MEMORY FUNCTIONING AFTER PEDI ATRIC TRAUMATIC BRAIN INJURY By MARIE D. SCHRODER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Marie D. Schroder
iii ACKNOWLEDGMENTS I would like to thank Dr. Shelley Heaton fo r her help and guidance on this project and for her generosity in allowing me to use la b resources. I would also like to thank the members of my committee, Dr. Michael Mars iske, Dr. James Rodrigue, and Dr. David Janicke, for their comments on my thesis. Finally, I would like to thank my fiancÃ©, family, and friends for their s upport during this process.
iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................vi ii CHAPTER 1 BACKGROUND AND SIGNIFICANCE....................................................................1 Pediatric Traumatic Brain Injury..................................................................................1 Incidence................................................................................................................1 Causes....................................................................................................................2 Classification.........................................................................................................2 Neuropathological and Neur obehavioral Sequelae...............................................3 Memory......................................................................................................................... 5 Models of Memory................................................................................................5 Neuroanatomical Structures of Memory...............................................................7 Development of Memory......................................................................................8 Assessment of Memory during Childhood..........................................................11 Memory after Pediatric Traumatic Brain Injury.........................................................12 Memory after Severe Traumatic Brain Injury.....................................................12 Memory after Mild Traumatic Brain Injury........................................................15 Less Frequently Studied Asp ects of Memory after TBI......................................17 Retention......................................................................................................17 Meaningfulness of material..........................................................................18 Other Factors Influencing Memory after TBI.....................................................20 Gender..........................................................................................................20 Age at injury.................................................................................................20 Time since injury..........................................................................................21 Purpose of Current Study............................................................................................21 2 METHODS.................................................................................................................23 Recruitment and Informed Consent............................................................................23 Participants.................................................................................................................24
v Assessment Procedures...............................................................................................28 Intellectual Assessment.......................................................................................28 Memory Assessment...........................................................................................29 3 RESULTS...................................................................................................................32 Intellectual Functioning..............................................................................................32 Memory Functioning..................................................................................................33 4 DISCUSSION.............................................................................................................39 Memory Functioning after Mild TBI..........................................................................39 Memory Functioning after Severe TBI.......................................................................40 Limitations..................................................................................................................45 Summary and Future Directions.................................................................................46 LIST OF REFERENCES...................................................................................................48 BIOGRAPHICAL SKETCH.............................................................................................53
vi LIST OF TABLES Table page 1 Classification System fo r Level of TBI Severity.....................................................24 2 Demographic Characteristics of Participants...........................................................27 3 Mechanism of Injury within the TBI Groups...........................................................27 4 ChildrenÂ’s Memory Scale: Core Battery..................................................................30 5 Descriptive Statistics for Primar y and Calculated Memory Indexes.......................34 6 ChildrenÂ’s Memory Scale Indexes a nd Limitations of Interpretation......................44
vii LIST OF FIGURES Figure page 1 Mean performance on ChildrenÂ’s Memory Scale indexes.......................................35 2 Mean performance on ChildrenÂ’s Memory Scale composite retention rate.............36 3 Mean performance on ChildrenÂ’s Memory Scale subtests grouped by meaningful versus relatively abstract content..........................................................37
viii Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MEMORY FUNCTIONING AFTER PEDI ATRIC TRAUMATIC BRAIN INJURY By Marie D. Schroder May 2005 Chair: Shelley C. Heaton Major Department: Clini cal and Health Psychology Memory impairments are common follo wing childhood traumatic brain injury (TBI), with severe injuries associated with more extensive impairment. Findings have been mixed, however, regarding the pattern of impairment acro ss specific dimensions of memory and types of material. The current study examined memory performance using the ChildrenÂ’s Memory Scale (CMS) in a pediatric TBI sample. Children ages 6-16 with mild TBI ( n = 9), severe TBI ( n = 16), and non-TBI controls ( n = 13) were evaluated within one year of injury. The groups were comparable in terms of age, gender, ethnicity, and intellectual func tioning. On the CMS, the severe TBI group performed significantly worse than both the mild TBI and control groups on the Learning, General Memory, and Recognition indexes. In contra st, the three groups demonstrated similar rates of retention, which averaged above 90% for each group . This latter finding suggests that children sustaining severe TBI have probl ems with the initial acquisition of material rather than its maintenance in memory. The severe TBI group was also noted to perform significantly worse than both the mild TBI a nd control groups on memory subtests with
ix relatively abstract content, but not on subtests involving more meaningful material. In summary, these findings indicate that memory skills following pediatric TBI vary as a function of injury severity, memory dimensions, and type of material. In a broader sense, these findings have implications for the design of interventions that target specific aspects of memory deficits in th e pediatric TBI population.
1 CHAPTER 1 BACKGROUND AND SIGNIFICANCE Traumatic brain injury (TBI) is a major public health problem and a leading cause of disability in children (Yeat es, 2000). It is important to study pediatric TBI because the neuropsychological sequelae may differ fr om adult TBI (Verger et al., 2000). Additionally, cognitive impairments after TBI can have serious implications for childrenÂ’s academic and developmental progres s (Fay et al., 1994). Pediatric TBI may result in various neurobehavioral conseque nces, and one of the most common outcomes is memory impairment (Yeates). Pediatric Traumatic Brain Injury Incidence The average incidence rate of TBI is approximately 180 per 100,000 children each year (Kraus, 1995). Estimates are that between 76-90% of injuries are mild in nature, 710% are moderate, and 5-13% are severe (K raus; Lescohier & DiScala, 1993). These figures likely underestimate the actual fre quency of pediatric head injury, as many individuals with milder injuries do not seek or receive medical treatment. Males are more likely to sustain a head injury than females, with gender ratio estimates ranging from 1.5:1 to 2:1 (Kraus; Lescohier & DiScala) . The incidence of TBI also varies with age. The overall incidence gradually in creases from age 5 until early adolescence and then displays rapid growth in adolescence (Kraus).
2 Causes Transportation-related accide nts and falls account for the majority of TBI. These sources typically account for be tween 75-80% of TBI in publis hed studies (Kraus, 1995). The distribution of causes vari es significantly by age. Y ounger children are most likely to be injured in falls (Lesc ohier & DiScala, 1993). In olde r children, however, most head injuries are caused by sports and recreation accide nts or pedestrian/bicycle collisions with motor vehicles. Transportation-related tr auma accounts for an increasing proportion of TBI in older children, with adolescents most likely to be injured in motor vehicle accidents (Yeates, 2000). Classification Children sustaining TBI are frequently clas sified on the basis of injury severity, which can range from mild to severe. Desp ite the use of severi ty categorizations, a significant amount of individua l variability remains within injury severity groups. Furthermore, there is poor consensus on the ex act criteria that should be used to classify severity. Commonly used measures of injury severity include the Glasgow Coma Scale, duration of loss of consciousness, length of posttraumatic amnesia, and MRI/CT scans. The Glasgow Coma Scale (Teasdale & Jennett, 1974), which is grounded in physiological assessment, is the most wide ly used measure of injury severity. Classification is based on a clinicianÂ’s ra ting of a patient in three domains: 1) eye opening Â– ranges from none to spontaneous, 2) motor response Â– ranges from none to obeys commands, and 3) verbal response Â– ranges from none to oriented. The total score for this scale ranges from 3 (no respon se across all three domains) to 15 (no abnormalities in areas of assessment). Using this scale, severity is typically classified
3 based on the total score as: Mild (13-15); M oderate (9-12); and Se vere (8 or less) (Lescohier & DiScala, 1993). Duration of impaired consciousness is typically defined as the amount of time following the injury until the child is able to follow simple commands (e.g., Â“open your eyesÂ”). An injury is usually considered se vere when loss of consciousness lasts more than 24 hours (Yeates, 2000). Mo st children sustaining a mode rate to severe TBI also experience fluctuations in arous al and orientation, as well as confusion and memory loss (Yeates). This period, which is referred to as posttraumatic amnesia, is also used as an index of injury severity. The ChildrenÂ’s Orientation a nd Amnesia test (Ewing-Cobbs, Levin, Fletcher, Miner, & Eisenberg, 1990) can be used to assess posttraumatic amnesia, and scores on this measure obtained during th e acute stages of injury are predictive of memory function at six months and one year after TBI (Ewing-Cobbs et al., 1990). Neuroimaging can be used to characterize in jury severity on the basis of the brainÂ’s structural integrity. Computed tomography im aging (CT) is frequently used to detect complications, such as swelling, blee ding, and skull fractures after TBI. Magnetic resonance imaging (MRI) has also been used to detect complications, such as focal brain lesions. One MRI study showed the presence of lesions in the majority of children sustaining severe TBI as defined by conventio nal measures, such as the Glasgow Coma Scale (Levin et al., 1993). Neuropathological and Ne urobehavioral Sequelae The consequences of pediatric TBI incl ude both neuropatholog ical as well as neurobehavioral sequelae. The injuries resul ting from TBI can be classified as either primary or secondary. Primary injuries, such as skull fractures, are a direct result of the trauma, while secondary injuries, such as swel ling, are an indirect result (Yeates, 2000).
