Long-Term Deficits in Episodic Memory after Ischemic Stroke: Evaluation and Prediction of Verbal and Visual Memory Performance Based on Lesion Characteristics

Long-Term Deficits in Episodic Memory after Ischemic Stroke: Evaluation and Prediction of Verbal and Visual Memory Performance Based on Lesion Characteristics Eveline A. Schouten, MSc,* Sven K. Schiemanck, MD, PhD,† Nico Brand, PhD,‡ and Marcel W. M. Post, PhD*x

We investigated the relationship between ischemic lesion characteristics (hemispheric side, cortical and subcortical level, volume) and memory performance, 1 year after stroke. Verbal and visual memory of 86 patients with stroke were assessed with Rey Auditory-Verbal Learning Test and the Doors Test, respectively. Lesion characteristics and presence of white matter lesions were assessed on magnetic resonance imaging early after stroke. Multiple regression analyses were used to investigate prediction of verbal and visual memory performance by lesion side (left v right hemisphere), lesion level (cortical v subcortical), and lesion volume. We controlled for the influence of demographic characteristics, language disability, and visuospatial difficulties on memory. The results demonstrated that poor verbal memory (immediate and delayed recall and recognition) could be predicted by lesion characteristics: patients with left hemispheric, subcortical, and large lesions showed poor memory performance. Poor visual recognition memory could not be predicted by lesion characteristics but only by low educational level. Our results suggest that lesion characteristics play an important role in episodic verbal memory poststroke if demographic and clinical characteristics are taken into account. Key Words: Cerebrovascular accident—magnetic resonance imaging—cognition disorders— memory disorders—neuropsychological tests. Ó 2009 by National Stroke Association

Cognitive impairment is a common sequel of stroke.1 It plays an important role in acute functional recovery2 and in long-term outcome such as independency and quality From the *Center of Excellence for Rehabilitation Medicine, Rehabilitation Center De Hoogstraat Utrecht, †Department of Rehabilitation, Academic Medical Center Amsterdam, ‡Department of Clinical and Health Psychology, Faculty of Social Sciences, Utrecht University; and xRudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, the Netherlands. Received June 2, 2008; revision received September 11, 2008; accepted September 12, 2008. S. K. Schiemanck was supported by a grant from the Netherlands Organization for Health Research and Development (ZON/MW) and the Scientific Foundation of Rehabilitation Center De Hoogstraat. Address correspondence to Eveline A. Schouten, Center of Excellence for Rehabilitation Medicine, Rehabilitation Center De Hoogstraat, Rembrandtkade 10, 3583 TM, Utrecht, the Netherlands. E-mail: [email protected]. 1052-3057/$—see front matter Ó 2009 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2008.09.017

128

of life.3,4 Memory is one of the cognitive domains frequently affected by stroke.5,6 Memory function–involving the ability to register, store, save, and retrieve information when needed7–is particularly important in the process of rehabilitation after stroke, as it is required for learning new skills and relearning old ones.8 There are many case studies on memory impairment after stroke and its relationship with lesion location. However, recently, group studies have been performed that investigated acute and chronic cognitive impairment (including memory impairment) after stroke and its relationship with a wide range of determinants, such as vascular risk factors, pre-existent neuropathology, and lesion characteristics of volume and location.1,9-12 Yet only few of these studies9,13 investigated memory impairment in the chronic phase of stroke in large samples, using specific indicators of verbal memory function and considering different stages of memory processing (i.e., encoding, storage, and retrieval). However, it has been

Journal of Stroke and Cerebrovascular Diseases, Vol. 18, No. 2 (March-April), 2009: pp 128-138

