The Influence of Memory Beliefs in Individuals with Traumatic Brain Injury. Karen Anne Kit B.Sc., University of Waterloo, 2000

Transcription

1 The Influence of Memory Beliefs in Individuals with Traumatic Brain Injury Karen Anne Kit B.Sc., University of Waterloo, 2000 A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of MASTER OF ARTS in the Department of Psychology O Karen Anne Kit, 2004 University of Victoria All rights reserved. This thesis may not be reproduced in whole or in part, by photocopying or other means, without permission of the author.

2 Supervisor: Dr. Catherine A. Mateer ABSTRACT Metamemory is a term used to refer to the beliefs individuals hold about aspects of their memory function. The purpose of this study was to explore the differences in metamemory between individuals with persistent complaints following a mild traumatic brain injury (MTBI) (n=26), individuals who had sustained a severe TBI (STBI) (n=16), and a control group of uninjured adults (n=42). Moreover, this study examined the role of various components of metamemory and the interactions of this construct with depression, stress, anxiety, neuroticism, and post-concussion symptoms. These constructs were measured by the Metamemory in Adulthood Questionnaire (MIA), Beck Depression Inventory - I1 (BDI-11), Perceived Stress Scale, Beck Anxiety Inventory (BAI), NEO PI-R, and Post-Concussion Symptom Checklist, respectively. It was found that the MTBI and STBI groups used external and internal memory strategies to a greater extent than the control group, believed their memory capacity was not as strong, and perceived a change in their memory abilities (although individuals with STBI did not perceive as much of a change as individuals with MTBI). In addition, the MTBI group indicated placing a higher importance on having a good memory than did the STBI or control groups. Other significant differences were found on the BDI and Post-Concussion Symptom Checklist in that both TBI groups scored higher than the control group on these measures.

3 There was a stronger relationship between the Change scale of the MIA and postconcussion symptoms for the two TBI groups than for the control group. Interestingly, the Change scale of the MIA was also found to mediate the relationship between TBI and depression, suggesting that negative affect may be, at least in part, the result of specific beliefs with respect to memory functioning. These findings emphasize the potential importance and impact of these beliefs in individuals with TBI, and suggest that efforts to address potentially maladaptive or negative beliefs about memory functioning and memory control might be valuable within a rehabilitative context.

4 TABLE OF CONTENTS TITLE PAGE... i.. ABSTRACT TABLE OF CONTENTS... iv... LIST OF TABLES... viii LIST OF FIGURES... ix ACKNOWLEDGEMENTS... x INTRODUCTION... 1 Repercussions of Severe TBI... 2 Neuropsychological Tests and their Relation to MTBI... 3 Affective Symptoms and their Relation to Subjective Evaluations... 4 Affective Symptoms and Severe TBI (STBI)... 7 Expectations Surrounding MTBI... 8 Memory Complaints in the Aging Population Definition of Metamemory Memory Self-Efficacy Assessment of Metamemory Metamemory Studies within the Aging Population Subjective Complaints and Metamemory Purpose of the Study: An Exploration of Metamemory in Individuals with TBI Awareness and Metamemory in Severe TBI (STBI) Metamemory Within the STBI Population... 18

5 METHOD Participants Participant Recruitment Measures Background Information Metamemory Questionnaire Postconcussion Syndrome Checklist Perceived Stress Scale Beck Depression Inventory-I Beck Anxiety Inventory NEO PI-R Procedure RESULTS Metamemory Findings Metamemory Components Analysis Correlations between Subscales of the MIA Findings on a Measure of Metamemory Correlations between Subscales of Affective Functioning Findings on Measures of Emotional Functioning and Personality TBI as a Moderator Metamemory as a Mediator DISCUSSION Differences in Metamemory... 42

6 Depression. Anxiety. Neuroticism. Stress and... Postconcussion Symptoms 46 The Interaction Between Metamemory and Negative Affect Metamemory as a Mediator Correlations between Metamemory Scales and Affective Scales Correlations between the Scales of the MIA Strengths and Limitations Future Research and Conclusions REFERENCES 58 APPENDIX A: Letter to Brain Injury Associations APPENDIX B: Flyer to Brain Injury Associations APPENDIX C: Advertisement in Local Newspaper APPENDIX D: Personal InjuryIHistory Questionnaire APPENDIX E: Metamemory in Adulthood Questionnaire APPENDIX F: Changes made to the... 'Metamemory in Adulthood' questionnaire 92 APPENDIX G: Postconcussion Syndrome Checklist APPENDIX El: Perceived Stress Scale APPENDIX 1: Beck Depression Inventory -. I APPENDIX J: Beck Anxiety Inventory APPENDIX K: NEO-R APPENDIX L: Participant Consent Form 103 APPENDIX M: Control Consent Form

7 APPENDIX N: The Rotated Component Matrix for the Revised MIA vii

8 LIST OF TABLES Table 1. Table 2. Table 3. Table 4. Demographic Characteristics of TBI and Non-TBI Groups Cause of Injury Correlations between Subscales of the MIA Means and Standard Deviations on the Metamemory in Adulthood Questionnaire (MIA) Table 5. Table 6. Correlations between Measures of Affective Functioning Means and Standard Deviations for Emotional Functioning and Personality Measures Table 7. Standardized Beta Coefficients for Interaction Terms and Dependent Variables Table 8. Correlations between a Metamemory Scale and Emotional Functioning for TBI Groups... 41

