1 LIFESPAN DIFFERENCES IN STRATEGY TRAINING BENEFITS: AN INVESTIGATION OF THE THETA AND ALPHA FREQ UENCY BANDS By KIMBERLY CASE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLM ENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Kimberly H. Case
3 To my Mom, who has never lost faith in me
4 ACKNOWLEDGMENTS I would like to thank the many research assistants and colleagues who assisted in all aspects of research development, execution and analysis A special thank you to my advisor, Dr. William Keith Berg, who helped me along the way in both academia and life; it couldnt have been done without him Teri DeLucca my lab mate and best friend is a shining star and a true example of what a friend should be. T o my sweet and very patient husband, you have inspired me to be a better person Your tranquility and strength carried me to the end ; a nd after all these long months of waiting Im back.
5 TABLE OF CONTENTS ACKNOWLEDGMENTS .................................................................................................. 4 page LIST OF FIGURES .......................................................................................................... 8 ABS TRACT ................................................................................................................... 10 CHAPTER 1 INTRODUCTION .................................................................................................... 12 Changes in Cognition and Across the Lifespan ...................................................... 12 Changes in Brain Function Across the Lifespan ..................................................... 16 Improving Working Memory in Younger and Older Adults ...................................... 18 Phy siological Changes Accompanying Working Memory Training ......................... 25 Why Do We See Varying Changes in Task Relevant Brain Activation After Training? .............................................................................................................. 27 The Use of Electroencephalogram Activity in the Study of Human Cognition ......... 28 Theta Changes in Power Across the Lifespan .................................................. 29 Alpha Changes in Power Across the Lifespan .................................................. 32 Current Study .......................................................................................................... 34 2 METHODS .............................................................................................................. 37 Participants ............................................................................................................. 37 Measures ................................................................................................................ 38 Digit Symbol Task ............................................................................................ 38 Tower of London Planning Task ....................................................................... 39 Selected problems for the Tower of London .............................................. 41 Working memory strategy training ............................................................. 43 Reclassification of Problems for Analysis ............................................................... 46 Procedure ............................................................................................................... 47 Electroencephal ogram (EEG) Data Acquisition ................................................ 48 EEG Data Processing and Quantification ......................................................... 49 3 RESULTS ............................................................................................................... 50 Behavioral Data ...................................................................................................... 50 Proportion Correct ............................................................................................ 51 Solution Time ................................................................................................... 53 Proportion of No Response .............................................................................. 56 Electroencephalogram (EEG) Data Analysis .......................................................... 56 The Theta Frequency Band .............................................................................. 58
6 Frontal and Parietal Activity in the Theta Frequency Band ............................... 59 Lateralized Frontal Activity in the Theta Frequency Band ................................ 61 Frontal and Parietal Activity in the Lower Alpha Frequency Band .................... 64 Parietal Activity in the Upper Alpha Frequency Band ....................................... 65 Individual Factors Contributing to Successful Strategy Application ........................ 66 Behavioral Analyses Between the High and Low Processing Speed Groups .. 69 Frontal and Parietal Activity in the Theta Frequency Band Between High and Low Processing Groups ......................................................................... 70 Lateral Frontal Activity in the Theta Frequency Band Between High and Low Processing Groups ................................................................................ 71 4 DISCUSSION ......................................................................................................... 72 Differences in the Benefits of Strategy Training Acros s the Lifespan ..................... 72 Changes in the Theta Frequency Across the Lifespan and the Impact of Strategy Training ................................................................................................. 76 Changes in the Alpha Frequency Across the Lifespan and the Impact of Strategy Training ................................................................................................. 80 An Integrative Look at the Theta and Alpha Frequency Bands and Its Impact Across the Lifespan ............................................................................................. 82 Impact of Individual Differences in Processing Speed on Behavior and EEG ........ 84 Implications, Suggestions for Future Research and Conclusion ............................. 88 APPENDIXES A LIST OF TOWER OF LONDON PROBLEMS ......................................................... 91 B CORRELATIONS BETWEEN BEHAVIORAL VARIABLES .................................... 92 C SIGNIFICANT FINDINGS FROM BEHAVIORAL, EEG, AND PROCESSING SPEED ANALYSES ................................................................................................ 94 LIST OF REFERENCES ............................................................................................... 96 BIOGRAP HICAL SKETCH .......................................................................................... 106
7 LIST OF TABLES Tabl e P age 2 1 Demographic and n europsychological d ata of the f inal sample .......................... 38 3 1 Means and s tandard d eviations for b ehavioral m easures .................................. 52 3 2 Demographic d ata for l ow and h igh p rocessing g roups ...................................... 67
8 LIST OF FIGURES Figure P age 2 1 Digit/Symbol t ask. ............................................................................................... 39 2 2 Three types of Tower of London p roblems ......................................................... 40 2 3 Depiction of m oves n e eded for the b asic strategy sequence ............................. 44 2 4 Depiction of 5move e xtended t echnique 1 ........................................................ 45 2 5 Depiction of 5move e xtended t echnique 2 ........................................................ 45 2 6 Depiction of 6move e xtended t echnique ........................................................... 46 2 7 Flow chart of p roblem t ype r eclassification ......................................................... 47 3 1 Mean proportion correct measure for older and younger adults ......................... 52 3 2 Change score measure for proportion correct for older and younger adults ....... 53 3 3 Mean solution time measures for older and younger adults ............................... 54 3 4 Change scores for solution time for older and younger adults ............................ 55 3 5 Mean proportion of no responses for older and younger adults .......................... 56 3 6 Topographic plots of t heta for o lder and y ounger a dults. .................................... 58 3 7 Theta power in the frontal and parietal leads during all three epoch times ......... 59 3 8 Theta power at the frontal and parietal sites during the problem completion epoch .................................................................................................................. 61 3 9 Lateral frontal theta power i n the problem presentation epoch ........................... 62 3 10 Lateral frontal theta power in the problem completion epoch ............................. 63 3 11 Topographic plots of l ower a lpha for o lder and y ounger a dults. ......................... 64 3 12 Lower alpha power in the frontal and parietal regions during all three epoch times ................................................................................................................... 65 3 13 Topographic plots of u pper a lpha for o lder and y ounger a dults. ......................... 66 3 14 Digit symbol means for LPS and HPS groups for younger and older adults ....... 68 3 15 The change score differences in solution time between LPS and HPS groups .. 69
9 3 16 Frontal and parietal theta power between LPS and HPS groups at all three epochs ................................................................................................................ 70 3 17 Lateral frontal theta activity between LPS and HPS groups at all three epochs ................................................................................................................ 71
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida i n Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LIFESPAN DIFFERENCES IN STRATEGY TRAINING BENEFITS: AN INVESTIGATION OF THE THETA AND ALPHA FREQ UENCY BANDS By Kimb erly H. Case May 2010 Chair: William Keith Berg Co Chair: Andreas Keil Major: Psychology Extensive research on the long lasting and beneficial effects of strategy training has demonstrated its importance in real world application While little work has been done to investigate the physiological changes following strategy training, the results are promising To further this work, younger and older adults were given strategy training on a difficult planning task (The Tower of London) while EEG was recorded. Both younger and older adults showed better performance following strategy training on directly trained problems as well as demonstrating transfer effects to other problems of a similar structure Only younger adults were found to improve on problems where the strategy could not be applied, demonstrating the older adults reliance on strategy training for cognitive improvement Frontal theta activity was found to increase following training replicat ing previous research that showed an increase in task re levant regions of the brain following strategy training Older adults were found to have larger increases than younger adults in frontal theta; a surprising but encouraging developmental difference. The alpha frequency band was found to be functionally rel evant to the task in both the younger and older adults show ing similar increases in the frontal regions in the post -
11 training condition. The strategy training was found to both increase behavioral performance and differentially affect the theta and lower alpha frequency bands Theoretical implications of the theta and alpha frequency bands are discussed in terms of their functional use during problem solving and how this changes across the lifespan.
12 CHAPTER 1 INTRODUCTION Across the human lifespan, th ere is a prolonged transition of advanced cognitive functions: a gradual onset of these throughout the childhood and teen years followed by a decline of these same functions starting in middle age (De Luca Wood, Anderson, Buchanan, Proffitt, Mahony, & Pateli s, 2003) The decline of these functions has been attributed to the shrinkage of the prefrontal cortex, a region thought to handle executive functioning/working memory tasks (Rypma & DEsposito, 2000) O lder adults who perform poorly on these tasks consist ently show altered brain activation patterns which may be indicative of a br eakdown of the neural components underlying these important functions (Rypma et al., 2000) While this decrease in cognitive function may be inevitable, there is a considerable amount of work demonstrating the benefits of cognitive training for the elderly in compensating for all or some of this loss (e.g. Friedman et al., 1998; Riis et al., 2008; Verhaeghen, Marcoen, & Goosens, 1992) The current study aims to further the understan ding of older adult cognitive plasticity through a training paradigm on a difficult planning/working memory task Brain activity will be collected before and after training to understand the neural differences associated with a ging and strategy acquisition. Changes in Cognition and Across the Lifespan The complex set of cognitive processes known as executive functions are thought to include the components of planning, working memory, inhibition and cognitive flexibility (Berg & Byrd, 2002; De Luca et al., 2 003; Welsh, Pennington & Groisser, 1991). These processes are believed to be controlled in large part by the prefrontal lobes (Welsh et al., 1991). Of the various executive functions, one of the most
13 complex is planning Planning is considered to be a highlevel cognitive component and involve s actions such as goal oriented behavior and strategy use (Berg & Byrd, 2002) Executive functioning abilities allow humans to succeed in a world of complex goals, plans and actions A decline of these functions starti ng in the 5th decade of life can have serious consequences for the individual in terms of stability and independent living S tudies have shown a clear connection between the ability for independent living and executive functioning skills ( C ahnWeiner, Boyl e, & Malloy, 2002; Grigsby Kaye, Baxter, Shetterly, & Hamman, 1998) Typically, many aspects of executive functioning decline across the lifespan which can leave an aging individual without the necessary components for successful living With a better und erstanding of this executive function decline we may be able to elucidate the necessary steps to take to slow down or remediate this inevitable regression of cognitive skills. Working memory is considered one of the fundamental components of executive func tioning ; without it one could not remember goals o r take action to complete them making it an essential aspect of executive functioning Working memory involves holding information in the mind and updating the currently stored information with new, more r elevant information (Baddley, 1992) T his ability is imper ative for independent living ; thus understanding its functional decline across the lifespan is inarguably worth further investigation and understanding De Luca and colleagues (2003) have carried out an important and comprehensive study of execut ive skills across the lifespan in participants ages 8 to 64. A battery of neuropsychological tests was administered including working memory and planning W orking memory was assessed using a task in which participants searched through an
14 increasing number of boxes to locate a hidden coin (trials contain 3 to 8 boxes) For successful completion of the task, participants must not return to a previously explored box, resulting in increased difficulty level with an increased number of boxes T he easier levels (in which the participant only needed to search either 3 to 4 boxes) resulted in all age groups having similar performance in completing the task Only on the more difficult trials (in which participants had to search through 6 to 8 boxes) did agerelated differences emerge. On these trials, older adults scored lower on a composite strategy score derived from task performance as compared to younger adults Mastery of less difficult levels in the older adult gr oup demonstrates an understanding of the task rules and at least a general strategy for completing the task During the most difficult level s of the task the participant must fully engage their cognitive skills in order to be successful It is possible that older adults were having difficulty in managing these more difficult levels because of a break down in cognitive skills needed to complete these more difficult tasks While older adults possess the ability for lower functioning skills they are taxed dur ing the more difficult levels of the task It is important to understand what cognitive functions and abilities are related to this difference in performance. Planning ability is another critical aspect of executive functioning and also declines with age ( Zook, Welsh, & Ewing, 2006) Planning involves an individual making goal oriented actions and achieving them through thoughtful and often strategically calculated moves Decline of this function has been thought to begin around the age of 60 ( Daigneault & Braun 199 3 ) A popular task used to assess planning is the Tower of London. This task involves participants moving balls on one board to match the color and orientation of balls on a goal board. Completing the task successfully involves
15 participants making goal oriented moves and applying a strategy for executing the most appropriate moves for completion The board contains three pegs and three balls which must be manipulated on the participants board to match that of the goal board. Increasing the number of moves required to match the goal board has been shown to increase difficulty in completing the task ( Berg, Byrd, McNamara, and Case, 2010) As was found in the working memory task, older participants are able to successfully complete easier planning problem s (De Luc a et. al, 2003) P erformance on the easier problems did not differ across age group with only the hardest difficulty level showing older adults perform ing worse on both measures of task performance ( number of perfectly solved problems and number of problems solved in maximum moves allowed; De Luca et al, 2003) A similar planning task, The Tower of Hanoi, has shown this pattern as well (Brennan, Welsh, & Fisher, 1997) It has been suggested that this drop i n optimal cognitive performance is due to the inability of older subjects to constrain their attention to the most task relevant features This ability, denoted as cognitive control, has been found to decline across the lifespan ( Persson, Lustig, Nelson, & Reuter Lorenz, 2007) A lack of additional cognitive resources would result in a decline or lack of cognitive control needed to successfully complete the task If older adults are unable to maintain cognitive control during a cognitive task interference could not be reduced and would result in poorer performance on the task This concept is useful in explaining why older adults are able to complete more simple cognition but struggle with more difficult concepts and tasks Neuroimaging work has demonstrated differential activation patterns of younger and older adults on tasks which require different levels of cognitive control ( Perss on et
16 al 2007) On the lowdemand task younger and older adults demonstrate d similar activation patterns O n the highdemand task however, older adults do not show as much of an increase in activation as the younger adults. It is suggested that cognitiv e control is handled by the anterior cingulated cortex, a region known to be impaired in older adults (Band & Kok, 2000) Electrophysiological work has replicated the findings of decreased activity in the older adult popula tion during highdemand tasks, ag ain suggesting an agerelated dysfunction in this cognitive ability (West & Moore, 2005) Changes in Brain Function Across the Lifespan Underlying the behavioral postulations of older adult cognitive decline is the consistent evidence of neural breakdown o f regions thought to handle the processing of executive functioning tasks Various methods have shown that older adults have decreases in frontal lobe white and gray matter, decreases in dopamine, and changes in fMRI BOLD and PET signals ( Venkatraman et al ., 2010; B rennan et al.,, 1997; Pradham, 1980; Gunning Dixon & Raz, 2003; Davis, Dennis, Buchler, White, Madden, & Cabeza, 2009) This neural breakdown has been correlated with the decrease in performance on executive functioning tasks (Raz, Dixon, William son, & Acker, 2002) R esearch on working memory has implicated the frontal regions, particularly the dorsolateral prefrontal cortex (DLPFC), as the main region underlying working memory processes (see Wager and Smith (2003) for a comprehensive metaanaly sis) This frontal region has also been implicated in contributing to planning ability, strategy application and goal orientation (Wagner Koch Reichenbach, Sauer, & Schlsser, 2006; Anderson, Albert & Fincham, 2005; Fincham, Carter, van Veen, Stenger & A nderson, 2002) In a study of healthy young adults, participants completed multiple trials of the Tower of London, and showed heightened activity in the DLPFC and other
17 regions known to be important in executive functioning abilities ( e.g. anterior cingulate cortex; Lazeron, Rombouts, Machielsen, Scheltens, Witter, Uylings, et al., 2000). Additionally the DLPFC has been implicated as being particularly susc eptible to aging (Romine & Reynold 2005) During memory tasks, older adult fMRI BOLD responses are co nsistently smaller than younger adults, particularly in the DLPFC ( Rypma et al., 2000) In a series of studies, Rypma and colle a g u es have demonstrated that the DLPFC is specifically affected by aging (Rypma, Berger, Genova, Rebbechi, & D'Esposito, 2005; Ry pma, Prabhakaran, Desmond, & Gabrieli, 2001; Rypma et al., 2000) Participants were presented with a string of letters (varying from 1 to 8 letters) and told to hold those letters in mind for a certain period of time (512 seconds depending on the study) After this delay period, participants were presented with a letter and asked if this letter was in the string of letters presented. Behavioral performance showed a decline in performance with increasing string lengths for both age groups Both the ventromedial prefron tal cortex and DLPFC were activ e during the task in both age groups, though only the DLPFC was found to have decreases in activation in the older compared to younger adult groups. While portions of the frontal regions seemed to be spared by aging, the DLPFC demonstrates declines in both function and structure ( Bae, MacFall, Krishnan, Payne, Steffens, & Taylor 2006) While the above mentioned literature reveals the decline of activity in important brain regions across the lifespan, there is e vidence to suggest that high performing older adults may engage compensatory neural mechanisms during difficult cognitive tasks. Correlation analyses have demonstrated differential activation patterns in the DLPFC across the lif espan. Rypma and colleagues ( 2005) showed that y ounger adults
