This article was downloaded by: [UQ Library] On: 01 November 2014, At: 17:49 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Clinical and Experimental Neuropsychology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/ncen19

Metamemory in multiple sclerosis William W. Beatty

a b

& Nancy Monson

a

a

Clinical Neuroscience Research Program Neuropsychiatric Research Institute , Fargo, ND b

North Dakota State University , Published online: 04 Jan 2008.

To cite this article: William W. Beatty & Nancy Monson (1991) Metamemory in multiple sclerosis, Journal of Clinical and Experimental Neuropsychology, 13:2, 309-327, DOI: 10.1080/01688639108401046 To link to this article: http://dx.doi.org/10.1080/01688639108401046

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sublicensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Journal of Clinical and Experimental Neuropsychology 1991,Vol. 13, NO.2, pp. 309-327

0168-8634/91/1302-03O9$3.O0 8 Swets & Zeitlinger

Metamemory in Multiple Sclerosis* William W. Beatty1S2 and Nancy Monson' Clinical Neuroscience Research Program Neuropsychiatric Research Institute Fargo, ND. North Dakota State University

Downloaded by [UQ Library] at 17:49 01 November 2014

ABSTRACT MS patients and age- and education-matched normal controls were administered several laboratory tests of metamemory and a questionnaire designed to measure subjects' capability to appraise their ability to remember events that might occur in everyday life. On laboratory tasks involving newly acquired information, MS patients with poor recognition memory abilities or poor performance on the Wisconsin Card Sorting Test (WCST) exhibited impairments on one test of metamemory; patients with deficits in both recognition and on the WCST showed more extensive impairments in metamemory. In contrast to their performance on tests involving newly acquired information, all groups of MS patients predicted their ability to recognize answers to general information questions that they could not recall as accurately as controls, and, like controls, they also searched their memories longer for answers to items that they believed they would recognize. In general, the results support the hypothesis that both trace-access arid inferential mechanisms, which are thought to involve the prefrontal cortex, contribute to metamemory, but the nature of the memory task importantly influences the accuracy of metamemory, as well. Results from the questionnaire indicated that many MS patients with demonstrable memory deficits do not acknowledge their memory difficulties. Hence, patient self-reports about memory are likely to be unreliable sources of information for clinical purposes.

Metamemory is a broad concept which refers t o an individual's knowledge about his or her own memory. Metamemory includes knowledge of task variables such as the difficulty of t h e subject matter, of mnemonic strategies, and of one's own ability as a learner. Early studies b y Hart (1965) required college students to predict whether o r not they would recognize answers to questions o f general knowledge that they could not recall. On the subsequent recognition test per-

* We thank Dr. Donald Goodkin for allowing us to study his patients, Dr. Arthur Shimamura for the use of the Sentence Memory Test, Dr. Susan McGlynn for the use of the Memory Questionnaire, Dr. Ruth Maki for the use of the General Information questions, and all of the patients who served in the study. Please address correspondence and reprint requests to: William W. Beatty. Ph.D., Alcohol Research Center, University of Oklahoma Health Sciences Center, Department of Psychiatry and Behavioral Sciences, PO Box 26901, Oklahoma City, OK 731 90, USA Accepted for publication: June 28, 1990.

Downloaded by [UQ Library] at 17:49 01 November 2014

310

WILLIAM W.BEA’ITY AND NANCY MONSON

formance was more accurate for questions that elicited a positive feeling-ofknowing (FOK) than for items that elicited a negative FOK. Subsequent research with normal subjects has replicated Hart’s results with questions that tap previously established general knowledge (Gruneberg & Monks, 1974; Nelson, Gerler, & Narens, 1984) and extended Hart’s findings to various measures of anterograde memory (Blake, 1973; Hart, 1967; Nelson, Leonesio, Shimamura, Landwehr, & Narens, 1982; Schacter, 1983). In addition, normal subjects can predict their memory span with considerable accuracy (Kelly, Scholnick, Travers, & Johnson 1976; Yussen & Levy, 1975). Under certain conditions, FOK ratings also predict the amount of time normal subjects will search their memories for the answers to items they are ultimately unable to recall. That is, subjects generally search longer for the answers to questions that evoke a high FOK rating than for answers to items associated with a low FOK. This positive relationship between FOK judgments and memory search time has been observed when the memory test involved general information questions (Lachman, Lachman, & Thronesbery, 1979; Nelson & Narens, 1980), but not for paired-associated learning (Nelson et al., 1982). The status of metamemory in neurological patients has not been studied extensively, but the available research indicates that amnesic patients with alcoholic Korsakoff‘s syndrome (AK) exhibit marked deficits in knowledge of memory strategies (Hirst, 1982), in predicting their performance on free recall (Bauer, Kyaw, & Kilbey, 1984), and in forecasting their performance on recognition tests of sentence memory and general information (Shimamura & Squire, 1986). In patients with amnesias of other etiologies (ECT. hypoxia, ischemia, trauma), however, Shimamura and Squire (1986) observed normally accurate FOK judgments even though these patients were as severely impaired in recalling and recognizing items on the anterograde (sentence) memory test as the AK patients. Since AK patients exhibit both neuroradiological and neuropsychological evidence of frontal lobe pathology and dysfunction (Shimamura, Jernigan, & Squire, 1988), which is not present in patients with amnesias of other etiologies, Shimamura and Squire (1986) suggested that the impairments in metamemory of the AK patients might be related to their additional disturbances in frontal lobe functioning. To test this idea, Janowsky, Shimamura, and Squire (1989) studied the performance of seven patients with frontal lobe lesions on the same sentence memory and general information tasks used by Shimamura and Squire (1986). The frontal patients were not amnesic or demented, but performed poorly on the Wisconsin Card Sorting Test (WCST) and on verbal fluency tests, classic measures of frontal lobe dysfunction. When required to make FOK judgments for newly learned material (i.e., the sentence memory test), the frontal patients were less accurate in predicting subsequent recognition performance than were controls. However, this effect was evident only when memory was weakened by imposing a delay of 1-3 days. With a shorter delay the frontal patients remembered the missing words in the sentences as well as the controls and their FOK judgments were equally accurate. The same pattern was evident on the general information