9 LIST OF FIGURES Figure 1. A Scree plot of the components of the revised MIA

10 ACKNOWLEDGEMENTS The very first person I would like to thank is my supervisor, Dr. Catherine Mateer, for her great support, knowledge, insight, and enthusiasm for this project. It is wonderful to be able to share very similar research interests and to explore these interests together. I would also like to thank my other committee members, Dr. Roger Graves and Dr. David Hultsch, as well as my External Examiner, Dr. Jane Milliken. Their expertise and insights into this research project helped a great deal in refining my thesis. Additionally, I would like to thank fellow graduate students/friends in the Department of Psychology for their support throughout the last couple of years. It was great to be able to experience this process with many of you. Moreover, a big thank-you to friends outside of the department to whom I could always turn to. Finally, I greatly thank my parents who provided me with all the tools necessary to complete this thesis. Their never-ending support and unwavering belief in me are very appreciated, and I couldn't have done it without the two of you.

11 INTRODUCTION Mild traumatic brain injury (MTBI) is one of the most common neurological disorders (Alexander, 1995). MTBI is characterized as an acute trauma to the brain due to contact forces or to acceleration/deceleration trauma. If there is duration of unconsciousness following the injury, it is brief, usually lasting seconds to minutes. Confusion, disorientation, andlor a period of post-traumatic amnesia, with loss of memory for events immediately preceding or following the injury may be present (Alexander, 1995). Many individuals who have sustained a MTBI will notice some symptoms within the acute phase of their injury, but most recover within days to weeks of the initial injury. However, some patients (approximately 5-15%) continue to report disabling symptoms a year or more post injury (Alexander, 1995). Common symptoms include memory impairment, difficulties with concentration, dizziness, reduced tolerance to light and noise, irritability, fatigue, insomnia, anxiety, depression, and headache (Lishman, 1988). These symptoms are grouped under the syndrome coined 'postconcussion syndrome' (PCS). There seems to be no agreed upon time span in which symptoms from a mild brain injury develop into PCS. Some studies refer to symptoms occurring immediately after injury as PCS, others only to symptoms that persist for many months after injury. A confounding factor within research and clinical practice in the area of MTBI is a lack of correspondence between the apparent severity of injury and the number, persistence, and severity of PCS symptoms (Bohnen & Jolles, 1992). In other words, there appears to be no clear physiological link between the injury and the resulting

12 postconcussion syndrome. In some, though not all cases, there are positive findings on neuroimaging measures, including magnetic resonance imaging, and functional imaging measures such as positron emission tomography (PET). Neuropsychological tests also tend to show contradictory findings. Some studies reveal subtle impairments in memory and attention (Barth, Macciocchi, Giordani, Rimel, Jane, & Boll, 1983; Bohnen, Jolles, & Twijnstra, Mellink, & Sulon, 1992) in this population, whereas other studies do not (Gentilini et al., 1985; Dikmen, McLean & Temkin, 1986; Newcombe, Rabbitt, & Briggs, 1994). In addition, many studies have indicated an impact of emotional functioning and stress on the severity of PCS symptoms and their impact on daily life. Thus, the relative impacts of biological and psychosocial factors on expression of and response to PCS symptoms are not well understood. Repercussions of Severe TBI Traumatic brain injuries occur along a spectrum, and a more severe brain injury (STBI) is typically associated with a longer duration of loss of consciousness and increased length of post-traumatic amnesia. Similar to those individuals with MTBI, those with STBI complain of poor concentration, poor memory, and depressed mood (Alexander, 1995). Additionally, Rapoport, McCauley, Levin, Song & Feinstein (2002) found that individuals with STBI displayed evidence of anxiety, depression, irritability, and mental fatigue. There is often diffuse injury to the frontalltemporal structures of the brain in STBI (Prigatano, 1992). Perhaps as a consequence of these injuries, individuals often demonstrate impaired insight with respect to changes in their behavior and cognitive abilities.

13 Neuropsychological Tests and their Relation to MTBI The literature on neuropsychological findings following MTBI has been mixed, and probably dependent on both the nature of the sample and the particular measure employed. Mateer (1992) found that MTBI individuals, who were symptomatic after an injury at least 12 months prior, did not differ from controls on the immediate recall portion of the Logical Memory subtest of the Wechsler Memory Scale-Revised (WMS- R) or the Rey Auditory Verbal Learning Test. However, the MTBI group did perform more poorly than controls on the Digit Symbol subtest of the Wechsler Adult Intelligence Scale-Revised (WAIS-R). The data were interpreted to suggest that some individuals with MTBI, even after a year or more post injury, may have intact memory, as measured by psychometric tests, but have difficulty with timed measures requiring attention and concentration. It was suggested that attentional problems may have been contributing to perceived memory failures. Raskin, Mateer, and Tweeten (1998) found that tests that require time-dependent complex attention, working memory, shifting mental sets (Test of Reading Speed, Symbol Digits Modalities Test, Trail Making Test-B, Picture Rapid Naming, Attention Process Test), and tests of verbal learning (Weschler Memory Scale-Revised - Paired Associates, Wechsler Memory Scale-Revised - Logical Memory, and California Verbal Learning Test - Trial 1) were impaired in mildly brain-injured individuals who continued to be symptomatic and were referred for neuropsychological assessment approximately 2 years after injury. Additionally, the Consonant Trigrams Test (Peterson and Peterson,