4 Usually, head injury involves both linear a nd rotational acceleration. Linear acceleration may result in skull fractures or contusion at th e site of impact. In rotational injuries, the brain continues to move afte r the skull is stopped by the im pact. This movement may cause focal contusions within the brain ti ssue by damaging blood vessels and may also lead to straining or shearing of white ma tter nerve fibers, resulting in diffuse axonal injury (Yeates). Regarding recovery after TBI, Kraus ( 1995) reported that between 75-95% of children sustaining TBI displayed Â“good rec overy,Â” and between 1-3% showed severe disability as rated by the Glasgow Outcome S cale (Jennett & Bond, 1975; cited in Kraus). However, a good recovery is not synonymous with full recovery and may still be associated with impairments. Such impairments could include transient or permanent problems in cognitive or physical domains. In general, studies have indicated that pe diatric TBI, especially when severe in nature, may be associated with impairment s in alertness and orientation, intellectual functioning, language skills, nonve rbal skills, attention and memory, executive functions, processing speed, corticosensory and motor skills, academic achievement, adaptive functioning, and behavioral adjustment (L ord-Maes & Obrzut, 1996; Yeates, 2000). Prospective, longitudinal studies show that children sustaining TBI display significant improvement in intellectual functioning, me mory, and other neurobe havioral functions over time, with the most rapid improvements occurring in the first one to two years postinjury (Yeates et al., 2002). However, for some children, particularly those who survive a severe TBI, memory impairments can pers ist and have a signifi cant impact on their
5 lives. The next sections provide a more de tailed examination of memory, which is the neurobehavioral outcome of TBI that is of interest in the current study. Memory Memory is a multidimensional cognitive constr uct, rather than a unitary ability. Current models of memory describe multi-co mponent systems, whereby information is registered, encoded, stored, and made available for future retrieval. Models of Memory Memory models generally recognize a dist inction between a short-term and a longterm store (Atkinson & Shiffrin, 1968). The di stinction is primarily based on duration of storage and capacity for the amount of inform ation to be stored. Short-term memory refers to temporary storage of limited in formation. Long-term memory, on the other hand, refers to more stable or permanent storage, which is relatively unaffected by capacity limitations or decay processes. The systems have traditionally been viewed as operating serially, such that short-term memory is the pr ecursor to long-term memory, although this idea has been challenged (Squire, Knowlton, & Musen, 1993). Long-term memory has been subdivided into nondeclarative and declarative memory systems (Squire et al., 1993). Declar ative memory can be further classified as semantic or episodic in nature (Squire et al .). Semantic memory concerns memory for facts and concepts and refers to an organi zed body of general knowledge, while episodic memory involves information that is situatio nand context-specifi c (Sohlberg & Mateer, 2001). Most neuropsychological memory tests primarily measure declarative, episodic memory. These tests commonly involve tasks such as learning and recalling lists of words or geometric patterns. Conversel y, nondeclarative memory typically involves learning that takes place without consci ous awareness (e.g., through conditioning).
6 Nondeclarative memories can be expresse d through performance, rather than by conscious recollection, and are manifested, for example, in learned motor skills or habits (Squire et al.). More recently, models of short-term memo ry have been supplemented or replaced by the concept of working memory. Working memory is defined as temporary and limited in capacity (Baddeley, 1992). However, in contrast to the conceptualization of passive short-term memory, working memory is viewed as an active process that both stores and manipulates information. Working memory refers to a more complex system that can direct attention, impl ement strategies, and control retrieval. As outlined in BaddeleyÂ’s model, working memory can be divided into three components; a central executive that is involved in attention contro l and manages the other systems. These two systems consist of 1) the phonological loop, wh ich stores and rehearses language-based information, and 2) the visuospatial sketch pa d, which manipulates visual images. It is important to note that, within BaddeleyÂ’s mode l, working memory would be an important predictor of individual differe nces in learning or the initia l acquisition of information. Memory can be characterized not only by various components, but also by the series of processes it entails . An initial stage of memory is encoding or acquisition, which describes the level of analysis perfor med on material and refers to the process whereby information is transformed into a mental representation (Sohlberg & Mateer, 2001). This stage is often referre d to as Â“learning.Â” Storage refers to the transfer of a transient memory to a form or location in the brain for retention (Sohlberg & Mateer). It is the storage of information that forms th e crux of what most think of as memory. However, measuring the degree of storage inherently involv es a retrieval process.
7 Retrieval or recall refers to searching for or activating existing memory traces and bringing the information into conscious awar eness (Sohlberg & Mateer). Two formats for evaluating the degree of storage and the efficiency of retrieval are free recall and recognition. These are closely linked functions of declarat ive memory, although there is some evidence that recognition may reduce the demands on retrieval processes and may depend on increased perceptual fl uency (Squire et al., 1993). Neuroanatomical Structures of Memory Memory relies on a complex system, involvi ng a number of brain structures. The anatomic basis of declarative memory includes structures in the medi al temporal lobes, diencephalic structures, parts of the basal forebrain, and the frontal lobes (Cohen, 1997; Frackowiak et al., 2004; Waxman, 2000). Damage to any of these structures may result in substantial memory impairment. Medial temporal lobe st ructures, particularly th e hippocampal formation and entorhinal cortices, are critic al for short-term and long-term memory (Scoville & Milner, 1957; Sohlberg & Mateer, 2001; Waxman, 2000) . Both experimental and clinical observations suggest that the encoding of long-term memory involves the hippocampus, adjacent cortex in the medial temporal lobes, and portions of the medial thalamus, with information likely stored in the higher-order association areas of the cerebral cortex (Sohlberg & Mateer; Waxman). Subcortical projections from the hippocampal formation move via the fornix, a compact fiber bundle (M artin, 2003). One projection of the fornix extends to the mammillary bodies of the hypothalamus and is part of an anatomical loop termed the Papez circuit, which is important in memory consolidation (Waxman). In sum, the hippocampus and other structures deep in the temporal lobes have an essential role in the encoding and consolidation of memory.
8 The frontal lobes are also important in memory processes. In neuroimaging studies, frontal activation has been shown to be a predictor of subsequent memory performance (Frackowiak et al., 2004). Areas th at are particularly re levant include left and right anterior and dorsolat eral prefrontal cortex, whic h may be involved in working memory and executive processing (Frackowiak et al.). The frontal cortex is implicated in retrieval processes, the ev aluation of retrieved inform ation, and memory monitoring (Frackowiak et al.). Lesions in the frontal lobes are associated with problems employing effective encoding and retrieval strategies, great er susceptibility to interference, difficulty monitoring recall for redundant or incorrect information, impaired decision processes involved in recognition memory, and poor recall for the source of information (Frackowiak et al.; Sohlberg & Mateer, 2001). In summary, areas of the temporal lo bes and frontal lobes form the main neuroanatomical basis for decl arative memory. Given the lo cation of these structures within the brain, it is not surprising that they are susceptible to damage from TBI. The skull is formed with irregularities or bony protrusions at the front al and temporal poles, which make the frontal and temporal lobes particularly susceptible to damage from acceleration-deceleration injuries (Cohen, 1997). Studies of pediatric TBI have indicated that memory impairments are associated with frontal lobe lesions (D i Stefano et al., 2000; Levin et al., 2000, 1993), left te mporal lobe damage (Levin & Eisenberg, 1979), and right hemisphere lesions (Woodward & Donders, 1998). Development of Memory The brain regions involved in human memory are not fu lly developed at birth. Hence, memory skills improve throughout ch ildhood, as the neuroanatomical structures of memory mature. Declarative memory, which draws broadly on limbic and cortical
9 structures, emerges during in fancy (Nelson, 1995) and continues to develop throughout childhood (Anderson & Lajoie, 1996; Pressley & Schneider, 1997). The ability to register information, learn, and remember all im prove with age, with the nature of this progress consistent with ongoing myelinati on and frontal lobe maturation (Anderson & Lajoie). Memory development can be conceptu alized as resulting from increases in 1) basic memory capacities and mechanisms, 2) the declarative knowledge base, 3) the use of memory strategies, and 4) metamemory. One possibility for improvement in memory could be that capacity increases. While it is possible that absolute differences in actual storage capacity exist at different ages, a number of theorists have attributed developmental improvements in memory to increases in functional capaci ty (Bjorklund, 2000; Kail, 1990). Research in this area suggests that processing resour ces and efficiency both incr ease with age. Processing resources refer to the amount of mental Â“effort Â” that a person can devot e to a task, and the efficiency of processing refers to how qui ckly and with what degree of elaboration information can be processed (Bjorklund; Kail). Increases in the breadth and accessibility of the knowledge base also contribute to developmental improvements in memory (Bjorklund, 2000). Even young children rely on already present knowledge, in the form of fr ameworks or schemas, to facilitate the encoding of information (Pressley & Schneider, 1997). Retrieval is also generally more efficient when more pathways or associati ons exist to access the information (Kail, 1990). These connections, in the form of new neuronal pathways, form through the natural course of development and thr ough environmental experiences. Also, an
10 increased knowledge base contributes to mo re efficient execution of strategies (Kail; Pressley & Schneider). With development, it is clear that the st rategies involved in learning and retrieval improve tremendously (Spreen, Risser, & Edge ll, 1995). The deliberate and spontaneous use of memorization strategies (e.g., rehear sal, organization, and elaboration) emerges during childhood and continues to develop through adolescence (Kail, 1990; Levin et al., 1988). Spontaneous rehearsal is first seen w ith some regularity at approximately 7 years of age (Kail), and the majority of 10-year-olds use rehearsal to learn serial lists (Pressley & Schneider, 1997). Some of the more co mplex strategies to develop include organization, which is found beginning around age 10, and elaboration, which is most likely to be observed in adolesce nts (Pressley & Schneider). The capacity for metamemory also expands during childhood (Pressley & Schneider, 1997). Metamemory refers to the sk ills used to evaluate the difficulty of a memory task, as well as the ability to m onitor progress towards a memory goal (Kail, 1990). Rudimentary metamemory is present in the preschooler, develops to a more complete level by age 11 or 12 years, and continues to mature through adolescence and adulthood. Pressley and Schne ider report that metamemory serves as a significant predictor of memory performance beginning around age 8 years. In summary, developmental increases in memory functioning may be observed over the course of childhood. These improveme nts may be attributed to a number of factors, including neuroanatomical developm ent, extensive growth of the knowledge base, increases in the speed and efficiency of information processing, the development of metamemory, and increased strategy use. Strategy use is uncommon before 6 years of
11 age, develops more fully between 7-11 years of age, and encompasses a flexible and broad repertoire in adolescen ce (Kail, 1990). Memory skill s required for complex tasks undergo continued development into adolescen ce, and are often impa ired in children who sustain a TBI while these skills are developing (Levin et al., 1988). Assessment of Memory during Childhood While studies have explored memory impa irments after pediatric TBI, inherent limitations of the measures that were used have limited the conclusions that can be drawn. Many measures are just downward extensions of instruments used to study memory in adults. Some of the early m easures examined only isolated aspects of memory functioning. In the 1990Â’s, measur es were first published that had been designed specifically to study me mory in children and that were more comprehensive in nature. Two of the most widely used measures ar e the California Verbal Learning Test Â– ChildrenÂ’s Version (CVLT-C; Delis, Kramer, Kaplan, & Ober, 1994; cited in Roman et al., 1998) and the Wide Range Assessment of Memory and Learning (WRAML; Sheslow & Adams, 1990; cited in Farmer et al., 1999). These comprehensive measures enable the comparison of different aspects of memory, although they still have limitations. For example, the CVLT-C focuses specifically on verbal memory. While the WRAML assesses memory in both auditory-verbal and visual domains, it does not provide standardized global indices for delayed r ecall and recognition. Co mparing performance on recognition versus free recall can provide va luable information on whether deficits are primarily in acquisition and storage or due to impaired retrieval. The current study proposes to use a relative ly new measure, the ChildrenÂ’s Memory Scale (CMS; Cohen, 1997), to explore differe nt dimensions of memory. The CMS