18 who performed better on the task (faster reaction time) showed decreased DLPFC activation as compared to slower younger adults For older adults this pattern w as opposite; the high performing older adult s (faster reaction time) showed increased DLPFC activity as compared to the slower older adults This result suggested that the younger adults demonstrat ed better performance because of a higher working memory capacity and thus needed fewer neural resources to complete the task. Older adults on the other hand, had worse behavioral performance because of a low working memory capacity (or possibly a lack of cognitive control) and thus needed to recruit additional neural resources in order to perform as well as the younger adults These highperforming older adults have adapted a method of neural compensatory activity which aides in the completion of difficult working memory tasks This compensatory activity has also been demonstrated in the Hemispheric Asymmetry Reduction in Older Adults model (HAROLD; Cabeza, Anderson, Locantore, & McIntosh 2002) Research concerning this theory has shown that high performing older adults have bilateral activation during cognitive tasks, while younger adults show more lateralized activity ( R ajah & McIntosh, 2008; Lee, Leung, Fox, Gao, & Chang, 2008; Cabeza, 2002) This finding indicates that older adults are recruiting additional areas to aid in task performance. The increase in activation is thought to be compensatory since those older adult s who recruited bilateral regions showed better performance on memory tasks as compared to other older adults who had more lateralized activity (Cabeza et al 2002). Improving Working Memory in Younger and Older Adults Given the large amount of literature documenting the decline of cognitive functions across the lifespan, it is not surprising that researchers are striving to understand how to
19 improve this A plethora of research has found positive effects of training for older adults demonstrating the plasticity of the brain in the later stages of life (Becic et al. 2008; Bottiroli, Cavallini, & Vecchi, 2007; Carretii, Borella, & Beni, 2007; Cavallini et al. 2003; Rapp, Brenes, & Marsh, 2002; McNamara & Scott, 2001) The variety of research in this area has led to some general conclusions about memory training enhancements and its agedifferentiated effects across the lifespan. First, older adults have the ability to increase memory performanc e demonstrating a continued plasticity of these functions for cognitive learning (Singer et al. 2003) Intensive cognitive training has been found to improve performance in older adults in a myriad of cognitive domains ( Sensory and cognitive training: Mahncke, Connor, Appelman, Ahsanuddin, Hardy, & Wood, et al., 200 6; Visual search training : Becic, Boot, & Kramer, 2008; Ecologically valid tasks: Cavallini, Pagnin, & Vecchi, 2003; Speed of processing: Edwards, Wadley, Myers, Roenker, Cissell, & Ball, 2002; Working memory: Dahlin, Nyberg, Bckman, & Neely, 2008) Train ing paradigms implemented with the oldest old (over 75 years old to 101 years old) have also demonstrated benefits through intensive training (Singer, Lindenberger, & Baltes, 2003; Yang & Krampe, 2009) This continued plasticity, even into the latest phases of life, shows the importance and value of this technique. A second important finding from these studies is the ability of these effects to be sustained over time If the effects from training last only for a short time, this greatly dim in ishes its impor tance for changing the life of an older adult But i n a comprehensive study of the duration of memory training effects, training induced improvements were still found at a 2 year follow up in both the younger and older adults ( Ball, Berch,
20 Helmers, Jobe, Leveck, & Marsiske et al., 2002 ) This finding has been replicated in other research, using more ecologically valid tasks ( Bottirol i et al. 2007). A critically important aspect of memory training is whether that training can and will be of benefit in circu mstances beyond those specifically trained. This cognitive process, termed transfer of training is important because it demonstrates that the participant is not only applying the procedure originally learned to the type of problem specifically trained, but can also generalize that knowledge to a similar or dissimilar problem or task. If the participant can transfer the knowledge to a different problem or task, it can be suggested that they understand the procedure or strategy involved and are not just g oing through the motions. Six levels of transfer have been outlined by Haskell (2001), ranging from simple learning and application (Levels 1,2 and 3) to more advanced transfer of training (Levels 4, 5 and 6) The more advanced levels are thought to evidence a true transfer of the learned strategy to a new situation. The main types of transfer studied in the aging literature are Near and Far transfer (Levels 4 and 5, respectively) Near transfer occurs when there is a transfer of previous knowledge to a n ew situations which is similar to the original situation Far transfer occurs when there is a transfer of previous knowledge to a new situation which is not similar to the original situation Near transfer would be if a participant were to be trained on a working memory task and were able to apply that learning to a different type of working memory task. Far transfer would be the ability to transfer the learning from the working memory task to a planning task The terms direct and indirect are applied h ere though these terms are analogous with the more popular terms near and far These terms are
21 applied as the methodolog y employed in this study are better describes as direct and indirect as compared to near and far. Younger adults have generally shown consistent transfer of strategies to problems similar to those trained, known as direct transfer problems (Li, Schmiedek, Huxhold, Rcke, Smith, & Lindenberger, 2008; McArdle & Prindle, 2008; Wright, Thompson, Gains, Newcombe, & Kosslyn, 2008), but show less consistent transfer to less related p roblems, known as indirect transfer problems (Dahlin et al. 2008; McArdle et al. 2008; for a review, see Rebok, Carlson, & Langbaum, 2007) Older adults often do not show transfer on the indirect problems, whic h concerns researchers that laboratory based memory training will not affect nonrelated tasks ( Rebok et al., 2007; Ball et al., 2003) In a comprehensive study of the transfer of working memory training to indirect and direct tasks, younger and older adul ts were trained on a spatial working memory task over several weeks (Li et al., 2007) The trained task, the spatial nback, taps updating working memory by requiring participants to indicate a match between the current stimuli and two stimuli previously presented (2 back) To measure the effect of transfer to similar (direct) and closely related tasks (indirect), two direct transfer tasks and one indirect transfer task w ere given to younger and older adults after training in the spatial working memory task Performance on these tasks was compared to a comparable group on nontrained younger and older adults The first direct transfer task, a harder version of the spatial nback task (3 back) showed better performance in the trained as compared to nontrained younger and older adults However this type of direct transfer only demonstrates that participants increased in a specific skill and not a generalized gain of knowledge. The second direct transfer task
22 was a 2back working memory task which utilized dif ferent stimuli though required the same cognitive skills for completion. Again, both age groups that were trained successfully complete d this task better as compared to the nontrained group. This shows that the training was able to be transferred to a tas k using similar cognitive processes but different enough to demonstrate that the cognitive training was effective in aiding performance on this task The two indirect transfer task s included tapping cognitive skills related to but not completely reliant on working memory These tasks required participants to use complex memory span, a construct that is not directly related to updating of working memory though can be argued to be in the same family of cognitive functions In both of these tasks, the nontrai ned and trained participants did not differ in their performance. T raining on the spatial working memory task did not improve performance on a different type of working memory task in either the younger or older adults Th e lack of transfer to indirect tas ks has been mirrored in other studies as well In a large study of adult memory training older adults trained in several cognitive skills did not demonstrate improvements in everyday memory tasks ( Ball et al., 2003) One area of research that has been abl e to demonstrate transfer effects is training in processing speed ( Ball, Edwards, & Ross, 2007) It is suggested that cognitive slowing is an important factor in cognitive decline across the lifespan ( Zimprich & Martin 2002) making the impact of processing speed training understandable. In a meta analysis of 6 studies which employed speed of processing training, it was found that age, education and mental status did not impact the level of improvement from training (Ball et al., 2007) The only factor whic h showed significant relation to the improvement due to speed of processing training was the initial speed of processing
23 performance of the participant Those with the lowest scores on the speed of processing test at the initial baseline showed the largest levels of improvement This indicates that this group of participants greatly benefits from the training and the construct of speed of processing is an important factor when measuring performance gains across the lifespan The importance of speed of processing in cognitive functions has been demonstrated in many cognitive paradigms (for a review, see Salthouse, 2005). Since processing speed is defined as the speed with which different cognitive operations can be executed it stands to reason that increase s in this ability would aid skill acquisition (Reichenberg & Harvey, 2007). Additionally, the task has been associated with working memory functions as those participants who are able to remember the pairing would complete more pairs (Koziol & Budding, 200 9). Older adults have been found to have consistently poorer behavior on this task compared to younger adults. While this test clearly measures processing speed, the impact of speed of writing may also play a role; the age difference relating to a slower motor ability rather than a cognitive one (Joy, Kaplan & Fein, 2004). It is likely though that both cognitive and motor slowing impact the performance on this task. In an adapted version of the digit symbol task, younger and older adults completed the task during f MRI scanning (Venkatraman et al., 2010). Both younger and older adults showed bilateral frontal activation and left parietal activation, regions both associated with working memory/executive control. Additionally, activation in these regions was associated with better performance and fewer structural abnormalities, as measured by white matter tracks across the brain (gained from the
24 measure of diffusion tensor imaging ) This suggests that completion of this task relies heavily on brain regions acti ve in cognitive, rather than just motor abilities. O lder adults have been found to have poorer performance pretraining and have fewer gains from training as compared to younger adults (Verhaeghen & Marcoen, 1996) This finding is possibly due to the ineffective strategy application of older adults on cognitive tasks Brigham and Pressley (1988) demonstrated this by teaching all participants two strategies ; one which was an effective process for completing the task and a less effective strategy The task learning a list of vocabulary words, was administered to younger and older adults before and after the strategy training All participants received training on both types of strategies All but one younger adult (out of an n of 30) was able to tell which strategy was superior and employ that strategy The older adults did not successfully choose the e ffective strategy and could not tell which strategy was more effective. Poor strategy choice in the older adult group indicated a reduced meta cognitive awar eness of the differences in strategy application. This same finding has also been replicated in more recent work (Lamson & Rogers, 2008; Yagoubi, Lemaire & Besson, 2005) In this light, it may be that older adults are struggling with working memory tasks, and executive tasks in general, because of a poor strategy choice. When older adults are taught a relevant strategy their performance increases dramatically because they are now exposed to the correct way of solving the task This was demonstrated in a stu dy which older adults who chose the wrong strategy were trained to choose more optimal strategies (Becic et al. 2008) P erformance was substantially improved in the older adults who were given strategy
25 choice training demonstrating that older adults are capable of completing these tasks when shown the most optimal strategy. Physiological Changes Accompanying Working Memory Training Studies of physiological changes after training have consistently shown increases in activation of task relevant areas in th e post training condition ( Olesen, Westerberg, and Klingberg, 2004) In a series of studies, Klingberg and colleag u es had younger adults complete an intensive training period; fMRI activity was measured before and after training ( Westerberg and Klingberg, 2007; Olesen et al., 2004 ) T raining included repetitive presentation of a spatial working memory task which was adapt ed to the participant s level of achievement As the participant improved on the task, it became increasingly more difficul t which allowed the participant to improve their working memory capacity Analyses investigated the changes in activity from the preto post training condition. The pattern of activity was similar in the preand post training condition, though the post training condit ion showed significantly higher levels of activation as compared to the pretraining condition Thus, there were no additional areas of activation in the post training condition, only an increase in previously activated regions. Physiological measures can also highlight differences in activation patterns across the lifespan (Rosano, Aizenstein, Cochran, Saxton, de Kosky, Newman et al., 2005) Brain activity measured before and after strategy training may indicate differential patterns of activation across t he lifespan, highlighting developmental differences not apparent in behavioral research. T he Method of Loci ( M.O.L ) training is beneficial to both younger and older participants as demonstrated in numerous behavioral studies employing this methodology ( Be cic et al. 2008; Bottiroli et al., 2007; Carretii et al.
26 2007; Cavallini et al. 2003; Rapp et al. 2002; McNamara et al., 2001) This technique has participants associate imagined places ( loci) along a walk with the stimuli to be remembered. The partici pant is then able to memorize the stimuli by associating them with more familiar stimuli. Nyberg and colleagues (Nyberg et al., 2003) trained younger and older adults in M .O.L, collecting PET images before and after training The post test condition showed increased activation of task relevant regions in all participants Behaviorally, participants were divided into three subgroups The first group was the younger adults, all showing substantial improvement on the behavioral task in the post training condi tion and showing the best performance in comparison to the other groups The older adult age group was then segmented into older adults who benefited from the strategy (Old +) and those who did not (Old ) While in the pretraining condition the Old + and Old did not differ in regions or strength of activation, several post training differences emerged The Y oung and Old + adults showed increased activity in the parietal/occipital regions, areas know n to be active during this type of task ( Cabeza & Nyber g, 2000) This is in contrast to the Old group who did not show any increases in parietal/occipital activation in the post training session Thus, though the Old + performed worse on the task as compared to the younger adults, their pattern of task relev ant brain activation was similar The Old + group was succ essfully employing the strategy which was demonstrated both behavioral ly and physiologically The Old group, on the other hand, did not dem onstrate strategy application and showed differential ac tivation patterns as compared to the Old +. A recent study aimed at training older adults in cognitive control demonstrated that training based on this construct is both possible and beneficial ( Braver, Paxton, Locke,
27 and Barch, 2009) Older adults were tr ained to the strategy of attending to the cue stimuli rather than the target stimuli for a continuous performance task Behavioral data demonstrated an increase in cuefocused behavior and evidence of increased cognitive control Activation of the prefront al cortex ( PFC ) was found to have increased in the second session, and this increase was related to performance on the task Better performance on the post training task was associated with higher activation in regions of the PFC These training studies have demonstrated that while older adults may show worse behavioral performance as compared to younger adults, they are activating many of the same brain regions Why D o W e S ee V arying C hanges in Task R elevant B rain A ctivation A fter T raining? As noted above, physiological changes after strategy training result in an increase in the taskrelevant brain regions Studies of motor strategy training have shed light onto the different phases of strategy acquisition and implementation that occur as training proceeds. When a participant is taught a motor strategy on a task there are several stages of acquisition and implementation that are required prior to mastery of the learned strategy (Doyan and Benali, 2005) During the first stage of motor strategy application, participants are bring ing new processes online for task completion that were not previously employed. This first stage is reflected by an increase in activity in task relevant regions during this time. Later phases of motor strategy application, in which participants may be consolidating information, are denoted by continual decreases in activ ation. This is thought to represent an increased efficiency in neural processing and thus a decreased need for neural resources If this same theoretical framework is applied to cognitively based strategy application, the participants in the studies
28 described would still be in the first phase of strategy application This is possible for two reasons Firstly, most of the studies looking at physiological changes after t raining are single session testing paradigms ( e.g. Nyberg et al., 2003) and even those that have extensive practice do not test the participant on the task for more than one session (e.g. Westerberg et al., 2007) Secondly, it is feasible that cognitive t asks may require more practice to become automated as compared to motor tasks which can have simpler demands As such, it might be expected that post training physiological signatures would show increased activation in the initial post training phase as comp ared to the pre training phase. The Use of Electroencephalogram Activity in the Study of Human Cognition While the literature discussed above has relied on fMRI techniques to distinguish regions of activation across training sessions, other physiological techniques m ay also be useful To date, my careful review has not been able to locate any studies of training investigating the changes in electroencephalogram ( EEG ) signatures pre to post training The use of this method can provide additional informat ion that is lacking in the previously conducted research. Of most importance is the temporal resolution afforded t o EEG that is lacking in fMRI It is likely that there are different regions of activation immediately following the presentation of the strategy (as in the first several seconds after the strategy is being explored) as compared to later in the trial when the strategy is more actively applied. While fMRI analysis is typically done by collapsing the BOLD signal across a trial or several trials E EG signatures can be examined in millisecond accuracy to understand fluctuations in power and topography as the early process of strategy analysis, the subsequent exploration and evaluation, and later strategy
29 development and application unfold. This technique is especially useful in cognitive paradigms that study long lasting processes, such as planning and problem solving In order to understand the complex EEG data collected during the task, researchers have employed various data reduction analyses The statistical procedure often used to extract frequency information from the EEG is a Fast Fourier transform (FFT) This method gives a measure of power in each of the frequency bands associated with traditional EEG analyses, such as delta (24hz), theta (47hz) and alpha (8 12hz) etc It allows the investigator to understand how the power of each frequency changes across independent variables For example, Sauseng and colleagues (Sauseng, Klimesch, Doppelmayr, Hanslmayr, Schabus & Gruber, 2004) found that activity in the theta frequency band increases with an increase in task diff iculty level. They report that increase s in theta power indicates that more resources are being put towards the task now that it is more difficult for the participant Theta Changes in Power A cross the Lifespan Theta power has been shown to increase during cognitive tasks, a sign that this frequency band is an appropriate indicator of mental effort Young adults performing a serial letter task (Sternberg task) showed marked increas es in theta power in the frontal regions (Onton et al. 2005) Increases in memory load (05 letters) and the maintenance period (2 4 seconds) resulted in significant increases in theta power Regions in the frontal cortex (frontal midline theta, discussed below) showed progressive increases in theta power with increasing difficulty level Other regions of interest did not show comparative theta increases, demonstrating this is specific to the frontal regions for task manipulation. Increases in theta power related to increases in the maintenance period demonstrate that theta power is not only linked to a number of
30 stimuli but also to mental effort In support of this finding several working memory paradigms have also demonstrated increases in theta power during cognitive tasks ( Ragavachari, Lisman, Tully, Madsen, Bromfield, & Kahana, 2005; Mizuhara, Wang, Kobayashi, & Yamaguchi, 2004; Sauseng, et al., 2004 ; Grunwald, Weiss, Krause, Beyer, R ost, & Gutberlet et al., 1999) As noted, the frontal theta power has been found to be concentrate d in the frontal midline (FM) region under higher task demands Researchers have thus termed this frontal theta power as FM theta. As discussed above, frontal theta was correlated to increases in memory load. When the impact of the FM theta was statistically removed from the signal, a correlation of task load to theta power was no longer significant This confirms that the association of theta power with mental effort is concentrated in the frontal midline region Source localization of the FM theta suggest s the cortical generator is the anterior cingulate cortex (ACC) (Onton et al. 2005 ; Gevins et al. 1997) The ACC has b een implicated in attentional systems an d significant cognitive effort, so its involvement in frontal mi dline theta is reasonable. Frontal midline theta has also been found to increase with increased practice and strategy application on a task (Gevins, Smith, McEvoy & Yu, 1997; Smith, McEvoy, & Gevins, 1999) Participants completing the nback task demonstrated increases in FM theta during the task and increases during more difficult levels of the task Interestingly, FM theta was higher in the last blocks of trials as compared to the first block of trials It is suggested by Gevins and coauthors that this increase in FM theta is due to increased mental effort on the task with more advanced strategy usage. As with the fMRI training literature, FM theta demonstrates an increase in activ ity after practice.