14 1959), a test of working memory, has also been found to be sensitive to MTBI (Stuss, Stethem, Hugenholtz, & Richard, 1989). These studies suggest that there may be subtle neuropsychological difficulties, particularly in the areas of concentration, attention, and working memory in some individuals with persistent PCS following MTBI. However, difficulty in these domains is not uncommon in other disorders, including disorders of mood involving depression and anxiety. Affective Symptoms and their Relation to Subjective Evaluations Although some neuropsychological deficits are commonly demonstrated in individuals with persistent PCS, there is concern that these might be secondary to emotional factors rather than to brain injury per se. Sawchyn, Brulot and Strauss (2000), for example, found significant correlations between the Postconcussion Syndrome Checklist (PCSC) and the Beck Depression Inventory (BDI); complaints of difficulties with concentration, fatigue and irritability were highly correlated with the BDI. The PCSC, however, did not appear to be related to measures of injury severity such as duration of loss of consciousness, or length of hospitalization. Thus, the injury characteristics of a MTBI alone were not strongly related to the experience of PCS symptoms. The authors suggested that the endorsement of PCS may be reflective of general distress, rather than the occurrence or the severity of concussion, per se. Further evidence for this is provided by Machulda, Bergquist, Ito, & Chew (1998). They found that in 'normal' undergraduate students, the intensity of PCS symptoms was associated with the level of perceived stress, thus reinforcing the idea that such symptoms may be

15 reflective of general distress. Indeed, Suhr and Gunstad (2002) concluded that the experience of depressive symptoms, not head injury, is responsible for the self-report of PCS symptoms, including cognitive symptoms. In another study highlighting the relation between affective symptoms and subjective ratings of post-concussion symptoms, Trahan, Ross and Trahan (2001) found that individuals who had sustained a MTBI (more than 12 months previous) reported significantly higher BDI scores than normal subjects. In addition, they found that depressed individuals reported significantly higher scores on the Postconcussional Index (a composite score from the Beaumont Postconcussional Index) and the General Symptom Index (which relates to physical and cognitive symptoms). These authors also reported moderate to high correlations between the Postconcussional Index and the Beck Anxiety Inventory, suggesting that anxiety and subjective ratings of PCS symptoms appear to have a direct relation. It remains unclear as to whether depression and anxiety causes individuals who have sustained a MTBI to perceive themselves as having more symptoms, whether experiencing persistent symptoms following a MTBI causes depression and anxiety, or whether PCS and depression/anxiety are simply characterized clinically by some of the same symptoms. There is some evidence, however, that depression and subjective ratings of PCS symptoms have a direct relation. Fann, Uomoto, and Katon (2001), for example, found that successful treatment of depression yielded significant improvements in cognition in individuals with MTBI. Leininger, Kreutzer and Hill (1991) found that subjects with MTBI reported elevated levels of depressive symptomatology, confusion, disorganized and ruminative

16 thinking, interpersonal alienation, and preoccupation with physical illness on the Minnesota Multiphasic Personality Inventory (MMPI). It is unclear, however, whether individuals who have sustained a MTBI have an emotional reaction to their symptoms, or have a pre-existing personality andlor coping style that impacts their response to the injury and its consequences. Mathias and Coats (1999) also found a relationship between the level of accident-related psychological stress and emotional symptoms, suggesting that these sequelae reflect a reaction to injury, rather than a direct consequence of organic injury. Karzmark, Hall and Englander (1995) found similar findings, in that the number of elevated MMPI scales was strongly associated with the subjective impact level. Thus, the self-reported emotional functioning of individuals with MTBI appears to have a influence on their perception of and reporting about PCS symptoms. Elaborating on the possible relationship between affective symptoms and PCS, Bohnen, Jolles, Twijnstra et al., (1992) found that individuals with MTBI complained more about cognitive problems associated with PCS, but did not complain disproportionately more about emotional symptoms than non-concussed controls. This was an interesting finding in that it suggests that, regardless of the reality of impairment, cognitive difficulties often become the target symptom of concern. It is important to recognize that checklists may be useful for acquiring information on perceived difficulties, but they do not directly measure areas of difficulty, as do neuropsychological tests. Moreover, it is important to recognize the role that affective functioning appears to have on responding to checklists of symptoms. When considering affective symptomatology in individuals with MTBI, a contentious issue is whether the symptoms reflect pre-morbid versus only post-morbid

