measure of verbal encoding abilities; whereas the Verbal Delayed Index reflects a general verbal memory score that is dependent on encoding, retention, and retrieval abilities. In addition, the following paragraph describes how the difference between the Verbal Delayed and Delayed Recognition Indices can be used to estimate verbal retrieval abilities. The CMS manual provided guidelines to calculate a contrast (i.e. difference) s core between the Verbal Delayed Recall and Delayed Recognition composite scores. In this study, the precise calculation for the difference score was Delayed Recognition Index

PAGE 131

131 Score minus Verbal Delayed Recall Index score. Although the CMS manual does not provide any group mean or standard deviations from a normative sample, the CMS publisher (NCS Pearson, Inc.) provided us with the agebased group means and standard deviations that were collected from original CMS normative sample. Table 22 below summari zes the CMS normative sample’s performance on the Verbal Delayed Contrast measure (NCS Pearson Inc., 1997). The normative data presented in Table A 1 has been transformed to use the same scale as the Verbal Immediate, Delayed, and Recognition Indices (i.e. mean = 100, standard deviation = 15). We used the normative group means and standard deviations to calculate standardized Recall/Recognition Contrast z scores, which can be used to compare patient performance to normal child/adolescent performance. Acc ording to the CMS manual, a Verbal Delayed (Recall) Index score that is significantly lower than Delayed Recognition Index (i.e. the raw contrast score is large, and the standardized contrast score is low) can indicate retrieval difficulties in the context of adequate encoding and retention abilities. Alternatively, if the Verbal Delayed (Recall) Index and Delayed Recognition Index are similar (i.e. the raw contrast score is small, and the standardized contrast score is within normal limits ), then the Rec all/Recognition Contrast score would indicate adequate retrieval abilities given the context of the individual’s encoding and retention abilities. Finally, the CMS manual provided guidelines for calculation of supplementary contrast scores which test administrators can use to further examine the memory process at the levels of encoding, retention, and retrieval. For the Stories and Word Pairs subtests, percent retention scores were calculated by dividing the delayed recall

PAGE 132

132 score by the immediate recall score and multiplying by 100. The CMS manual provided percent retention agebased group mean and standard deviations from a normative group of children so that standardized percent retention z scores can be calculated for each subtest (Cohen, 1997, Appendix D, Table D.5). Additionally, we calculated a composite Verbal Percent retention score by averaging together the Stories and Word Pairs percent retention scores. Although the CMS manual warns that the percent retention approach depends on the assumption that the immediate recall score is an accurate measure of encoding abilities, low percent retention scores generally indicate a rapid rate of forgetting.

PAGE 133

133 Table A 1 . Normative d escriptive statistics of the v erbal delayed c ontrast s core Age Group (in years) Skewness Kurtosis N with Score <81.9 Statistic N Min. Max Mean Std Stat. Std. Error Stat. Std. Error 5 100 66 134 100.61 13.31 0.11 0.24 0.20 0.48 12.00 6 100 63 136 99.13 16.04 0.02 0.24 0.47 0.48 14.00 7 99 54 137 100.80 13.53 0.33 0.24 1.10 0.48 5.00 8 100 69 125 98.56 12.38 0.07 0.24 0.33 0.48 7.00 9 99 69 137 100.30 14.64 0.11 0.24 0.30 0.48 9.00 10 99 66 137 100.35 14.75 0.20 0.24 0.21 0.48 8.00 11 98 44 134 100.96 13.43 0.55 0.24 2.47 0.48 5.00 12 100 66 128 100.31 12.59 0.11 0.24 0.04 0.48 8.00 13 50 64 134 99.90 13.92 0.05 0.34 0.52 0.66 3.00 14 50 60 128 99.66 15.08 0.26 0.34 0.16 0.66 7.00 15 50 72 134 103.00 13.74 0.14 0.34 0.35 0.66 4.00 16 50 57 134 98.34 13.39 0.23 0.34 1.12 0.66 4.00 Reference: Standardization data from the Children’s Memory Scale (CMS) . Copyright 1997 NCS Pearson, Inc. Used with permission. All rights reserved.