12 includes standardized index scores for learning (initial acquisition), general memory (immediate and delayed recall), and delaye d recognition, and asse sses memory within auditory-verbal and visual domains. A small study in the test manual illustrated the scaleÂ’s sensitivity to memory impairments after pediatric TBI. Memory after Pediatric Tr aumatic Brain Injury Memory impairment is a frequent cons equence of pediatri c TBI (Yeates, 2000). Memory problems are a prominent concern for children sustaining TBI who return to school, and research indicates that children w ith severe TBI are at high risk for problems in the acquisition of new academic skills (Fay et al., 1994; Jaffe et al., 1992). Additionally, memory impairment s may be a difficult domain to rehabilitate (Lord-Maes & Obrzut, 1996). These factors make the study of memory after TBI critical. While it is well-established that children with more seve re head injuries tend to experience more extensive memory impairments (e.g., Catroppa & Anderson, 2002), much is left to be learned about the nature and extent of im pairment across levels of severity. Memory after Severe Traumatic Brain Injury Children with severe TBI have been show n to perform more poorly than children with mild TBI and controls across a variety of memory tasks. With regard to memory as a multidimensional construct, aspects that may be impaired after childhood TBI include learning, delayed recall, and recognition (Farmer et al., 1999; Roman et al., 1998). As described earlier, memory is a step-wis e process, with initial learning forming the foundation for subsequent memory performa nce. Impaired learning has consistently been demonstrated after severe TBI (e.g., Farmer et al., 1999). Children sustaining severe TBI have shown impairments on both ve rbal learning (Bassett & Slater, 1990; Di Stefano et al., 2000; Fay et al., 1994; Le vin et al., 1994, 1993; Ma ssagli et al., 1996;
13 Roman et al., 1998; Yeates, Blumenstein, Patterson, & Delis, 1995) and spatial learning (Ewing-Cobbs et al., 1990; Lehnung et al., 2003). Impairments have also been reported for learning that takes place over multiple tria ls, despite the successive opportunities to encode information and to develop and im plement strategies (Catroppa & Anderson, 2002; Ewing-Cobbs et al., 1990; Levin, Eise nberg, Wigg, & Kobayashi, 1982; Levin et al., 1988). Fay and colleagues examined a number of outcomes following pediatric TBI and suggested that deficits in a variety of cognitive domains combine to create a pervasive disability in learning. Retrieval or recall refers to searching fo r or activating existi ng memory traces. Recall can be assessed both immediately af ter learning and after a delay, and children with severe TBI have demonstrated impairm ents on both immediate and delayed recall (Lowther & Mayfield, 2004). For example, a number of studies have demonstrated impaired immediate and delayed recall for wo rd lists (Jaffe et al., 1992; Levin et al., 1993, 1982; Roman et al., 1998) and impaire d recall for stories both immediately (Catroppa & Anderson, 2002) and af ter a delay (Donders, 1993). In contrast to studies utili zing verbal tasks, recall for visual-spatial information has been shown to be impaired in short-term tasks, though recall did not deteriorate after a delay (Catroppa & Anderson, 2002; Donders, 1993; Lowther & Mayfield, 2004). Farmer and colleagues (1999) indicated that partic ularly poor performance on tasks requiring visual scanning and immediate visual reca ll is shown after severe TBI, paralleling intellectual weaknesses seen on Performance IQ measures. In summary, research indicates that children with severe TBI te nd to be impaired across most measures of recall, though intact recall has b een demonstrated on some visual/design memory tasks.
14 Some researchers posit that the impaired learning and recall in pediatric TBI may result from poor or inefficient use of stra tegies (Catroppa & Anderson, 2002; Farmer et al., 1999). Harris (1996) demonstrated that children with severe TBI tended to use inefficient, passive rehearsal strategies, wh ile children with mild TBI and controls utilized more active rehearsal strategies. She reported data showing that severely injured participants primarily used repetition, while children with mild injuries used rehearsal strategies including both repetition and organization of mate rial. In contrast, the control participants used elaborative rehearsal and verbal-visual association in addition to the more basic techniques utilized by the TBI groups. It has also been suggested that child ren with severe TBI may not process information deeply or elaborately enough, leading to poor recall (Donders, 1993). For instance, children with TBI may be more lik ely to rely on the temporal order of items during retrieval than controls (Vakil, Blach stein, Rochberg, & Vardi, 2004). During recall, children with TBI may al so tend to produce more intrusions (i.e., items that were not included in the stimuli to be learned) , which could be evid ence of poor monitoring (Levin et al., 1993; Vakil et al.; Yeates et al., 1995). However, these findings have not always been consistent, with some studies in dicating normal levels of intrusions (Roman et al., 1998). In conclusion, ineffici ent production and/or implementation of memorization strategies are likely to cont ribute to the severity effects observed on learning and retrieval tasks. Recogniti on, however, may be less dependent on active strategy use by the child. During recognition tasks, child ren are presented with stimuli from the memory task imbedded in non-learned distracters. In this paradigm, children are asked to correctly
15 identify (or Â“recognizeÂ”) the previously pr esented material and reject the incorrect distracters. On recognition test s, children with severe TBI tend to make fewer Â“hitsÂ” (i.e., true positive/correct id entification) and more false positive responses (i.e., incorrectly saying material was from that in itially learned). This patter n of performance is indicative of poor recognition skills and may also allu de to a metamemory or monitoring deficit. Children sustaining severe TBI have shown impaired recognition for verbal material relative to controls (Fay et al., 1994; Massagli et al., 1996; Roman et al., 1998; Yeates et al., 1995) and relative to children sustaining mild TBI (D i Stefano et al., 2000). They have also demonstrated poorer performance than children with mild TBI or controls on a short-term visual recognition memory ta sk (Hannay & Levin, 1988; Levin et al., 1982 ) . Memory after Mild Traumatic Brain Injury In comparison to severe TBI, there are relati vely few studies of pe diatric mild TBI. Design issues in studies of mild TBI furt her limit our understanding of this group. A number of researchers have reported that ch ildren with mild TBI show memory skills that are similar to controls (Bassett & Slater, 1990 ; Max et al., 1999; Roma n et al., 1998). For example, children sustaining mild TBI and or thopedic controls have been reported to show comparable performance on a comprehe nsive memory battery at 1 week and 3 months post-injury (Ponsford et al., 1999). Pediatric mild TBI samples have also been reported to perform similarly to controls on tests assessing learning (Levin et al., 2000; Roman et al.) and recall (Bassett & Slater). In contrast, other studies ha ve indicated subtle impairments in children sustaining mild TBI. One of the first studies to s uggest problems after mild TBI was conducted by Levin & Eisenberg (1979). They reported th at approximately one-four th to one-fifth of adolescents who received mild injuries show ed impairments on composite measures of
16 memory. Some research has indicated that children sustaining mild TBI have specific difficulty learning new material. One study repo rted that adolescents sustaining mild TBI displayed poor initial learni ng, with improvement to the performance level of controls after repeated exposure to the material (Ba ssett & Slater, 1990). When assessed acutely after injury, another study show ed that children with mild TBI performed slightly below age expectations for verbal learning (C atroppa & Anderson, 2002). In the nonverbal domain, specific problems in learning asso ciations between sounds and symbols and in visual reproductions have also be en reported (Farmer et al., 1999). While some studies of mild TBI have indi cated subtle deficits in learning, others have pointed to problems in retrieval. Yeat es and colleagues (1995) showed that children with mild TBI did not differ from controls in the rate or manner in which they learned a word list. The mild TBI group showed comp arable performance on learning trials and subsequent recognition, but reca lled proportionally fewer words after a delay. Recall of stories has also been shown to be impaired after mild TBI relative to controls (Anderson, Catroppa, Morse, Haritou, & Rosenthal, 2001; Catroppa & Anderson, 2002). In general, studies on mild TBI in chil dhood suggest relatively intact memory, with some evidence of subtle deficits in learning and retrieval (e.g., Farmer et al., 1999, Yeates et al., 1995). However, findings are inconsis tent across studies, and potential reasons for this lack of uniformity include: 1) inconsistent criteria for mild TBI, 2) different methods of recruitment, 3) differences in time since injury, 4) varied exclusion/inclusion criteria, 5) control group composition, and 6) sensitivit y of measures. Mild TBI is not always classified in the same manner across studies , so it is possible that studies showing impaired memory after mild TBI used groups w ith relatively more severe injuries. Also,
17 studies that recruit patients referred for neuropsychological assessment or rehabilitation after mild TBI (e.g., Yeates et al., 1995) may be more likely to include children exhibiting memory impairments than community-based samples. Additionally, studies vary as to how long after injury the assessment takes place. Deficits that tend to resolve with time may be most readily detected in th e acute stage after injury. Furthermore, the composition of the control group is a factor to consider, as some groups may be better matched on variables known to affect memo ry performance such as intelligence. However, it is important to not e that the findings regarding mi ld TBI pertain to relatively subtle deficits, such as slowed learning that im proves to the level of c ontrols after trials. Another reason, then, for the inconsistent findings could pertain to the sensitivity of the measures involved, and whether they allow fo r an in-depth examination of different aspects of memory. Less Frequently Studied Aspects of Memory after TBI Learning, recall, and recognition have been examined in a number of studies focusing on mild and severe TBI. The research described in the preceding sections generally served to show that children with severe head injuries demonstrate memory impairments on these dimensions across a wide variety of tasks, while those sustaining mild head injuries exhibit little to no deficit. Less frequently studi ed aspects of memory after TBI include retention and the meaningf ulness or Â“contextual relevanceÂ” of the material to be learned. Examining these dimensions of memory may provide more insight into the memory problems experienced after pediatric TBI. Retention Retention refers to the storage aspect of memory and can provide valuable information about rates of forgetting. Rete ntion is usually examined by comparing the
18 amount of information recalled after a long de lay to the amount of information initially learned (i.e., the amount recalled immediately af ter the learning trials are completed). In general, studies have neglected to examin e retention. The few studies that have investigated retention rates in TBI have yielded conflicting findings. One study of severe TBI reported normal retention on a word list le arning task (Roman et al., 1998). Other studies have reported impaired retention usi ng the same list learning task (Yeates et al., 1995; Levin et al., 1994, 1993). However, these findings may be moderated by differences in intellectual functioning be tween the groups. For example, Levin and colleagues (1993) covaried for Full Scale IQ, which served to attenuate the effects, though they were still significant. In a second study, covarying for IQ attenuated the effects such that they were no longe r significant (Levin et al., 1994). In studies utilizing tasks other than wo rd list learning, there has also been suggestion of intact retention capabilities. On a comprehensive memory battery, children with TBI had difficulty after the delay on most subtests only if they showed initial problems with immediate recall, which suggests intact retention that is not impacted by rapid forgetting or information loss (Farmer et al., 1999). Finally, Bassett and Slater (1990) looked at an adolescent sample and f ound that retention for verbal material was comparable across0 controls, mild TBI, and severe TBI, but that retention for visual material was impaired in severe TBI. Meaningfulness of material Another area that has been relatively ne glected in prior studies concerns the meaningfulness or Â“contextual relevanceÂ” of th e material to be learned. For example, verbal memory can be evaluated using storie s (which provide inhe rent organization and context) or word lists (which typically do not contain contextual cues to aid in memory).