17 functioning. It is difficult to ascertain whether the affective sequelae were present before the head injury, or occurred as a result of the head injury and attempts to deal with it (Dikmen & Levin, 1993). Another complication is the fact that post-injury personality changes are frequently viewed as being influenced by a person's pre-morbid personality characteristics. In trying to study both pre- and post- morbid functioning, Kurtz, Putnam, and Stone (1998) found that significant others rated individuals with MTBI as increasing in neuroticism over a time period of 6 months post-injury, as measured by the revised NEO Personality Inventory (NEO PI-R). Thus, stress, anxiety, depression, and neuroticism all appear to influence PCS symptoms. It cannot be determined, however, if these personality traits were present before injury and contributed to the increase of PCS symptoms, or whether they are either a consequence of dealing with PCS symptoms post-injury, or a direct result of the injury itself. Affective Symptoms and Severe TBI (STBI) Affective and emotional symptoms, including anxiety and depression, are common sequelae of both mild and moderate to severe TBI (e.g., Rutherford, 1989; Van Zomeren & Van den Berg, 1985). However, several studies have found that individuals with MTBI have, paradoxically, higher elevated MMPI-2 profiles, suggestive of more extensive and severe subjective symptoms than individuals with more severe TBI (Millers & Donders, 2001). The authors also stated that it was unlikely that this could be attributed to denial or complete lack of awareness. Affective disturbances that begin soon after severe brain damage are thought to be closely tied to pathological changes in

18 the brain, whereas those that are in the years following the injury are more likely to reflect reactionary responses to the injury. In other words, depression and anxiety often occur as reactionary responses in chronic stages following severe injury as the individual continues to experience losses and frustration as helshe attempts to re-enter society. Prigatano (1992) hypothesized that individuals with brain injury demonstrate worsening emotional symptoms over time as the individual has enhanced awareness of hislher deficits. From this viewpoint, individuals with severe TBI who are unaware of their own cognitive limitations experience little psychological distress and are less likely to develop affective symptoms. In contrast, those individuals who have insight into their cognitive limitations may develop a reactionary affective disturbance. Expectations Surrounding MTBI To understand how affective symptomatology changes after a MTBI, it is useful to examine the expectations surrounding MTBI. Gunstad and Suhr (2001) found that individuals expect negative consequences, including memory problems, following a head injury. Similarly, Mittenberg, DiGuilio, Perrin, and Bass (1992) found that individuals who have sustained a MTBI may underestimate the normal prevalence of certain symptoms and tend to reattribute these 'everyday' physical and cognitive difficulties or inefficiencies to the head trauma. These expectations may produce selective attention to symptoms and anxiety about their significance. Consistent with this, Ferguson, Mittenberg, Barone, and Schneider (1999) found that 'normal' individuals tended to substantially overestimate the number of PCS symptoms likely to follow mild head trauma. Furthermore, head-injured athletes

19 underestimated pre-injury symptoms compared with uninjured athletes. This suggests that individuals expect more symptoms than actually occur and 'forget' about the occurrence of such symptoms beforehand. However, the mean number of postconcussion symptoms reported by MTBI individuals did not differ significantly from the average number of symptoms reported at baseline by normal controls, suggesting that MTBI individuals may have experienced such symptoms in the past, but are now reattributing them to the head injury (Ferguson et al., 1999). Along this line of reasoning, it follows that treatment targeted at alleviating the negative expectations about MTBI may produce a more positive outcome. Mittenberg, Tremont, Zielinski, Fichera, and Rayls (1996) gave subjects who had sustained a mild head injury or concussion a 10-page manual that normalized common post concussive symptoms, and highlighted the need to reduce stress and anxiety. These individuals also met once with a therapist to review the nature and incidence of expected symptoms, and techniques for reducing symptoms. The control group, on the other hand, was given standard discharge information and was told to contact the doctor if any persisting problems were encountered. Treated patients who received the manual and met with a therapist reported significantly shorter symptom duration and significant remission of memory difficulty. Both groups showed significant reductions in anxiety and depression, but the treatment group also experienced significant remission of headache, fatigue, memory difficulty, concentration impairments, and visual disturbances (Mittenberg et al., 1996). Thus, it appears that expectations about the repercussions of a mild brain injury, as well as the experience of cognitive difficulties early after injury, may play a large role in the persistence of PCS.

20 Overall, it appears that a negative affective state can influence PCS symptoms, including cognitive symptoms. If individuals are made aware of the potential symptoms following MTBI, and assisted in coping with symptoms, as well as helped to deal with the interplay between negative affect and symptoms, a more positive outcome, including remission of cognitive deficits, can be facilitated. Those who do not receive such treatment may develop an expectation of negative change, and have selective attention to common brain injury symptoms. They may also over-attribute cognitive inefficiencies to the head trauma, instead of to everyday stressors, potential emotional factors, or to the normal variability of daily cognitive functioning. Thus, individuals' beliefs about the repercussions of a head injury may play a large role in whether they develop PCS. Memory Complaints in the Aging Population An analysis of parallel research in the aging domain may help in further clarifying the role of beliefs in the post-injury functioning of individuals with MTBI. Older individuals, in general, as compared to younger adults, view themselves as having less memory capacity, report a decline in memory, and believe there is little they can do to enhance their memory or prevent deterioration (Hultsch, Hertzog, Dixon, & Davidson, 1988). Moreover, older individuals oftentimes adhere to the belief of failing memory in old age, regardless of any real or perceived age-related changes in memory. Similar to individuals who have sustained a MTBI, older individuals may focus on normal incidents of forgetting and attribute these everyday memory failures to age-related loss (Hertzog & Dixon, 1994).