Downloaded by [UQ Library] at 17:49 01 November 2014

METAMEMORY IN MS

31 1

test, on which the frontal patients showed normal recall of factual information and normal ability to predict their performance on recognition tests for items they could not recall. Taken together, these findings suggest that impairments in metamemory for newly learned information occur only when memory is somewhat degraded and frontal lobe functioning is impaired. In apparent contradiction to this conclusion are the findings of Prevey, Delaney, and Mattson (1988) who reported that patients with temporal lobe epilepsy were less than normally accurate in estimating their performance on tests of short-term memory for newly acquired verbal or visual information. The same patients also made less accurate FOK judgments for items that they could not recall on a general information test and they recalled fewer facts on the same test. These findings may indicate that metamemory is impaired in some amnesias that do not involve structural damage to the frontal lobes. However, Herman, Wyler, and Ritchey (1988) have reported abnormally high perseverative error rates on the WCST in patients with temporal lobe seizures. Hence, the deficits in metamemory observed by Prevey et al. (1988) may have arisen because their memory-impaired temporal lobe epileptics also suffered frontal lobe dysfunction. Studies of metamemory in patients with diseases that primarily affect subcortical structures (e.g., Huntington’s disease (HD), Parkinson’s disease (PD), multiple sclerosis (MS)) might help to clarify some of these issues, since these patients often perform poorly on tests of memory or frontal lobe function (e.g., the WCST) and these deficits may occur alone or in combination. In both of the published studies of metamemory in subcortical disease (Brandt, 1985 with HD patients, Coulter, 1989 with PD patients) the patients were able to make accurate FOK judgments of their ability to recognize items that they could not recall. However, in both studies only general information was studied and metamemory for this type of well established knowledge may be less sensitive to disturbances in frontal lobe functioning than is metamemory for newly acquired information (Janowsky et al., 1989). Further, Brandt (1985) found that HD patients did not exhibit the usual positive correlation between FOK and memory search time for items that ultimately could not be recalled, which he attributed to disruption in frontal lobe functioning. In an effort to clarify the contributions of memory impairment and frontal lobe dysfunction to disruption of normal metamemory, we studied patients with multiple sclerosis (MS). When considered as a group, MS patients perform more poorly than controls on various measures of anterograde and remote memory (Beatty, Goodkin, Monson, Beatty, & Hertsgaard, 1988; Beatty, Goodkin, Beatty, & Monson, 1989; Beatty, Goodkin, Monson, & Beatty, 1989) as well as on tests that are sensitive to acquired lesions of the frontal lobes such as the WCST and verbal fluency (Beatty, Goodkin, Beatty, & Monson, 1989; Beatty, Goodkin, Monson, & Beatty, 1989). However, the cognitive changes in MS are extremely variable: some patients exhibit no cognitive deficits at all; others show focal disturbances in some cognitive domains but not in others, while a few patients

312

WILLIAM W.B E A T AND NANCY MONSON

Downloaded by [UQ Library] at 17:49 01 November 2014

exhibit more global cognitive impairments. We attempted to exploit the heterogeneity of cognitive performance of MS patients by forming groups of patients that were characterized by deficits in recognition memory, poor WCST perfonnance, the combination of deficits on both measures, or normal performance on both measures. Since the literature indicates that task variables may influence the conclusions about the effects of brain lesions on metamemory, we studied subjects’ ability to forecast their performance on three different laboratory tasks. Finally, since patients’ knowledge of their own memory ability has important implications for rehabilitation and for their adjustmentto their disease, we also studied subjects’ estimation of their ability to remember important events and obligations that might occur in their everyday lives.

METHOD

Subjects A total of 45 patients (16 male, 29 female) with clinically definite multiple sclerosis (Poser et al., 1983) participated in the study. Eighteen patients carried the current diagnosis of chronic progressive (CP) and 10 were diagnosed as relapsing remitting (RR).These disease types were based on the results of the neurological examination closest in time to the neuropsychological testing and were assigned according to operational definitions that have been published elsewhere (Beatty et al., 1988; Beatty, Goodkin. Beatty, & Monson, 1989; Beatty. Goodkin, Monson, & Beatty, 1989). Seventeen patients had shown no significant change in neurological status during the two years preceding this study and were considered to have a stable (S)disease course. Level of disability was taken as the score on the Ambulation Index (AI, Hauser et al., 1983), obtained at the time of the current neuropsychological tests. In this population, scores on the A1 are highly correlated with scores on Kurtzke’s (1983) Expanded Disability Status Scale ( r = 0.96). At the time of testing, 10 patients were receiving ongoing treatment with muscle relaxants, 4 patients were receiving antidepressants, 7 patients were receiving the immunosuppressant, azathioprine. and 7 patients were receiving anticholinergic medications for bladder control. No patient was tested while receiving corticotropin or steroid treatment for an exacerbation. Normal controls (8 male, 14 female) were recruited from the community and paid for their participation. All subjects gave written informed consent after a thorough explanation of the procedures. Patients or controls with a history of alcohol or drug abuse, serious head injury, learning disability,recent or complicated heart attack, uncontrolled hypertension or metabolic disease, central nervous system diseases (other than MS for the patients) or major psychiatric illness were excluded.

Materials and Procedures Background Measures - As a measure of global cognitive functioning, all subjects were given the Screening Examination for Cognitive Impairment in Multiple Sclerosis (SECIMS). This test is a modification of the Mini-Mental State Exam (MMSE, Folstein, Folstein, & McHugh, 1975). The latter has proved insensitive for identifyhgzognitive impairment in MS patients (Beatty & Goodkin, 1990). The SECIMS is identical to the MMSE except that the number of to-be-remembered words is increased from 3 to 7 and the two confrontation naming items are replaced by a 15-item version of the Boston Naming Test (Kaplan. Goodglass, & Weintraub. 1983). A sample of 52 normal controls, closely matched in terms of average age, education and gender distribution to the population of MS patients

Downloaded by [UQ Library] at 17:49 01 November 2014

METAMEMORY IN MS

313

served by our clinic averaged 48.3 f 1.6 (out of 51) on the SECIMS. A score below 46 (the 5th centile for controls) on this test is strongly suggestive of global cognitive impairment (Goodkin, Beatty, & Monson, 1990) as defined by performance on an extensive neuropsychological battery. All of the patients and 18 of the controls had served in previous studies of cognitive functioning in MS (Beatty et al., 1988; Beatty, Goodkin, Beatty, & Monson, 1989; Beatty, Goodkin, Monson, & Beatty, 1989). During the course of these studies their performance on the WCST (Heaton, 1981) and a delayed yes-no recognition test of verbal memory (Beatty et al., 1988, Beatty, Goodkin, Monson. & Beatty, 1989) had been measured. The 4 control subjects not previously studied received the WCST at the time of the present tests, but they did not receive the recognition memory test. Our previous studies indicate that scores below 86% correct on the recognition memory test and below 4 categories on the WCST fall below the 5th centile for normal controls. These cut-off scores were used to classify patients into four subgroups that represented all possible combinations of normal and impaired (i.e., “low”) performance on the recognition memory test and the WCST. As a second measure, known to be sensitive to frontal lobe lesions (Benton, 1968). we included a word fluency test. On this test subjects were allowed 60 s to generate (orally) as many different words as possible that began with the letter “F”. Proper names were not allowed. The procedure was then repeated with the letter “A” and “S”. In a previous study (Beatty, Goodkin, Beatty, & Monson, 1989) with chronic progressive MS patients we observed that performance on the FAS test was only weakly correlated with performance on the WCST. Because lesions anywhere in the language-dominant hemisphere may affect scores on verbal fluency tests we felt that performance on the WCST was a ‘purer” measure of frontal lobe dysfunction. Hence performance on the word fluency test was not considered in assigning patients to subgroups. Metnrnemory Measures - Four measures of metamemory were employed: the sentence memory test developed by Shimamura and Squire (1986), predicting the number of words that would be recalled on a verbal learning task, a general information test modified from the one devised by Nelson and Narens (1980), and a memory questionnaire modified from the one developed by McGlynn, Schacter, and Glisky (1989).