31 This suggests that FM theta is an appropriate tool for investigating physiological changes due to strategy training. S everal studies have found decreased power in frontal midline region in older adults ( Cummins & Finnigan, 2007; McEvoy, Pell ouchoud, Smith & Gevins, 2001) D uring a working memory task (n back) both younger and middle aged adults demonstrated increases in FM theta with i ncreasing difficulty level The older participants, h owever, did not show an increase in FM theta during increased task difficulty which is thought to be indicative of a breakdown of processing mechanisms in the frontal regions This result has been replicated in other studies of older adult FM theta During a modified Sternberg task (in which the memorized set was words rather than letters), older adults showed lower FM theta l evels as compared to younger adults in both the retention and recognition intervals of this task ( Cummins et al. 2007; Karrasch, Laine, Rapinoja, & Krause, 2004) Results again suggest disruption in the frontal regions in older adult groups Based on thes e two studies, many further questions can be asked. Is the decrease in FM theta due to age related changes in brain structure or possibly due to differences in task performance and mental effort? As suggested by the training literature, one would expect an increase in FM theta with increased mental effort and strategy application Further investigations of FM theta and its changes in strategy training may answer some questions about its underlying functional significance to cognitive tasks While there is currently no work investigating the differences in theta power before and after training, it is expected that this change will follow suit with the fMRI research
32 ( e.g. Nyberg et al., 2003) Frontal midline theta was been found to increase in power with in creased exposure and practice on a task (Gevins et al 1997) Younger adults performed a spatial nback task, with varying difficulty levels, over a prolonged period of time In addition to demonstrating increases in theta power with increased difficulty of the task, practice effects were also investigated. Comparison of frontal midline theta between the first and last block of problems showed a clear increase in power with increased practice Interestingly this interacted with difficulty level with larger increases in theta power with practice for more difficult problems The authors infer that this is related to task performance and the increasing demands and difficulty of maintaining attention on this task for an extended period of time. Clearly, fronta l midline theta is an appropriate tool to study developmental changes in strategy application due to its sensitivity to age, difficulty level and practice effects. Alpha Changes in Power A cross the Lifespan Although it has been postulated in the past that the function of alpha was solely for periods of inactivity or brain idling ( Jensen, Gelfand, Kounious & Lisman, 2002) more recent research has begun to uncover a very different picture of what alpha brain activity may represent C ontrary to the notion that alpha activity represents a relatively inactive brain, several studies have shown increases in alpha activity during cognitive tasks ( Sauseng, Klimesch, Schabus, et al., 2005; Sauseng, Klimesch, Stadler, et al., 2005; Silberstein, Danieli & Nunez, 200 3; Jensen, et al. 2002) and this increase has been associated with better performance (Hoptman & Davidson, 1998) The alpha frequency band has been divided and designated as lower (810hz) and upper (1113hz) alpha frequency bands
33 The lower alpha frequency band has been associated with changes in cognitive tasks and found to be generated in the frontal and parietal regions ( McEvoy, et al. 2001; Gevins et al. 1997; Klimesch, Schimke, & Pfurtscheller, 1993) Younger adults completing a difficult working memory task showed clear increases in the lower alpha frequency band during the task as compared to the rest condition (Gevins et al. 1997) The upper alpha frequency band has been associated with basic visual processing by some researchers ( McEvoy et al., 2001; Gevins & Smith, 2000) though it has also been attributed to being involved in long term memory processing (Klimesh, Freunberger, & Sauseng, 2009) When investigating power changes between task conditions, there seems to be little increase in the upper alpha frequency band (McEvoy et al., 2001) However, Klimesh and colleagues who have investigated the long range communication of this frequency band suggest a more cognitive role. Klimesch and colleagues have developed a theoretical framework to s uggest a functional role of the alpha frequency band which spans both lower and upper ranges The Klimesch framework suggests that alpha is playing a functional role of long range communication between frontal and parietal brain regions necessary to complete working memory tasks. Thus, the alpha band is no longer seen as an indication of idling but as a possible index of functional relevance to certain task s. T his view would suggest an increase in alpha with increased mental effort similar to the function of the theta frequency band However, other research would suggest a different functional role for the lower alpha band. Jensen and colleagues have shown an increase in lower alpha in non task relevant regions suggest ing alpha is inhibiting regions not nec essary for task completion (Jensen et al. 2002) Clearly, the function of the alpha band is not
34 fully understood. While the ex act role of alpha is not well defined, it clearly has a functional relationship with cognitive tasks Further investigation into the alpha band will help to more fully understand its role in the completion of cognitive tasks. Research in older adults investigating the lower and upper alpha frequency band have shown differential patterns of alpha activi ty across the lifespan (McEvoy et al., 2001) During an nback task, younger adults showed increased deactivation of the lower alpha frequency band during task performance in both the frontal and parietal regions This was interpreted as younger adults showing increased task engagement in these regions, following the more traditional view of alpha as a measure of cortical idling Older adults were found to show decreases in only the parietal regions without a decrease in the frontal regions The upper alpha frequency band did not show c lear changes across age or task epochs suggesting a less cognitive role. Clearly, this result could be interpreted several ways, depending on the applied theoretical model Further investigation of what the lower and upper alpha bands may represent and its changing function across the lifespan is needed to interpret the presently available evidence. Current Study T he study employs a wo rking memory training technique on a difficult variant of a spatial planning task (the Tower of London) with younger and old er adults in order to understand the changes in behavior and physiology as a consequence of training The Tower of London task offers several advantages over more commonly used training tasks. Firstly, the complexity of the task allows for large differenc es in performance, shedding light on factors important for individual ability to perform on the task and effects of the trained strategy Secondly, while strategies employing more simplistic training techniques are widely present in the literature (e.g. MO L) little work has been
35 done to understand complex strategies involving spatial manipulation (Li et al. 2008) The Tower of London makes significant demands on working memory allowing comparison of these results to previous working memory studies while adding new information to the current knowledge of strategy training across development (Phillips, Wynn, Gilhooly, Sala, & Logie, 1997) The Tower of London activates similar regions of the brain as do standard working memory tasks, such as the DLPFC, and incorporates regions known to handle spatial manipulation (e.g. the parietal lobe ; Anderson et al., 2005) As discussed above, the developmental performance trajectory of the Tower of London has similar timing to working memory paradigms (De Luca et al., 2 003) Thus, while this is the first study to employ strategy training with the Tower of London, the theoretical justification and application to previous work is clear. Based on previous research, it is expected that i nitial performance on the TOL will be better in the younger age group as compared t o the older age group. This will be shown by the younger adults showing more correct responses and faster solution times on the TOL. Training is expected to result in increased performance for younger and older adults, though previous research would suggest a greater benefit of training for the younger age group (L ustig, Shah, & Seidler, 2009) Investigations of theta power will be limited to the frontal and parietal regions to further explore the decrease in FM theta in older age. It is expected that older adults will exhibit lower theta power as compared to younger adults specifically in the FM theta region (McEvoy et al., 2001; Cummins et al., 2007) with the older adults having a relatively smaller increase i n theta in the post training session as compared to the younger adults Hemispheric differences in the frontal regions will be explored as this has been found to be import ant in TOL
36 performance ability ( Goethals, Audenaert, Jacobs, van der Wiele, Pyck & H am, 2004 ; Morris, Ahmed, Syed, & Toone, 1993) Since the alpha band has not been found to have a unifying theoretical basis of change across age or task, exploratory analyses will concentrate on regions of interest that may shed more light on its functiona l importance in this cognitive task The lower alpha frequency band will be investigated at the frontal and parietal sites, as interesting developmental differences were found in a cognitive task (McEvoy et al., 2001) The upper alpha frequency band will be investigated in the parietal regions, as it has been found to be maximal in this region (McEvoy et al., 2001; Gevins et al., 1 997) And finally the impact of individual variations in processing speed will be explored to assess its importance and impact on brain changes.
37 CHAPTER 2 METHODS Participants Participants were tested in a pre/post design, with all participants completing two sessions. The second session, spaced one to four days following the first session, includ ed a strategy training program All participants analyzed completed both sessions successfully, indicated they were right handed and use English as their primary language. All participants were asked of head trauma or neurological diseases likely to affect the brain (e.g. Parkinsons di sease, stroke), in which all indicated they had not. The younger adult group ( n = 32) was between the ages of 18 and 24 and recruited from the Gainesville, Florida area. Most younger adult participants (87%) were recruited from the general psychology cour se held at the University of Florida and received credit for their participation. The remainder of the young adult group was also recruited from Gainesville, Florida and had similar educational backgrounds The older adult age group ( n = 2 4 ) was between the ages of 63 and 79 and recruited from the Gainesville, Florida area. All older adults who participated in the study were deemed cognitively capable by the administration of the Telephone Interview for Cognitive Status (TICS; Folstein, Folstein, & McHugh, 1975) The TICS is an 11 question assessment of orientation, concentration, short term memory and delayed recall done over the phone to screen participants for mild cognitive impairment The use of the TICS as an assessment of cognitive status has been val idated in comparison to more indepth cognitive interviews demonstrating the clinical use of this tool for assessment of mental stability in an older adult population (Knopman, Roberts, Geda, Pankratz, Christianson, Petersen, et al., 2010; Duff, Beglinger, & Adams, 2009) In addition, the TICS has been found to be have
38 similar diagnostic outcomes as compared to the Mini Mental State Examination (MMSE; Ferrucci, Del Lungo, Guralnik, Bandinelli, Benvenuti, Salani, et al., 1998). As suggested in comparison to other more extensive tools of assessment, a cut off score of 28 was used to screen out participants with cognitive deficits (Barber & Stott, 2004) Only one screened participant had a score of 28 but this person was excluded due to missing the second appointment. Table 1 outlines the demographic information of the final N for this study Of the ninety one participants that were originally screened to participate in this study, fifty six are included in the final N The main reason for rejection for these p articipants was failure to return for the second appointment, found mostly in our younger subjects The second most common reason for rejection of a participant was poor quality EEG The term poor quality EEG is an umbrella for many problems which happen ed with the EEG including poor signal (n=2), too many motor artifacts in the data (n=12), and equipment failure (n=7) Only two participants were excluded due to experimenter error. Table 2 1 Demographic and Neuropsychological Data of the Final Sample Age (st.dev) Digit symbol Score (st.dev) TICS Score (st.dev) Percentage of Males Percentage of Caucasian Younger Adult s 20. 0 4 (1.64) 70.00 (8.41) N/A 54% 74% Older Adult s 70.81 (4.55) 50.68 (11.75) 36.20 (2.77) 57% 91% N = 5 6 Measures Digit Symbol Task To assess individual variations in processing speed, the d igit symbol substitution task was administered during the first testing session (Wechsler Intelligence Scale for Adults 4th Edition) This speeded task has part icipants completing pairs, based on
39 presented s timuli to assess processing speed The applicability of this task has been shown in models which indicate processing speed as an important predictor of strategy usage and memory performance (Verhaeghen and Mar coen, 1996) In this task, participants are shown the numbers 19 and their matched symbol (see Figure 21, block A for a fabricated example similar to that used in this test) During the test, participants must correctly write in the symbol below the numb er (Figure 21, block b) There are a total of 93 pairs to complete within a total time of 90 seconds No participants were able to complete the entire board in the allotted time. The score on the digit symbol was derived by taking the number of correct responses minus the number of incorrect responses Average performance on this task can be found in Table 1. Figure 21. Digit/Symbol Task a task of processing speed assessed during the pretraining session in both younger and older adults. (A) The legend of the test which tells participants the matching pairs, (B) The portion for the participant to fill in during testing Tower of London Planning Task The TOL is a planning task in which participants are asked to manipulate the balls on the move board to match the balls in color and orientation to the goal board (See Figure 22) There are several rules to the TOL: only one ball can be moved at a time,
40 balls must always be placed on a peg, and there a is maximum of three balls that can fit on the tall peg, two balls on the middle peg, and only one ball on the small peg. A modified version of the TOL was administered to participants in which the participant was only able to view the picture of the move and goal boards and was no t able to actually manipulate the balls In the typical version of the TOL, participants are able to move the balls on the move board to correctly match the ball configuration on the goal board. In this modified version, participants were asked to covertly manipulate (in their heads) the balls into the correct order This modified version has been successfully used in previous fMRI research (e.g., van den Heuvel, Groenewegen, Barkhof, Lazeron, van Dyck, and Veltman, 2003) and was employed here for several reasons Firstly, the manual manipulation of the balls would create excess artifacts that could have rendered the EEG data useless Secondly, this covert solution paradigm results in a substantial increase in the demand on working m emory, including both t he short term storage and the manipulation aspects of working memory Throughout the solution, participants must keep in mind what balls they have moved and where the current ball placements are. With this change in task presentation, the working memory st rategy training can be expected to become far more critical. Figure 22. Three types of Tower of London Problems Presented problems could either be categorized as switch (Tall or Middle) or control problems.
41 Participants were presented with 54 trials of the TOL during each session to adequately measure performance on this task before and after training TOL difficulty level was measured by number of moves needed to complete the problem Each problem required the participant to move five or six balls for correct completion Number of moves has been used in the past as an indication of the difficulty of the problem (Berg et al., 2002) De Luca, et al. (2003) found that older adults struggled with 5move problems making this pr oblem selection ideal with detection of strategy application. During each TOL trial, a problem was presented for a maximum of 45 seconds During this time participants attempted to solve the TOL problem covertly and, whenever ready respond with the color of the last ball put into place using a simple keypad provided (see Figure 22 ) Using this response technique, a correct answer would indicate that the participant solved the problem in the fewest number of moves or guessed the correct answer Responding above chance level (33 percent) indicates that participants are solving the problems If the participant did not respond within the allotted 45 seconds they were automatically moved on to the next problem Once the participant responded or the 45 second maximum was reached, there was a brief inter trial interval of five seconds where the participant viewed a small red circle in the middle of the screen After this rest period, the next problem was presented. Feedback on task performance was not provided during the testing procedure. Selected p roblems for the T ower of L ondon Training of a selective application of the strategy to some, but not all, presented problems was accomplished by constraining the types of problems presented to participants Two out of the three problem types presented allowed participants to either
42 directly or indirectly apply the learned strategy The other type of problem (control problems) could not be solved by applying the strategy and served as a control condition. Switch Problem s: Problems that were able to be solved by applying the trained strategy are identified as switch problems These problems are characterized by having two balls on a single peg in opposite orientation for the goal and move board. For example, in Figure 22 middle panel, balls on the tall peg for the move board are ordered red green (red on top, green on bottom) but are on the goal board as greenred Since this opposite orientation is occurring on the tall peg, these are identified as tall switch problems Those referenced as middle switch problems, shown in the right panel, are analogous with balls being reversed in order on the middle peg Participants received training on only one type of problem either tall or middle switch training The training could then be applied directly to trained problems (e.g. tall switch training directly applied to all tall problems) or potentially transferred to assist in indirectly solving trained problem s (e.g. tall switch training indirectly applied to middle switch probl ems) Control Problems: These problems are identified by a lack of balls being in the switch orientation from the move to goal boards ( Figure 22, left panel) Ball placement for these problems was selected based on the same number of moves required ( fiv e or six ) as the switch problems These problems were presented to effectively measure the effect of practice alone with the task rather than specific strategy training The switch strategy could not be successfully applied to these problems, and thus any increase in
43 performance on the second session would indicate general practice gains, not increase in performance due to strategy application. The 54 problems presented per session were divided into 18 control, 18 middle switch, and 18 tall switch problems Problems were presented in nine blocks, six problems per block Each block contained two problems of each type (a list of presented problems can be found Appendix A ) and order was pseudorandomized within the block Performance measures included proportion correct, solution time and proportion of no responses ( not responding by the end of the 45 second interval ) Working m emory s trategy t raining During strategy training, all practice problems and discussion of the application of the strategy were performed on either the tall or middle switch problems No mention of the other type of switch problem was discussed during training Participants were randomly selected into either the middle or tall switch training conditions Training was presented in PowerPoint format on a screen sitting in front of the participant First, training consisted of the participants being able to correctly recognize a switch problem Once participants were able to complete this step, demonstration and application of the strategy is a pplied to 4 5 and 6 move switch problems with appropriate corrective feedback from the researcher provided as needed. No participants needed to complete the training more than one time. During the first part of the training, participants were trained on the basic sequence for 4move switch problems As shown in Figure 23, this sequence involves four moves to correctly reverse the balls and solve a 4 move switch problem Participants were trained on solving this type of problem two times and then given two chances to apply the strategy with corrective feedback as needed.
44 During the second part of the training, extensions of the basic strategy were demonstrated by building upon the basic sequence. Participants were taught three such extensions These ad ditional strategies are involved in solving the 5and 6move problems For the 5move problems, a ball can be moved prior to the switch taking place ( Figure 24) or after the sequence has taken place ( Figure 25) depending on the specific problem Partici pants are trained to recognize when these additional moves are necessary and also given training on applying this extension to solving the problem The researcher demonstrated each of these extensions two times with the participant given two times in whic h to apply each extension while receiving corrective feedback as necessary For the 6 move problems, a n additional ball move is needed both prior to as well as after the basic sequence ( Figure 26) As with the previous extension, participants are given th e opportunity to see a demonstration and try an application of this technique. This ensures that all participants not only have a chance to see the board being solved but also to solve it on their own. Figure 23. Depiction o f Moves Needed for the Basic Strategy Sequence, Participants were first taught this strategy sequence and then trained on more advanced sequences.