19 In normally developing children, context facilitates learning and memory (Kail, 1990). In the TBI literature, only one study has directly compared tasks that differ in meaningfulness, and this study showed impaired recall after severe TBI for material both with and without contextual cues (Lowther & Mayfield, 2004). In this study, memory impairments were shown on tasks utilizing ab stract items (e.g., geometric figures) as well as meaningful items (e.g., faces). Impairments were also shown across varying task demands, from free recall to tasks where the it ems to be remembered were associated or presented sequentially. It is difficult to draw conclusions from existing studies that have a narrow focus on one type of material or task, but there is some evidence that context or meaning aids memory in children sustaining severe TBI. On a story task, which is inherently meaningful and involves the cont extual recall of verbal inform ation, children with severe TBI have been shown to perform simila rly to those with mild-moderate TBI on immediate recall (Donders, 1993) and recogniti on testing (Farmer et al., 1999), though in the latter study children with severe TBI s howed impaired story recall. Farmer and colleagues contrasted the inta ct recognition for stories with the impaired recognition generally shown for word lists (e.g., Yeates et al., 1995), and suggested that children with severe injuries may be able to encode and st ore meaningful verbal material more readily than a rote word list. Using word lists, one study showed that children sustaining TBI, inclusive of all severity levels, were ab le to use their knowledge of conceptual relationships to improve cued recall (Jaffe et al., 1992). Further support comes from studies showing that children with severe TBI did not differ from controls in their use of
20 semantic clustering strategies (Roman et al., 1998; Yeates et al., 1995), although this finding has not been consistent across studies (see Levin et al., 1993). Other Factors Influencing Memory after TBI Memory after pediatric TBI is likely infl uenced by a number of factors in addition to injury severity. Memory performance af ter TBI has been linked to demographic and injury-related variables. Of these factors, some of th e most relevant to the study of memory after pediatric TBI include gender, age at injury, and time since injury. Gender Gender may moderate the effects of TBI on childrenÂ’s memory. Boys sustaining TBI have been reported to perform more poorly than girls with TBI and matched controls (Donders & Hoffman, 2002; Donders & Wood ward, 2003). After controlling for the effects of injury severity and age, gender e xplained an additional 59% of the variance in memory performance in these studies. Ho wever, gender differences were largely attenuated when speed of processing was used a covariate. These results suggest that male gender may be associated with an incr eased risk for memory deficits after TBI, possibly because boys need longer to effectively process information. Age at injury Experiencing a TBI can significantly alter the developmental trajectory (Lord-Maes & Obrzut, 1996). Children who sustain an injury at a younger age (generally preadolescent) have shown greater impairments on memory tasks relativ e to older children (Donders & Hoffman, 2002; Donders & W oodward, 2003; Levin et al., 1993, 1982; Verger et al., 2000). Little research supports the idea that memory in older children is more vulnerable to TBI (Roman et al., 1998). However, a substantial number of studies have failed to find evidence of age effects (Di Stefano et al., 2000; Levin et al., 1988;
21 Yeates et al., 1995, 2002). Though inconsistent , these results may suggest an increased vulnerability of emerging skill s in young children, with so me age differences perhaps attributable to disruption of frontal lobe maturation (Ande rson et al., 1997; Lord-Maes & Obrzut, 1996). Memory may fail to progr ess in accordance with developmental expectations, with Â“emerging de ficitsÂ” appearing as expected developmental gains are not achieved (Anderson et al., 1997). Time since injury Another important variable to consider when evaluating memory after pediatric TBI is time since injury. Memory deficits have been shown to persis t after severe TBI at one year (Levin et al., 1982; Massagli et al., 1996) and at three years post-injury (Fay et al., 1994). However, substantial recovery in memory skills has been described in a number of studies. Memory has been shown to gradually improve following head injury over three months (Knights et al., 1991; Roman et al., 1998) a nd over the course of a year (Catroppa & Anderson, 2002; Y eates et al., 2002). Purpose of Current Study The goal of the current study was to exam ine memory functioning in children and adolescents who sustained a traumatic brain inju ry within the past ye ar. Different aspects of memory were explored as th ey relate to severity of inju ry. Objectives of the current study were to examine the pattern of memory performance after pe diatric TBI, with respect to memory as a multidimensional construct, and to address areas that have been relatively neglected in prior stud ies. Particular gaps in the literature exist for examining retention rates and how the meaningfulness of material impacts memory performance. This type of detailed examination facilitate s the identification of memory functions that are most susceptible to impairment after TBI an d those that may be re latively spared.
22 Based on the literature, it was hypothesized that participants sustaining severe TBI would demonstrate more impairment in learning, general memory, and delayed recognition than participants sustaining mild TBI and controls. Additionally, we planned to compare different aspects of retrieval, specifically free re call versus recognition, within the severe TBI group in order to asce rtain whether a recall de ficit was present. Given the inconsistencies in the literature, it was difficult to generate hypotheses about the mild TBI group. In view of the hetero geneous nature of the current sample (e.g., recruited from the community a nd per referral), it was hypothesi zed that participants with mild TBI would generally perform similarly to controls. Regarding retention rates, it was hypothesi zed that the groups would show similar retention rates. Past studies have served to highlight encoding problems, which do not necessarily imply a problem w ith retention. Also, several studies have demonstrated intact retention for verbal material. The current study sought to re plicate those findings using a broader measure of memory that inco rporates both visual-s patial and auditoryverbal memory tasks. The current study also sought to examine how the meaningfulness of the material influenced memory performance. As descri bed earlier, some materi al, such as a story, provides an inherent structure and organization that could ai d memory. Given the extent of their impairments, it was hypothesized that children with severe TBI would perform significantly worse than those with mild TBI and controls on all types of tasks. However, within the severe TBI group, it was hypothesized that perf ormance would be relatively better on meaningful as compared to less meaningful (or abstract) tasks.
23 CHAPTER 2 METHODS This study focused on memory performance after pediatric traumatic brain injury (TBI). Participants ranged in age from 6-16 years and completed a comprehensive assessment of learning and memory. Memory performance was compared between those sustaining mild TBI, those sustaining severe TBI, and control partic ipants. The current study took place within the context of a la rger, longitudinal study of attention and memory in children sustaining TBI. Recruitment and Informed Consent Participants were recruited through the Univ ersity of Florida Pediatric Brain Injury Program (Co-Directors: Drs. Shelley Heaton and Eileen Fe nnell), which recruits through a variety of medical centers and community sites in north-central Florida. Multiple methods of recruitment were utilized, includi ng fliers and physician referrals. A number of participants were recruited through Shands Hospital at the University of Florida, with colleagues in the Emergency Department, Pedi atric Neurology, Pediat ric Intensive Care Unit, and Orthopedic Clinic referring potential participants to the pr ogram. According to Yeates (2000), samples recruited prospectiv ely from consecutive admissions to a large hospital, are more likely to yield represen tative groups of children with TBI than retrospective recruitment methods. The current study is subsumed within the larger study, which obtained approval from the University of Florida Health Center Institutional Review Board. Written informed consent to participate was obtaine d from the parent/guardian, and the child
24 provided written assent. Informed consent to gain access to medical records pertaining to the injury was also obtained. For their part icipation, families receiv ed a feedback letter detailing the childÂ’s performance on neuropsyc hological measures. In addition, families received a $20.00 check as compensation for their time. Participants The total sample size ( N = 38) included 25 participants who sustained a documented TBI and 13 control participants. Sixteen children were classified as sustaining severe TBI, and 9 children were cl assified as sustaining mild TBI. Injury severity levels were determined on the basis of CT or MRI scan results, Glasgow Coma Scale scores, period of loss of consciousness, and duration of post-traumatic amnesia. The classification system is outlined in Table 1, and these indices and cut-off points are commonly used in the literature to group indi viduals sustaining TBI according to injury severity (Yeates, 2000). Table 1. Classification System for Level of TBI Severity Indices of Severity Mild TBI Severe TBI CT or MRI Scan Abnormalities Absent Present Glasgow Coma Scale total score 13-15 upon admission 8 upon admission Loss of Consciousness <1 hour >24 hours Post-Traumatic Amnesia <24 hours >7 days The information needed to make this classification was abstracted from participantsÂ’ medical records. Those partic ipants falling between th e classifications of mild and severe were labeled as moderate , of which there were 3. Their data was collected as part of the larger research project, but was not used in the current study as there were not enough participants with mode rate injuries to form a separate group. Additionally, to keep the m ild and severe groups as homogenous as possible, the
25 moderately injured participants were not co mbined with another group as has been done in other studies (e.g., Donders, 1993). The control group consisted of 8 participants who had sustained an orthopedic injury not involving th e head or brain and 5 participants who had not sustained any injury. The children within the control gr oup received injuries resulting from falls ( n = 3), motor vehicle accidents ( n = 1), and sports-related incidents ( n = 4). Ideally, the control group would have been composed solely of children sustaining orthopedic injuries because these children have experien ced a traumatic injury. To the extent that traumatic injuries are not random, children with orthopedic injuries may control for premorbid characteristics influencing injur y, such as premorbid behavioral problems (Yeates, 2000). However, in order to augmen t the sample size, uninjured children were combined with children who had orthopedic injuries to form a single control group. Inclusion criteria for participation in the study included (1a) sustained a documented TBI within the past year, (1b) su stained an orthopedic injury not involving the head or brain within the past year, or (1c) healthy uninjured child; (2) between the ages of 6-16; (3) English speaking; and (4) me dically stable at the time of evaluation. Exclusion criteria applicable to all partic ipants included history of neurological or developmental disorder and Verbal-Scale IQ less than 70 or greater than 130. These values were selected because they represent scores that are greater than two standard deviations from the mean, with more than 95% of cases in the test normative sample falling between these two points. Since memory relates to intellectual functioning both in normally developing children (e.g., Cohen, 1997) and in pediatric TBI samples (e.g., Jaffe et al., 1992), it is important to ensure that participants ha d fairly equivalent levels of
26 intellectual functioning. Two pot ential participants with se vere TBI were excluded due to Verbal IQ less than 70, and five potential control participants were excluded due to Verbal IQ greater than 130. Two TBI part icipants were excluded, one due to an unresolved question of whethe r this individualÂ’s mild TBI followed a seizure and a second due to psychiatric history and the natu re of the head injury (which involved a penetrating, open injury due to self-inflicted gun shot). An additional exclusion criterion for the control group was history of previous head injury, and one potential control was excluded due to a possible history of mild TBI. Demographic characteristics of the participants are detailed in Table 2. Participants ranged in age from 6 to 16 years, with an average age of approximately 12 years across all three groups. There were more males in the head-injured groups, which is consistent with epidemiological data showing that male s are more likely to sustain a head injury (Kraus, 1995). Participants were predomin antly Caucasian. Othe r ethnicities included African-American ( n = 8), Hispanic ( n = 1), and other ( n = 1). The percentage composition of the sample for ethnicity and race is similar to the demographics of Alachua County, Florida, in which the major ity of recruitment t ook place (White 73.5%; African American 19.3%; and Hispanic or Latino origin: 5.7%; U.S. Census Bureau, 2000). A one-way analysis of variance was used to compare the groups on age [ F (2, 35) = .69, p = .51], and chi-square tests were us ed to compare the groups on gender [ 2(2, N = 38) = 4.28, p = .12], and ethnicity [ 2(6, N = 38) = 4.23, p = .64]. These analyses demonstrated that the groups were compar able on age, gender, and ethnicity.