21 Overall, elderly individuals and individuals with MTBI appear to have similar concerns and complaints regarding memory, and similar issues exist within both populations in regard to how to objectively measure complaints, thus endorsing the idea that examining parallel research within the aging literature may help shed light on similar issues that exist within the MTBI literature. Definition of Metarnemory Research in the aging domain has included investigating the concept of metamemory and how it relates to memory performance. Metamemory is defined as one's knowledge, awareness, and beliefs about the functioning, development, and capacities of one's own memory (Lovelace, 1990). The extent to which one places oneself in memory demanding situations and the amount of effort put into the tasks, as well as expectations of memory performance, appears to be determined, at least in part, by metamemory (Bandura, 1989). Research on metamemory has established it as a multidimensional construct. Hultsch et al. (1988) identified four broad dimensions of metamemory. The first is memory knowledge, which reflects factual knowledge about memory tasks and processes. Examples of this dimension include the realization that recall tasks are more difficult than recognition tasks. The second dimension, memory monitoring, involves knowing how one's memory is typically used and the current state of one's memory. An example of this is evaluations of the accuracy of one's performance. The third dimension, memory-related affect, is the emotions surrounding memory situations, examples of which include depression and anxiety with respect to real or potential

22 memory failures. The fourth dimension, memory self-efficacy, refers to beliefs about one's ability to use memory effectively. Examples include the degree to which one feels memory performance is under personal control. Memory Self-EfJicacy Memory self-efficacy (MSE) incorporates an individual's beliefs with respect to his or her memory abilities, strengths, and weaknesses. If these beliefs are strongly negative, they may easily lead to a self-fulfilling prophecy (Bandura, 1989). For example, negative self-efficacy beliefs may result in selective attention to everyday forgetfulness. In addition, as a consequence of low memory performance expectations, individuals may not engage in memory-demanding situations or may decrease the effort expended on everyday memory tasks. This may lead to an increased perception of memory failure and further enhance negative memory self-efficacy beliefs. Poor memory self-efficacy can also interfere with the development and use of task strategies, and with motivation and may result in negative affect, especially anxiety. Anxiety, in turn, may influence the extent to which individuals effectively monitor their performance. If individuals are preoccupied with feelings of anxiety, they may be less likely to attend to the challenges of the task, and may miss opportunities in which new strategies should be used (Zimmerman & Schunk, 1989). In general, self-efficacy influences expectations, which, in turn, may influence memory functioning (Perlmutter, Adams & Berry, 1987).

23 Assessment of Metamemory Metamemory has been studied through two approaches. The first involves selfassessment of memory function in specific memory tasks. In the most general sense, it requires individuals to make judgments on future memory performance. The second approach relies heavily on questionnaires. One of the most commonly used questionnaires is the 'Metamemory in Adulthood' (MIA) questionnaire, which consists of 8 scales and is based upon the Hultsch et al. (1988) four dimension theory of metamemory. These scales include Internal and External Strategies, which are knowledge and use of information about one's remembering abilities such that performance in given instances is potentially improved. The second scale, which reflects knowledge of basic memory processes, is Task. Capacity samples beliefs about memory capacities as evidenced by predictive report of performance on given tasks. Beliefs about memory abilities (including one's own) as generally stable, improving, or declining is reflected through the Change scale. The Anxiety scale reflects feelings of stress related to memory performance. The sixth scale, Achievement, involves the importance placed on performing well on memory tasks. The last scale, Locus, reflects perceived personal control over remembering abilities. Studies have shown that three of the MIA scales (Capacity, Change, and Locus) form a Memory Self-Efficacy (MSE) dimension, although the Capacity scale is sometimes used as a marker of MSE (Hertzog & Dixon, 1994). Furthermore, two higher order factors have been ascertained from factor analyses - a memory knowledge factor and a memory self-efficacy factor (Hertzog, Dixon, Schulenberg, & Hultsch, 1987). This

24 means that an individual may have extensive and accurate knowledge about how memory functions, but may also believe that hislher ability to remember in a given context is poor. Metarnernory Studies within the Aging Population The previously defined concepts of metamemory and MSE have been studied within aging populations. Dixon and Hultsch (1983a) found that young adults scored significantly higher on the Task and Capacity subscales than older adults. This suggests that young adults have better knowledge of basic memory processes, and a better perception of their own memory capacities. Similarly, young adults tended to perceive their memory abilities as generally stable, as compared to older adults, who perceived their memory abilities as subject to long-term decline. Finally, younger individuals perceived more personal control over their remembering abilities. Loewen, Shaw and Craik (1990) reported similar findings in that young adults scored higher than older adults on the Capacity scale of the MIA. Luszcz (1993) also found that the young displayed greater memory self-efficacy than did the older subjects. Other work on metamemory has indicated that an anxious attitude towards memory tasks was associated with decreased self-perceived memory capacity and an increased motivation to do well on memory tasks (Jonker, Smits, & Deeg, 1997). In addition, control of memory efficiency was associated with self-perceived memory capacity (Jonker et al., 1997). Hertzog and Dixon (1994) also noted that high selfefficacy limits negative affect, including depression and anxiety, whereas low selfefficacy beliefs lead to increased levels of performance anxiety in the testing situation.