PAGE 134

134 APPENDIX B SOURCES OF DATA As described in the Participants section of this document, data was collected from both the UF/Shands Psychology Clinic as well as Dr. Heaton’s Pediatric Neuropsychology research laboratory. As summarized below in Table B 1 below, data analyzed in Aims 1 and 2 were collected only from the Psychology Clinic. However, Aim 3 analyzed combined data from the Psychology Clinic as well as the Research Laboratory. The Discussion section (beginning on page 110) describes in detail the caveats of combining data from the Psychology Clinic and Research Lab in Aim 3.

PAGE 135

135 Table B 1 . Breakdown of data s ources Aim Data Source (N) Samples Analyzed (N) Aim 1 Psychology Clinic (234) heterogeneous group of medical and psychological disorders (234) Aim 2 Psychology Clinic (122) ADHD (99) Clinical Control (23) Aim 3 Psychology Clinic (16) TBI (16) Research Lab (49) TBI (28) Orthopedic Injury (21)

PAGE 136

136 LIST OF REFERENCES APA (2000). Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM IV TR)., Washington, DC: American Psychiatric Association. Arbuckle, J. L. (2008). Amos 17.0 [computer software]. Spring House, PA: SPSS. Arnsten, A.F.T. (2006). Fundamentals of AttentionDeficit/Hyperactivity Disorder: circuits and Pathways., Journal of Clinica l Psychiatry, 2006 (suppl 8): pp.712. Atkinson, R.C., & Shiffrin, R.M. (1971). The control processes of short term memory. Scientific American, 8290. Baddeley, A. & Wilson, B. (1988). Frontal Amnesia and the Dysexecutive Syndrome. Brain & Cognition , 7: 212 – 230. Baddeley, A. (2000). Short term and working memory. In E. Tulving & F.I.M. Craik (Eds.), The Oxford Handbook of Memory , pp. 7792. New York: Oxford University Press. Bauer, R.M., Grande, L., & Valenstein, E. (2003). Amnesic Disorders. I n K.M. Heilman & E. Valenstein (Eds.). Clinical Neuropsychology , pp. 495 – 573. New York: Oxford University Press. Bennett, T.L., Davalos, D.B., Ho Turner, M., & Banz, B.C. (2011). Seizure Disorders. In S. Goldstein & C.R. Reynolds (Eds.), Handbook of Neurodevelopmental and Genetic Disorders in Children, pp. 460483. New York: The Guilford Press. Bollen, K. & Long, S. (1993). Testing structural equation models. Newbury Park: Sage. Borden, K,A., Burns, T.G., & O’Learly, S.D. (2007). A Comparison of Children with Epilepsy to an Ageand IQ matched control group on the Children’s Memory Scale, Child Neuropsychology, 12: pp. 165 172. Bowler, D.M., Gardiner, J.M., & Berthollier, N. (2004). Source Memory in Adolesc ents and Adults with Asperger’s Syndrome., Journal of Autism and Developmental Disorders, 34(5): pp. 533542. Broadbent, N.J., Clark, R.E., Zola, S., & Squire, L.R. (2002). The Medial Temporal Lobe and Memory. In L. Squire & D. Schacter (Eds.), Neuropsych ology of Memory, Third Edition , pp. 3 23. New York: The Guilford Press. Brooks, B.L., Iverson, G.L., Sherman, E.M.S., & Holdnack, J.A. (2009). Health children and adolescents obtain some low scores across a battery of memory tests. Journal of the Internat ional Neuropsychological Society, 15: 613 617.