45 The entire training session lasted an average of 20 minutes In total participants received four problems for identifying the switch problems, five problems of demonstration of the training sequences, and eight problems in which they were able to apply the trained sequences Participants gave verbal confirmation that they understood each strategy trained and wer e ready to move on to the testing problems. Figure 24. Depiction of 5 move Extended Technique 1 This technique adds one move to the beginning of the problem (5 move problem) Figure 25. Depicti on of 5 move Extended Technique 2 This technique adds one move to the beginning of the problem (5move problem).
46 Figure 26. Depiction of 6 move Extended Technique, This technique adds one move to the beginning of the proble m (5 move problem). Reclassification of Problems for Analysis There is a reclassification of the switch problems after the post training session for the purpose of simplifying the analyses Figure 27 displays a flow chart explanation of the problem reclassification. Control problems do not require reclassification since these problems are not affected by the training procedure. Switch problems are reclassified as either direct problems or indirect problems. Direct problems are those problems that match th e training received (e.g., all tall switch problems presented during the post training session would be classified as direct if participant received the tall training condition). The indirect problems are those problems that do not match the training recei ved but are still switch problems (e.g. all middle switch problems on the post training session would be reclassified as indirect if participants received the tall training condition). Problems will be reclassified based on training received in the second session, though pretraining problems will also be reclassified as control, indirect or direct.
47 Figure 27. Flow Chart of Problem Type Reclassification Procedure Participants came to the laboratory at the University of Flor ida for two sessions 14 days apart During the first session, participants were familiarized with the testing equipment and asked to sign the informed consent form. P articipants were then given instructions on the d igit symbol task Administration of this task took an average of five minutes Following this participants were escorted into a soundattenuated booth, seated in a comfortable chair and familiarized with the EEG equipment and computer interface (electrode placement and recording discussed below). For the pretraining session, following electrode application, participants were given instruction on the basic rules of the TOL. Practice problems for the TOL consisted of the participants solving six problems presented on a 15inch monitor and having the research assistant give appropriate feedback and correction on their responses Only control problems were included in these pretraining practice problems Once the session was completed, the participant was escorted from the room and asked about the ir general strategy on solving the TOL problems.
48 For the post training session, participants received a brief, 5minute presentation on the rules of the TOL Following this, each participant received either middle or tall switch strategy training. The set of problems used for post training TOL was identical to those administered in the pretraining session Previous research in our laboratory has suggested that identical problem presentation over a long delay (a day or more) does not increase problem solving efficiency compared to novel problems Once completed, the participant was escorted from the room and asked about what strategy they employed while solving the TOL problems. Electroencephalogram (EEG) Data Acquisition Electroencephalogram ( EEG ) was coll ected from a 32electrode NeuroScan Sintered electrode cap designed for the International 10/20 electrode arrangement and connected to a NeuroScan Synamp system The standard 10/20 arrangement was modified by replacing two scalp leads (O1 and O2) with two drop leads (leads not connected to the cap) used for the collection of heart rate and respiration. The EEG cap was also modified from the standard 1020 arrangement to have an additional lead, FCz, which is located between Fz and Cz All EEG electrodes wer e referred to the mastoid electrode placed behind the right ear Blinks and vertical eye movements were measured from leads placed above and below the right eye. H orizontal saccades and other horizontal eye movements were measured from leads placed just lateral to the outer canthus of each eye. EEG signals were amplified 150,000 times and collected with an A/D sampling rate of 1000 hz Online filter settings were placed at DC to 100 hz with a 60 Hz notch filter turned on to lessen the interference of noise from standard electrical wiring in the room.
49 EEG Data Processing and Quantification Data was offline filtered with a low pass of 30hz and transferred into Matlab for eye blink correction Eye blink correction was performed with the assistance of Independen t Components Analysis on the continuous EEG files collected from each 5 minute block (Delorme and Makeig, 2004) After blink removal, three separate epochs were extracted for each problem presentation. Each epoch contained 4098 milliseconds of EEG activity with a 200 mill second baseline The epoch length was chosen to allow for adequate sampling of the period of interest without being too long to allow excessive eye blink or motor artifact to interfere. Trials in which EEG amplitude in any of the leads exceeded 100mv were excluded from further analysis The three epochs chosen allow adequate sampling across the trial to focus on two critical phases of planning and strategy application and a rest period ( Figure 28; Rest, Problem Presentation and Problem Com pletion) The Rest epoch was sampled during the ITI between TOL problems while participants were sitting with eyes open viewing a red dot on the monitor The Problem Presentation epoch was drawn from the EEG sampled following the presentation of the TOL T his epoch was taken to capture the participants examination of the problem, and initial selection of a strategy This epoch began 800ms post problem presentation to ensure the frequency data collected did not include visual response to the stimuli The Pr oblem Completion epoch was sampled such that the epochs end was just prior to the participants response at the end of the trial This epoch was taken to capture the final stages of strategy application and TOL problem solving EEG power ( V2) data for all electrodes was calculated for the theta band (47hz), lower alpha band (810hz) and upper alpha band (1113hz) via Fast Fourier Transforms
50 using a cosine window (length: 10%). Fast fourier transforms were computed with Neuroscan, version 4.1 using the FFT batch program. Figure 28 Epoch selection for Rest, Problem Presentation and Problem Completion CHAPTER 3 RESULTS Behavioral and EEG data were analyzed using repeated measures mixed factor ANOVAs Post hoc follow up analy ses were conducted using estimated marginal means and Bonferroni corrected methods to ensure that the type I error was not violated1Behavioral Data GreenhouseGeisser tests are reported as there were significant deviations from sphericity on many of the variables, thoug h the Sphericity Assumed degrees of freedom are presented for convenience. A table of correlations involving the behavioral variables can be found in Appendix B. In order to evaluate performance preand post tra ining the variables of prop ortion of correct ly solved problems ( portion correct) the time used to correctly solve problems 1 Post hoc follow up tests were also completed with the Least Squares Difference procedure (essentially, no pvalue correction) and similar results were obtained.
51 (solution time), and the proportion of trials in which the participant did not make a response (proportion no response) were examined. A mixed factor t hreeway ANOVA was conducted for each dependent variable with age (older, younger), session (pretraining, post training) and problem type (direct, indirect, and control) entered. To better understand the effect of training across age group and problem type, a me asure of performance change was calculated. For each pr oblem type, a measure of the percentage of change between the preand post training scores was calculated [ ( Post score Pre score)/Pre score ] 100 This score allows a comparison of benefit from tra ining across the age groups without a confound of the differential performance between age groups found in the pretraining session2Proportion Correct A larger change score indicates larger improvement on the post training session in relation to the pre training session. F or each performance score, r esults and analysis with the performance change score are presented following the results and analysis with the raw sessionby session means Group means and standard deviations can be found in Table 31 for all behavioral measures. Means of the raw performance scores showed, as expected, better performance for the younger adults as compared to the older adults and this was confirmed in the analysis by a main effect of Age ( F (1, 54) = 28.379, p < .001). Additi onally, a main effect of Session ( F (1, 54) = 24.931, p < .001) was found and the means demonstrated that participants had better performance in the post training session as compared to the 2 A measure of performance gain in which the formula was (Post score Pre score) was calculated. Similar results were obtained as the presented variable.
52 pre training session (Figure 31). A main effect of Problem type ( F (2, 108) = 5.554, p = .006) was further examined in follow up analyses with means demonstrating that control problems had a lower score on proportion correct as compared to the indirect ( p = .032) and direct problems ( p = .026). Indirect and direct probl ems do not differ from each other. Table 31 Means and Standard Deviations for Behavioral Measures Pre Training Post Training Percent Change Older Younger Older Younger Older Younger Proportion Correct Control 00.56 (0.16) 00.80 (0.17) 00.61 (0.18) 00.85 (0.20) 18.0%(.48) 07.0%(.17) Indirect 00.59 (0.17) 00.82 (0.20) 00.69 (0.14) 00.89 (0.20) 24.2%(.37) 09.8%(.18) Direct 00.57 (0.20) 00.85 (0.19) 00.71 (0.20) 00.87 (0.20) 44.2%(.78) 3.0%(.18) Solution Time (sec) Control 24.75 (4.52) 17.60 (3.6 1) 24.62 (5.09) 13.20 (3.52) 0 .1%(.19) 24.3%(.12) Indirect 24.42 (5.35) 14.20 (3.89) 19.87 (5.23) 08.35 (2.08) 17.3%(.24) 39.1%(.16) Direct 24.30 (5.42) 14.50 (3.25) 18.39 (5.14) 07.83 (1.57) 21.1%(.16) 47.5%(.13) Proportion No Response Control 00.1 4 (0.10) 00.03 (0.05) 00.11 (0.13) 00.01 (0.02) --Indirect 00.12 (0.12) 00.01 (0.03) 00.04 (0.06) 00.00 (0.00) --Direct 00.14 (0.14) 00.01 (0.03) 00.03 (0.07) 00.00 (0.00) --Note: Scores are presented with standard deviations in parentheses For the percent difference score, a higher number indicates more improvement Solution time scores were revered in sign so that higher scores indicate better performance in the post training session Figure 3 1 Mean proporti on correct measure for older and younger adults The change score of proportion correct yielded a main effect of Age ( F (1, 54) = 8.929, p = .004) with the means of the older adults showing larger benefits as compared
53 to younger adults on all problem types (Figure 3 2) Although, this measure should be interpreted with caution, as pretraining means on the measure of proportion correct for the younger adults suggest s a ceiling effect with little room for improvement on the post training session3 ( Figure 3 1 ) In light of this possible ceiling effect, older adults were tested separately and although t he means suggest differential improvement across problem type this is not supported statistically Figure 3 2 Change score measure for proportion correct for older and younger adults Solution T ime Raw solution time results indicated younger adults were, as expected, faster at solving the TOL problems as compared to the older adults, confirmed by a main effect of Age ( F (1, 54) = 143. 829, p > .001) This was qualified by a threeway interaction of Session x Problem type x Age ( F (2, 108) = 3.579, p = .038) Follow up analyses indicated that though the younger adult group significantly improved following training on all three problem types in the post training session ( p s < .001), the older adult group only improved on the in d irect and direct problems in the post training session ( p s < 3 The ceiling effect was not statistically supported with proportion correct values being significantly different from 1 for both age groups at all problem types ( p s < .003).
54 .001) This demonstrates the importance of the training for the older adult group as they showed improvement in only the trained problems and not the control probl ems ( p = .983) As suggested in Figure 3 3, younger adults were faster at solving the indirect and direct problems in the pretraining session as compared to the control problems and this was sta tistically supported in follow up analyses with younger adults only when comparing solution time means for each problem type. While the direct and indirect problems did not differ from each other ( p = .999 ) both means showed better performance than the control problems ( p s < .001) on the pretraining session. It is possible that younger adults had discovered the strategy needed to solve switch problems and were therefore, able to apply the strategy to the indirect and direct problems before formal training occurred. Figure 3 3 Mean solution time measures for older and younger adults The change score for solution time yielded a main effect of Age ( F (1, 54) = 50.583, p < .001) with younger adults demonstrating more improvemen t as compared to the older adults ( Figure 3 4) A main effect of Problem type ( F (2, 108) = 41.867, p < .001) was further investigated in post hoc analyses showing that all three problem types
55 were significantly different from each other with direct probl ems showing the most improvement ( p s < .011) Thus, both age groups showed a clear improvement on the direct problems and were able to successfully transfer the strategy to the indirect problems in the post training session. While the interaction of Probl em type x Age was not significant, the graph of their means suggests an interesting difference between age groups The younger adult group improves in their performance on all three problem types, with up to 50% faster solution times for the direct problem s. Older adults, on the other hand, only improve on the indirect and direct problems Both these findings are consistent with raw score results. This again highlights the importance of the older adults receiving the strategy training While the younger adults are benefitting more from the strategy training than control training improvements occur for all problems types the lack of improvement on the nontrained problems indicates that older adults only improved from the training and not from general practi ce effects. The fact that older adults only improved on the trained problems demonstrates the importance of the training since improvement from general practice effects was not witnessed for the older adults group. It was, however, shown for the younger adult group. Figure 3 4 Change scores for solution time for older and younger adults
56 Proportion of No Response The proportion of no responses across age groups and problem types was quite low4 (Table 31 ) but results were abl e to further illuminate the importance of training for the older adult group. A three way interaction of Session x Problem type x Age ( F (2, 108) = 4.608, p = .0 20 ) was further examined in post hoc analyses The older adults demonstrated a significant decrease in the number of no responses in the post training session for only the indirect and direct problems ( p s < .001) with no drop in the number of no responses for the control problems ( Figure 3 5; p = .153) Th is pattern is showing a transfer of benefit to the indirect problems in the post training session. Younger adults did not demonstrate significant differences between problem type or session in this measure, though the pre training scores suggest little room for improvement on this measure. Figure 3 5 Mean proportion of no responses for older and younger adults Electroencephalogram ( EEG ) Data Analysis The frequency bands of theta, lower alpha, and upper alpha were investigated across age, session, lead, and epoch time to understand the effects of training and age .4 A one way t test against 0 demonstrated a lack of floor effect for either age group on any problem type ( p s < .009)
57 on brain activity A repeated measures ANOVA was conducted for each frequency band with age (older, younger ) session (pretraining, post training) and problem type (direct indirect, and control ). For t he theta f requency band the m idline lead set of Fz (frontal) and Pz (parietal) and the frontal lateral lead set of FC3 (left frontal) and FC4 (right frontal) were separately assessed The lateral set is used to examine laterality effects in the theta band Based on research reviewed in the introduction, the lower alpha frequency band was assessed with the midline lead set of Fz (frontal) and Pz (parietal) For the upper alpha band, the lead of Pz (parietal) was tested Lead sets analyzed at each frequency band were c hosen by adopting methodology previously used in comparison literature ( Cummins et al., 2007; McEvoy et al., 2001; Gevins et al., 1997) The lower alpha band is investigated at the frontal and parietal sites to understand its functional significance to the task and the topographic changes expected during TOL problem solving ( Sauseng, Klimesch, Schabus, et al., 2005) ANOVA t ests were completed separately for each epoch time ( r est, p roblem p resentation, p roblem completion) and are presented below for each fr equency band and lead set The organization of this section is as follows: The theta band will be presented at the frontal and parietal midline sites for each epoch time followed by the frontal lateral set for each epoch time The alpha band will follow with comparisons between the frontal and parietal regions for each epoch time for the lower alpha band and investig ations of the parietal region for the upper alpha frequency band. For convenience, an outline of the main findings from each of these ANOVAs is presented in Appendix C At the beginning of each frequency band section, topographic maps for that frequency are presented to provide a general overview of activity in that band.
58 The Theta Frequency Band Topographic maps of the theta frequency band for both younger and older adults are presented in Figure 3 6 For the older adult participants, there does not seem to be a large difference in theta power between the three epoch types There does, however, seem to be an increase in theta power in the post training as compared to the pretraining session especially for older adults For the younger adults, there is also an increase in theta power from the rest to the problem presentation and problem completion epoch. Note the difference in scales between g roups employed to highlight the differences for each group separately. The older adults show significantly less theta power as compared to the younger adults. Figure 3 6. Topographic plots of Theta for Older and Younger Adults Note the different axis values used to maximize the visual depiction of the topographical changes across the head for each age group individually.