25 Subjective Complaints and Metamemory Ponds and Jolles (1996) described research findings which link subjective reports of memory to metamemory and to actual memory performance in older individuals. Two groups were present in this study; a high-memory-complaint group and a no-complaint group. The high-complaint group, as predicted, reported significantly more memory failures, rated their memory as worse than others, said their memories had declined considerably, and expressed concern about this. The high-complaint group also had a poorer perception of their memory capacities, as measured by the Capacity scale of the MIA, perceived a greater decline in memory over time, and reported more feelings of stress and anxiety related to memory performance. High-complaint individuals did report significantly higher levels of motivation to achieve well in everyday memory tasks, had more knowledge of basic memory processes, and reported using memory strategies more frequently than the low-complaint group. Thus, the presence of memory complaints in older individuals was not directly related to an actual decline in memory. The only factor of the MIA that distinguished the two groups differing in memory complaints was memory self-efficacy beliefs (Ponds & Jolles, 1996). Purpose of the Study: An Exploration of Metarnemory in Individuals with TBI It has previously been demonstrated that individuals with MTBI expect negative repercussions from such an injury. The MTBI literature also supports a definite relation between affective symptoms and subjective rating of cognitive symptoms in individuals with PCS. However, the results of these studies have been varied, with some studies showing differences in depression, anxiety, stress and neuroticism after a MTBI, and others showing no differences. Since metamemory constructs play a large role in

26 defining differences between young and older adults, the purpose of this study is to explore the differences in metamemory between individuals with persistent complaints following MTBI, individuals who have sustained a severe TBI, and a control group of uninjured adults. This study will examine the role of metamemory, including the different components of metamemory, and the interactions of this construct with depression, stress, anxiety and neuroticism. It may be that, as in older adults, negative expectancies with respect to memory function result in lower levels of memory selfefficacy. Other analyses that will be conducted include examining the relationship between metamemory and affective functioning. It is hypothesized that beliefs about memory will mediate the relationship between TBI and affective symptomatology. In other words, sustaining a TBI leads to changes in beliefs about one's memory, which, in turn, leads to negative affective symptoms. This proposed pathway will be examined through mediation analyses. Additionally, it is hypothesized that there will be a stronger relationship between certain scales of the MIA and affective measures for the MTBI and STBI groups than for the control group. This hypothesis will be studied through moderator analyses. Drawing upon the metamemory literature in aging, hypotheses are made in the present study with respect to individuals with MTBI. First, it is predicted that individuals who have sustained a MTBI and who demonstrate persistent complaints 6 months or more post injury, will express the belief that their memory has declined, have decreased MSE, and a decreased feeling of personal control over their remembering abilities. Similar to older adults, individuals with PCS are expected to believe that their memory is

27 not as good as it once was and that it is not something they can change or control. Additionally, it is hypothesized that individuals with MTBI will score lower on the Capacity scale than control subjects. In other words, individuals with MTBI will have a poorer perception of their own memory capacities. The second set of hypotheses examines the relation between different scales of the MIA. It is predicted that increased Anxiety is associated with increased Motivation and decreased Capacity. Thus, individuals who are more anxious will have greater motivation to perform well on memory tasks. However, the increased anxiousness will be associated with decreased beliefs in ability. Awareness and Metamemory in Severe TBI (STBI) Metamemory, and beliefs about memory functioning, within the more severely brain-injured population, may be very different than those in the MTBI population. As previously mentioned, many individuals who have experienced STBI experience limited or partial knowledge of their disabilities or impairments. For example, an individual may report having a memory impairment but be unable to describe adequately the extent of it. To date, insight and awareness within the severely brain-injured population has been most commonly investigated by comparing self-reports with the report of a significant other. Sbordone, Seyranian, & Ruff (1998) found that TBI patients complained of significantly fewer problems than their significant others had observed. More specifically, significant others observed more cognitive and emotional/behavioura1 problems than individuals with severe TBI reported. Fleming and Strong (1999) also found great discrepancies between the reports of individuals with TBI and significant

28 others less than 12 months post-injury, but this discrepancy was not as great at 12 months. However, in contrast to this, Port, Willmott, & Charlton (2002) found that significant others reported impairments as more severe than patients with TBI for executive functioning, but found a high level of agreement in relation to a number of other physical, cognitive, behavioural, emotional and memory functions. Metamemory Within the STBI Population It has previously been discussed that individuals with more severe TBI appear to have a reactionary response to their injury, and as their awareness and insight into their injury increases over time, so do the affective symptoms. Because little research has been conducted into the belief systems of individuals with TBI, perhaps as a result of concerns about their level of insight, few hypotheses can be made as to what differences might be seen between metamemory scales for these individuals and a control group. Some research (e.g., Kennedy, 2001; Kennedy & Yorkston, 2000), however, has indicated that individuals with severe TBI have difficulty in accurately gauging their success on memory tasks and tend to be 'passive' in their approach to new learning activities. It is thus predicted that they will score lower on the Strategy scale of the MIA than individuals with MTBI or control individuals. MSE, however, is expected to be high, as a result of their poor insight and their demonstrated belief that their memory is better than what those around them believe it to be (e.g. Port et al., 2002). In addition, because individuals with STBI often demonstrate reduced motivation, which leads to deficits in activity, interest, initiative, and perseverance, they are predicted to score lower on the Achievement scale of the MIA than those individuals with MTBI or controls.