STROKE LESION AND VERBAL AND VISUAL MEMORY DEFICITS

frequently assumed that different neuroanatomic stroke sites lead to different types of memory impairment in terms of the type of information and stages of processing.5,6,14 Verbal memory deficits have been consistently demonstrated to occur primarily after left hemispheric stroke, and nonverbal (visuospatial) memory deficits principally after right hemispheric stroke.7 However, there is increasing evidence suggesting that this distinction is not absolute: left hemispheric lesions have been reported to produce visual memory impairments, whereas verbal memory deficits are also observed after right hemispheric stroke.5,6,14 Results of functional imaging studies on the hemispheric lateralization of memory also diverge. Several studies have shown that verbal encoding of words produced left hemispheric activations, whereas nonverbal encoding of unfamiliar faces produced right hemispheric activations.15,16 However, some recent studies demonstrated either no lateralization of verbal and visual memory function, or only minimal lateralization of verbal memory function to the left hemisphere.17-19 Cortical and subcortical lesion levels have been shown to influence memory deficits in patients with brain damage, including patients with stroke. However, only few studies have compared the relative influence of cortical versus subcortical lesions on memory function. These studies showed that patients with stroke with a cortical lesion have more severe memory deficits than patients with a subcortical lesion.2,20 This finding is not surprising, because several cortical areas play an essential role in episodic memory function. The posterior cortical region of the medial temporal lobe is widely recognized to mediate the associative, contextual, and recollective aspects of episodic encoding and retrieval.21-23 The hippocampus is the crucial structure in encoding of ongoing information, and the multimodal association areas of the posterior cortex are generally assumed to be the site for long-term storage of episodic memories.24-27 Furthermore, results from imaging studies have shown that regions of the anterior (prefrontal) cortex also take part in the network mediating episodic memory processing,28,29 probably facilitating (re)constructive and search processes of encoding and retrieval by its inherent executive functions (i.e., attentional processes, monitoring, and coordination).7,24 Functional imaging studies have demonstrated a hemispheric asymmetry in the memory processes of the prefrontal cortex, first postulated by Tulving et al30 in the hemispheric encoding/retrieval asymmetry (HERA) model, which involves an increased activity in left prefrontal regions during intentional encoding, whereas episodic retrieval primarily activated right prefrontal brain regions.17,28,31 However, although the essential role of cortical structures in episodic memory is evident, recent studies increasingly show the important role of subcortical structures as well.12,32-36 Although lesions of the subcortical structure of the thalamus are long associated with

129

memory disorders, results of recent imaging studies have suggested that memory deficits may also result from focal lesions of the anterior and medial cortical portions of the thalamus, particularly if there is involvement of subcortical white matter tracts such as the mammillothalamic tract.33,37,38 Furthermore, studies12,32,34,35 have reported an impairment of short-term and long-term memory, particularly of encoding and recall, after stroke in the basal ganglia. Therefore, it seems that more research is needed to evaluate the influence of cortical and subcortical stroke lesions on memory impairment after stroke. Only a few studies have examined the influence of lesion volume on poststroke memory function.13,34 However, lesion volume is known to correlate moderately to strongly with long-term functional outcome and quality of life39 after stroke. This study investigates the relationship between ischemic lesion characteristics (side, level, volume) and verbal and visual episodic memory performance 1 year after stroke in a relatively large patient sample. Based on the previously discussed findings of episodic memory deficits after stroke, we expected that patients with left rather than right hemispheric lesions would manifest more verbal memory dysfunctions, and that a larger volume would give rise to more severe disturbances.

Methods Procedure The study population consisted of patients with a firstever ischemic stroke admitted to one of 6 participating stroke departments in the Netherlands. Stroke has been defined as a rapidly developing sign of focal or global disturbance of cerebral function with symptoms lasting 24 hours or longer or leading to death, with no apparent origin other than vascular.40 The research protocol was approved by our medical ethics committee. All patients gave their informed consent. Patients included had a single first-ever supratentorial nonlacunar ischemic infarction of the anterior, medial, or posterior cerebral artery; were aged between 18 and 85 years; had a premorbid Barthel Index greater than or equal to 18; and had a stable neurologic condition 1 week after stroke. They did not have any mental comorbidity (e.g., dementia or psychiatric disorder) that might influence neuropsychological outcome. Lacunar infarctions were defined as infarctions of the deep white matter of the brain, caused by an occlusion of small perforating arteries, with a diameter ranging from 3 to 4 mm to a diameter of 15 to 20 mm and located at the site of the basal ganglia, internal capsule, or corona radiata.41,42 Patients with premorbid cognitive limitations or beginning dementia, assessed by (hetero-)anamnesis with patient and/or family, were excluded. Participants received normal drug treatment and were not treated with thrombolysis or neuroprotective agents, because this was not yet

130

a common treatment at the time of inclusion. Patients received common hospital treatment with regard to language and other cognitive problems. Data about stroke severity were taken at hospital admission. Disability level was assessed about 6 days poststroke. Magnetic resonance imaging (MRI) scan was obtained at a mean of 11 days (SD 3.5) after the stroke incident. Participants showed a visible lesion on the MRI scan as diagnosed by an independent neuroradiologist. Measurement of infarct localization and volume was performed blinded for the patients’ clinical status.43 After 1 year poststroke, patients were tested by an experienced neuropsychologist.