Sentence Memory Test During the input phase of the sentence memory test subjects were presented with 38 index cards on which a sentence (e.g. “Patty’s garden was full of marigolds”) was printed. They were told to try to remember each sentence. During presentation the examiner read each sentence aloud while the subject studied it. Two seconds elapsed between reading a sentence and presentation of the next sentence. The first 2 sentences were fillers designed to reduce primacy effects. Of the remaining 36 cards, 12 contained sentences that were presented only once and 12 contained sentences that were presented twice. Subjects were told that some sentences would be repeated while others would be presented only once. One set of sentences was repeated twice to half of the subjects in each group; for the remaining subjects the other set of sentences was repeated twice. Following a delay (60-90 min long for most subjects) filled with other memory tests, recall was tested by presenting the 24 sentences studied earlier along with 8 sentences that were not presented. Each sentence was missing a key word which was indicated by a blank underlined space (e.g.. “Patty’s garden was full of ”). Subjects were told to read the sentence aloud and say the missing word if they could remember it. If they could not remember the missing word or thought that they had never seen the sentence, they were to complete the sentence by saying a word that made sense in the context of the sentence. The recall phase was tape recorded for later analysis of latencies. Next, for all sentences that they did not recall correctly, subjects were asked to predict how strongly they felt that they could recognize the missing word in each sentence if they

314

WILLIAM W.BEAT” AND NANCY MONSON

Downloaded by [UQ Library] at 17:49 01 November 2014

were shown the comct word among some other choices. They made their ratings using the following scale: 1 = pretty sure I would, 2 = maybe, about 50-50 chance, 3 = not very likely, but I might, and 4 = pure guess. After subjects completed their FOK ratings, recognition memory was tested for all 32 sentences by presenting each sentence with the underlined blank space indicating the missing word along with 8 alternatives (the correct word and 7 plausible distractors). Subjects were told to select the correct word if they knew it and to guess if they were unsure or thought they had not seen the complete sentence before on the test. Free Recall Test The free recall test employed two 14-word lists known from previous work (Beatty, Butters, & Janowsky, 1986; TrOster, Beatty, Staton, & Rorabaugh, 1989)to be of equivalent difficulty. Before the first trial, subjects were told that they would be read a list of 14 words and askedto recall them immediately after presentation. They were also told that the average person gets 4 to 10 correct and asked to predict how many words they would recall. Then the first word list was read aloud to the subjects and they recalled as many words as possible in any order. The examiner told the subjects how many words they had recalled correctly. After a brief delay, during which the letter fluency test was presented (see above), the entire procedure was repeated using the second word list. Before they predicted how many words they would recall, subjects were told that a different 14-word list would be presented. Two lists were tested in this way so that subjects’ ability to predict their own recall could be determined under two conditions: first, when they had only vague information about the performance of the “average” person and second. when they had explicit and recent knowledge about their own performance. General Information Test For the general information test, 88 questions were selected from the battery constructed by Nelson and Narens (1980) and printed on index cards. Only questions on which the recall accuracy by males and females differed by no more than 1596 were used. Within the limits of this constraint, the range of difficulty (measured in terms of the percentage of college students who correctly recalled each item) of 83 of the questions varied from 19 to over 90% correct based on norms published by Nelson and Narens (1980).The remaining 5 questions were recalled correctly by fewer than 2% of the subjects in the same normative sample. A single order of presentation of the questions was used for all subjects. The difficulty of the questions in the sequence varied in a quasi-random manner with a slight tendency for easier questions to be overrepresented in the first half of the test. During the recall phase of the test, subjects read the question aloud and then attempted to recall the correct answer. They were allowed a maximum of 30 s to answer each question and responses were tape recorded for later analysis. Testing continued until the subject made 25 errors (i.e.. incorrect answers or don’t know responses) or answered all 88 questions. Four control subjects and three MS patients answered 88 questions without making 25 errors. Hence, the data are slightly biased against fmding differences between controls and MS patients in terms of their recall of general knowledge. Next, subjects made FOK ratings for all questions they had not recalled correctly. Using the same 4-point scale employed on the sentence memory test, they were asked to indicate how certain they were that they would recognize the correct answer to each question if given some choices. After completing the FOK ratings recognition memory was tested. Subjects were shown a card on which the question and 8 alternatives (the correct answer and 7 plausible distractors) were printed. They were encouraged to guess if they didn’t know the correct answer.

METAMEMORY IN MS

315

Downloaded by [UQ Library] at 17:49 01 November 2014

Memory Questionnaire There were three parts to the memory questionnaire. On the fiist part, subjects were asked to think of a friend of about the same age, education and gender who did not have MS or any other chronic disease. Using their friend as a comparison standard, they were asked 18 questions of the general form “How much trouble are you having with ___ ?” Ratings were made on 7-point Likert type scales (0 = a great deal of trouble, 6 = no trouble at all). One question concerned motor functions (walking, lifting), two questions concerned sensory functions (vision, hearing), two questions Concerned attention and concentration, three questions concerned language (speaking, understanding, reading), one question required a global estimate of memory, one question concerned remote autobiographical memory (remembering what happened when you were a child), and eight questions Concerned various specific aspects of recent memory (e.g.. remembering where you put things, names of people several minutes after being introduced, conversations with family members, etc.) On the second part of the questionnaire subjects were read 10 hypothetical scenarios describing problems in memory that occur in everyday life (e.g., remembering to call the doctor to make an appointment). For each scenario subjects were asked to estimate how long they would remember the critical aspect of the problem. Alternatives and the point values assigned to each in the scoring were: less than 10 min = 0, 10 min = 1, 1 hr = 2, 1 day = 3 , l week or longer = 4. Total points on the 10 scenarios was the dependent measure on this part of the questionnaire. The third part of the questionnaire was intended to measure subjects’ knowledge of important variables that affect memory. They were asked to imagine that they were to remember 10 items that differed in some way (e.g., 10 concrete words vs 10 abstract words). Then they were asked to state which of the two conditions would be easier to remember or if the two conditions would be equally easy to remember. Each of the 8 questions was scored correct or incorrect; a subject’s score was the total number of questions answered correctly. Analysis of the FOK data - In many previous studies of the FOK phenomenon subjects were required to rank order (from 1-n) the probability that they would recognize correct answers to missed questions. Although this procedure has anumber of desirable psychometric properties (Nelson, 1984), pilot studies with undergraduates indicated that these students found the rank order task extremely difficult and frustrating even when only 12 questions had to be ranked. Accordingly, we felt that the task of rank ordering 25 or more questions would prove overwhelming to many of the patients and we elected to employ the 4-point rating scale described above for the FOK judgments. As might be expected, many patients and controls did not utilize all 4 scale values in making their FOK ratings. To analyze the data, for each subject we divided responses into two categories (high and low FOK). This division was made so that the number of items that contributed to the high and low FOK categories was as similar as possible. In the rare situation in which the items could be grouped in two different ways, the division was made between scale points 2 and 3. For example, suppose that the number of items assigned various FOK ratings was as follows: FOK rating “1”=0, “2”=6, “3”=13, “4”=6. In this case the high FOK category would include the 6 items assigned a FOK rating of “2” and the low FOK category would include the remaining 19 items. If, however, the number of questions assigned FOK ratings was: FOK rating “l”=O, “2”=5. “3”=13, “4”=7, then the high FOK category would include the 18 items rated “2” or “3” and the low FOK category would include the 7 items rated “4”. Two types of analyses of the FOK data were conducted. In the first analysis the percent correct for high and low FOK categories was computed for each subject. In the second analysis gamma coefficients (Nelson, 1984) were computed for each subject according to the following formula:

316

WILLIAM W.BEAT" AND NANCY MONSON

Gamma = #concordances - # discordances #concordances + # discordances where a concordance = a correctly recognized item with a high FOK rating or an incorrectly recognized item with a low FOK rating. The gamma coefficient has the same limits as Pearson r and can be interpreted in a similar fashion (i.e.. if gamma = 1.0, then the subject predicted his recognition performance with maximal accuracy). Unlike Pearson r, gamma requires only an ordinal level of measurement, and, in addition, possesses a number of other desirable psychometric properties for use in metamemory research (Nelson,

Downloaded by [UQ Library] at 17:49 01 November 2014

1984).

Separate analyses of each type were performed for the sentence memory and general information tests. As a check on the possibility of systematic differences among groups in the use of the FOK scale we computed the mean FOK rating for each subject. Mean FOK values were determined separately for the sentence memory and general information tests. Response latencies for recall were measured from the tape recording to the nearest second with a digital stopwatch with the constraint that the minimum latency recorded was 1 s. The mean latency of correct responses and for incorrect responses assigned high or low FOK ratings as described above was determined. These latencies were computed separately for the sentence memory and general information tests. For items that could not be recalled correctly we did not differentiate between incorrect and no response answers. Two observers measured the latencies for the first few subjects tested. Since the interrater reliability was high ( r = 0.99). only a single observer measured the latencies for subsequent subjects. RESULTS

Background Variables Table 1 summarizes the demographic, clinical, and psychometric characteristics of the control and MS patient groups. There were no significant differences among the five groups on age or education (Fs < 1.29), or among the four patient groups on level of disability as measured by the A1 (F < 1). Likewise, the proportion of patients of different disease types did not differ significantly among groups (x2 (6) = 5.13,p > SO). Performance on the two variables used to divide patients into groups indicates that the intended group characteristics were achieved. Thus, controls and patients in the two groups with normal delayed recognition memory (Groups NM, NW and NM, LW) achieved high and similar recognition memory scores while patients in the low memory groups (Groups LM, NW and LM, LW) attained much lower recognition memory scores, but these latter two groups did not differ significantly. Likewise, the three groups with normal performance on the WCST (controls and Groups NM, NW and LM, NW) were well matched in terms of the number of categories achieved, the number of perseverative errors and the number of trials needed to master the first concept, whereas patients in the low WCST condition (i.e., Groups, NM, LW and LM, LW) performed equally poorly on all three measures derived from the WCST. As expected, there were significant differences among groups on the SECIMS, a measure of overall mental status (F (4. 62) = 12.97. p < .OOl>. Subsequent

317

METAMEMORY IN MS

Table 1. Demographic, Clinical and Psychomemc CharaLTeristics of Control and MS Patient Groups: Mean (SD) Controls

MS NM, NW

Patients LM, NW

Nh4, LW

22

18

8

10

40.8 (9.0) 14.7 (2.4) 3.2 (2.6)

44.8 (10.6) 14.4 (2.8) 4.5 (3.6)

Downloaded by [UQ Library] at 17:49 01 November 2014

N Age ( Y d Education (yr) A1 Diagnostic Type CP/RR/S Recognition Memory (% Cor) WCST No. Categories No. Persev. Errors Trials to First Category SECIMS FAS No. Words

49.0 (16.7) 14.3 (2.0)

Groups' LM, LW 9

51.0 (16.7) 46.4 (9.9) 13.7 (1.9) 14.1 (2.4) 3.6 (2.9) 5.0 (3.5)

5/3/10

31312

96.0 (4.6)2

95.6 (4.0)

83.0 (4.2)

96.4 (3.8) 80.2 (5.4)

5.7 (1.1) 10.3 (7.1)

5.9 (0.5) 10.1 (5.9)

5.9 (0.4) 13.8 (5.7)

1.8 (1.0) 1.6 (1.2) 44.9 (19.4) 43.3 (24.5)

11.4 (1.1) 48.7 (1.4)

12.4 (2.6) 48.0 (2.6)

16.1 (10.8) 45.9 (2.9)

40.0 (43.4) 40.8 (49.9) 45.4 (3.7) 41.9 (3.1)

44.9 (1 I .8)

35.7 (7.0) 31.6 (16.0)

33.4 (12.0) 21.1 (8.2)

1. NM = Normal Memory, NW =Normal WCST, LM = Low Memory, LW = LOW WCST 2. N = 18.

analyses demonstrated that controls differed from all patient groups except the NM, NW group ( t s > 2.70, p s < .02). Further, patients in the LM, LW group attained lower SECIMS scores than subjects in any of the other groups ( t s > 2.23, p s < .05). No other comparisons reached statistical significance. Analysis of performance on the FAS test also revealed significant differences among groups ( F (4,62) = 8.18, p < .OOl). Subsequent tests showed that controls generated more correct words than any of the patient groups (rs > 2.13, p s < .05). In addition, patients in the LM, LW group differed significantly from patients in the NM, NW group and from patients in the NM, LW group (ts > 2.56, p s < .05). N o other comparisons reached statistical significance.

General Information Table 2 summarizes performance on the general information test. Omnibus ANOVAs indicated significant differences among groups in the number of items correctly recalled (F (4,62) = 6.81, p < .OOl), but not in the percentage of nonrecalled items that were correctly recognized. Subsequent analyses indicated that controls and patients in Group NM, NW recalled significantly more items than patients in the other three groups ( t s > 2.12, p s < .05). No other significant differences between groups were noted.