29 It is recognized that having individuals with STBI complete self-reports on perceptions of memory functioning may be subject to confounding variables, including both limited insight and poor memory functioning. However, the belief systems of individuals with STBI have been rarely studied, and this study may help elucidate such beliefs. Moreover, individuals with STBI in the current study were at least 1 -year postinjury, suggesting that their insight may not be as compromised as those individuals who sustained a brain injury more recently. The proposed study would provide valuable information in that there is little research with respect to metamemory beliefs in the TBI population. Rather, the focus of most investigation has been on emotional functioning in a more general sense. Moreover, as of yet, few studies have examined whether differences exist in the belief systems of individuals with TBI and controls. If individuals have expectations of a negative change in memory and other cognitive functions after a MTBI, and have poor awareness after a severe TBI, metamemory may well be affected. If awareness of memory functioning is negatively affected by a TBI, this may have important implications for memory performance. If memory performance on neuropsychological tests is lower as a result of negative memory beliefs and low memory self-efficacy, everyday aspects of memory may also be influenced by metamemory. Thus, this study provides the groundwork for future work on understanding and perhaps altering negative metamemory beliefs in individuals who have sustained a TBI and continue to be symptomatic. METHODS

30 Participants Participant Recruitment Eighty-four adults ranging in age from 18 to 60 years of age, were recruited for the study. A lower limit of age 18 was imposed for subject selection criteria as research suggests that adolescents and children may respond differently to a brain injury than do adults. An upper age limit of 60 was imposed as individuals beyond this age may experience age-related cognitive decline and memory difficulties. This strict age criteria helped ensure that beliefs about age-related memory changes did not affect responses. Numerous Brain Injury Associations throughout the province of British Columbia were contacted. The associations were initially contacted through a phone call and a subsequent letter and flyer were ed to them (see Appendix A and Appendix B, respectively). Individuals belonging to these associations were given information about the study and the option of participating. Other participants were informed about the study and recruited through the Psychology Clinic at the University of Victoria. Other means of recruiting participants included posting advertisements in local newspapers (see Appendix C for advertisement), and contacting practicing Psychologists, some of whom agreed to inform potential participants of the study. All participants with a TBI had, in some context, expressed concern about persistent cognitive andor emotional symptoms following their injury. Each potential participant was screened for a history of significant neurological, medical, or psychiatric difficulties through completion of a history questionnaire. Both participants with a brain injury and controls were excluded if they had significant pre-injury psychiatric or neurologic history or complaints. This was to

31 ensure that differences between the TBI group and control group reflected beliefs about memory with respect to TBI, and not other factors. The participant sample was divided into two groups; individuals who had sustained a TBI (42 individuals) and adults without a history of brain injury (42 individuals). Many adults without a history of brain injury (the control participants) were friends or family members of individuals who had sustained a brain injury. Previous research has shown that head-injured individuals may not represent a typical crosssection of the population, but may be more representative of people who were previously maladjusted or risk-takers. These factors are likely to affect neuropsychological and social-vocational status of subjects, and friends and family members of individuals with MTBI may be more closely matched on these characteristics. There were no significant differences between the TBI and control groups for age, gender and education. Also, after controlling for gender, there were no differences in further statistical outcomes. Table 1 provides the means and standard deviations for these demographic characteristics. The TBI group was heterogeneous with respect to severity of injury. All participants had sustained closed head injuries and more than half of the accidents were vehicular-related injuries (see Table 2). The injury parameters for the MTBI and STBI groups were fairly varied. Little information was provided with respect to GCS scores as few individuals knew their score. Furthermore, exact length of PTA was difficult to ascertain as many individuals had difficulty pinpointing how long after injury their memory appeared intact. Thus, the major marker of severity of injury was LOC. The STBI group had a range of length of LOC spanning from 0 minutes to 84 days. A LOC

32 Table 1. Demographic Characteristics of TBI and Non-TBI Groups Variable TBI Group (a = 42) Non-TBI Group (a = 42) Gender (% of females) 60% 69% Education (2.31) (2.73)

33 Table 2. Cause of Injuvy MVA Fall Other Hit by Bike Assault Unknown Car Accident 52.4% 9.5% 7.1 % 14.3% 2.4% 4.8% 9.5%