Materials Stroke severity of the patient at hospital admission was assessed with the National Institutes of Health Stroke Scale44 and the Glasgow Coma Scale.45 Demographic variables (age, sex, educational level) were obtained from the patient. Educational level was scored using the revised Dutch classification system of Verhage46: levels 1 and 2 indicate a low educational level; levels 3 through 5 an intermediate level; and levels 6 and 7 a high level. Disability level was assessed with the Barthel Index47,48 about 6 days poststroke. After 1 year poststroke, general cognitive functioning was assessed with the Mini Mental State Examination.49,50 The presence of depression was assessed with the Center for Epidemiologic StudiesDepression Scale (CES-D),50,51 because the presence of depression might influence long-term memory function.52 The CES-D measures the presence and severity of depressive symptoms, with scores greater than or equal to 16 (range 0-60) indicating a clinically significant level of psychological distress. As deficits of language comprehension and production might also influence verbal memory performance, the presence of language deficits was assessed with a short form of the Token Test53 and the Boston Naming Test (BNT).54 The short form of the Token Test55 consists of 21 spoken assignments and the subject is asked to manipulate cards with printed ‘‘tokens’’ varying in form, color, and size. One point is scored for every correct response. Normative data56 are based on cutting points that vary according to age, sex, and intelligence. Five practice assignments were administered in advance: subjects were asked to select a certain shape and color. If subjects responded incorrectly on 3 practice items, the Token Test was not administered. The Token Test has proven to be useful as a screening instrument for the presence of aphasia.53 The BNT54 consists of 60 large ink drawings of familiar and unfamiliar objects and animals. Subjects are asked to tell the common name of the presented items. The BNT is designed to evaluate naming impairments in aphasics after left-sided brain damage, but has been proven to effectively assess word-finding problems in both aphasics and nonaphasics and in patients with left-sided and

E.A. SCHOUTEN ET AL.

right-sided lesions. Scores from 0 to 3 are given for each answer, with a score of 0 indicating a completely nonfitting name or description and a score of 3 indicating a fully accurate or specific name or description. Deficits in visuospatial functioning might influence visual memory performance. To control for this type of deficit the Judgment of Line Orientation (JLO) test57 was administered. This test, consisting of 30 items, examines the ability to estimate angular relationships between line segments by visually matching angled line pairs to 11 numbered radii arranged in a semicircle.7 Although left hemispheric involvement in this test has been shown using neuroimaging techniques the right hemisphere plays a more important role in line orientation.57 As an additional control test for visuoperceptual difficulties a Letter Cancellation Task was administered. This task consists of a sheet of paper with 40 ‘‘O’s printed on it (20 on the left half and 20 on the right) that have to be found between a large number of distractor letters and cancelled with a pencil as fast and accurately as possible. The total score of the test was used as a measure of visual search. The memory tests used in this study assessed episodic anterograde memory. Verbal learning and memory were assessed with the Dutch version of the Rey AuditoryVerbal Learning Test (RAVLT),52,58-60 which is widely used among Dutch neuropsychologists and has good reliability and validity. This test comprises a list of 15 words, presented to the subject at 5 trials. Recall is tested immediately after each trial and the total number of correctly remembered words in all 5 trials is taken as the score for immediate recall. After a delay of 20 minutes the patient is asked to recall again as many words as possible (delayed recall), followed by a recognition test in which the 15 words have to be recognized from a mix with 15 distracter words.7 In the current study a recognition score was comprised of correct responses on both targets and distracters. According to the normative data of the Dutch RAVLT, raw scores on immediate and delayed recall are converted into decile scores, corrected for age, sex, and educational level.52,59 According to Rasquin et al,10 scores equal to or below the 10th percentile of the norm group are considered to indicate clinically relevant memory deficits. Normative data are not available for verbal recognition. However, according to the guidelines52,59 raw scores below 27 fall into the 10th percentile and were considered as impaired performance. In our study, we used the raw scores of the RAVLT for analysis only. Visual memory was assessed with the Doors Test, a subtest of the Doors and People Test.61 The Doors Test evaluates visuospatial recognition memory by presenting two sets (easy and difficult) of 12 pictures of doors (e.g., front doors, church doors) that have to be memorized and recognized from arrays of 4 pictures (one target and 3 distracters) of similar doors. For memorization, each picture is shown for 3 seconds. The test has a wide range of applicability and is

STROKE LESION AND VERBAL AND VISUAL MEMORY DEFICITS

sensitive to effects of Alzheimer disease, stroke, normal ageing, and schizophrenia.61 In addition, validation studies have proven the capacity of the test to distinguish between modality-specific memory impairment in patients with left or right temporal lobectomy62 and in normal ageing.63 For reasons of time limitation patients in this study were tested with the easy set only. To assess impaired performance, raw scores are converted into age-scaled scores (mean 10; SD 3), which are available for both sets separately. According to Davis et al64 scaled scores less than or equal to 5 are considered impaired. In the current study raw scores were used for analysis.