27 Table 2. Demographic Charact eristics of Participants Demographics Control ( n = 13) Mild TBI ( n = 9) Severe TBI ( n = 16) Group Comparison Age (Years: Months) 11:8 (3:7 ) 11:5 (3:5) 12:11 (3:10) NS Gender (% Male) 46% 89% 56% NS Ethnicity (% Caucasian) 69% 89% 69% NS Time since injury (weeks) 7.8 (8.6) 19.0 (13.8) .02 Note . Group data represent means, with stan dard deviations in parentheses, unless otherwise noted. Since not all control participants sustai ned an orthopedic injury, time since injury is not reported for this group. NS = not significant The range of time since injury across groups was 2 to 49 weeks. Children in the mild TBI group were assessed about two months post-injury, while children in the severe TBI group were assessed on average at four to five months post-injury. Due to a significant LeveneÂ’s test for equality of variances, a ttest with unequal variances assumed was used, and it revealed that th e groups differed significantly on time since injury, t (22.64) = -2.51, p = .02. Participants were not as sessed until medically stable, and this factor may help to account for th e group difference since the trajectory for recovery after severe injury tends to be longer. Regardi ng mechanism of injury, children sustaining severe TBI were most likely to be injured in a motor ve hicle accident, while children sustaining mild TBI we re often injured in falls. The mechanism of injury is specified by severity level in Table 3. Table 3. Mechanism of Inju ry within the TBI Groups Mechanism of Injury Mild TBI ( n = 9) Severe TBI ( n = 16) Fall 67% 19% Motor Vehicle Accident 22% 69% Sport 11% 12% Note . Values denoted as percentage of participants within each TBI group. In the control group, participants did not ha ve a history of head injury, which was one of the exclusion criteria. Within the TBI groups, several particip ants had a history of
28 a prior head injury (1 partic ipant in the mild TBI group and 2 participants in the severe TBI group). Several participan ts also had a history of A ttention-Deficit Hyperactivity Disorder (ADHD; 3 controls, 1 participant in the mild TBI group, and 5 participants in the severe TBI group). These children we re not excluded from the study because a clinically relevant sample was desired. It has been argued that children sustaining TBI have a higher incidence of premorbid behavior al deficits (Anderson et al., 1997; Jaffe et al., 1992). Children who have a history of head injury may also be at higher risk of sustaining another head injury (Ponsford et al., 1999). Excluding these participants would have resulted in a sample that wa s less representative of the pediatric TBI population. Assessment Procedures Participants were individually administ ered a neuropsychological battery, assessing intellectual functioning, academic achievement, attention, executive function, memory, and psychosocial functioning. Trained research assistants, supervised by a doctoral-level psychologist, administered the tests in priv ate rooms. Total test administration time averaged about 4 hours. Caregivers co mpleted a demographic questionnaire and behavioral rating scales for the particip ating children and adolescents. The study demographic questionnaire took approximate ly 15 minutes to complete. Basic information was obtained regarding the childÂ’s gender, date of birt h, ethnicity or race, academic grade, handedness (right/left), me dical conditions, and current medication. Intellectual Assessment In addition to age and demographic factors, intellectual skill level is an important factor to consider when assessi ng children after TBI. It is critical to estimate premorbid level of intellectual functioning and to en sure that study groups are comparable.
29 However, impairments in intellectual func tioning are common following TBI (Catroppa & Anderson, 2002), so it is also important to asse ss how injury has impacted intellectual functioning. Verbal IQ appears less vulnerab le to head injury than Performance IQ, which makes more demands for speeded re sponses and motor control (Catroppa & Anderson; Knights et al., 1991; Massagli et al., 1996; Yeat es, 2000). For this reason, Verbal IQ is probably a bette r estimate of pre-injury inte llectual functioning and was used as an indicator of intellectual functioning in the current study. The Wechsler Abbreviated Intellig ence Scale (WASI; The Psychological Corporation, 1999) was used to estimate inte llectual functioning for the majority of participants. The WASI yields a Full Scale IQ, a Performance IQ, and a Verbal IQ. It is comprised of four subtests, which take a tota l of approximately 30 minutes to administer. An estimated Verbal IQ can be computed fr om two subtests, Vocabulary (orally define words) and Similarities (identify conceptual similarities between two words). The WASI has been widely examined in the developmen tal literature and its psychometric properties indicate high reliability and validity. Memory Assessment Various facets of memory functioning were examined as they relate to injury severity. The ChildrenÂ’s Memory Scale (CMS ; Cohen, 1997) was selected as the primary outcome measure because it provides a co mprehensive evaluation of learning and memory. The psychometric properties of the CMS indicate adequate to high reliability and validity (Cohen). Memory functions ev aluated by the CMS include learning, recall after a short-delay and a longdelay, and recognition. Four s ubtests comprise the Core Battery and can be administered in approximately 30-35 minutes. The four core subtests in order of administration include Dot Locat ions, Stories, Faces, and Word Pairs. A
30 description of what the subtests entail can be viewed in Table 4. E ach subtest includes a delayed recall component, which is given a pproximately 30 minutes after completion of the immediate (i.e., short dela y) recall. Delayed recogniti on is also assessed for the Stories and Word Pairs subtests. The CMS provides a nice dichotomy betw een tasks that are meaningful and relevant to the child and those that are mo re abstract. The meaningful tasks include Stories, a verbal task that involves remembering semantically related verbal material, and Faces, a visual task that util izes inherently meaningful stimuli. The abstract, less meaningful tasks include Word Pairs, a verbal task that involves le arning mainly arbitrary word pairs, and Dot Locations, a visual task that involves learning a random array of dots. Table 4. ChildrenÂ’s Memory Scale: Core Battery Meaningful Abstract Verbal STORIES Listen to two stories and then retell each story; recognition component involves answering factual yes/no questions about the stories WORD PAIRS Learn a list of unrelated word pairs such as Â“nursefireÂ” and Â“rice-chairÂ” over three learning trials; examiner reads first word of each pair and asks examinee to provide second word; recognition component Visual FACES Asked to remember a series of individually presented faces; then asked to identify faces as one seen previously or a new one DOT LOCATIONS Learn the spatial location of an array of dots over three learning trials; reproduce the dot pattern after viewing and recalling a distracter array The CMS was nationally standardized on 1000 healthy children and yields scores that are standardized for age. Seven index scores ( M = 100, SD = 15) can be derived from performance, and the ones used in the current study include Learning, General
31 Memory, and Delayed Recognition. The Learni ng Index is a composite of immediate memory, and the General Memory Index incorp orates aspects of im mediate and delayed memory. Percent retention scores can also be calcu lated as an additional measure of delayed recall capabilities. Percent retention is calculated in the following manner: (delayed recall raw score Ã· immediate reca ll raw score) * 100. The per cent retention score reflects the proportion of items a child correctly identifies during i mmediate recall relative to the number of items remembered during the dela yed recall portion of a subtest. For the purposes of the current study, percent retent ion was calculated for each of the four subtests, and then an average composite pe rcent retention score was derived based on performance on all subtests. Measures of performance on meaningful versus abstract subtests were also calculated. Age-corrected scaled scores ( M = 10, SD = 3) were averaged across subtests to create composite scores. The mean of scaled scores on the Stories and Faces subtests represents a mean ingful or Â“contextualÂ” performance score, and the mean of scaled scores on Word Pairs and Dot Locations represents an abstract or Â“noncontextualÂ” performance score.