59 Frontal and Parietal Activity in the Theta Frequency Band In the separately conducted ANOVAs for each epoch time there was a main effect of Lead ( Fs (1, 54) > 12.629 ps < .001) and of Age ( Fs (1, 54) > 36.227, ps < .001) Follow up analyses conducted show that in all epochs the frontal lead has more theta power as compared to the parietal lead ( p s < .001) and younger adults had more theta power as compared to the older adults ( Figure 3 7 ) Since this main effect of age was found in all three epoch times, the younger adults show more theta power as compared to the older adults Additionally, all three epochs had a Session x Lead interaction ( Fs (1, 54) > 5.517, ps < .023) This interaction was f ollow ed up with analyses conducted separately for each lead type, and these show ed that only the frontal regions significantly increased in theta power between the preand post training sessi on ( p s < .014) T he parietal region did not s how a change in theta power between the preand post training session. As demonstrated in Figure 3 7 this is found in both age groups. The fact that this was found in both age groups shows a lack of developm ental change across the lifespan. Figure 3 7 Theta power in the frontal and parietal leads during all three epoch times
60 During the p roblem p resentation and p roblem completion epochs, there was also a Age x Lead interaction ( F (1, 54) = 21.52, p < .001) This interaction was followed up separately for each age group which show ed that the younger adults had a greater difference in theta power between t he frontal and parietal regions ( p s < .001) as compared to the older adults ( p s < .018; Figure 3 7 ) Y ounger adults are demonstrating larger differentiation between frontal and parietal regions, with greater theta power at the frontal leads during both epochs taken during the TOL presentation and problem solving As expected, o ld er adults are greatly lacking in frontal midline theta as compared to the younger adults While younger adults have larger theta power at both the frontal and parietal leads, the frontal difference between age groups is larger than the parietal difference in both the preand post training sessions. During only the Problem Completion epoch, a three way interaction of Session x Prob lem t ype x Lead ( F ( 2 108 ) = 5.94 p =.004 ) demonstrated training differences between problem types in the theta band ( Figure 3 8 ) Post hoc analyses conducted separately for each problem type indicate that both the control and indirect problems increase in t heta power between the preand post training conditions but only in the frontal regions ( ps < .038) with direct problems approaching significance (p= 10) There were no changes in theta power in the parietal regions across problem types or session Figure 3 8 demonstrates that this increase is apparent in both the older and younger adults though the means clearly indicate a l arger effect in the older adults Inspection of raw mean increases in the post training session shows that older adults have a larger increase in each problem type as compared to younger adults though follow ups with each age group separately did not supp orted this claim
61 In review, the theta band at the frontal and parietal leads showed clear age differences with the older adults having significantly less theta power as compared to the younger adults, specifically in the frontal regions A differential in crease in the frontal regions in the post training session was found for both younger and older adults The effect of problem type, while small, showed larger increases in theta power in the indirect and control problems as compared to the direct problems during the problem completion epoch. Figure 3 8 Theta power at the frontal and parietal sites during the problem completion epoch Lateral ized Frontal Activity in the Theta Frequency Band In the separately conducted ANOVAs for each epoch time, there was a main effect of lead left versus right ( F s (1, 54) > 55.03, p s < .001) at all epochs F ollow up analyses show ed that the right frontal region shows larger theta power as compared to the left frontal region ( p < .001) at al l three epochs Similar to the midline theta analysis all three epochs showed a main effect of Age with the younger adult group showing larger theta power as compared to the older adult group in both lateral frontal leads ( F s (1, 54) > 16.40 p s < .001) A main effect of S ession was only found in the problem
62 presentation and problem completion epochs with larger theta power in the post training as compared to the pretraining session ( F s (1, 54) > 4.96, p s < .030) During only the Problem Presentation epoch, a threeway interaction of Session x Lead, and x Age ( F (1, 54) = 4.34, p =.042 ) was followed up separately for each age group. T he older adult group showed an increase in theta power in the left frontal lead in the post training session as compared to the pretraining session ( p = .016) with no increase found in the right frontal lead ( p = .401) The younger adults do not show any significant increase across session in theta power in either the right or left frontal leads ( Figure 3 9 ) Thus, during the initial TOL planning, older adults were showing an increase in activity in the post training session as compared to the pretraining session for the left frontal lead while the younger adults were not displaying a similar increase. O lder adults sh ow less overall activity in the left frontal hemi sphere as compared to the right and there is an increase in this region in the post training session. Thus, as with the frontal region analysis, older adults are showing more increases in the post training session as compared to the younger adults. This is an interesting finding as it would be expected that the younger adults show larger increases since their performance was better as compared to the older adults. Figure 3 9 La teral frontal theta power in the problem presentation epoch
63 During the Problem Completion epoch, a twoway interaction of Session x Problem type ( F (2, 108) = 3.45, p =.037) was further investigated by an analyses of each problem type. As shown in Figure 3 10, only the control problems show ed an increase in theta power in the lateral electrodes in the post training session ( p = .032) Though the indirect and direct problems are not significantly increasing in the post training condition, the means suggest a similar pattern as found in the control problems This pattern is similar to the increases seen in the midline lead analysis. Given that the importance of age is to be highlighted here, an exploratory follow up was conducted which showed that the increase in theta power in the control problems was found only for the older adult group ( p = .025) The younger adult group did not show differences between the preand post training condition in any problem type. The lateral frontal theta analysis showed similar changes across age, with the younger adults showing larger theta values as compared to the older adults at all three epoch times Also similar to the midline lead analysis, the older adults showed larger increases in the post training session as compared to the younger adults (Figure 3 7 and Figure 39) The control problems showed the largest increases in theta power, though as was found in the midline lead analysis, the figure suggests increases in all three problem types is evident (Figure 38 and Figur e 310). Figure 3 10. Lateral frontal theta power in the problem completion epoch
64 Frontal and Parietal Activity in the Lower Alpha Frequency Band Topographic maps of the lower alpha frequency band for both younger and older adults are presented in Figure 31 1 A clear decrease in power between the rest and TOL epoch times are present in both younger and older adults though possibly stronger for younger adults Figure 31 1 Topographic plots of Lower Alpha for Older and Young er Adults In the separately conducted ANOVAs for each epoch time, there was no significant main effect of Age While t he theta frequency band showed large age differences this alpha band did not follow a similar pattern. A main effect of S ession was fo und in rest and in the problem completion epoch, with an increase in lower alpha in the post training session as compared to the pretraining session ( F s (1, 54) > 4.470, p s < .0 39 ) Acr oss the three epochs, there i s a change in topography of lower alpha In the rest epoch, there was a L ead x Ag e interaction ( F (1, 54) = 4.552, p = .037) Analyses
65 separately conducted for each age group reveal that younger but not older adults have larger alpha power in the parietal region as compared to the frontal regi ons ( p = .044) Older adults had a non significant difference in the opposite direction, as shown in the left panel of Figure 3 1 2 During the problem presentation and problem completion epochs, t he frontal regions have larger lower alpha power as compared to the parietal regions ( p s < .003) Both the lac k of a Lead x Age interaction and Figure 3 1 2 show that this is true for the younger and older adults Unlike the theta frequency band, there was not a significant different in age, though Figure 312 suggests a pattern in the direction of younger adults showing more lower alpha as compared to the older adults The change in topography noted across epoch times is strongest for the younger adults as the older adults show a similar pattern of lower alpha across all three epoch times Figure 3 1 2 Lower alpha power in the frontal and parietal regions during all three epoch times Parietal Activity in the Upper Alpha Frequency Band Topographic maps of the upper alpha frequency band for both younger and older adults are presented in Figure 313. A clear decrease in power between the rest and TOL epoch times are present in both younger and older adults and this appeared stronger for older than younger adults
66 Figure 313. Topograph ic plots of Upper Alpha for Older and Younger Adults Analyses conducted with the upper alpha frequency band did not have any significant interactions or main effects This demonstrates an important contrast to the other frequency bands While both the l ower alpha and theta band were modified by the independent variables, the upper alpha band was not Thus, the changes seen in the lower alpha and theta frequency bands are not due to general changes in all frequency bands It is also not surprising that th e upper alpha frequency band did not produce changes across age or task as this band is thought to be associated with visual stimulation ( Gevins et al., 1997 ) something that was consistent across both sessions and all problem types in the present study Individual Factors Contributing to Successful Strategy Application The effect of processing speed on cognitive variables (Salthouse, 2005; Riccio, Wolfe, Romine, Davis and Sullivan, 2003) and its predictive ability on successful strategy application acros s the lifespan has been previously explored (Verhaeghen et al., 1996) Individual patterns of cognitive style have been related to performance
67 differences as well as differences in EEG patterns ( Gevins et al., 1997) Employing similar methods to Gevins and Smith (2000), participants were split into high processing speed (HPS) and low processing speed (LPS) groups based on their performance on the digit symbol task Within each age group, a median split determined the placement of participants into either th e HPS or LPS group. Table 32 presents pertinent information for these subgroups of participants This exploratory analysis was conducted to further understand the importance of processing speed in TOL performance, strategy application and EEG patterns acr oss the lifespan. Table 32 Demographic Data for Low and High Processing Groups Age N Digit Symbol Score Older Younger Older Younger Older Younger Low Processing Speed 71.87(5.31) 20.30 (1.65) 12 1 6 41.58 (6.65) 64.13 (5.03) High Processing Speed 68.64 ( 3.65) 20.11 (1.48) 12 16 59.42 (6.87) 76.75 (3.64) Two preliminary ANOVAs were conducted to verify the median split procedure. The first univariate ANOVA was run separately for each age group and confirmed significantly different digit symbol mean score s between the high and low processing speed groups for each age group ( Figure 3 1 4 ; F s (1, 52) > 178.88, p s < .001). The second univariate ANOVA was also run separately for each age group and confirmed the lack of age difference between the HPS and LPS groups ( p s > .097) Thus, differences found between these groups cannot be attributed to group differences in age. This is important to demonstrate that if there are differences between these groups it is not due to cognitive development which occurs within each of the age groups. For the younger adults this would be cognitive improvements and for the older adults this would be cognitive decline.
68 Figure 3 1 4 Digit symbol means for LPS and HPS groups for younger and older adults Primary analyses were conducted t o evaluate differences in HPS and LPS across behavior and EEG variables A repeated measures mixed factor ANOVA was performed with the variables of processing speed (high, low), age (older, younger), session (pretraining, post training) and problem type (direct, indirect, and control) To confine these exploratory analyses, depe ndent variable choices were based on findings from the ANOVA analyses completed in the previous section. The behavioral measures explored are soluti on time both within each session and the change score Proportion correct was not included due to a ceiling effect for younger adults in the pretraining session. The measure of proportion of no responses was omitted due to floor effects in the preand post training session for the younger adult group. EEG measures include the theta frequency band at midline frontal/parietal sites and lateral frontal sites at each epoch t ime Lower alpha frequency band was explored but did not produce significant findings related to processing speed. The high alpha band was not included in these analyses due to insignificant results as reported above. In the following section, o nly significant findings involving the difference between processing speed groups will be highlig hted. All of the significant effects found above were replicated in these analyses
69 Behavioral Analyses B etween the High and L ow Processing Speed Groups The measure of solution time across session did not show any significant effects involving the processi ng speed group variable The measure of the change score for solution time showed a threeway interaction of Age x Problem Type x Processing speed group, F (2,108) = 3.371, p = .046, which was further investigated in follow up analyses conducted separately for each age group. As shown in Figure 3 1 5 the younger adults do not differ in performance changes between the LPS and HPS groups on any of the problem types For the older adult group, the LPS group has more improvement on the indirect problems as comp ared to the HPS group ( p = .007) There were no significant differences between these older adult groups on the other problem types While this is a surprising finding, it may be interpreted as the LPS group benefiting from the strategy training more because this allows them to maximize their resources and gain better cognitive control This is especially important for older adults and those with low processing spee d ability Further elaboration on this point will be highlighted in the discussion. It is an interesting difference which may only be attributed to a small number of subjects involved in each group. Figure 3 1 5 The change score differences in solution time between LPS and HPS groups
70 Frontal and P arietal A ctivity in t he T heta F requency B and Between High and Low Processing Groups In the separately conducted ANOVAs for each epoch time there was an interaction of Lead x Age x Processing speed group ( Fs (1, 52) > 5.934, ps < .018) Post hoc follow up analyses conducted f or each epoch time were done separately for each age group, and mean results are shown in Figure 316. As was found with the behavioral data, the younger adult group shows no significant differences between the LPS and HPS groups for any of the epochs Int erestingly, in older adults only the HPS group shows a significant difference between the frontal and parietal leads ( p < .001) ; there is no significant difference between these leads for the LPS older adults Thus, both younger adult groups as well as the HPS older adults group showed differential activity between the frontal and parietal regions during all three epoch times During the problem presentation and problem completion epochs, the HPS older adults show larger frontal theta power as compared to t he LPS older adults ( p s < .0 5 ) During the rest epoch there was a difference in the same direction, but it was not significant It is possible then that this difference is inherent and found to be more reliable or exaggerated during difficult cognitive st imulation. Figure 3 1 6 Frontal and parietal theta power between LPS and HPS groups at all three epochs
71 Lateral Frontal Activity in the Theta Frequency Band Between High and Low Processing Groups In the separately conducted AN OVAs there was an interaction of Lead x Processing speed group in each epoch time ( Fs (1, 52) > 7.210, ps < .010) Post h oc analyses conducted separately for each processing speed group show ed that in all three epoch times the HPS group showed larger t heta power in the right as compared to the left lead ( ps < .001) The LPS group demonstrated a similar pattern with less differentiation between hemispheres While Age was not a part of the interaction, exploratory post hoc analyses conducted separately f or each age group confirmed that t his pattern between HPS and LPS groups is reflected in both age groups ( Figure 3 17). Figure 3 1 7 Lateral frontal theta activity between LPS and HPS groups at all three epochs
72 CHAPTER 4 DIS CUSSION The strategy training designed to reduce working memory load in younger and older adults engaged in a difficult planning task was successful in improving performance on both directly and indirectly related problems. Behavioral measures indicate that the strategy training produced larger gains in the younger as compared to older adults. As expected, the older adults showed decreased FM theta as compared to the younger adults, a difference that was somewhat different in the post training session as the older adults showed larger increases in FM theta in the post training condition as compared to younger adults. The alpha frequency band as measured from the parietal cortex was not affected by task, training, or development. The lower alpha band showed topographic shifts from the parietal to frontal cortices during epoch times in younger, but not older, adults. Individual differences in processing speed in older adults only was found to affect both performance and EEG signatures during task and rest con ditions, suggesting this affects both state and trait functions of the EEG signatures. The discussion that follows will focus on several key issues which were raised by this research to gain further understanding of changes across development in strategy application and neuronal activity. A discussion of the changes across the lifespan in strategy benefits, EEG signature changes and the impact of processing speed will be highlighted. Differences in the Benefits of Strategy Training Across the Lifespan The beneficial impact of memory training, even into very old age, has been clearly demonstrated in a multitude of paradigms (Becic et al., 2008; Bottiroli et al., 2007;
73 Carretii et al., 2007; Singer et al., 2003; Cavallini et al., 2003; Rapp et al., 2002; McN amara et al., 2001). As expected, in the current study both age groups were able to apply the trained strategy and demonstrated substantial improvement in both behavioral measures. Many previously conducted memory training paradigms have employed tasks whi ch have not taxed the individual cognitively. The current study employed a task which required participants not only to spatially rearrange the balls on a peg, but to consistently update their memory for the most current locations of those balls without ac tually seeing the change made. Du ring the pretraining condition, older adults exhibited significantly poorer performance as compared to the younger adults while the increase in older adults was substantial, they did not r each the levels of the younger adults in the post training condition. In fact, the older adult participants, even after training, were not able to reach the performance levels of younger adults prior to training This lack of improvement to the level of a younger adult does not however, r educe the importance of the substantial improvements made by the older adults in the post training session. Additionally, this training paradigm was quite brief and the older adults demonstrated large improvements The evidence of transferring the knowledge gained by training in one task to the benefit of another task has been difficult to demonstrate in previous research (Karbach & Kray, 2009). Based on behavioral results of the current study, both younger and older adults were able to transfer the trained strategy on one type of problem to another; to transfer this strategy from the directly trained problem to the indirectly related one. There are several reasons why transfer may have been evident here and lacking in other studies. First of all, the indir ectly related task had many similar qualities to the
74 directly trained task; it was not transferred to a new task, but to a related problem in the same task. The process to solve and the goal of the indirectly related task w ere very similar to the directly trained task. Nonetheless, the adaptation of the trained strategy to the indirectly trained problems was not a trivial one. A participant had to recognize a switch problem on a different peg than the one in training, and to rearrange the order of several m oves and peg positions in which balls were placed in order to adapt the trained st rategy to this new peg Other studies involving similar indirect tasks have also succeeded in showing transfer effects (Li et al., 2008). Importantly, the transfer in the pre sent study occurred after just a single short training session. The Li et al (2008) study included 45 days of practice prior to testing for transfer T he process of training and the stepby step method employed may have helped participants understand the training and apply it to a similar problem on a deeper level as compared to more holistic methods (Becic et al. 2008) or compared to extended training (Li et al 2008) Of particular importance for a training study is demonstrating the training for the intended population. Previous research investigating developmental differences found in strategy training have consistently shown larger effects for younger, rather than older, adults (Verhaeghen et al., 1992) The differential application of the trained s trategy to two out of the three problems aids in demonstrating the importance of memory training for the older adult population. While the younger adults in the present study did demonstrate larger increases in solution time between the preand post train ing condition, this increase was found for all problem types. The older adults, on the other hand, only showed improvements on the switch problems, those which are able to be solved using or adapting the trained strategy. This lack of increase in performance
75 demonstrated on control problems by older adults suggests a failure to benefit either from general experience with the task or from transferring some general benefits of the training to nonswitch problems Further, it very clearly demonstrates the importance of the strategy training for the older adults at least for short term training Unlike older adults, younger adults were able to successfully apply the strategy and in addition improve on the control problems. This developmental difference between age groups could be due to several factors. First, younger adults may have self generated a strategy which allowed for faster problem solving. Previous research has demonstrated that younger adults naturally choose optimal strategies (Becic, Boot, & Kramer 2008) and are more likely to generate self developed optimal strategies as compared to older adults. Secondly, older adults may have shown poor performance on the control problems because of a misapplication of the trained strategy. It is possible that the older adults were attempting to apply the strategy to the control problems resulting in a longer solution time on these problems alone. This is an unlikely outcome, however, as the older adults also did not show increases in the proportion of correctly solved control problems which is unrelated to longer problem solving Thus, even if the older adults did initially attempt to apply the strategy to the control problems, this would only result in a longer solution time rather than an incorrect response. Ad ditionally, the strategy training included identification of the directly trained problems so as to avoid a misapplication to the control problems. It is possible that the older adults found the task very difficult and performance on the pretraining sessi on reached an asymptote of performance. Training on the switch problems promoted increased performance on only
76 the trained problems Thus, strategy training on this paradigm can be considered to be more beneficial and more critical for the older than the y ounger adults. Changes in the Theta Frequency Across the Lifespan and the Impact of Strategy Training Theta as a marker of cognitive effort has been well established in previous research on working memory tasks (Ragavachari et al., 2005; Onton et al., 2005 ; Mizuhara et al., 2004; Sauseng et al., 2004; Grunwald et al., 1999). In the current study the older adult group had significantly less theta power as compared to the younger adults, both during the rest and TOL problem solving epochs. This suggests the older adults were less effective or less focused in their cognitive effort during the entire task situation Previous research conducted with older adults during cognitive tasks have also shown decreased theta power, particularly in the frontal midline reg ion (Cummins et al., 2007; McEvoy et al., 2001). Since this FM theta has been found to be primarily generated in the frontal lobes it suggests this region may be beginning to show less effective or aberrant functioning in the older adult group, a view cons istent with a variety of other evidence (Onton et al., 2005; GunningDixon et al., 2003; Davis et al., 2009) There may also be another view relevant to these age differences It is possible that this decrease in FM theta in the older adults may additionally be attributed to a shift in reliance on somewhat different brain structures If this were the case, however, the older adult group would have shown increases in other regions during problem solving (e.g. parietal), a finding that was not supported here However, the current analysis was necessarily limited in spatial scope, and future work should include larger topographical observations across the brain to investigate this possibility (though imaging research looking at the entire brain has not support ed this suggestion; Nyberg, et al., 2003)
77 A unique aspect of the current study was the ability to investigate the changes in theta after strategy training, something which has not been previously investigated. In the post training session, both younger and older adults demonstrated increases in theta power in the frontal but not parietal regions. Since the trained strategy employed here was primarily a frontally driven, topdown processing strategy, the differential increase in the frontal regions fits well in having increases in task relevant regions in the post training condition (Westerberg et al., 2007; Olesen et al., 2004). The older adult group was found to show larger increases in the post training session as compared to the younger adults. While th e older adults did not reach the theta power levels of the younger adults, following training there were substantial and significant increases in that power, particularly in the frontal regions. In only a short, 20minute training session older adults show ed increases in theta frequency in the frontal regions. Altered frontal functioning of the theta band, as evidenced by a decrease in power, has been associated with task performance decrements (Cummins et al., 2007). The theta frequency band, over and abov e other frequency bands investigated, was found to be abnormally low in patients with mild cognitive impairment (MCI; Ueda, Musha, and Yagi, 2009). Training studies with MCI and Alzheimers disease patients is scarce, as outlined in a detailed metaanalysi s of behavioral interventions by van Paasschen, Clare, Woods, and Linden (2009). A preliminary study of name/face training in one MCI patient produced promising results with increases in fMRI BOLD signal found in task relevant areas (inferior frontal regions; van Paasschen et al., 2009).