PAGE 137

137 Brown, R.D. & Chiu, C.Y. P. (2006). Neural Correlates of Memory Development and Learning: Combining Neuroimaging and Behavioral Measures to Understand Cognitive and Developmental Processes., Developmental Neuropsychology, 29(2): 279 291. Brown, S.C. & Craik, F.I.M. (2000). Encoding and retrieval of information. In E. Tulving & F.I.M. Craik (Eds.), The Oxford Handbook of Memory , pp. 93 107. New York: Oxford University Press. Cahn, D.A. & A.C. Marcotte (1 995). Rates of forgetting in attention deficit hyperactivity disorder. Child Neuropsychology,1(2):pp. 158 163. Carreno, M., Donaire, A., Sanchez Carpintero, R. (2008). Cognitive Disorders Associated with Epilepsy: Diagnosis and Treatment., Neurologist , 14(6S): S26 S34. Corbett, B. & Gunther, J. (2011). Autism Spectrum Disorders. In S. Goldstein & C.R. Reynolds (Eds.), Handbook of Neurodevelopmental and Genetic Disorders in Children , pp. 228247. New York: The Guilford Press. Corbett, B.A., Constantine, L.J., Hendren, R., Rocke, D., & Ozonoff, S. (2009). Examining executive functioning in children with autism spectrum disorder., Psychiatry Research , 166: pp. 210222. Cohen, M. (1997). Children’s Memory Scale Manual . San Antonio: The Psychological Corpor ation. Craik, F.I.M. & Lockhart, R.S. (1972). Levels of Processing: A Framework for Memory Research. Journal of Verbal Learning and Verbal Behavior , 11i: 671 – 684. Dawson, G., Webb, S., Schellenberg, G.D., Dager, S., Friedman, S., Aylward, E., & Richards, T. (2002). Defining the broader phenotype of autism: Genetic, brain, and behavioral perspectivies. Develpmental and Psychopathology, 14: pp. 581611. de Haan, M., Mishikin, M., Baldewed, T., & VarghaKhadem, F. (2006). Human memory development and its dysfunction after early hippocampal injury. TRENDS in Neurosciences, 29(7): 374381. Delis, D.C., Kramer, J.H., Kaplan, E., & Ober, B.A. (1989). California Verbal Learning Test Children’s Version (CVLT C) , San Antonio, Texas: The Psychological Cor poration. Delis, D.C., Kramer, J.H., Kaplan, E., & Ober, B.A. (2000). California Verbal Learning Test Second Edition (CVLT II) , San Antonio, Texas: The Psychological Corporation.

PAGE 138

138 Dehn, M.J. (2010). Long Term Memory Problems in Children and Adolescents: A ssessment, Intervention, and Effective Instruction., Hoboken, NJ: John Wiley & Sons, Inc. Denckla, M. (1996). Biological Correlates of Learning and Attention What is Relevant to Learning Disability and AttentionDeficit Hyperactivity Disorder., Journal of developmental and behavioral pediatrics, 17(2): pp.114119. Diamond, A. (2006). The Early Development of Executive Functions. In E. Bialystok & F.I.M. Craik (Eds.), Lifespan Cognition: Mechanisms of Change, pp. 7095. Oxford: Oxford University Press. D oyle, A.E. (2006). Executive Functions in AttentionDeficit/Hyperactivity Disorder. Journal of Clinical Psychiatry, 67(suppl 8): pp.2126. Enns, J.T. & Trick, L.M. (2006). Four Modes of Selection. In E. Bialystok & F.I.M. Craik (Eds.), Lifespan Cognition: Mechanisms of Change, pp. 43 XX. Oxford: Oxford University Press. Felix, L. & Hunter, S.J. (2010). Pediatric aspects of epilepsy. In J. Donders & S.J. Hunter (Eds.), Lifespan Developmental Neuropsychology , pp. 359369. Cambridge: Cambridge University P ress. Field, A. (2009). Discovering Statistics Using SPSS, Third Edition ., London: SAGE. GoldmanRakic, P.S., Cools, A.R., & Srivastava, K. (1996). The Prefrontal Landscape: Implications of Functional Architecture for Understanding Human Mentation and the Central Executive [and Discussion]. Philosophical Transactions of the Royal Society, 351: 1445 1453. Goldstein, S. (2011). Attention Deficit/Hyperactivity Disorder. In S. Goldstein & C.R. Reynolds (Eds.), Handbook of Neurodevelopmental and Genetic Disorders in Children , pp. 131150. New York: The Guilford Press. Hannay, H.J., Howieson, D.B., Loring, D.W., Fischer, J.S., & Lezak, M. (2004). Neuropathology for Neuropsychologists. In M.D. Lezak, D.B. Howieson, and D.W. Loring (Eds.) Neuropsychological Assessment: Fourth Edition , pp. 157285. New York, New York: Oxford University Press. Hayes, A.F. (2009). Beyond Baron and Kenny: Statistical Mediation Analysis in the New Millenium. Communication Monographs , 76(4): pp. 408420. Heaton, S.C., Reader, S.K., Preston, A.S., Fennell, E.B., Puvana, O.E., Gill, N., & Johnson, J.H. (2001). The Test of Everyday Attention for Children (TEA Ch): Patterns of Performance in Children with ADHD and Clinical Controls. Child Neuropsychology , 7(4): pp. 351264.