318

WILLIAM W.BEATIY AND NANCY MONSON

Table 2. Performance on the General Information Test by Control and MS Patient Groups: Mean (SD) MS No Recalled % Recognized % Recognized, High FOK % Recognized

Downloaded by [UQ Library] at 17:49 01 November 2014

Low FOK

Gamma FOK Rating Latency (sec) Correctly Recalled High FOK Low FOK

Patients LM,NW

Groups * NM.LW LM.LW

Controls

NM.NW

41.7 (19.2) 57.8 (13.9)

38.5 (19.3) 17.3 (6.3) 52.9 (13.7) 47.5 (7.5)

24.7 (15.7) 15.7 (8.4) 44.8 (10.3) 50.1 (15.1)

68.7 (18.1)

61.6 (16.5) 59.8 (9.6)

57.2 (13.6) 56.9 (17.2)

42.9 (20.1) 40.1 (16.5) 36.1 (11.9) 32.2 (10.2) 44.8 (16.2) 0.26 (0.29) 0.20 (0.18) 0.21 (0.12) 0.26 (0.17) 0.18 (0.13) 2.79 (0.3) 2.69 (0.9) 2.15 (0.4) 2.20 (0.5) 2.68 (0.4) 2.6 (1.1) 11.4 (5.3) 6.6 (3.4)

3.1 (1.2) 3.0 (0.8) 12.8 (4.2) 9.4 (4.6) 7.2 (3.8) 4.9 (2.1)

2.9 (1.2) 4.1 (2.3) 8.5 (2.7) 8.8 (3.0) 4.9 (2.2) 4.4 (3.0)

1. NM = Normal Memory, NW = Normal WCST, LM = Low Memory, LW = LOW WCST

Despite differences in the recall of general information items, subjects in each of the five groups were equally able to forecast their ability to recognize the correct answers to items that they could not recall. A 5 (Groups) X 2 (levels of FOK) ANOVA on the percentage of items correctly recognized revealed only a significant effect of FOK (F (1,62) = 69.55, p c .OOl); the Groups effect and the Groups X FOK interaction were not statistically significant ( F s < 1.42). Likewise, mean gamma coefficients for each of the groups were positive and significantly greater than 0 (ps < .05), but there were no significant differences among the five groups (F < 1) in the magnitude of the gamma coefficients. Analysis of the mean FOK ratings revealed significant differences among groups ( F (4,62) = 5.12, p < .002), which arose because controls and patients in Group NM, NW tended to assign lower scale values (i.e., greater probability of accurate recognition) than patients in the other three groups. Pairwise comparisons indicated that the differences between controls or patients in the NM, NW group and patients in the NM, LW or LM, NW groups were significant ( t s > 2.75, p s c .02). Average ratings by patients in the LM, LW group were highly variable and not significantly different from those of any other group. No other pairwise comparisons attained statistical significance. The pattern of differences among groups in terms of the average value of their FOK ratings closely parallels differences among groups in the number of general information items that were correctly recalled. Hence, it seems likely that the average FOK ratings reflect subjects’ overall appraisal of their fund of general information.

319

METAMEMORY IN MS

Downloaded by [UQ Library] at 17:49 01 November 2014

Examination of the latency data suggests that subjects searched their memories longer for the answers to items that they believed that they knew. A 5 (Groups) X 2 (levels of FOK) ANOVA confirmed this impression, revealing a highly significant main effect of FOK ( F (1,62) = 119.77, p < .OOl) and a modest main effect of Groups (F 4,62) = 2.58, p < .05), but the Groups X FOK interaction was not significant ( F < 1). Additional analyses revealed significant effects of FOK on the memory search times of each of the five groups ( F s > 6.66, p s < .05). There was no significant difference among groups in the latency scores for items that were correctly recalled. Free Recall Table 3 shows the number of words that subjects predicted they would recall as well as the number of words actually recalled for both of the trials. Inspection of the data suggests that the accuracy of prediction by all groups improved from Trial 1 to Trial 2 in the sense that the difference between predicted and actual performance was smaller in magnitude on the second trial, and, further, that patients in the LM, LW group consistently overestimated their performance. A 5 (Groups) X 2 (Trials) ANOVA on the difference between predicted and actual performance confirmed both of these apparent trends, revealing significant main effects of Groups ( F (4,62) = 2.65, p < .05) and Trials ( F (1,62) = 8.20, p < .01) and the absence of significant Groups X Trials interaction ( F = 1.21). Subsequent analyses demonstrated that the discrepancy between predicted and actual recall was greater in magnitude for patients in the LM, LW group than for any other group (ts > 2.78, p s < .02). No other differences were significant in pairwise tests. Analysis of the number of words actually recalled revealed a significant effect of Groups ( F (4,62) = 8.19, p < .001), but neither the Trials effect ( F = 1.63)

Table 3. Performance on the Free Recall Task by Control and MS Patient Groups: Mean (SD)

Trial 1 Predicted Actual Trial 2

Predicted Actual

Predicted-Actual, Overall

Groups* NM, LW LM, LW

MS

Patients

Controls

NM, NW

LM, NW

7.0 (1.7) 7.3 (2.2)

6.8 (1.1) 5.9 (1.6)

5.3 (1.4) 4.8 (1.6)

5.8 (1.9) 5.2 (1.9)

6.2 (1.5) 3.9 (1.9)

7.2 (1.6) 7.4 (2.2)

6.2 (1.1) 6.5 (1.6)

4.6 (1.4) 4.9 (1.8)

5.4 (1.9) 5.8 (1.7)

5.1 (1.6) 3.7 (1.4)

4 . 4 (4.3)

0.6 (2.6)

0.3 (3.1)

0.2 (2.4)

3.8 (2.9)

1. NM = Normal Memory, NW = Normal WCST, LM = Low Memory, LW = LOW WCST

Downloaded by [UQ Library] at 17:49 01 November 2014

320

WILLIAM W.BFATIY AND NANCY MONSON

nor the Groups X Trials interaction (F < 1) was significant. Subsequent comparisons showed that controls recalled more words than subjects in any of the patient groups (ts > 2 . 1 2 , < ~ .05). ~ Further, patients in the NM, NW group recalled more words than patients in the LM, NW or LM, LW groups ( t s > 2 . 1 4 , ~ < s .05). Finally, patients in the NM, LW group recalled more words than patients in the LM, LW group (t (16) = 2.39, p < .05). Since patients in the LM, LW group attained lower scores on the SECIMS than patients in the other groups, their inability to forecast their performance on the free recall task might have arisen from a general decline in cognitive performance rather than from a specific constellation of deficits in recognition memory and on the WCST. To examine this possibility, we selected a group of patients who scored below 46 on the SECIMS but did not display impairment on both the recognition memory and WCST tasks. Four of these patients were drawn from Group NM, LW, two were from Group LM, NW, and two were from Group NM, NW. This subgroup of 8 patients averaged 42.0 f 2.2 on the SECIMS, which was not significantly different from the performance of patients in the LM, LW group. Over the two recall trials their predicted recall averaged 9.9 f 2.8 words and they actually recalled 10.0 f 2.3 words. Thus, this subgroup of patients recalled only slightly but not significantly more words than patients in the LM, LW group (t = 1.91), but they predicted their performance more accurately. The overall predicted-actual discrepancy score (-0.13 f 2.95) for these patients was significantly lower than the comparable score for patients in the LM, LW group (t (15) = 2.77, p < .02). The results of this analysis indicate that the striking tendency of patients in the LM, LW group to overestimate their free recall performance cannot be attributed to their generalizedcognitive impairment as measured by the SECIMS. Instead the specific combination of poor memory and impaired WCST performance seems to have been responsible for this metamemory deficit.