32 CHAPTER 3 RESULTS All statistical procedures were conducted using the SPSS-11 statistical package. One-way analyses of variance (ANOVAs) were used to compare scores on the cognitive measures across severity groups, and Bonferroni adjusted post-hoc anal yses were utilized to examine specific group differences. All data was analyzed using an alpha level of .05. Intellectual Functioning Intellectual functioning was examined to ascertain if groups were comparable on estimates of pre-injury intellectual functioning and to document the impact of injury on IQ. Intellectual functioning was evaluated usi ng Wechsler IQ scores. Most participants ( n = 36) completed the Wechsler Abbrevia ted Scale of Intelligence (WASI; The Psychological Corporation, 1999), while 2 of the TBI participants completed the expanded Wechsler Intelligence Scale for Ch ildren Â– Third Edition (W ISC-III; Wechsler, 1991) in the context of a clinical assessment during the same week of their research visit. The WASI is comparable to the full-length WISC-III as demonstrated by high correlations (ranging from .76 to .87) between the IQ scores obtained on these measures (The Psychological Corporation). A preliminary examination of the data suggested approximate normality of the distribution of scores (i.e, all IQ scores had skewness a nd kurtosis estimates less than one). The assumption of homogeneity of va riances was supported by LeveneÂ’s test for Performance IQ and Full-Scale IQ. However, LeveneÂ’s test was si gnificant for Verbal
33 IQ, F (2, 35) = 4.85, p = .01. Thus the Brown-Forsythe test, a test of equality of means robust to heterogeneous variances, was used. Groups were not found to diffe r significantly on Verbal IQ [ F (2, 33.70) = 2.64, p = .09, 2 = .11], with scores falling in th e average range for severe TBI ( M = 95.25, SD = 16.79), mild TBI ( M = 99.22, SD = 8.86), and controls ( M = 105.92, SD = 11.53). Groups also did not demonstrate signif icant differences on Performance IQ [ F (2, 35) = 1.10, p = .34, 2 = .06], with scores in the average range observed for severe TBI ( M = 93.25, SD = 17.29), mild TBI ( M = 99.67, SD = 12.25), and controls ( M = 100.69, SD = 11.84). Finally, no significant differences be tween groups were detected for Full Scale IQ [ F (2, 35) = 2.50, p = .10, 2 = .12], with mean scores again falling in the average range for Severe TBI ( M = 93.44, SD = 14.61), Mild TBI ( M = 99.00, SD = 7.86), and Controls ( M = 103.69, SD = 11.66). These analyses indi cate comparable intellectual functioning between groups. Memory Functioning Memory performance on the ChildrenÂ’s Memory Scale (CMS; Cohen, 1997) was compared between the severe TBI, mild TBI, and control groups. Composite standardized index scores were examined for General Memory, Learning, and Delayed Recognition. Additionally, scores were derived to explore other facets of memory functioning, including Percent Re tention, Meaningful subtest performance, and Abstract subtest performance. Group means, standard deviations, and sample sizes are presented for all memory scores in Table 5. First, the assumptions of univariate normality were checked for each of the dependent variables. Values of skewness and kurtosis suggested approximate normality (i.e., all estimates less than one). LeveneÂ’s test supported the assumption of homogeneity
34 of variances for all memory scores, except Percent Retention, F (2, 35) = 3.78, p = .03. Thus Percent Retention was examined using the Brown-Forsythe test. Table 5. Descriptive Statistics for Pr imary and Calculated Memory Indexes Memory Score Control ( n = 13) Mild TBI ( n = 9) Severe TBI ( n = 16)a Post-Hoc Group Differences General Memory 103.54 (13.43) 99.67 (15.91) 77.75 (20.91) Severe < Control, Mild Learning 97.46 (13.98) 96.00 (12.73) 75.75 (16.85) Severe < Control, Mild Recognition 101.54 (17.76) 105.44 (13.88) 83.67 (21.07) Severe < Control, Mild Percent Retention 91.94 (6.07) 101.34 (9.42) 91.74 (14.43) N.S. Meaningful Subtests 9.94 (1.87) 9.75 (2.62) 8.33 (2.44) N.S. Abstract Subtests 10.67 (2.18) 10.00 (2.17) 6.41 (2.93) Severe < Control, Mild Note. Group data represent means, with st andard deviations in parentheses. aOne participant sustaining severe TBI was not administered the re cognition subtests. N.S. = Not significant The groups were found to differ signifi cantly on composite index scores for General Memory [ F (2, 35) = 8.97, p = .001, 2 = .34], Learning [ F (2, 35) = 9.17, p = .001, 2 = .34], and Recognition [ F (2, 34) = 5.10, p = .01, 2 = .23]. Effect sizes ranged from moderate to large. Bonferroni-corrected post-hoc test s revealed that the severe TBI group performed significantly worse than bot h the mild TBI and control groups on all three memory indexes (see Table 5). Qua litatively, the control and mild TBI groups performed in the average range on the com posite index scores, while the severe TBI group performed in the mildly impaired to low average range rela tive to the normative test sample. This pattern of performance is illustrated in Figure 1. Within the severe TBI group, memory performance was further analyzed to compare recall versus recognition. In th e verbal domain, average performance on delayed recall subtests ( M = 7.50, SD = 3.19) was similar to average delayed recognition
35 ( M = 7.37, SD = 3.44). A paired samples t-test indicated that performance was not disproportionately facilitated by recogniti on testing within the severe TBI group [ t (14) = .39, p = .70]. 0 20 40 60 80 100 120General MemoryLearningRecognition CMS IndexesStandard Score Control Mild TBI Severe TBI ** ***Avera g e Im p aired Figure 1. Mean performance on Ch ildrenÂ’s Memory Scale indexes.1,2 1Significant differences are represented by: * p < .05; ** p = .001. 2Error bars are based on st andard deviation values. In contrast to the impaired performan ce of the severe TBI group on the primary memory indexes, the three groups sh owed similar rates of retention, [ F (2, 28.39) = 2.92, p = .07], which averaged above 90% for each gr oup. These rates are comparable to the average retention rate of approximately 90% for the normative samp le of 12-year-olds, which is close to the mean age (12 years, 2 months) of participants in the current study (Cohen, 1997). Retention rates are depicted in Figure 2, showing that groups retained proportionately similar amounts of material. Taken together, the results indicate that the severe TBI group learned and recalled less th an the mild TBI and control groups, but retained information at a comparable rate.
36 0 20 40 60 80 100 120Retention Rate (%)No Significant Group Differences 92% >100% 92% Control Severe TBI Mild TBI Figure 2. Mean performance on ChildrenÂ’s Memory Scale compos ite retention rate.1,2 1Error bars are based on st andard deviation values. 2Dashed line represents average retention rate across subtests in the normative sample of 12-year-olds. The groups were also compared on subtes ts that differed in content and task demands. Subtests were divided into relatively more meaningful versus abstract subtests. On the more meaningful subtests, no signi ficant differences between groups were detected, [ F (2, 35) = 2.07, p = .14, 2 = .11]. All groups showed average performances on the meaningful subtests, which assessed immediate and delayed recall for material with an inherently meaningful structure, su ch as short stories and faces. In contrast, significant differences between groups were detected on the more abstract subtests [ F (2, 35) = 11.73, p < .001, 2 = .40], which assessed immediate and delayed recall for dot arrays and mainly arbitrary word pairs. The eta-squared coefficient suggests a large effect size. The mean of the severe TBI group fell more than one standard deviation below the normative mean on the abstract subtests, while other groups showed an average performance. This pattern of results is illust rated in Figure 3. Bonferroni-corrected post-
37 hoc analyses (see Table 5) indicated that the severe TBI group performed significantly worse than either the mild TBI or control groups on the abstract subtests. Within the severe TBI group, a paired samples t-test in dicated poorer performance on the abstract subtests relative to the meaningful subtests, [ t (15) = -3.44, p = .004, r2 pb = .44]. This pattern suggests that the abstract tasks were probably more difficult for the severe TBI group. 0 2 4 6 8 10 12 14MeaningfulAbstract Type of SubtestScaled Score Control Mild TBI Severe TBI* Average Impaired Figure 3. Mean performance on Children Â’s Memory Scale subtests grouped by meaningful versus relatively abstract content.1,2 1Significant differences are represented by: * p < .001. 2Error bars are based on st andard deviation values. A comparison of memory performance in mild TBI relative to the control group showed small, nonsignificant effects. On t hose memory scores for which the mild TBI group performed slightly lower than controls, effects ranged from negligible to small, General Memory [ t (20) = .51, p > .05, d = .28], Learning [ t (20) = .22, p > .05, d = .11],
38 Abstract Subtests [ t (20) = .61, p > .05, d = .32], and Meaningful Subtests [ t (20) = .19, p > .05, d = .09]. In summary, the mild TBI group performed similarly to controls on all aspects of memory functioning assessed. The severe TBI group performed more poorly than mild TBI and control groups on all standardized index scores of the CMS, which measure aspects of learning, recall, and recognition. However, when memory performance was further explored using derived scores, memory functions that may be relatively intact after severe TBI were highlight ed. More specifically, rete ntion rates and performance on tasks involving meaningful materials that pr ovide a context for lear ning seem relatively preserved after TBI.
39 CHAPTER 4 DISCUSSION The results of the current study indicate th at children sustaining severe TBI show wide-ranging memory impairments that imp act performance on learning, recall, and recognition tasks. In general, memory func tioning across various dimensions fell in the mildly impaired to low average range for th e severe TBI group, a nd solidly within the average (i.e., unimpaired) range for both th e mild TBI and control groups. Reports of impaired memory functioning after severe TBI have been found in numerous previous studies (e.g., Farmer et al., 1999; Roman et al., 1998). The current study went beyond the existing research, however, by condu cting a detailed analysis of the pattern of impairment across various dimensions of memory. Memory Functioning after Mild TBI In the current study, children with mild TBI showed similar performance to controls on all facets of memo ry functioning assessed. These findings are consistent with other published studies (Max et al., 1999; Pons ford et al., 1999; Ro man et al., 1998). The pattern of findings was not suggestive of m eaningful differences between the mild TBI and control groups. Effects sizes were sma ll and nonsignificant for those memory scores on which the mild TBI group perfor med worse than controls. Other studies have reported evidence of subtle impair ments in learning and recall following mild TBI (Farmer et al., 1999; Yeates et al., 1995). These previous findings of subtle deficits after mild TBI suggest that a dose-response relati onship exists between injury severity and memory, with severity of TBI associated with relative level of
40 memory impairment. In the current study, the children with mild TBI were seen on average about two months post-injury, and acu te impairments may have resolved by that time. Also, mild TBI tends to be a heterogeneous group. While most children with mild TBI may demonstrate no residual impairments , a small subset may exhibit persistent problems in cognitive or behavioral functio ning (Levin & Eisenberg, 1979; Ponsford et al., 1999). In the typical heterogeneous mild TBI sample, these deficits may be washed out in the larger group. Future research w ould benefit from targeting this subsample by specifically recruiting children with mild TBI who have memory complaints. Memory Functioning after Severe TBI The current study also found that children with severe TBI performed more poorly than children with mild TBI and controls on learning, recall, and recognition. These results support our initial hypotheses and are co nsistent with a number of previous studies (e.g., Catroppa & Anderson, 2002; Farmer et al., 1999; Fay et al ., 1994; Levin et al., 1994; Lowther & Mayfield, 2004; Massagli et al., 1996; Roman et al., 1998; Yeates et al., 1995). In addition, the current study found that within the severe TBI group, performance was relatively better on reten tion and on tasks involving more meaningful content. These findings are particularly intere sting because they show that some facets of memory functioning may be relatively intact after severe TBI, despite significant impairments in other memory dimensions. Our hypothesis that TBI and control groups would be comparable on retention rates is supported. The severe TBI group showed normal retention rates, which averaged above 90% and were consistent with the normative average fo r 12-year-olds (Cohen, 1997). This finding suggests comparable rates of Â“forgettingÂ” across groups. In other words, the children sustaining severe TBI re tained what was learned in comparable
41 proportion to controls (e.g., 90% of initial lear ning retained after a delay); however, they simply learned less material. When viewed in light of the fairly equivalent levels of impairment noted on learning, recall, and rec ognition, the intact re tention rate in the severe TBI group suggests that a mild encodi ng deficit best characterizes their memory difficulties. Thus impaired recall and recogni tion after severe TBI lik ely reflect impaired initial encoding and learning processes, rather than a retention or re trieval deficit. The normal retention rates shown after seve re pediatric TBI are consistent with prior studies that suggested normal retention (Farmer et al., 1999), particularly on verbal tasks (e.g. Bassett & Slater, 1990; Roman et al ., 1998). The current study served to show intact retention across a variety of subtests , including not only aud itory-verbal but also visual-nonverbal tasks. In contrast, some prior studies have found evidence of impaired retention after severe TBI; however, these findings may have been driven by differences in intellectual functioning between TBI and control groups (e.g., Le vin et al., 1994). The current participants showed comparable rate s of intellectual functi oning, with all groups performing in the average range for Verb al, Performance, and Full-Scale IQ. The clinical implication of intact reten tion is that children sustaining severe TBI may benefit from interventions that target th e initial stages of learning rather than the maintenance of material in memory. Children could be taught strategi es that facilitate encoding, including rehearsal, organization, and elaboration (Kail, 1990). Interventions could also incorporate multiple presentations of the material to be learned or increased length of exposure to the mate rial, thereby helping to bypass memory problems related to attentional difficulties (e.g., distractibility) or limitations in functional capacity (e.g., slowed processing speed and efficiency).