78 Differences in laterality have been identified in previous research investigating TOL problem solving with increased right activation in the frontal regions being previously reported (Lazeron et al., 2000). Specifically, t hose with better planning ability have been found to have parametrically increased right frontal activity as compared to the left suggesting a differential role of the hemispheres in problem solving (Unterrainer et al., 2004). In the present study both younger and older adults were found to have greater right hemisphere theta power as compared to the left hemisphere in the preand post training conditions but the increase in theta following training older adults was in the left frontal regions. This coul d be interpreted as older adults engaging bilateral frontal regions to assist with the newly learned strategy. Older adults have been found to show bilateral activation in tasks which younger adults only show lateral activity, known as the HAROLD model (Hemispheric Asymmetry Reduction; Cabeza, 2002). This increased activation has been interpreted as compensatory as those older adults who engage bilateral regions also demonstrated better performance on the task (Cabeza, 2002). This fits well with the behavioral data which shows older adults are greatly benefiting from the strategy training. Few differences in power or topography of theta activity were evident between problems types after training. One possible suggestion is that the increase in theta power in the post training session was due to continued task exposure rather than an increase specific to the strategy training. Such an effect was demonstrated in previous research that had extended practice on a working memory task and found increased theta without a specific strategy being taught (Gevins et al., 1997) A second possible suggestion is that the post training session problem solving was attenuated by
79 participants checking each problem, regardless of type, to see if the strategy could be applied. Nonetheless, the differential increase found in the frontal but not parietal regions would suggest that the increase is not merely a general process resulting from the continued exposure to the task because completion of the task requires both frontal and parietal involvement. The application of the strategy, however, can be argued to be a more frontally driven cognitive process The similar increase in the rest epoch does not consistent with this theoretical claim, though it could be suggested that participa nts are reviewing the strategy during the rest epoch in preparation of the upcoming trials Regardless, the behavioral data, if not the theta findings, for both the younger and older participants clearly demonstrates differential improvement on the problem types showing that different cognitive processes were underlying problem solving of the different problem types. While the participant may have been differentially applying the strategy to certain problems this was not able to be differentiated in the theta power. Thus, while the theta band was unable to differentiate strategy application, the behavioral data demonstrates the application of the strategy to the different problem types. One additional issue that arises in the reported findings is the lack of the anticipated increase in theta from the rest to TOL problem solving epochs in the older adult group. Increases in theta power from the rest to task conditions has been previously reported, suggesting a difference in the presented task and those previously conducted (Cummins et al., 2007; McEvoy et al., 2001). The task given in this study was considerably more difficult than the tasks previously studied (nback tasks of reasonably easy difficulty) which may have impacted the way in which the older adults
80 approached the task Additionally, the current task has participants solving a problem over a long period of time where other tasks are more quickly completed. The drawn out problem solving time period may have washed out effects or smeared effects across the solving period. The task employed in this study was clearly found to be difficult for older adults, as indicated by their longer solution times and low proportion correct scores on the pretraining session as compared to the post training session. In previous studies, older adults have evidenced a lack of increase in frontal theta power on more difficult levels while younger adults show increases in theta power with more difficult cognitive levels (Cummins et al., 2007). Similarly, older adults in a di fficult reading task were found to have of less of an increase in the most difficult condition as compared to easier conditions (Persson et al., 2007) These failures to show any increase in task relevant regions for more difficult problems has been interpreted as a breakdown in cognitive control as the task becomes too difficult Interestingly, older adults were found to show increases in theta power in the post training condition. This may be due to better cognitive contr ol with the aid of the strategy. T o take this into account, future research with older adults should administer a variety of difficulty levels to understand the lack of increase during the most cognitively challenging condition. Cognitive training on the most difficult conditions may also help to shed light on the impact of the training on the brain activity. Changes in the Alpha Frequency Across the Lifespan and the Impact of Strategy Training Past research on the alpha frequency band has reported that different aspects of processing is b ased on the lower or upper limits. The upper alpha frequency band has generally been depicted as being involved in visual processing and found to be primarily
81 in the occipital/parietal regions (Gevins et al., 2000). As expected, the upper alpha frequency band in the current study did not show agerelated or task related changes during rest or TOL epochs. Previous research has also found that the upper alpha frequency band did not show agerelated changes across the lifespan (McEvoy et al., 2001). This suggests a similar functional identity for the upper alpha frequency band across the lifespan. The lower alpha frequency band has been found to increase during cognitive tasks and have a functional role during working memory tasks (Gevins et al., 2000) Present research showed an interesting change in topography across rest and TOL epochs which suggests a functional role of the lower alpha frequency band for the younger adult group. In the rest condition, larger lower alpha values were found at the parietal as compared to the frontal region. In contrast, during the problem presentation epochs, larger alpha power was found at the frontal as compared to the parietal regions. This topographic shift of the lower alpha band to functionally relevant regions during TOL problem solving suggests functional indexing of the problem solving process. Sauseng and colleagues note a similar shift of the lower alpha frequency band to functionally relevant regions during the task as compared to the rest condition ( Sauseng, Klimesc h, Schabus, et al., 2005). In their task, younger adults were found to have increases in frontal alpha power during a difficult working memory task. This work suggests a more functional role of alpha, at least during a difficult planning task Further, thi s finding is not consistent with the claim that alpha is representative of brain idling or inhibiting nontask relevant regions (Jensen et al., 2002) Lower alpha can be seen as indicative of functional involvement of portions of the brain in task relev ant processes. The
82 differences in the rest epoch between the younger and older adults could be attributed to different preparation styles between these groups It is possible that older adults were more actively anticipating the next TOL problem which resulted in larger frontal alpha involvement during the rest epoch. An Integrative Look at the Theta and Alpha Frequency Bands and Its Impact Across the Lifespan The results of the present study and previous sections of discussion focusing separately on the theta and alpha bands have demonstrated both similarities and differences across these frequency bands As with the theta frequency band, the lower alpha frequency band had increases in the post training condition suggesting a functional role of the underlyi ng regions as increases in activity that have been previously reported were interpreted as compensatory activity (e.g. Westerberg et al., 200 7 ) Unlike the theta frequency band, alpha power values were found to be similar across age groups in both the frontal and parietal regions. An integrated evaluation of these will help to clarify the different functional roles of these bands during working memory tasks. An interesting theoretical model proposed by Klimesch and colleagues suggests complementary roles of the theta and alpha frequency bands, with each being associated with different aspects of memory and cognitive processes (Klimesch et al., 2009) The model suggests that the alpha frequency band is associated with the processing of a knowledge system a construct that describes general processing of meaningful stimuli The theta frequency band is argued to reflect processing associated with working memory a construct used to describe controlled access to new information. Under this model, these two m emory systems are integral to memory
83 processes and are integrated through a larger network which can be studied by investigating the alpha and theta frequency bands The importance of this theoretical model rests on the integrative nature of the different frequency bands working in unison across the brain. One way to measure this interaction among brain regions or the combined effort of several regions across the brain is looking at the coherence values in different frequency bands For example, during tas k completion the theta frequency band has been shown to increase in coherence between the frontal and parietal regions (Sauseng, Klimesch, Schabus, et al., 2005) Alpha, in the same light, demonstrates increases in coherence across widespread regions Base d on this model, the importance of a single region is diminished as a more integrative role across the brain is identified for task completion Though the current study did not employ methods which investigate coherence across regions of interest, the pow er values obtained at the frontal and parietal regions are consistent with this theoretical model The theta frequency band was found to increase in the post training session which involved increased use of working memory processes The application of this model would suggest that power values measured at one region may reflect processing happening at that region or processing happening among several regions As demonstrated in fMRI work, the completion of a task heavily relies on many regions for completion (e.g., Lazeron et al., 200 0 ) The decrease in FM theta observed across the lifespan and in previous work (Cummins et al., 2007; McEvoy et al., 2001) may be indicative of frontal breakdown or possibly due to a widespread deterioration of long range commu nication across the
84 brain. A compelling analysis of this idea was suggested by Greenwood (2000) in which work was presented to show all regions of the brain are demonstrating deterioration as humans age and that not only the frontal regions are affected by time This idea is presented to demonstrate the importance of further work in coherence analysis across the entire brain. Diffusion tensor imaging work has demonstrated a clear decrease in structural connectivity in older adults, supporting the suggestion of widespread neural breakdown (Venkatraman, et al., 2009) While work has shown this widespread breakdown, it is currently unknown what the impact of training is on this neuronal network A coherence analysis of the theta and alpha frequency bands in a pre and post training design would provide valuable information to help understand these localized increases and how this changes across the lifespan. Impact of Individual Differences in Processing Speed on Behavior and EEG The task employed to measure processing speed, the digit symbol task, has been commonly used to measure this construct in previous research (e.g., Venkatraman, et al., 2009; Ball, Edwards and Ross, 2007; Joy, Kaplan and Fein, 2004; Verhaeghen and Marcoen, 1996). Individual performance on the digit symbol task has been found to be a predictor in strategy application (Kliegl, Smith, and Baltes, 1990). It is thought that the ability to quickly process information will have an impact on the ability to acquire a new strategy. Kliegl et al ( 1990) trained younger and older adults on a serial recall task demonstrated differential ability to apply the strategy with the digit symbol score becoming more important as training progressed. The digit symbol score did not correlate with pretraining performance but only emerged as a significant predictor as training continued. This was interpreted as the importance of processing speed becoming more influential as the application of the strategy was attempted.
85 In the current study, the younger adults di d not demonstrate behavioral differences between the LPS and HPS groups, whereas older adults did do so suggesting the role of processing speed or the application of the strategy to the TOL was not the same in the younger and older adult groups. It is poss ible that the younger adults did not find this task overly challenging and thus regardless of initial ability, application of the strategy to the trained problem did not tax the younger adults. Consistent with this, even in the pretraining session the younger adult group was solving nearly all problems correctly. For the older adult group the impact of processing speed was only found to be a factor on the indirect problems Contrary to expectations, the LPS older adults were more likely to transfer the str ategy to the indirect problems as compared to the HPS older adults. The LPS and HPS older adults did not differ in level of performance on the task in the pretraining session showing that the LPS group did in fact show more improvement in the post trainin g condition. A further inspection of individual change scores for each problem type showed that while all older adults are improving on the direct problems, four older adults in the HPS group were not improving on the indirect problems. If those four older adults are removed from the analysis, both the LPS and HPS older adult groups have similar improvement on the indirect and direct problem types. It is surprising that the four older adults who are not able to transfer the strategy are in the HPS group, though other cognitive factors not relating to processing speed may be interacting with the ability to apply and transfer a strategy. Both processing speed groups showed similar performance gains for the direct problems in the post training session.
86 Individual differences in working memory capacity have been found to affect electrophysiological signals in adults (Gevins et al., 2000; Nittono, Nageishi, Nakajima, and Ullsperger, 1999) For example, when G e vens and Smith (2000) divided younger adults by scores on a test of general cognitive ability clear differences in FM theta frequency band were found. The high performing younger adults showed greater increases in FM theta with increased task exposure while the low performing younger adults showed no increase across task exposure. This was interpreted as the highperforming younger adults showing greater ability to sustain attention and allocate resources. Thus, these high performing younger adults were showing greater cognitive control as compared to the other groups In the current work, processing speed ability was found to be an important factor in EEG signatures. The theta frequency band illuminated several interesting differences between the LPS and HPS groups The HPS older adults were found to have simil ar EEG topography to the younger adults This was found in the midline lead analysis in which the HPS older adults showed larger FM theta as compared to the LPS older adults While the HPS older adults had lower FM theta as compared to the younger adults g roup they were more similar in power and topography as compared to the LPS older adults. Additionally, both younger and older HPS groups showed larger right frontal theta activity as compared to the LPS groups. Given the differences in EEG processing between the LPS and HPS older adults, it is surprising that the HPS older adults do not show correspondingly better TOL behavioral performance as compared to the LPS older adults There are several possible reasons for this lack of performance difference between these groups First, as
87 noted earlier, it is possible that the task is very difficult and this masks the more efficient problem solving, as evidenced by increased brain activity, in the HPS older adults Further examination of this finding with problems of varying difficulty level would be a way to investigate this possibility Secondly, individual differences in processing speed and its effect on brain activity may not be involved in the problem solving process Thus, while HPS older adults show more FM theta and increased right activation that is indicative of better problem solving ability, this may not be seen in the performance measures studied here These HPS older adults may be applying more efficient strategies or catching on to the strategy quic ker, something not investigated in the current study Future research should strive to investigate the changes in performance and EEG across the preand post training session to understand differences in learning rates, which may be different across these processing speed groups In support of this, the Kliegl study demonstrated that the effect of individual differences in processing speed was only evident in the later stages of training as compared to the beginning (Kliegl et al. 1990) The collapse of t hese measures across blocks of trials within a session may be masking a behavioral difference5The lack of findings in the lower alpha frequency band between LPS and HPS groups indicates that the underlying cognitive functions associated with the lower alpha frequency band are not affected by processing speed. As described above, the functional role of alpha is associated with the processing of meaningful stimuli, a cognitive process possibly not affected by processing speed. 5 T his analysis could not reasonably be carried out for the current study as the number of trials needed to compute the EEG frequency measures was too small to allow for additional separation across time.