PAGE 139

139 Hill, D.E., Ciesielski, K.T., Sethre Hofstad, L., Duncan, M.H., Lorenzi, M. (1997). Journal of Pe diatric Psychology, 22(6): pp. 861870. Hitch, G.J. (2006) Working Memory in Children: A Cognitive Approach. In E. Bialystok & F.I.M. Craik (Eds.), Lifespan Cognition: Mechanisms of Change, pp. 112127. Oxford: Oxford University Press. Yeates, K. O. (2 010). Traumatic Brain Injury, In K.O. Yeates, M.D. Ris, H.G. Taylor, & B.F. Pennington (Eds .). Pediatric Neuropsychology: Research, Theory, and Practice , pp.112 146, New York: The Guilford Press. Kail, R. (1984). The Development of Memory in Children, S econd Edition., New York: W.H. Freeman and Company. Kirkwood, M.W. (2010). Traumatic brain injury in childhood. In J. Donders & S.J. Hunter (Eds.), Lifespan Develpmental Neuropsychology , pp. 299313. Cambridge: Cambridge University Press. Korkman, M., Kirk, U., Kemp, S.L. (2007). NEPSY II. Clinical and Interpretative Scoring Manual. San Antonio, TX: Psychological Corporation. Kopelman, M.D. & Stanhope, N. (2002). Anterograde and Retrograde Amnesia Following Frontal Lobe, Temporal Lobe, or Diencephalic Lesions. In L. Squire & D. Schacter (Eds.), Neuropsychology of Memory, Third Edition, pp. 3546. New York: The Guilford Press. Kourakis, I.E., Katachanakis, N., Vlahonikolis, I.G., Paritsis, N.K. (2004). Examination of Verbal Memory and Recall Time in Chi ldren with Attention Deficity Hyperactivity Disorder, Developmental Neuropyschology, 26(2): pp. 565570. Krull, K.R. & Jain, N. (2010). Leukemia and lymphoma across the lifespan. In J. Donders & S.J. Hunter (Eds.), Lifespan Develpmental Neuropsychology , p p. 379391. Cambridge: Cambridge University Press. Lezak, M.D., Howieson, D.B., & Loring, D.W. (2004). Neuropsychological Assessment: Fourth Editi on , New York, New York: Oxford University Press. Manly, T., Anderson, V., NimmoSmith, I., Turner, A., Watso n, P., & Robertson, I.H. (2001). The Differential Assessment of Children’s Attention: The Test of Everyday Attention for Children (TEA Ch), Normative Sample and ADHD Performance. Journal of Child Psychology and Psychiatry, 42(8): pp. 10651081. Manly, T., Robertson, I.H., Anderson, V., Nimmo Smith, I. (1999). The Test of Everyday Attention for Children (TEA Ch)., England: Thames Valley Test Company.