Sentence Memory As shown in Table 4, there were significant differences among groups in the number of sentences correctly recalled (F (4, 62) = 10.78, p < .001) and recognized (F (4,62) = 10.87,p < .001). Subsequent analyses showed that both controls and patients in the NM, NW group attained higher memory scores on both measures than did patients in the other three groups ( t s > 2.12, p s < .05). In addition, patients in the NM, LW group recalled more sentences than did patients in the LM, LW group, but the difference between these two groups on sentence recognition fell short of significance ( t = 2.02). These differences in performance on the recall and recognition tests occurred despite the fact that the delays between sentence presentation and the start of recall testing were longer for controls and patients in the NM, NW group than for the other patients (F (4,62) = 2.67, p < .05). A 5 (Group) X 2 (Levels of FOK)ANOVA on the percentage of nonrecalled items that were correctly recognized revealed a main effect of FOK (F (1,62) = 15.80, p < .001), but no effect of Groups (F < 1). Despite the appearance of differences among groups in the magnitude of the FOK effect, the Groups X FOK

321

METAMEMORY IN MS

Table 4. Performance on the Sentence Memory Test by Control and MS Patient Groups: Mean (SD) MS No Recalled % Recognized Delay (min) % Recognized, High FOK % Recognized

Downloaded by [UQ Library] at 17:49 01 November 2014

Low FOK

Gamma FOK Rating Latency (sec) Correctly Recalled High FOK Low FOK

Patients LM, NW

Groups'

Controls

NM, NW

16.1 (4.1) 20.2 (3.2) 85.1 (16.7)

14.1 (3.7) 8.6 (5.7) 19.1 (3.7) 14.3 (4.4) 83.3 (14.1) 70.5 (4.8)

11.3 (3.1) 6.9 (4.5) 15.8 (2.9) 11.9 (5.1) 74.5 (12.8) 73.0 (16.1)

42.8 (20.4)

40.4 (20.6) 27.2 (14.7)

33.9 (17.3) 32.8 (19.7)

NM, LW

LM, LW

18.3 (17.1) 25.3 (12.2) 27.1 (10.2) 27.6 (17.7) 16.6 (20.1) 0.24 (0.32) 0.14 (0.25) -0.04 (0.33) 0.05 (0.31) -0.02 (0.35) 2.54 (0.55) 2.44 (0.48) 2.61 (0.29) 2.38 (0.56) 2.20 (0.78)

1.8 (0.8) 4.9 (2.6) 4.4 (3.3)

1.9 (1.1) 5.0 (3.6) 4.5 (3.8)

1.6 (0.6) 2.4 (0.9) 2.8 (1.2)

1.7 (0.6) 3.9 (2.6) 3.6 (2.3)

2.8 (1.0) 5.1 (2.5) 5.6 (3.5)

1 . NM = Normal Memory, NW = Normal WCST, LM = Low Memory, LW = LOW WCST

interaction was not significant ( F = 2.03). Nevertheless, separate analyses showed that the FOK effect was significant for normal controls ( F (1,21) = 2 0 . 1 7 , ~< .OOl) and for patients in the NM, NW group (F (1, 17) = 9.28, p < .Ol), but not for patients in the LM, NW group (F < l), the NM, LW group (F < 1) or the LM, LW group (F 1 , 8) = 2.98, p > .lo). Analysis of the gamma coefficients revealed a similar pattern. Although there was no significant main effect of Groups ( F (4, 62) = 2.06, p < .lo), the gamma coefficients for controls ( t (21) = 3.56, p < .Ol) and patients in the NM, NW group ( t (17) = 2 . 4 9 , < ~ .05) differed significantly from 0. while the gamma coefficients for the other three groups did not (ts < 1). Despite the absence of a significant F for groups in the omnibus analysis, pairwise t-tests comparing controls to each of the four patient groups were performed. These analyses seemed useful because of the a priori hypothesis that patients with poor recognition memory and poor WCST performance should show impairment in this metamemory task (Janowsky et al., 1989; Shimamura & Squire, 1986), an hypothesis that was supported by the data on the free recall task described above. Results of the subsequent tests indicated that the normal controls had a significantly higher mean gamma coefficient than did patients in the LM, NW, NM, LW or LM, LW groups (ts > 2.20, p s < .05). Patients in the NM, NW group did not differ from controls. A 5 (Groups) X 2 (Levels of FOK) ANOVA on the latency data revealed no significant effects of Group, FOK or the interaction of these variables ( F s < 1.44).

322

WILLIAM W.BEATTY AND NANCY MONSON

Although patients in the LM, LW group tended to have longer latencies for correctly recalled items than did subjects in the other four groups, the effect of Groups on this measure fell short of statistical significance (F (4,62) = 2.52, p >

Downloaded by [UQ Library] at 17:49 01 November 2014

05). Memory Questionnaire Results on the various measures from the memory questionnaire are shown in Table 5. There was a significant effect of Groups on the Motor Scale ( F (4,62) = 11.73,p < .001) which arose because control subjectsreported less difficulty walking and lifting objects than any of the patient groups (ts > 4.22, ps < .oOl). Differences among the four patient groups on the Motor Scale were not significant. Analyses of the Sensory, Attention, General Memory, Remote Memory and Recent Memory Scales disclosed no significant differences among groups ( F s < 2.21, p s > .05). However, there was a significant effect of Groups on the Scenarios portion of the questionnaire (F (4,62) = 2.88, p c .05). Subsequent analyses showed that controls differed from patients in the NM, LW and LM, LW groups (ts > 2.1 1, ps c .05), but not from patients in the LM, NW group ( f = 1.29). No other comparisons were statistically significant. There were no significant differences among groups on the 8 items intended to measure knowledge of memory (F = 1.10). Twenty of the patients were taking one or more medications (See Methods) that could potentially affect their performance on neuropsychological tests, while

Table 5. Performance on the Memory Questionnaire by Control and MS Patient Groups: Mean (SD) MS Motor (6)2 Sensory (12) Language (18) Attention (12) Memory, General (6) Remote Memory (6) Recent Memory (48) Scenarios (40) Knowledge of Memory (8)

'

Groups

Controls

NM.NW

Patients LM, NW

5.1 (1.2) 9.7 (1.6) 15.6 (1.8) 9.9 (2.0)

2.3 (1.8) 9.6 (2.4) 14.7 (2.4) 8.7 (2.7)

2.1 (2.0) 8.5 (3.7) 13.5 (4.2) 7.5 (2.6)

8.5 (3.1) 14.4 (3.8) 8.7 (2.7)

1.6 (2.1) 7.4 (3.5) 14.9 (2.5) 7.4 (3.3)

4.9 (1.1)

4.4 (1.5)

3.9 (2.0)

4.0 (1.7)

4.1 (1.3)

4.6 (1.5)

4.3 (1.8)

3.5 (2.0)

4.1 (1.2)

4.1 (1.5)

36.7 (6.5) 37.6 (2.1)

32.5 (9.1) 36.8 (3.3)

30.5 (9.8) 34.3 (7.3)

32.4 (11.0) 32.4 (7.8)

29.4 (8.5) 35.2 (3.2)

6.7 (1.4)

7.0 (1.5)

6.1 (1.7)

7.5 (0.7)

6.7 (1.9)

NM.LW 2.1 (2.1)

LM. LW

1. NM = Normal Memory, NW = Normal WCST, LM = Low Memory, LW = LOW WCST 2. Maximum score shown in parentheses.

METAMEMORY IN MS

323

25 patients were not taking any of these medications at the time of testing. TO examine the possible influence of medications on test performance we compared these two groups of patients on all of the measures of cognitive performance, not including the memory questionnaire. Of 29 analyses that were performed only one (percent correctly recognized on general information) revealed a significant (p < .OS) effect of medication status. It is therefore unlikely that the use of medications influenced the results of the study.