42 There are several reasons why initial learni ng may be impaired after severe TBI. First, attentional functioning may be impaired, such that children w ith severe TBI do not attend to material to the same degree as controls or children with mild TBI. They may be less able to sustain attention or to selectively devote attention to a task. Second, capacity limitations may confine the amount of informati on that can be learned or the efficiency with which it can be processed. Lastly, st rategy use, particularly as it pertains to encoding, may be deficient in children with severe TBI. In the current study, results also indicate d that performance with in the severe TBI group varied based on task cont ent and demand. It was hypoth esized that children with severe TBI would perform more poorly than ch ildren with mild TB I and controls on all types of tasks; but they w ould show relatively better pe rformance on meaningful as compared to abstract tasks. It appears that the effects of meani ngful content may have actually been more substantial than initially hypothesized. Results showed that children sustaining severe TBI performed more poorly on abstract compared to meaningful subtests, and this deficit was demonstrat ed both in between group and within group comparisons. This pattern suggests that ch ildren sustaining seve re TBI demonstrated more effective learning when material was i nherently meaningful or presented within a context. This relatively better performance of th e severe TBI group on the meaningful tasks is consistent with the limited literature s howing that context fa cilitates learning and memory in pediatric TBI samples (Donders, 1993; Farmer et al., 1999; Jaffe et al., 1992). When information fits into a framework that is already in place, it may be more easily learned and subsequently accessed or retr ieved (Bjorklund, 2000; Kail, 1990). The
43 clinical implication is that children sustai ning severe TBI may benefit from interventions that provide a contextual fram ework for material that is to be learned. For example, academic interventions could focus on teaching history using a story format rather than through less integrated dates and facts. Interventions could also focus on teaching strategies to improve the in itial acquisition of material that lacks a meaningful framework. One reason for the severe TBI groupÂ’s relatively better pe rformance on the meaningful subtests could relate to specific de mands associated with each subtest. It is somewhat artificial to divide tasks and labe l them as Â“abstractÂ” or Â“meaningful,Â” when the distinction is clearly not as pronounced as the labels woul d seem. It is possible that the abstract tasks may simply be more complex than the meaningful tasks. Nonetheless, some requirements of the abstr act tests (e.g., learning over multiple trials or the nature of the stimuli) resulted in them being more difficu lt for children with severe TBI. It will be important for future researchers to create better-matched analogues for meaningful and abstract tasks. Another expl anation for the poorer performa nce of the severe TBI group on the abstract subtests could be a primar y deficit in employing internally driven organization strategies. The ability to inte ntionally apply and flex ibly use memorization strategies may facilitate learning for complex, abstract tasks that do not relate to already present schematic frameworks. The pattern of findings observed in the current study has implications for the interpretation of CMS memory index scores a nd the design of future memory tests. In the current study, a traditional examination of the standard index scores on the CMS showed fairly consistent impairments across all composite indexes after severe TBI.
44 However, calculating derived scores served to highlight aspects of memory functioning that may be preserved after severe TBI and th at have been relatively neglected in many studies using conventional memory meas ures and performance indexes. Additionally, children with severe TBI were shown to perform poorly on the abstract or rote tasks, which are represented in each of the composite indexes. Thus in the current study, group differences on these two abstract subtests may have driven the findings related to the composite indexes, wh ich represent performance across subtests. By only examining the composite indexes, the nature of impairment after severe TBI may be misinterpreted and characterized as too gl obal. Limitations in the design of the CMS are outlined in Table 6. In th e creation of future measures , developers should seek to remedy these drawbacks by designing more i ndependent indexes (e.g., separate indexes for immediate and delayed recall). Table 6. ChildrenÂ’s Memory Scale Inde xes and Limitations of Interpretation Index (and composite subtests) Limitation Learning Dot Locations Word Pairs Composed only of abstra ct (noncontextual) tasks General Memory Dot Locations Stories Faces Word Pairs Combines both immediate and long delay performance and does not allow for the separation of these factors Delayed Recognition Stories Word Pairs Composed only of tasks from the verbal domain The memory problems observed in this study are likely due to patterns of frontal pathology, temporal lobe injury, and damage to diencephalic structur es (Yeates et al., 1995), leading to poor encoding and consolid ation of material in memory. These memory processes may be particularly disr upted by damage to medial temporal lobe
45 structures, including the hippocampal forma tion (Sohlberg & Mateer, 2001; Waxman, 2000). Damage to the frontal lobes may be most eviden t in impaired encoding and retrieval strategies and difficulty in mon itoring memory processes (Frackowiak et al., 2004). Limitations The major limitations of this study in clude the small sample size and the heterogeneity of the sample. Sample heteroge neity is inherent in studying the pediatric TBI population, which is diverse and differs on numerous factors, such as mechanism and locus of injury. A confounding variable in the current study may be time since injury, which showed a large range across participants and differed between the mild and severe TBI groups. However, the direction of the findings does not suggest that this difference substantially impacted the results. An additional limitation of the study is that the control group included both healthy controls and orthopedic injury controls. It would have been ideal to have a control group com posed entirely of children, who sustained an orthopedic injury, to control for preinjury variables that may predispose children to trauma as well as injury-related variables (Yeates, 2000). A further limitation is that participants were examined at a single point in time within the relatively acute stages of injury. Obviously this type of examination is not informative of recovery over time, with a longitudinal study being the only means of obtaining such information. Our future goals include examining TBI using longitudinal methodology. Finally, the ChildrenÂ’s Me mory Scale (CMS; Cohen, 1997) is a comprehensive test and improves upon measures used in past studies. However, as discussed earlier, the scale st ill contains confounds between variables, making it difficult to examine certain aspects of memory independently.
46 Summary and Future Directions When a child recovers physically from TB I, people often expect that the childÂ’s cognitive and behavioral func tioning should also return to normal (Lord-Maes & Obrzut, 1996). However, the present study highlighted the memory problems following severe TBI that persist after the acute phase of in jury. The results provided some evidence of task-specific impairment on abstract or noncont extual tasks after seve re TBI. The study also highlighted intact retent ion abilities across severity le vels. The pattern of memory impairments noted in the current study likely re flects problems with the initial acquisition of material. The current findings have strong clini cal relevance. The observed memory difficulties are particularly significant as they relate to academic functioning, which often demands fast-paced learning of complex material in an environment full of distractions. Results of this study could be used in the design of academic interventions and rehabilitation programs. Interventions could be tailored to target pr edicted deficits (e.g., initial encoding and learning) and to take advantage of re latively spared aspects of memory functioning (e.g., retention, memo ry for contextual information). In the present study, memory was assessed using a relatively new and comprehensive measure, the CMS, and results provide evidence for the validity of this measure in examining memory outcome after childhood TBI. A comprehensive evaluation instrument, like the CMS, can illustrate the pattern s of strengths and weaknesses across a broad range of memory functioning. Interestingly, the current study indicated that the index or composite sc ores might obscure important aspects of performance, which can best be observed thr ough the pattern of performance on specific subtests. The current findings could be used in the development of new memory
47 measures and scoring methods. It will also be important for future researchers to address the relationship between clinical memory measures, such as the CMS, and memory needed for everyday tasks. The broader goal of future re search in this area is to examine factors that could influence memory outcome after pediatric TBI in addition to injury severity. Developmental factors include age at injury and recovery over time. Also, it would be useful to study other cognitive skills, such as attention and executive functioning, and how they relate to memory after TBI. With a larger sample size, it would be possible to develop a predictive model of memory impa irment after pediatric TBI. A regressionbased model would account for the variance in memory outcome explained by each of these predictors. In conclusion, future rese arch directions incl ude exploring memory after TBI in a developmental context across seve rity levels, across time, and in relation to other cognitive functions.
48 LIST OF REFERENCES Anderson, V., Catroppa, C., Morse, S., Harit ou, F., & Rosenfeld, J. (2001). Outcome from mild head injury in young children: A pros pective study. Journal of Clinical and Experimental Neuropsychology, 23 , 705-717. Anderson, V. A., & Lajoie, G. (1996). Deve lopment of memory and learning skills in school-aged children: A neurops ychological perspective. Applied Neuropsychology, 3/4 , 128-139. Anderson, V. A., Morse, S. A., Klug, G., Catr oppa, C., Haritou, F., Rosenfeld, J., et al. (1997). Predicting recovery from head injury in young childre n: A prospective analysis. Journal of the International Neuropsychological Society, 3 , 568-580. Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. T. Spence (Eds.), The psychology of learning and motivation: Advances in research and theory: Vol. 2 (pp. 89-195). New York: Academic Press. Baddeley, A. (1992, January 31). Working memory. Science, 255 (5044), 556-559. Bassett, S. S., & Slater, E. J. (1990). Neuropsychological func tion in adolescents sustaining mild closed head injury. Journal of Pediatric Psychology, 15 , 225-235. Bjorklund, D. F. (2000). ChildrenÂ’s thinking: Developmental function and individual differences (3rd ed.). Belmont, CA: Wadsworth. Catroppa, C., & Anderson, V. (2002). Recove ry in memory function in the first year following TBI in children. Brain Injury, 16 , 369-384. Cohen, M. J. (1997). ChildrenÂ’s Memory Scale manual . San Antonio: The Psychological Corporation. Di Stefano, G., Bachevalier, J., Levin, H. S., Song, J. X., Scheibel, R. S., & Fletcher, J. M. (2000). Volume of focal brain lesi ons and hippocampal formation in relation to memory function after closed head injury in children. Journal of Neurology, Neurosurgery, & Psychiatry, 69, 210-216. Donders, J. (1993). Memory functioning af ter traumatic brain injury in children. Brain Injury, 7 , 431-437.