88 Clearly, future work investig ating processing speed as an important predictor of strategy application and EEG signatures would be a valuable extension of this work First, performance on the measured task should be investigated in several ways, including more advanced cognitive constr ucts, such as problem solving efficiency or rate of learning abilities Additionally, coherence measures of neural networks across the brain may yield differential efficiency and widespread processing in the HPS groups, as did the power analysis conducted in the current study. Implications, Suggestions for Future Research and Conclusion The inclusion of physiological measurements in future strategy training research supplies critical information to the understanding of the underlying changes accompanied wi th strategy training This is especially important when investigating changes in aging populations as the underlying structural regions involved are damaged more so than in younger populations The way in which underlying neural regions respond to training may vary depending on the degree of breakdown due to age or disease. This would be expected, of course, since systems severely damaged have been found to behave differently as compared to intact systems ( e.g., Prvulovic, Van de Ven, Sack, Maurer, & Linde n, 2005 ) Additionally, different types of strategic training may be associated with differing post training activity differences An increase in efficiency may be associated with an increase in the task relevant regions A strategy in which participants use a different set of cognitive abilities would possibly result in different regions becoming active (Jonides, 2004 ) Additionally, measures of coherence across brain regions and widespread neuronal interaction may help to pinpoint other regions of decline across adulthood as well as the type of decline An understanding of
89 the changes in the neural systems across the brain may aid in tailoring the cognitive intervention to tap regions less affected by aging. The integration of physiological methods in behavioral training research can help to identify what is actually changing i n the post training condition. As evidenced in the current study, a more stepby step approach may aid in the transfer of the learned strategy to new problems A n approach similar to this step by step training goal management training (GMT), has been applied effectively in both normal and clinical populati ons with excellent success (van Hooren, Valentijn, Bosma, Ponds, van Boxtel, Levine, et al., 2007) Th e theory of GMT is that any activity requires a list of actions to complete until the goal in achieved. Similarly, the strategy training employed in the current study concentrated on participants understanding the stepby step actions needed to take in order to reach the goal While the GMT method of training was originally designed for brain injured patients (Levine, Robertson, Clare, Carter, Hong, Wilson, et al., 2000), its application to normal populations has demonstrated its use in everyday life (van Hooren et al., 2007) I n a st udy which employed GMT younger and older adults received stepby step instructive training on several executive function measures (Levine et al., 2000) Outcome measures of ability to perform day to day activities showed that the trained group far exceeded the performance of the wait list control group. It is suggested that the stepby step procedure of the training aided in increased performance in the elderly. Through innovative research and clear goals on the future of aging research, training paradigms that impact the lives of older adults can become a reality Through
90 dedicated researchers the lives of older adults can be impacted and the last years of life can be associated with independence, growth and sustainability
91 APPENDIX A LIST OF TOWER OF LONDON PROBLEMS Problem numbers refer to the number designated in Berg and Byrd (2002) Problem Number Difficulty Level Problem Type Problem Number Difficulty Level Problem Type Block 1 Block 5 23:32 5 tall switch 32:62 5 control 26:44 5 middle switch 36:66 6 middle switch 26:64 6 C ontrol 32:23 5 tall switch 22:32 6 tall switch 16:34 6 control 22:52 5 C ontrol 13:62 5 tall switch 24:66 5 middle switch 16:54 5 middle switch Block 2 Block 6 56:14 6 C ontrol 14:36 5 middle switch 22:33 5 tall s witch 12:42 5 control 56:34 5 middle switch 12:62 6 tall switch 22:52 5 C ontrol 16:46 6 middle switch 26:56 6 middle switch 12:63 5 tall switch 53:42 5 tall switch 12:42 5 control Block 3 Block 7 56:26 6 middle switch 43:52 5 tall switch 52:4 3 5 tall switch 46:64 5 middle switch 62:22 5 C ontrol 46:24 6 control 54:16 5 middle switch 42:52 6 tall switch 62:22 5 C ontrol 42:12 5 control 52:42 6 tall switch 42:26 5 middle switch Block 4 Block 8 33:22 5 tall switch 42:12 5 control 36:5 4 6 C ontrol 42:53 5 tall switch 36:14 5 middle switch 46:16 6 middle switch 32:22 6 tall switch 66:44 6 control 34:56 5 middle switch 66:24 5 middle switch Block 9 64:46 5 middle switch 62:12 6 tall switch 62:32 5 control 66 :36 6 middle switch 62:32 5 control 62:13 5 tall switch
92 APPENDIX B C ORRELATIONS B ETWEEN B EHAVIORAL V ARIABLES Abbreviations: PC = proportion correct; ST=solution time, NR=proportion of no response Young adult only correlations on top portion of table, Older adult only correlations on lower portion of table ** = p<.001; = p <.01 Variable Name 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Pre Training 1 Control PC 1.000 .695 ** .778 ** .019 .449 .341 .188 .146 .063 .757 ** .812 ** .740 .268 .233 .109 .210 --2 Indirect PC .309 1.000 .814 ** .079 .411 .102 .223 -.357 .005 .680 ** .846 ** .722 ** .264 .045 .070 .181 --3 Direct PC .105 .774 ** 1.000 -.103 .327 .274 .313 -.229 -.191 .766 ** .833 ** .739 ** .184 .097 -.041 .229 --4 Control ST -.064 -.034 -.087 1.000 .664 ** .548 ** .379 .317 .280 -.102 .043 -.051 .757 ** .437 .341 .177 --5 Indirect ST .287 .402 .446 .663 ** 1.000 .704 ** .391 .124 .247 .321 .382 .313 .658 ** .441 .296 .212 --6 Direct ST .073 .306 .410 .517 ** .813 ** 1.000 .483 ** .414 .081 .222 .280 .220 .558 ** .568 ** .348 .121 --7 Control NR .138 .026 .090 .079 .317 .391 1.000 .019 .171 .135 .297 .146 .356 .268 .028 .159 --8 Indirect NR .041 .514 .655 ** .353 .551 ** .620 ** .194 1.000 250 .156 .244 .189 .312 .123 .040 .161 --9 Direct NR .117 .584 ** .795 ** .227 .534 ** .500 .050 .696 ** 1.000 .031 .084 .022 .466 ** .178 .161 .161 --Post Training 10 Control PC .227 .319 .390 -.223 -.433 -.513 -.152 -. 531 ** -.274 1.000 .820 ** .794 ** .239 .170 .054 -.001 --11 Indirect PC .473 .671 ** .446 -.043 -.374 -.338 -.158 -.467 -.289 .653 ** 1.000 .866 ** .262 .159 -.007 .155 --12 Direct PC .153 .496 .418 .056 .257 .257 .451 .346 .315 .522 ** .658 ** 1.000 .208 .100 .128 .217 --13 Control ST .316 .212 .163 .434 .676 ** .609 ** .269 .421 .212 .346 .189 .018 1.000 .633 ** .382 .012 --14 Indirect ST .053 .222 .403 .143 .482 .552 ** .177 .569 ** .296 .409 .276 .387 .440 1.000 .628 ** .049 --15 Direct ST .089 .332 .416 .136 .526 ** .711 ** .207 .565 ** .405 .498 .520 ** .489 .468 .830 ** 1.000 .033 --16 Control NR .082 .219 .255 .077 .422 .480 .342 .609 ** .396 .650 ** .433 .351 .499 .443 .532 ** 1.000 --17 Indirect NR .000 .444 .506 .359 .443 .406 .243 .708 ** .501 .387 .458 .467 .294 .592 ** .538 ** .485 1.000 18 Direct NR .000 .391 .446 .146 .396 .287 .458 .689 ** .397 .372 .401 .630 ** .273 .652 ** .454 .531 ** .718 ** 1
93 Var iable Name 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Pre Training 1 Control PC 1.00 .542 .480 -.018 .114 .162 .013 .038 .130 .545 .691 .503 .007 .062 .019 .106 .009 .010 2 Indirect PC .688 ** 1.00 .793 .039 .029 -.087 .081 -.356 -.326 .540 .784 .633 .046 -.103 -.152 -.093 -.267 -.234 3 Direct PC .668 ** .856 ** 1.00 -.080 -.065 -.089 .181 -.452 -.541 .600 .676 .599 .026 -.210 -.253 -.124 -.340 -.298 4 Control ST -.446 ** -.358 ** -.467 ** 1.00 .649 .502 .152 .267 .164 -.137 .019 -.001 .54 8 .212 .154 .046 .258 .100 5 Indirect ST -.411 ** -.393 ** -.487 ** .840 ** 1.00 .757 .309 .400 .404 -.021 .065 .016 .645 .438 .415 .320 .347 .309 6 Direct ST -.395 ** -.461 ** -.505 ** .777 ** .902 ** 1.00 .389 .523 .369 -.131 .008 -.042 .553 .524 .586 .377 .335 .231 7 Control NR -.352 ** -.259 -.231 .507 ** .613 ** .657 1.00 .135 .021 -.013 .055 -.210 .261 .177 .135 .292 .210 .404 8 Indirect NR -.314 -.547 ** -.632 ** .553 ** .640 ** .706 .419 ** 1.00 .650 -.320 -.273 -.248 .320 .470 .450 .557 .684 .666 9 Direct NR -.289 -.543 ** -.706 ** .526 ** .672 ** .657 .373 ** .767 ** 1.00 -.136 -.142 -.190 .162 .231 .319 .347 .483 .379 Post Training 10 Control PC .698 ** .676 ** .730 ** -.476 ** -.437 ** -.500 -.336 -.529 ** -.427 ** 1.00 .760 .687 -.021 -.156 -.2 40 -.376 -.230 -.221 11 Indirect PC .779 ** .841 ** .767 ** -.344 ** -.347 ** -.384 -.256 -.474 ** -.401 ** .824 ** 1.00 .786 .078 -.072 -.246 -.193 -.240 -.209 12 Direct PC .589 ** .689 ** .661 ** -.254 -.262 -.301 -.368 -.390 ** -.356 ** .732 ** .814 ** 1.00 .110 -.206 -.269 -.198 -.305 -.411 13 Control ST -.506 ** -.422 ** -.481 ** .807 ** .868 ** .841 .608 ** .610 ** .578 ** -.472 ** -.378 ** -.239 1.00 .462 .392 .365 .231 .212 14 Indirect ST -.464 ** -.482 ** -.577 ** .665 ** .786 ** .823 .561 ** .679 ** .599 ** -.524 ** -.440 -.400 ** .824 1.00 .800 .370 .536 .588 15 Direct ST -.517 ** -.524 ** -.612 ** .669 ** .796 ** .858 .563 ** .669 ** .652 ** -.579 ** -.537 ** -.433 ** .827 .935 1.00 .457 .503 .420 16 Control NR -.279 -.369 ** -.419 ** .434 ** .606 ** .638 .530 ** .697 ** .573 ** -.573 ** -.422 ** -.355 ** .641 .640 .682 1.00 .475 .522 17 Indirect NR -.241 -.423 ** -.491 ** .455 ** .515 ** .508 .395 ** .745 ** .597 ** -.400 ** -.392 ** -.405 ** .455 .617 .584 .588 ** 1.00 .717 18 Direct NR -.214 -.378 ** -.438 ** .324 .463 ** .417 .517 ** .717 ** .500 ** -.372 ** -.349 ** -.487 ** .411 .615 .511 .606 ** .758 1.00 Abbreviations: PC = proportion correct; ST=solution time, NR=proportion of no response Age partial correlations with Younger and Older Adults on top portion of table, All participants together, no age partial on lower portion of table ** = p<.001; = p <.01
94 APPENDIX C SIGNIFICANT FINDINGS FROM BEHAVIORAL, EEG AND PROCESSING SPEED ANALYSES S ignificant effects are listed below for each type of analysis completed. Main Effects Inte raction Behavioral Proporti on Correct Age Session PT -Change Score Age -Solution Time Age PT Age*Session*PT Age* Session Session* PT Change Score Age Age* PT Proporti on of No Respons es Age Session PT Age*Session*PT Age* Session Age* PT EEG Rest P.P P.C Rest P.P. P.C Theta Fz Pz Age Lead Age Session Lead Age Lead Session*Lead Session* Lead Age*Lead Session PT Lead Session* Lead Age*Lead Theta FC3 FC4 Age Lead Age Session Lead Age Session Lead -Age* Session* Lead Session* PT Lower Alpha Fz Pz Se ssio n Lead Session Lead Lead*Age --Processing Speed Groups Solution Time --Change Score -Age* PT PS G Rest P.P. P.C Rest P.P. P.C Theta Fz Pz ---Age*Lead*PSG Age* Lead* PSG Age* Lead *PSG Theta FC3 FC4 ---Lead* PSG Lead*PSG Session*PSG*P T Lead* PSG Session* PSG *PT Lower Alpha Fz Pz ----Session*Lead* PSG -Note: Problem presentation (P.P.); Problem completion epoch (P.C.); Processing speed g roup (P.S.G.); Problem type(P.T.)
96 LIST OF REFERENCES Anderson, J.R., Alb ert, M.V., & Fincham, J.M. (2005). Tracing problem solving in real time: fMRI analysis of the subject paced Tower of Hanoi. Journal of Cognitive Neuroscience, 17(8), 12611274. Baddeley, A. (1998). Working memory. Science, 255, 556 559. Bae, J.N., MacFall J.R., Krishnan, K.R.K, Payne, M.E., Steffens, D.C. & Taylor, W.D. (2006). Dorsolateral Prefrontal Cortex and Anterior Cingulate Cortex White Matter Alterations in LateLife Depression. Biological Psychiatry, 60(12), 13561363. Ball, K., Berch, D.B., Helm ers, K.F., Jobe, J.B., Leveck, M.D., & Marsiske, M., et al., (2002). Effects of cognitive training interventions with older adults: A randomized controlled trial. The Journal of the American Medical Association. 288(18), 22712281. Ball, K ., Edwards, J.D ., & Ross, L.A (2007) The impact of spe ed of processing training on cognitive and everyday functions. The Journals of Gerontology: Series B, Psychological Sciences and Social Sciences, 62(1), 19 31. Band, G.P. & Kok, A. (2000). Age effects on response monitoring in a mental rotation task. Biolo gical Psychology, 51(23), 201221. Barber, M., & Scott, D.J. (2004). Validity of the Telephone Interview for Cognitive Status (TICS) in post stroke subjects. International Journal of Geriatric Psychiatry, 19, 75 79. Becic E., Boot, W.R., & Kramer, A.F. (2 008) Training older adults to search more effectively: S canning strategy and visual search in dynamic displays. Psychology of Aging, 23(2), 461 466. Berg, W.K, & Byrd, D. (2002). The Tower of London spatial problem solving task: enhancing clinical and research implementation. Journal of Clinical and Experimental Neuropsychology, 24(5), 586604. Berg, K., Byrd, D., McNamara, J., & Case, K. (2010) Deconstructing the Tower: Parameters and Predictors of Problem Difficulty on the Tower of London Task, Brain a nd Cognition, 72 (3), 472482. Bottiroli, S., Cavallini, E., & Vecchi, T. (2008) Long term effects of memory training in the elderly: A longitudinal study. Archives of Gerontology and Geriatrics, 47(2), 277289. Braver, T.S., Paxton, J.L., Locke, H.S., & B arch, D.M. (2009). Flexible neural mechanisms of cognitive control within human prefrontal cortex. Proceedings of the National Academy of Sciences of the United States of America, 106(18); 73517356.
97 Brigham, M.C., & Pressley, M. (1988). Cognitive monitori ng and strategy choice in younger and older adults. Psychology of Aging, 3(3), 249 257. Brennan, M., Welsh, M.C., & Fisher, C.B. (1997) Aging and executive function skills: A n examination of a community dwelling older adult population. Perceptual and Motor Skills, 84 (3), 1187 1197. Cabeza, R., Anderson, N.D., Locantore, J.K., & McIntosh, A.R. (2002). Aging gracefully: Compensatory brain activity in highperforming older adults. Neuroimage, 17(3), 13941402. Cabeza R. (2002). Hemispheric asymmetry reduction in older adults: The HAROLD model. Psychology of Aging, 17(1), 85 100. Cabeza, R., & Nyberg, L. (2000) Neural bases of learning and memory: Functional neuroimaging evidence. Current Opinion in Neurology, 13(4), 415 421. CahnWeiner, D.A., Boyle, P.A., & Malloy, P.F. (2002). Tests of executive function predict instrumental activities of daily living in community dwelling older individuals. Applied Neuropsychology, 9(3), 187 191. Carretti, B., Borella, E., & de Beni, R. (2007). Does strategic memory training improve the working memory performance of younger and older adults? Experimental Psychology, 54 (4), 311 320. Cavallini, E., Pagnin, A., & Vecchi, T. (2003). Aging and everyday memory: The beneficial effect of memory training. Archives of Gerontology and Geriatrics 37(3), 241257. Cummins, T.D., & Finnigan, S. (2007). Theta power is reduced in healthy cognitive ageing. International Journal of Psychophysiology, 66, 10 17. De Luca, C.R., Wood, S.J., Anderson, V., Buchanan, J.A., Proffitt, T.M., Mahony, K., & Pantelis, C. (2003). Normative data from the CANTAB I: Development of executive function over the lifespan. Journal of Clinical and Experimental Neuropsychology, 25(2), 242 254. Delorme, A. & Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single trial EEG dynamics. Journal of Neuroscience Methods, 134, 9 21. Davis, S.W., Dennis, N.A., Buchler, N.G., White, L.E., Madden, D.J., & Cabeza, R. (2009). Assessing the effects of age on long white matter tracts using diffusion tensor tractography. Neuroimage, 46, 530 541. Dahlin, E., Nyberg, L., Bckman, L., & Neely, A.S. (2008) Plasticity of executive functioning in young and older adults: I mmediate training gains, transfer, and long term maintenance. Psychology and Aging, 23(4), 720730.
98 Daign eault, S., & Braun, C.M. (1993). Working memory and the Self Ordered Pointing Task: further evidence of early prefrontal decline in normal aging. Journal of Clinical and Experimental Neuropsychology, 15(6), 881 895. Doyon, J., & Benali, H. (2005). Reorgani zation and plasticity in the adult brain during learning of motor skills. Current Opinion in Neurobiology, 15(2), 161 167. Duff, K., Beglinger, L.J., & Adams, W.H. (2009). Validation of the modified telephone interview for cognitive status in amnestic mild cognitive impairment and intact elders. Alzheimer Disease and Associated Disorders, 15(2), 161 167. Edwards, J.D., Wadley, V.G., Myers, R.S., Roenker, D.L., Cissell, G.M., & Ball, K.K. (2002). Transfer of a speed of processing intervention to near and far cognitive functions. Gerontology, 48(5), 329 340. Ferrucci, L., Del Lungo, I., Guralnik, J.M., Bandinelli, S., Benvenuti, E., & Salani, B., et al. (1998). Is the telephone interview for cognitive status a valid alternative in persons who cannot be evaluat ed by the Mini Mental State Examination? Aging (Milano), 10(4), 332338. Fincham, J.M., Carter, C.S., van Veen, V., Stenger, V.A., & Anderson, J.R. (2002). Neural mechanisms of planning: A computational analysis using event related fMRI. Proceedings of the National Academy of Sciences of the United States of America, 99(5), 33463351. Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975) Mini mental state: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychia triatry Research, 12(3), 189 198. Friedman, L., Kenny, J.T., Wise, A.L., Wu, D., Stuve, T.A., Miller, D.A, Jesberger, J.A., & Lewin, J.S. (1998). Brain activation during silent word generation evaluated with functional MRI. Brain Language, 64(2), 231 256. Gevins, A., Smith, M.E., McEvoy, L., & Yu, D. (1997). High resolution EEG mapping of cortical activation related to working memory: Effects of task difficulty, type of processing, and practice. Cerebral Cortex, 7(4), 374 385. Gevins, A. & Smith, M.E. (2000 ) Neurophysiological measures of working memory and individual differences in cognitive ability and cognitive style. Cerebral Cortex, 10(9), 829 839. Goethals, I., Audenaert, K., Jacobs, F., Van de Wiele, C., Pyck, H., Ham, H., Vandierendonck, A., van He eringen, C., & Dierckx, R. (2004). Application of a neuropsychological activation probe with SPECT: T he 'Tower of London' task in healthy volunteers. Nuclear Medicine Communications, 25(2), 177182.