Downloaded by [UQ Library] at 17:49 01 November 2014

DISCUSSION The present findings demonstrate that MS patients with impairments in either memory or on the WCST also display deficits on some measures of metamemory, but not on others. These results indicate that metamemory is a multidimensional construct, and suggest that future investigators should employ a variety of tasks in order to advance understanding of this broad concept. Nelson et al. (1984) suggested two different mechanisms by which FOK judgments might be made. One method requires access to specific item information in memory (trace-access mechanisms). The other depends upon inferential mechanisms that utilize contextual and other related knowledge. Assuming that the postulated inferential mechanisms depend upon the integrity of the same frontal circuits that are essential for normal performance on the WCST, then the pattern of deficits in metamemory displayed by the MS patients on the anterograde memory tests is in good agreement with the model of Nelson et al. (1984), provided that one allows for differences in task familiarity and difficulty. The sentence memory test was completely unfamiliar to all subjects while the free recall task was reasonably familiar to all subjects since they had attempted to recall the seven-word list on the SECIMS only a few minutes before being required to estimate their performance on the 14-word lists used on the free recall test. Further, since the sentence memory test required recall after at least a 60min delay while on the free recall task, memory was tested immediately after presentation of the to-be-remembered words, it is reasonable to suppose that the sentence memory test placed a greater demand on long-term memory. If one accepts this analysis of the two anterograde memory tasks, then patients in the LM, NW group may have been able to make reasonably accurate predictions of their performance on the free recall task by means of their intact inferential mechanisms, but these mechanisms were inadequate to allow accurate FOK judgments on the more demanding, less familiar sentence memory task. Likewise, patients in the NM, LW group could have forecast their performance on the free recall task by utilizing their nearly normal short-term memory capacities, but they could not make accurate FOK judgments on the sentence memory task because of their impairments in inferential mechanisms mediated by the frontal lobes and because of their inferior long-term memory. The inability of patients in the LM, LW group to make accurate predictions of their performance on either of

Downloaded by [UQ Library] at 17:49 01 November 2014

324

WILLIAM W.BEATIY AND NANCY MONSON

the anterograde memory tasks may have resulted from the fact that both traceaccess and inferential mechanisms appear to be deficient in these patients. Because the between-group differences on the sentence memory task were only marginally significant, this interpretation should be regarded as tentative until confirmed by future research. The normal ability of patients in the LM, LW group to predict their recognition performance on the general information test appears to pose some problems for the model of Nelson et al. (1984) since these patients are presumed to have deficits in both of the mechanisms postulated to underlie accurate FOK judgments. However, as Janowsky et al. (1989) pointed out, the general information test taps topics that were learned long ago and are likely to be well integrated into a person’s store of knowledge. Perhaps more extensive damage to trace-access or inferential systems than occurs in MS is required to disrupt ability to make accurate FOK judgments for this type of material. Brandt (1985) reported that patients with HD were able to predict their ability to recognize answers to general information questions that they could not recall with nearly normal accuracy, but they did not exhibit the normal correlation between FOK ratings and memory search time. In contrast, we found that all of the groups of MS patients searched their memories longer for the answers to questions that they thought they were likely to recognize than for answers to questions that they thought they would not recognize. Differences in the nature and severity of damage to the caudate nucleus in MS and HD may account for the apparently differential ability of MS and HD patients to direct their memory searches, although obviously numerous other explanations are possible. Because many of the MS patients tested in the present study displayed unambiguous memory deficits and impairments on laboratory measures of metamemory, the results of the memory questionnaire are of particular interest. The fact that none of the groups of MS patients reported significantly greater difficulty in remembering events or obligations that occurred in their daily lives might be interpreted as evidence that these patients lack awareness of their memory impairments. However, ratings on the memory items on the questionnaire by controls and MS patients were quite variable. The failure to detect differences might, therefore, simply reflect the insensitivity of the questionnaire rating scales. Alternatively, the highly variable performance of the groups may have arisen because some patients acknowledge their memory deficits while others do not. On the Scenarios portion of the memory questionnaire, patients in the LW, LM and NM, LW groups produced significantly lower estimates of their abilities to remember events that might occur in their everyday lives than did the normal controls, This findings suggests that at least some of these patients had some awareness of their memory deficits. However, there appeared to be marked differences in the degree to which patients had “insight” into their memory disturbances. For example, one patient who was unable to recall a single sentence correctly attained a score of 38 out of 40 on the Scenarios scale. In an effort to quantify the extent of individual variation in ability to appraise memory ability,

Downloaded by [UQ Library] at 17:49 01 November 2014

METAMEMORY IN MS

325

we counted the number of patients in each of the four groups who scored at or below the 5th percentile of controls on sentence recall and above the 5th percentile of controls on the Scenarios scale. The results of this tally by groups were as fOllOWS: NM, NW - 2/18, LM, NW - 3/8, NM, LW - 2/10, and LM, LW - 5/9. These findings indicate that many MS patients are likely to overestimate their abilities to remember events and obligations that occur in their everyday lives. In a previous study (Beatty, Goodkin, Hertsgaard, & Monson, 1989) we demonstrated that the deficits in memory and other cognitive functions exhibited by some MS patients cannot be predicted reliably on the basis of demographic measures or from disease type, disease duration, or extent of physical disability. The present findings suggest that asking MS patients to appraise their own memory capabilities will also prove to be unreliable. Consequently, estimating MS patients’ cognitive abilities will require some sort of neuropsychological testing. Because many MS patients do not have significant cognitive impairments, administering an extensive and costly neuropsychological battery to all patients may be difficult to justify. Fortunately, the SECIMS,which takes only about 10minutes to administer and score, appears to have high sensitivity and specificity for predicting cognitive dysfunction in MS, and may prove to be a suitable screening examination for routine clinical use with MS patients (Goodkin, Beatty, & Monson, submitted).