49 Donders, J., & Hoffman, N. M. (2002). Gende r differences in learning and memory after pediatric traumatic brain injury. Neuropsychology, 16 , 491-499. Donders, J., & Woodward, H. R. (2003). Gender as a moderator of memory after traumatic brain injury in children. Journal of Head Trauma Rehabilitation, 18 , 106-115. Ewing-Cobbs, L., Levin, H. S., Fletcher, J. M., Miner, M. E., & Eisenberg, H. M. (1990). The ChildrenÂ’s Orientation and Amne sia Test: Relationship to severity of acute head injury and to recovery of memory. Neurosurgery, 27 , 683-691. Farmer, J. E., Haut, J. S., Williams, J., Kapila , C., Johnstone, B., & Kirk, K. S. (1999). Comprehensive assessment of memory functioning following traumatic brain injury in children. Developmental Neuropsychology, 15 , 269-289. Fay, G. C., Jaffe, K. M., Polissar, N. L., Liao, S., Rivara, J. B., & Martin, K. M. (1994). Outcome of pediatric traumatic brain in jury at three year s: A cohort study. Archives of Physical Me dicine and Rehabilitation, 75 , 733-741. Frackowiak, R. S. J., Friston, K. J., Frith, C. D ., Dolan, R. J., Price, C. J., Zeki, S., et al. (2004). Human Brain Function (2nd ed.). London: Academic Press. Hannay, H. J., & Levin, H. S. (1988). Visu al continuous recognition memory in normal and closed-head-injured adolescents. Journal of Clinical and Experimental Neuropsychology, 11 , 444-460. Harris, J. R. (1996). Verbal rehearsal and me mory in children with closed head injury: A quantitative and qualitative analysis. Journal of Communication Disorders, 29 , 7993. Jaffe, K. M., Fay, G. C., Polissar, N. L., Marti n, K. M., Shurtleff, H., Ri vara, J. B., et al. (1992). Severity of pediatric traumatic brain injury and ea rly neurobehavioral outcome: A cohort study. Archives of Physical Me dicine and Rehabilitation, 73 , 540-547. Kail, R. (1990). The development of memory in children (3rd ed.). New York: W. H. Freeman. Knights, R. M., Ivan, L. P., Ventureyra, E. C., Bentivoglio, C., Stoddart, C., Winogron, W., et al. (1991). The effects of head injury in children on neuropsychological and behavioral functioning. Brain Injury, 5 , 339-351. Kraus, J. F. (1995). Epidemiological featur es of brain injury in children: Occurrence, children at risk, causes and manner of injury, severity, and outcomes. In S. H. Broman & M. E. Michel (Eds.), Traumatic head injury in children (pp. 22-39). New York: Oxford University Press.
50 Lehnung, M., Leplow, B., Ekroll, V., Benz, B., Ritz, A., Mehdorn, M., et al. (2003). Recovery of spatial memory and persisten ce of spatial orientat ion deficits after traumatic brain inju ry during childhood. Brain Injury, 17 , 855-869. Lescohier, I., & DiScala, C. (1993). Blunt trauma in children: Causes and outcomes of head versus intracranial injury. Pediatrics, 91 , 721-725. Levin, H. S., Culhane, K. A., Fletcher, J. M., Mendelsohn, D. B., Lilly, M. A., Harward, H., et al. (1994). Dissociation between delayed alternati on and memory after pediatric head injury: Rela tionship to MRI findings. Journal of Child Neurology, 9 , 81-89. Levin, H. S., Culhane, K. A., Mendelsohn, D., L illy, M. A., Bruce, D., Fletcher, J. M., et al. (1993). Cognition in relation to magne tic resonance imaging in head-injured children and adolescents. Archives of Neurology, 50 , 897-905. Levin, H. S., & Eisenberg, H. M. (1979). Neuropsychological impairment after closed head injury in children and adolescents. Journal of Pediatric Psychology, 4 , 389402. Levin, H. S., Eisenberg, H. M., Wigg, N. R ., & Kobayashi, K. (1982). Memory and intellectual ability after head inju ry in children and adolescents. Neurosurgery, 11 , 668-672. Levin, H. S., High, W. M., Jr., Ewing-Cobbs, L., Fletcher, J. M., Eisenberg, H. M., Miner, M. E., et al. (1988). Memory f unctioning during the first year after closed head injury in children and adolescents. Neurosurgery, 22 , 1043-1051. Levin, H. S., Song, J., Scheibel, R. S., Fletcher , J. M., Harward, H. N., & Chapman, S. B. (2000). Dissociation of frequency and r ecency processing from list recall after severe closed head injury in children and adolescents. Journal of Clinical and Experimental Neuropsychology, 22 , 1-15. Lord-Maes, J., & Obrzut, J. E. (1996). Ne uropsychological consequences of traumatic brain injury in children and adolescents. Journal of Learning Disabilities, 29 , 609617. Lowther, J. L., & Mayfield, J. (2004). Me mory functioning in children with traumatic brain injuries: A TOMAL validity study. Archives of Clinical Neuropsychology, 19 , 105-118. Martin, J. H. (2003). Neuroanatomy: Text and Atlas (3rd ed.). New York: McGraw-Hill. Massagli, T. L., Jaffe, K. M., Fay, G. C., Polissa r, N. L., Liao, S., & Rivara, J. B. (1996). Neurobehavioral sequelae of severe pedi atric traumatic brain injury: A cohort study. Archives of Physical Medicine and Rehabilitation, 77 , 223-231.
51 Max, J. E., Roberts, M. A., Koele, S. L., Li ndgren, S. D., Robin, D. A., Arndt, S., et al. (1999). Cognitive outcome in children and adolescents following severe traumatic brain injury: Influence of psychosocial, ps ychiatric, and injury-related variables. Journal of the International Neuropsychological Society, 5 , 58-68. Nelson, C. A. (1997). The neurobiological basi s of early memory development. In N. Cowan (Ed.) & C. Hulme (Series Ed.), The development of memory in childhood (pp. 41-82). Hove East Susse x, UK: Psychology Press. Ponsford, J., Willmott, C., Rothwell, A., Ca meron, P., Ayton, G., Nelms, R., et al. (1999). Cognitive and behavioral outcome fo llowing mild traumatic head injury in children. Journal of Head Trauma Rehabilitation, 14 , 360-372. Pressley, M., & Schneider, W. (1997). Introduction to memory development during childhood and adolescence . Mahwah, NJ: Lawrence Erlbaum. The Psychological Corporation. (1999). Wechsler Abbreviated Scale of Intelligence manual . San Antonio: Author. Roman, M. J., Delis, D. C., Willerman, L., Magul ac, M., Demadura, T. L., de la Pena, J. L., et al. (1998). Impact of pediatric traumatic brain injury on components of verbal memory. Journal of Clinical and Expe rimental Neurosychology, 20 , 245258. Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20 , 11-21. Sohlberg, M. M., & Mateer, C. A. (2001). Cognitive rehabilitation: An integrative neuropsychological approach . New York: Guilford Press. Spreen, O., Risser, A. H., & Edgell, D. (1995). Developmental neuropsychology . New York: Oxford University Press. Squire, L. R., Knowlton, B., & Musen, G. (1993). The structure and organization of memory. Annual Review of Psychology, 44 , 453-495. Teasdale, G., & Jennett, B. (1974, July 13). Assessment of coma and impaired consciousness: A practical scale. Lancet, ii , 81-84. United States Census Bureau. (2000). United States Census 2000. Retrieved December 12, 2003, from http://quickfacts.cens us.gov/qfd/states/12/12001.html Vakil, E., Blachstein, H., Rochberg, J., & Vard i, M. (2004). Charact erization of memory impairment following closed-head injury in children using the Rey Auditory Verbal Learning Test (AVLT). Child Neuropsychology, 10 , 57-66.
52 Verger, K., JunquÃ©, C., Jurado, M. A., Tresserras , P., Bartumeus, F., NoguÃ©s, P., et al. (2000). Age effects on long-term neurops ychological outcome in paediatric traumatic brain injury. Brain Injury, 14 , 495-503. Waxman, S. G. (2000). Correlative neuroanatomy (24th ed.). New York: Lange Medical Books/McGraw-Hill. Wechsler, D. (1991). Wechsler Intelligence Scale for Children manual (3rd ed.). San Antonio: The Psychological Corporation. Woodward, H., & Donders, J. ( 1998). The performance of ch ildren with traumatic head injury on the Wide Range Assessment of Memory and Learning-Screening. Applied Neuropsychology, 5 , 113-119. Yeates, K. O. (2000). Closed-head injury. In K. O. Yeates, M. D. Ris, & H. G. Taylor (Eds.), Pediatric neuropsychology: Re search, theory and practice (pp. 92-116). New York: Guilford Press. Yeates, K. O., Blumenstein, E., Patterson, C. M. , & Delis, D. C. (1995). Verbal learning and memory following pediat ric closed-head injury. Journal of the International Neuropsychological Society, 1 , 78-87. Yeates, K. O., Taylor, H. G., Wade, S. L., Dr otar, D., Stancin, T., & Minich, N. (2002). A prospective study of shortand long-te rm neuropsychological outcomes after traumatic brain injury in children. Neuropsychology, 16 , 514-523.
53 BIOGRAPHICAL SKETCH Marie Schroder graduated from the University of the South with a bachelorÂ’s degree in psychology. Ms. Schroder was awar ded a Fulbright Fellowship to spend a post-graduate year in Munich, Germany, st udying clinical psychology. She then spent two years working at the Yale University Ch ild Study Center as a research assistant. There she studied the effects of prenatal cocaine exposure on cognitive and social development in children. Ms. Schroder is currently working towards a doctorate in clinical and health psychology with a speci alization in clinical neuropsychology at the University of Florida.