PAGE 140

140 Marks, D., Trampush, J. & Chacko, A. (2010). Attention deficit hyperactivity disorder in children and adolescents. In J. Donders & S.J. Hunter (Eds.), Lifespan Devel o pmental Neuropsychology , pp. 83 96. Cambridge: Cambridge University Press. Martin, A. & Chao, L.L. (2001). Semant ic memory and the brain: structure and processes. Current Opinion in Neurobiology, 11: 194201. Max, J.E. (2005). Children and Adolescents. In J.M. Silver, T.W. McAllister, & S.C. Yudofsky (Eds.). Textbook of Traumatic Brain Injury , pp.477 508. Arlington, VA: American Psychiatric Publishing, Inc. Mayes, A.R. (2002). Exploring the Neural Bases of Complex Memory. In L. Squire & D. Schacter (Eds .), Neuropsychology of Memory, Third Edition, pp. 2434. New York: The Guilford Press. McDermott, K.B. and Buckner, R.L. (2002). Functional Neuroimaging Studies of Memory Retrieval. In L. Squire & D. Schacter (Eds.), Neuropsychology of Memory, Third Edition, pp. 1661173. New York: The Guilford Press. Miller, G.A. (1956). The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological Review, 63 (2): pp. 81 – 97. Moscovitch, M. (1994). Memory and working with memory : Evaluation of a component process model and comparisons with other models. In D. Schacter & E. Tulving (Eds.) Memory Systems ., Cambridge, MA: The MIT Press: pp. 407 – 451. Motamedi, G. & Meador, K. (2003). Epilepsy and cognition. Epilepsy & Behavior , 4: pp. S25 S38. Mottram, L. & Donders, J. (2006). Cluster Subtypes on the California Verbal Learning Te st Children’s Version After Pediatric Traumatic Brain Injury. Developmental Neuropsychology, 30(3): 865883. Nagel, B.J., Palmer, S.L., Reddick, W.E., Glass, J.O., Helton, K.J., Wu, S., Xiong, X., Kun, L.E., Gajjar, A., & Mulhern, R.K. (2004). Abnormal Hi ppocampal Development in Children with Medulloblastoma Treated with Risk Adapted Irradiation. American Journal of Neuroradiology, 25: pp. 15751582. Nyberg, L. (2002). Where Encoding and Retrieval Meet in the Brain. In L. Squire & D. Schacter (Eds.), Neur opsychology of Memory, Third Edition, pp. 193 215. New York: The Guilford Press. O’hearn, K., Asato, M., Ordaz, S., & Luna, B. (2008). Neurodevelopment and executive function in autism. Development and Psychopathology,20: pp. 11031132.

PAGE 141

141 Ornstein, P.A., Haden, C.A., & Elischberger, H.B. (2006) Children’s Memory Development: Remembering the Past and Preparing for the Future. In E. Bialystok & F.I.M. Craik (Eds.), Lifespan Cognition: Mechanisms of Change, pp. 143161. Oxford: Oxford Univers ity Press. Paivio, A. (1991). Dual Coding Theory: Retrospect and Current Status., Canadian Journal of Psychology, 45(3): pp. 255287. Paller, K.A. (2002). Cross Cortical Consolidation as the Core Defect in Amnesia. In L. Squire & D. Schacter (Eds.), Neuropsychology of Memory, Third Edition, pp. 8387. New York: The Guilford Press. Patrikelis, P., Angelakis, E., & Gatzonis, S. (2009). Neurocognitive and behavioral functioning in frontal lobe epilepsy: A review, Epilepsy & Behavior, 14: pp. 1926. NCS Pearson, Inc. (1997). Standardization data from the Children’s Memory Scale (CMS). Used with permission. All rights reserved. Posner, M.I. & Petersen, S.E. (1990). The Attention System of the Human Brain. Annual Reviews in Neuroscience, 13: 2542. Prea cher, K.J. & Hayes, A.F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Method, 40(3): pp. 879891. Rodgers, J., Britton, P.G., Morris, R.G., Kernahan, J., & Craft, A.W. (1992). Memory after treatment for acute lymphoblastic leukemia. Archives of Disease in Childhood, 67: pp. 266 268. Rugg, M.D., Herron, J.E., & Morcom, A.M. (2002). Electrophysiological Studies of Retrieval Processing. In L. Squire & D. Schacter (Eds .), Neuropsychology of Memory, Third Edition, pp. 154165. New York: The Guilford Press. Rzezak, P., Guimaraes, C., Fuentes, D., Guerreiro, M.M., & Valente, K.D.R. (2011). Episodic and semantic memory in children with mesial temporal sclerosis. Epilepsy & Behavior, 21: pp. 242 247. Sattler, J.M. (2008). “Chapter 7. Historical Survey and Theories of Intelligence.” In Assessment of Children: Cognitive Foundations 5th edition, pp.215 244. La Mesa, CA: Jerome M. Sattler, Publisher, Inc. Schneider, W. & Pr essley, M. (1989). Memory Development Between 2 and 20., New York: Springer Verlag New York Inc. Shimamura, A.P. (2002). Relational Binding Theory and the Role of Consolidation in Memory Retrieval. In L. Squire & D. Schacter (Eds.), Neuropsychology of Mem ory, Third Edition, pp. 6172. New York: The Guilford Press.