99 Greenwood, P.M. (2000). The frontal aging hypothesis eval uated. Journal of the Internation al Neuropsychological Society, 6( 6 ) 705726. Grigsby, J., Kaye, K., Baxter, J., Shetterly, S.M., & Hamman, R.F. (1998). Executive cognitive abilities and functional status among community dwelling older persons in the San Luis Valley Health and Aging Study. Journal of the American Geriatrics Society, 46(5), 590 596. Gunning Dixon, F.M., & Raz, N. (2003). Neuroanatomical correlates of selected executive functions in middleaged and older adults: A prospective MRI study. Neur opsychologia, 41(14).19291941. Grunwald, M., Weiss, T., Krause, W., Beyer, L., Rost, R., Gutberlet, I., & Gertz, H.J. (1999). Power of theta waves in the EEG of human subjects increases during recall of haptic information. Neuroscience Letters, 260(3), 18 9 192. Haskell, E.H. (2001). Transfer of learning: Cognition, instruction, and reasoning. New York: Academic Press. Hoptman, M.J., & Davidson, R.J. (1998). Baseline EEG asymmetries and performance on neuropsychological tasks Neuropsychologia, 36(12), 13431347. Jensen, O., Gelfand, J., Kounios, J., & Lisman, J.E. (2002). Oscillations in the alpha band (912 Hz) increase with memory load during retent ion in a short term memory task. Cerebral Cortex, 12(8), 877 882. Jonides J. (2004 ) How does practice makes perfect? Nature Neuroscience, 7(1), 10 11. Joy, S., Kaplan, E., & Fein, D. (2004). Speed and memory in the WAIS III Digit Symbol Coding subtest across the adult lifespan. Archives of Clinical Neurospychology, 19(6), 759 767. Karbach, J., & Kray, J. (2009) How useful is executive control training? Age differences in near and far transfer of task switching training. Developmental Science, 12(6), 978990. Karrasch, M., Laine, M., Rapinoja, P., & Krause, C.M. (2004). Effects of normal aging on event related desynchronization/synchronization during a memory task in humans. Neuroscience Letters, 366(1), 18 23. Kliegl, R., Smith, J., & Baltes, P. B. (1990). On the locus and process of magnification of age differences during mnemonic training. Developmental Psychology 26 894 904. Klimesch, W., Schimke, H., & Pfurtscheller, G. (1993) Alpha frequency, cognitive load and memory performance. Brain Topography, 5(3), 241 251.
100 Klimesch, W., Freunberger, R., & Sauseng P. (in press). Oscillatory mechanisms of process binding in memory Neuroscience and Biobehavioral Reviews. Knopman, D.S., Roberts, R.O., Geda, Y.E., Pankratz, S.V., & Christianson, T.J.H, Petersen, R.C., & Rocca, W.A. (2010). Validation of the telephone interview for cognitive status modified in subjects with normal cognition, mild cognitive impairment. Neuroepidemiology, 34, 34 42. Koziol, L.F., & Budding, D.E. (2009) Subcortical structures and cognition, New York, New York: Springer. Lamson, N., & Rogers, W.A. (2008). Assessing agerelated patterns in s trategy selection on a mathematical problem solving task. Journal of Gerontology, 63(3), 146155. Lazeron, R.H., Rombouts S A Machielsen, W C Scheltens P Witter M P Uylings H B & Barkhof F. (2000). Visualizing brain activation during plannin g: the tower of London test adapted for functional MR imaging. American Journal of Neuroradiology, 21(8), 14071414. Lee, T.M., Leung, A.W., Fox, P.T., Gao, J.H., & Chan, C.C. (2008). Age related differences in neural activities during risk taking as revealed by functional MRI. Social Cognitive Affective Neuroscience 3(1), 7 15. Levine, B., Robertson, I.H., Clare, L., Carter, G., Hong, J., & Wilson, et al. (2000). Rehabilitation of executive functioning: an experimental clinical validation of goal management training. Journal of International Neuropsychology, 6(3), 299 312. Li, S.C., Schmiedek, F., Huxhold, O., Rcke, C., Smith, J., & Lindenberger, U. (2008) Working memory plasticity in old age: P ractice gain, transfer, and maintenance. Psychology and Aging 23 (4), 731742. Lustig, C., Shah, P., Seidler, R., & Reuter Lorenz, P.A. (2009) Aging, training, and the brain: A review and future directions. Neuropsychology Review, 19(4), 504 522. Mahncke, H.W., Connor, B.B., Appelman, J., Ahsanuddin, O.N., Hardy, J .L., & Wood, R.A., et al. (2006) Memory enhancement in healthy older adults using a brain plasticity based training program: A randomized, controlled study. Proceedings of the National Academy of Sciences of the United States of America, 103(33), 1252312 528. McArdle, J.J., & Prindle, J.J. (2008) A latent change score analysis of a randomized clinical trial in reasoning training. Psychology and Aging, 23(4), 702 719. McEvoy L., Pellouchoud E., Smith M., & Gevins, A. (2001) Neurophysiological signals of w orking memory in normal aging. Cognitive Brain Research, 11(3), 363 376.
101 McNamara, D.S., & Scott, J.L. (2001). Working memory capacity and strategy use. Memory and Cognition, 29(1), 10 17. Mizuhara, H., Wang, L.Q., Kobayashi, K., & Yamaguchi, Y. (2004) A long range cortical network emerging with theta oscillation in a mental task. Neuroreport, 15(8), 12331238. Morris, R.G., Ahmed, S., Syed, G.M., & Toone, B.K. (1993) Neural correlates of planning ability: Frontal lobe activation during the Tower of London test. Neuropsychologia, 31(12),13671378. Newman, S.D., Greco, J.A., & Lee, D. (2009). An fMRI study of the Tower of London: A look at problem structure differences. Brain Research, 1286, 123 132. Nittono, H., Nageishi, Y., Nakajima, Y., & Ullsperger P (1999) Event related potential correlates of individual differences in working memory capacity. Psychophysiology, 36(6), 745 754. Nyberg, L., Sandblom, J., Jones, S., Neely, A.S., Petersson, K.M., Ingvar, M., & Bckman, L. (2003) Neural correlates of t raining related memory improvement in adulthood and aging. Proceedings of the National Academy of Sciences of the United States of America, 100(23), 1372813733. Olesen, P.J., Westerberg, H., & Klingberg, T. (2004). Increased prefrontal and parietal activi ty after training of working memory. Nature Neuroscience, 7(1), 7579. Onton J., Delorme A., & Makeig S. (2005). Frontal midline EEG dynamics during working memory NeuroImage, 27(2), 341 356. Phillips, L.H., Wynn, V., Gilhooly, K.J., Della Sala, S., & Log ie, R.H. (1999). The role of memory in the Tower of London task. Memory, 7 (2), 209 231. Persson, J., Lustig, C., Nelson, J.K., & Reuter Lorenz, P.A. (2007). Age differences in deactivation: A link to cognitive control? Journal of Cognitive Neuroscience, 19(6), 10211032. Pradhan, S.N. (1980). Central neurotransmitters and aging. Life Science,26(20), 16431656. Prvulovic, D., Van de Ven, V., Sack, A.T., Maurer, K., & Linden, D.E. (2005) Functional activation imaging in aging and dementia. Psychiatry researc h, 140(2), 97113. Raghavachari, S., Lisman, J.E., Tully, M., Madsen, J.R., Bromfield, E.B., & Kahana, M.J. (2006) Theta oscillations in human cortex during a working memory task: Evidence for local generators. Journal of Neurophysiology, 95(3), 1630 1638. Rajah, M.N., & McIntosh, A.R. (2008). Age related differences in brain activity during verbal recency memory. Brain Research, 1199, 111 125.
102 Rapp, S., Brenes, G., & Marsh, A.P. (2002). Memory enhancement training for older adults with mild cognitive impairment: A preliminary study. Aging and Mental Health, 6(1), 5 11. Raz, N., Lindenberger, U., Rodrigue, K.M., Kennedy, K.M., Head, D., Williamson, A., Dahle, C., Gerstorf, D., & Acker, J.D. (2005) Regional brain changes in aging healthy adults: General trends, individual differences and modifiers. Cerebral Cortex, 15 (11), 16761689. Rebok, G.W., Carlson, M.C., & Langbaum, J.B. (2007) Training and maintaining memory abilities in healthy older adults: Traditional and novel approaches. The Journals of Gerontology. Series B, Psychological sciences and social sciences, 62(1), 53 61. Reichenberg, A. & Harvey, P.D. (2007) Neuropsychological impairments in schizophrenia: Integration of performancebased and brain imaging findings. Psychological Bulletin, 133(5), 8 33858. Riccio, C.A., Wolfe, M.E., Romine, C., Davis, B., & Sullivan, J.R. (2003) The Tower of London and neuropsychological assessment of ADHD in adults. Archives of Clinical Neuropsychology : The Official Journal of the National Academy of Neuropsychologists, 19 (5), 661 671. Riis, J.L., Chong, H., Ryan, K.K., Wolk, D.A., Rentz, D.M., Holcomb, P.J., & Daffner, K.R. (2008). Compensatory neural activity distinguishes different pat terns of normal cognitive aging. Neuroimage, 39(1), 441 454. Romine, C.B., & Reynolds, C.R. (2005). A model of the development of frontal lobe functioning: Findings from a metaanalysis Applied Neuropsychology, 12(4), 190 209. Rosano, C., Aizenstein, H., Cochran, J., Saxton, J., De Kosky, S., & Newman, A.B., et al., (2005) Functional neuroimaging indicators of successful executive control in the oldest old. Neuroimage, 4, 881 889. Rypma, B., & D'Esposito, M. (2000). Isolating the neural mechanisms of agerelated changes in human working memory. Nature Neuroscience, 3(5), 509 515. R ypma, B., Prabhakaran, V., Desmond, J.E., & Gabrieli, J.D. (2001). Age differences in prefrontal cortical activity in working memory. Psychology and Aging, 16(3), 371384. Rypma, B., Berger, J.S., Genova, H.M., Rebbechi, D., & D'Esposito, M. (2005). Dissoc iating agerelated changes in cognitive strategy and neural efficiency using event related fMRI. Cortex, 41 (4), 582594. Salthouse, T.A. (2005). Relations between cognitive abilities and measures of executive functioning. Neuropsychology, 19(4),532545.
103 Sa useng, P., Klimesch, W., Doppelmayr, M., Hanslmayr, S., Schabus, M., & Gruber W.R. (2004) Theta coupling in the human electroencephalogram during a working memory task. Neuroscience Letters, 354(2), 123 126. Sauseng, P., Klimesch, W., Schabus, & Doppelmay r, M. (2005). Fronto parietal EEG coherence in theta and upper alpha reflect central executive functions of working memory. International Journal of Psychophysiology, 57(2), 97 103. Sauseng, P., Klimesch, W., Stadler, W., Schabus, M., Doppelmayr, M., & Han slmayr, S., et al. (2005). A shift of visual spatial attention is selectively associated with human EEG alpha activity. The European Journal of Neuroscience, 22(11), 29172926. Silberstein, R., Danieli, F., & Nunez,P., (2003) FrontoParietal Evoked Potent ial Synchronization is Increased During Mental Rotation. Brain Imaging, 14(1), 67 71. Singer, T., Lindenberger, U., & Baltes, P.B. (2003). Plasticity of memory for new learning in very old age: A story of major loss? Psychology and Aging, 18(2), 306 317. Smith, M.E., McEvoy, L.K., & Gevins, A. (1999) Neurophysiological indices of strategy development and skill acquisition. Cognitive B rain R esearch, 7(3):389 404. Ueda, T., Musha, T., & Yagi, T. (2009) Research of the characteristics o f Alzheimer's disease using EEG. Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 1, 49985001. Unterrainer, J.M., Rahm, B., Kaller, C.P., Ruff, C.C., Spreer, J., Krause, B.J., Schwarzwald, R., Hautzel, H., & Halsband, U. (2004) When pl anning fails: I ndividual differences and error related br ain activity in problem solving. Cerebral Cortex, 14 (12), 13901397. van den Heuvel, O.A., Groenewegen, H.J., Barkhof, F., Lazeron, R.H., van Dyck, R., & Veltman, D.J. (2003) Frontostriatal system i n planning complexity: A parametric functional magnetic resonance version of Tower of London task. Neuroimage, 18(2), 367 374. van Paasschen, J., Clare, L., Woods, R.T., & Linden, D.E. (2009) Can we change brain functioning with cognitionfocused interventions in Alzheimer's disease? The role of functional neuroimaging. Restorative neurology and neuroscience, 27(5), 473 491. van Hooren, S.A., Valentijn, S.A., Bosma, H., Ponds, R.W., van Boxtel, M.P., & Levine, B., et al. (2007) Effect of a structured cour se involving goal management training in older adults: A randomised controlled trial. Patient Education and Counseling, 65(2), 2006.
104 Venkatraman, V.K., Aizenstein, H., Guralnik, J., Newman, A.B., Glynn, N.W., & Taylor, C., et al.(2010) Executive control f unction, brain activation and white matter h yperintensities in older adults. Neuroimage, 49(4), 34363442. Verhaeghen, P., Marcoen, A., & Goossens, L.(1992). Improving memory performance in the aged through mnemonic training: A meta analytic study. Psychol ogy of Aging, 7 (2), 242 251. Verhaeghen, P., & Marcoen, A. (1996). On the mechanisms of plasticity in young and older adults after instruction in the method of loci: Evidence for an amplification model. Psychology and Aging, 11(1), 164178. Yagoubi, R.E., Lemaire, P., & Besson, M. (2005) Effects of aging on arithmetic problem solving: An event related brain potential study. Journal of Cognitive Neuroscience, 17(1), 37 50. Yang, L., & Krampe, R.T. (2009) Long term maintenance of retest learning in young ol d and oldest old adults. The Journals of Gerontology, Series B, Psychological Sciences and Social Sciences, 64(5), 608611. Wager, T.D., & Smith, E.E. (2003). Neuroimaging studies of working memory: A metaanalysis. Cognitive, Affective & Behavioral Neuros cience, 3 (4), 225 274. Wagner, G., Koch, K., Reichenbach, J.R., Sauer, H., & Schlsser, R.G. (2006). The special involvement of the rostrolateral prefrontal cortex in planning abilities: An event related fMRI study with the Tower of London paradigm. Neuropsychologia, 44(12), 23372347. Wechsler, D. (1997). Wechsler Adult Intelligence Scale Third Edition, The Psychological Corporation, San Antonio, TX. Welsh, M C Pennington, B F., & Groisser, D.B (1991). A normativedevelopmental study of executive functi on: A window on prefrontal function in children. Developmental Neuropsychology, 7(2), 131 149. West, R. & Moore, K. (2005) Adjustments of cognitive control in younger and older adults Cortex, 41(4), 570 581. Westerberg, H., & Klingberg, T. (2007) Changes in cortical activity after training of working memory: A single subject analysis. Physiology and B ehavior, 92(1 2), 186192. Wright, R., Thompson, W.L., Ganis, G., Newcombe, N.S., & Kosslyn, S.M. (2008) Training generalized spatial skills. Psychonomic Bu lletin and Review, 15(4), 763771.
105 Zimprich, D. & Martin, M. (2002) Can longitudinal changes in processing speed explain longitudinal age changes in fluid intelligence? Psychology and Aging, 17(4), 690 695. Zook, N., Welsh, M.C., & Ewing, V. (2006). Performance of h ealthy, o lder a dults on the t ower of l ondon r evised: Associations with v erbal and n onverbal a bilities Aging, Neuropsychology \ and Cognition, 13(1), 1 19.
106 BIOGRAPHICAL SKETCH Kimberly was born in Ipswitch, Massachusetts and raised in Miam i, Florida After high school Kimberly attended the University of Florida to pursue her undergraduate degree in psychology During her time at UF, Kimberly worked in several laboratori es on studying cognition, development and neuropsychology After workin g in Dr. William Keith Berg s laboratory for several months, Kimberly decided to stay at UF to continue her study of neuropsychological development Kimberlys g raduate research focused on children, with and without Attention Deficit Hyperactivity Disorder ( ADHD) and the developmental spectrum of planning and learning until adulthood. During her study at UF, Kimberly became interested in the later aspects of development, specifically ways to improve the declining memory and planning abilities of older adul ts. This change in focus has helped her to more fully understand the changes that occur throughout life, why those occur, and its impact on different ages. Kimberly hopes to continue her work in older adult memory and cognition improvement by becoming more involved in research with Alzheimers disease and mild cognitive impairment These two devastating diseases often claim the last years of a person s life and only through intense research and medical practice can they be alleviated. In her spare time, Ki mberly enjoys reading and spending time with the love of her life, Christopher Case The lovebirds reside in Gainesville, Florida in their first home.