REFERENCES Bauer, R.H., Kyaw, D., & Kilbey, M.M. (1984). Metamemory of alcoholic Korsakoff patients. Society for Neuroscience Abstracts, 10, 318. Beatty. W.W., Butters, N., & Janowsky, D.S. (1986). Patterns of memory failure after scopolamine treatment: Implications for cholinergic hypotheses of dementia: Behavioral and Neural Biology, 45, 196-211. Beatty, W.W., Goodkin. D.E. (1990). Screening for cognitive impairment in multiple sclerosis: An evaluation of the Mini-Mental State Exam. Archives of Neurology, 47, 297-301. Beatty, W.W., & Goodkin. D.E., Beatty, P.A., & Monson, N. (1989). Frontal lobe dysfunction and memory impairment in patients with chronic progressive multiple sclerosis. Brain and Cognition, 11, 73-86. Beatty, W.W., Goodkin, D.E., Hertsgaard, D., & Monson, N. (1 990). Clinical and demographic predictors of cognitive performance in multiple sclerosis: Do diagnostic type, disease duration, and disability matter? Archives of Neurotogy, 47, 305-309. Beatty, W.W., Goodkin, D.E., Monson, N., & Beatty, P.A. (1989). Cognitive disturbances in patients with relapsing remitting multiple sclerosis. Archives of Neurology, 46, 11131119. Beatty, W.W., Goodkin, D.E.,Monson, N., Beatty. P.A., & Hertsgaard, D. (1988). Anterograde and retrograde amnesia in patients with chronic progressive multiple sclerosis. Archives of Neurology, 45, 61 1-619. Benton. A.L. (1 968). Differential effects of frontal lobe disease. Neuropsychologia, 6,5360. Blake, M. (1973). Prediction of recognition when recall fails: Exploring the feeling-ofknowing phenomenon. Journal of Verbal Learning and Verbal Behavior, 12, 31 1-319.

326

WILLIAM W.BEATTY AND NANCY MONSON

Brandt, J. (1985). Access to knowledge in the dementia of Huntington’s disease. Developmental Neuropsychology, I , 335-348. Coulter, L. (1989). The feeling of knowing in depressed and nondepressed patients with Parkinson’s disease. Journal of Clinical and Experimental Neuropsychology, 11, 91. Folstein, M.F., Polstein. S.E..,& McHugh, P.R. (1975). “Mini-Mental State”: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatry Research, 12, 189-198. Goodkin, D.E., Beatty, W.W., & Monson, N. (May, 1990). Screening examination for detecting cognitive impairment in patients with multiple sclerosis. Presented at the annual meeting of the Consortium of Multiple Sclerosis Centers, Denver, CO. Gruneberg, M.M., &Monks, J. (1974). Feeling of knowing and cued recall.Acra Psychologia,

Downloaded by [UQ Library] at 17:49 01 November 2014

41,257-265.

Hart, J.T. (1965). Memory and the feeling-of-knowing experience. Journal of Educational Psychology, 56, 208-216. Hart, J.T. (1967). Memory and the memory-monitoring process. Journal of Verbal Learning and Verbal Behavior, 6, 685-691. Hauser, S.L., Dawson, D.M., Lehrich, J.R., Beal, M.F., Kevy, S.V.,Propper, R.D., Mills, J.A., & Weiner, H.L. (1983). Intensive immunosuppression in progressive multiple sclerosis. New England Journal of Medicine, 308, 173-180. Heaton. R.K. (1981). Wisconsin Curd Sorting Test Manual. Odessa, FL: Psychological Assessment Resources. Herman, B.P., Wyler, A.R., & Richey, E.T. (1988). Wisconsin Card Sorting Test performance in patients with complex partial seizures of temporal-lobe origin. Journal of Clinical and Experimental Neuropsychology, 10, 467-476. Hirst. W. (1 982). The amnesic syndrome: Descriptions and explanations. Psychological Bulletin, 91,435-460. Janowsky, J.S. Shimamura, A.P., & Squire, L.R. (1989). Memory and metamemory: Comparisons between patients with frontal lobe lesions and amnesic patients. Psychobiology, 17, 3- 11. Kaplan, E., Goodglass, H., & Weintraub, S. (1983). Boston Naming Test. Philadelphia: Lea & Febiger. Kelly, M., Scholnick, E.K., Travers, S.H., & Johnson, J.W. (1976). Relations among memory, memory appraisal and memory strategies. Child Development, 47,648-659. Kurtzke, J.F. (1983). Rating neurological impairment in multiple sclerosis: An expanded disability status scale (EDSS). Neurology, 33, 1444-1452. Lachman, J.L., Lachman, R., & Thronesbery, C. (1979). Metamemory through the adult lifespan. Developmental Psychology, 15. 543-551. McGlynn, S.M., Schacter, D.L., & Glisky, E.L. (1989). Unawareness of deficit in organic amnesia. Journal of Clinical and Experimental Neuropsychology, 11, 50. Nelson, T.O. (1984). A comparison of current measures of the accuracy of feeling-ofknowing predictions. Psychological Bulletin, 95, 109-133. Nelson, T.O., Gerler, D., & Narens, L. (1984). Accuracy of feeling-of-knowing judgments for predicting perceptual identification and relearning. Journal of Experimental Psychology: General, 113, 282-300. Nelson, T.O.,Leonesio, R.J., Shimamura. A.P., Landwehr R.F., & Narens, L. (1982). Overlearning and the feeling of knowing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 8, 279-288. Nelson, T.O., & Narens, L. (1980). Norms of 300 general-information questions: Accuracy of recall, latency of recall, and feeling-of-knowing ratings. Journal of Verbal Learning and Verbal Behavior, 19, 338-368. Poser, C.M., Paty, D.W.. Scheinberg, L., McDonald, W.I., Davis, F., Ebers, G.C., Johnson, K.P., Sibley, W.A., Silberberg, D.H.. & Tourtellotte, W.W. (1983). New diagnostic

METAMEMORY IN MS

327

criteria for multiple sclerosis: Guidelines for research protocols. Annals of Neurology,

Downloaded by [UQ Library] at 17:49 01 November 2014

13,227-231. Prevey, M.L., Delaney, R.C., BE Mattson, R.H. (1988). Metamemory in temporal lobe epilepsy: Self-monitoring of memory functions. Brain and Cognition, 7 , 298-311. Schacter, D.L. (1983). Feeling of knowing in episodic memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9, 39-54. Shimamura, A.P., Jemigan, T.L.,& Squire, L.R. (1988).Korsakoff's syndrome: Radiological (CT) findings and neuropsychological correlates. Journal of Neuroscience, 8,4400-4410. Shimamura, A.P., & Squire, L.R. (1986). Memory and metamemory: A study of the feeling-of-knowing phenomenon in amnesic patients. Journal of Experimental Psychology: Learning, Memory,and Cognition, 12,452-460. Tri)ster, ALBeatty, W.W., Stat0n.R.D.. &Rorabaugh, A.G.(1989).Effects of scopolamaine on anterograde and remote memory in humans. Psychobiology, 17, 12-18. Yussen, S.R., & Levy, V.M. (1975). Developmental changes in predicting one's own span of short-term memory. Journal of Experimental Child Psychology, 19,502-508.