PAGE 142

142 Sowell, E.R., Delis, D., Stiles, J., & Jernigan, T.L. (2001). Improved memory functioning and frontal lobe maturation between childhood and adolescence: A structural MRI study., Journal of the International Neuropsychological Society , 7: pp. 312 – 322. Stancin, T. & Alyward, G.P. (2008). Chapter 7c. Assessment of Development and Behavior. In M. L. Wolraich et al. Developmental Behavioral Pediatrics: Evidence and Practice, pp. 144 177. Philadelphia, PA: Mosby Elsevier. Verfaellie, M. & Keane, Margaret M. (2002). Impaired and Preserved Memory Processes in Amnesia. In L. Squire & D. Schacter (Eds.), Neuropsychology of Memory, Third Edition, pp. 3546. New York: The Guilford Press. Stuss, D.T., & Benson, D.F. (1984). Neuropsychological studies of the frontal lobes. Psychological Bulletin , 95(1): 3 28. Verfaellie, M. & Keane, Margaret M. (2002). Impaired and Preserved Memory Processes in Amnesia. In L. Squire & D. Schacter (Eds.), Neuropsychology of Memory, Third Edition, pp. 3546. New York: The Guilford Press. Wagner, A.D. (2002). Cognitive Control and Episodic Memory: Contributions from Prefrontal Cortex. In L. Squire & D. Schacter (Eds .), Neuropsychology of Memory, Third Edition, pp. 174192. New York: The Guilford Press. Waugh & Norman (1965). Primary Memory. Psychological Review, 72(2): 89 104. Wecshler, D. (1999). The Wecshler Abbreviated Scale of Intelligence (WASI), San Antonio, Texas: The Psychological Corporation. Wecshler, D. (1999). The Wecshler Abbreviated Scale of Intelligence (WASI), San Antonio, Texas: The Psychological Corporation. Wecshler, D. (2003a). The Wecshler Intelligence Scale for Childrenfourth edition (WISC IV), San Antonio, Texas: The Psychological Corporation. W ecshler, D. (2003b). WISC IV Technical and Interpretive Manual. San Antonio., Texas: The Psychological Corporation. Wecshler, D. (2009). Wechsler Memory Scale – Fourth Edition (WMS IV) Administration and Scoring Manual. San Antonio., Texas: The Psychologi cal Corporation. Westerveld, M. (2010). Childhood Epilepsy. In K.O. Yeates, M.D. Ris, H.G. Taylor, & B.F. Pennington (Eds .). Pediatric Neuropsychology: Research, Theory, and Practice ,pp. 71 91 . New York: The Guilford Press.

PAGE 143

143 Williams, P.E., Weiss, L.G., & Rolfhus, E.L. (2003). WISC IV Technical Report #2 Psychometric Properties. In The Psychological Corporation’s WISC IV Technical Manual #2. Wilson, B. (2009). Memory Rehabilitation: Integrating Theory and Practice, New York, New York: The Guilford Press. Wolfe, J.M. & Paterson, S.J. (2010). Lifespan aspects of PDD/autism spectrum disorders (ASD). In J. Donders & S.J. Hunter (Eds .), Lifespan Develpmental Neuropsychology , pp. 239250. Cambridge: Cambridge University Press. Zola Morgan, S. & Squire, L.R. (1993). Neuroanatomy of Memory. Annual Reviews in Neuroscience, 16: 547 5463.

PAGE 144

144 BIOGRAPHICAL SKETCH Liz abeth L. Jordan will receive her Ph.D. from the Department of Clinical and Health Psychology in Summer 2014. She received her Bachelor of Science degree in Brain and Cognitive Sciences from the Massachusetts Institute of Technology and her Master of Science degree in Clinical and Health Psychology from the University of Florida. She is currently completing her predoctoral internship in pediatric neuropsy chology and pediatric psychology at Kennedy Krieger Institute / Johns Hopkins Hospital. Lizabeth’s current graduate research is supervised under the mentorship of Dr. Shelley Heaton at the University of Florida. The overall goal of Lizabeth’s graduate r esearch is to understand memory dysfunction in pediatric patients