34 of 0 minutes was included in this group as the length of PTA fit into the severe TBI category. The average LOC was hours (20.6 days)(median = 276 hours; SD = ). It is difficult to determine if this average length of LOC is comparable to other studies as few studies cite detailed information pertaining to LOC. However, in the few that do cite this information, it appears that participants in the current study had a longer duration of LOC. Severity of injury was rated according to participants' reported account of medical information, using either Glasgow Coma Scale (GCS) (Teasdale & Jennett, 1974), length of post-traumatic amnesia (PTA) (Russell, 197 I), or length of unconsciousness. These are commonly accepted methods of estimating injury severity in adults. It is important to note that medical records of participants were not consulted so it cannot be definitively concluded that the reported injury parameters are accurate. In this study, injuries were classified as 'mild' when GCS was 13-15, PTA was less than an hour, or loss of consciousness was less than 30 minutes. Injuries were 'moderate' when GCS was 9-12, PTA was greater than an hour, but less than 24 hours, or loss of consciousness was greater than 30 minutes but less than 6 hours. Injuries were classified as 'severe' when GCS was less than 8, PTA was greater than 24 hours, or loss of consciousness was greater than six hours. In the present study, injuries falling in the moderate to severe range were grouped together under the term 'STBI'. It is important to note that Busch and Alpern (1998), upon review of some of the MTBI literature, found that many studies used differing criteria for the definition of a MTBI. The current study, in comparison, is slightly more liberal in the length of LOC (i.e., many studies used a

35 criteria of LOC less than 20 minutes), but is more conservative in the length of PTA (i.e., some studies use a criteria of PTA less than 24 hours). According to participants' recount, 26 individuals in this study had sustained a mild brain injury, however four of these individuals were unsure about injury parameters (i.e., LOC and PTA) and it is possible that they may have sustained a more moderate brain injury. Two other individuals were classified as moderate head injuries but were grouped under STBI. Fourteen individuals had sustained a severe brain injury. CT or MRI findings were unknown for most of the individuals. The mean age at injury was years (SD = 11.81). When broken down into MTBI and STBI, the mean age at injury for the MTBI group was years (SD = 10.63) and the mean age at injury for the STBI group was years (SD = 1135). There was a significant difference between these two groups; t(40) = 2.53, p < On average, participants were 7.95 years (SD = 7.60) post-injury at the time of the current study. However, a significant difference was found on the length of time post-injury between the MTBI and STBI group; t(40) = -2.60, p < The MTBI group was, on average, 5.25 years post-injury (SD = 2.41). The mean length of time post-injury for the STBI group was years (SD = 10.73). This is a longer length of time post-injury for both MTBI and STBI groups than what is typical for other research studies. Nine individuals had sustained multiple brain injuries. Three of these individuals were in the MTBI group, and the other six were in the STBI group.

36 Measures Background Information Participants were asked to complete a questionnaire about demographics, the details of the brain injury, and to note any significant psychiatric/medical/alcohol problems (see Appendix D). Metamemory Questionnaire The Metamemory in Adulthood questionnaire (Dixon and Hultsch, 1983b) (see Appendix E) was administered to provide a measure of adults' self-perceptions of their everyday memory functioning. The questionnaire consists of 108 items and eight subscales. However, the questionnaire was adapted for the current study in order to better fit with a brain-injured population. The adapted questionnaire contains 105 items, in a 5-point Likert response format such that a score of 1 indicates 'agree strongly' and a score of 5 indicates 'disagree strongly'. The revised questionnaire also consists of eight scales; Internal Strategies, External Strategies, Task, Capacity, Change, Anxiety, Achievement, and Locus. A high value on each of the scales indicates high use of strategies, high memory knowledge, high capacity, belief in stability of memory, high anxiety, high achievement, and high internal locus of control, respectively. Changes to questions and deleted items are listed in Appendix F. Prior work with the original questionnaire suggests that the scales are internally consistent (Cronbach's alpha range =.61 to.92).

37 Postconcussion Syndrome Checklist This checklist (Gouvier, Cubic, Jones, Brantley, & Cutlip, 1992) (Appendix G) allows subjects to rate the frequency, intensity, and duration of each of nine separate symptoms most commonly associated with PCS; headaches, dizziness, irritability, memory problems, difficulty concentrating, visual disturbances, aggravation by noise, judgment problems, and anxiety. Subjects are told to rate the frequency, intensity, and duration of each symptom, over the past week, on a Likert-type scale from 1 'not at all' to 5 'all the time', 'crippling', or 'constant'. Four symptom scores are derived for each part; frequency total, intensity total, duration total, and total scores across dimensions. Perceived Stress Scale This scale (Cohen, Kamarck, & Mermelstein, 1983) (Appendix H) contains 14 items that assess the level of perceived stress in a day. However, the questionnaire was modified so as to assess the level of stress during the previous month (e.g., in the last month how often have you been upset because of something that happened to you unexpectedly?) which are rated from 1 (never) to 4 (very often), yielding an overall measure of level of perceived stress. Beck Depression Inventory-11 The BDI-I1 (Beck, Steer, & Brown, 1996) (Appendix I) is a face-valid self-report 21 -item measure of the common symptoms of depression and the severity of symptoms. The individual checks 21 four-choice statements for the choice most appropriate to him or her. A high value on this scale indicates a greater degree of depression. The BDI-I1