Lesion Analysis All MRI scans were performed about 11 days poststroke (range 5–23 days; SD 3.6). MRI studies were performed using 0.5-, 1.0-, or 1.5-T MRI scans (Philips Medical Systems, Eindhoven, The Netherlands). A standard scanning protocol was performed in all hospitals: a sagittal T1-weighted sequence (survey), an axial T2weighted sequence, slice thickness 6 mm, gap 1.2 mm, a transversal fluid-attenuated inversion recovery sequence, and a transversal gradient echo T2-weighted sequence. Scans were stored and further analyzed using a workstation (Easy Vision, Philips Medical Systems). Surfaces of areas of abnormal hyperintensity were summed and multiplied with slice thickness (6 mm) and interslice gap (1.2 mm) to calculate lesion volumes.43 Lesions were defined as cortical if they were located for 50% or more in the cortex and as subcortical if they were located for 50% or more in subcortical areas. Cortical lesion localization was further divided into frontal, parietal, temporal, or occipital dependent on the location of the largest part of the lesion, using the atlas of Damasio and Damasio.65 Based on this classification, two groups of patients were obtained: the anterior group consisting of patients with frontal (cortical) lesions and a posterior group consisting of patients with parietal, temporal, or occipital cortical lesions. As pre-existent white matter lesion (WML) could influence memory function, its presence was also determined on the MRI scans.66

Data Analyses Bivariate relationships between variables were obtained by Pearson product-moment correlations. Correlation coefficients less than 0.3 were considered weak, between 0.3 and 0.5 as moderate, and greater than 0.5 as strong.67 Differences in scores on immediate recall, delayed recall, and recognition in discrete groups based on lesion side (left v right hemisphere), lesion level (cortical v subcortical), and location (anterior v posterior) were tested using univariate analysis of variance corrected for the demographic variables (age, sex, education) as covariates. Hierarchic multiple regression analyses were performed to evaluate the independent predictive value of lesion side

131

(left and right hemisphere), lesion level (cortical and subcortical), and lesion volume on patients’ verbal (RAVLT) and visual (Doors Test) memory performance. Because lesion volume was strongly skewed to the right, this variable was transformed logarithmically. To control for the influence of demographic and related clinical characteristics, the analysis consisted of 3 steps. In the first step the demographic variables of age, sex, and educational level were entered as independent predictors. To control for the influence of depression, WML, word finding problems and visuospatial difficulties, CES-D scores, presence of WML, and scores on Token Test, BNT, JLO, and the Letter Cancellation Task were added in the second step. Although scores on the Token Test and BNT were strongly correlated (r 5 .72 in our research) we included both indices in the regression analyses, because in separate analyses the Token Test was a better predictor of RAVLT immediate recall, and the BNT score was a better predictor of RAVLT delayed recall. The 3 lesion characteristics (side, level, volume) were entered in the final step. Cortical lesion location was not included in these regression analyses because this variable was only applicable to persons with cortical lesions, which would greatly reduce the number of degrees of freedom.

Results Of 105 included patients, 19 patients could not be evaluated after 1 year poststroke (mean 377 days; SD 22): 9 patients had died; 4 patients had recurrent stroke; two patients developed comorbidity seriously affecting functional outcome; and 4 patients refused further examination. RAVLT test results of two patients were incomplete and in 8 patients the RAVLT could not be administered because of severe aphasia. Finally, 76 patients had complete data on the RAVLT. Results on the Doors Test were available from 83 patients, because 3 patients had incomplete test results.

Group Characteristics Demographic, clinical, and lesion characteristics of the resulting 86 patients (42 men; 44 women) are presented in Table 1. In all, 46 patients had a left and 40 patients had a right hemispheric lesion. A total of 48 patients had subcortical lesions (32 purely subcortical and 16 mixed). In all, 38 patients had cortical lesions that were all mixed lesions. In all, 33 (61.1%) of all mixed lesions (N 5 54) were located in anterior regions. Mean lesion volume was 58.6 mL (SD 76.4). Lesion volume was not normally distributed (median 29.3; 25%-75% interquartile range 6.4-81.2). At hospital admission, all patients were conscious (mean Glasgow Coma Scale score of 14.6 [range 11-15]) and had a mean National Institutes of Health Stroke Scale score of 10.7 (SD 5.5), indicating mild stroke. About 6 days poststroke, patients had a mean score of 10.3 (SD 6.4) at the Barthel Index.

E.A. SCHOUTEN ET AL.

132

Table 1. Demographic, clinical, and lesion characteristics of patients with first-ever ischemic stroke at 1 year poststroke (N 5 86) N Demographics Age, y Sex Male Female Handedness Right Left or ambidexterity Unknown Educational level Low (2)* Intermediate (3-5) High (6) Lesion characteristics Side Left hemisphere Right hemisphere Level Cortical Pure Mixed, .50% in cortex Subcortical Pure Mixed, .50% in subcortex Cortical location Anterior Posterior Volume, mL Additional clinical characteristics Depressive mood, CES-D total score (0-60)y CES-D $ 16 MMSE MMSE , 24 Token Test Boston Naming Test Judgment of Line Orientation Letter Cancellation Task total score (0-40)y White matter lesions present

Percent

86 42 44

48.8 51.2

72 11 3

83.7 12.8 3.5

19 50 17

22.1 58.1 19.8

46 40

53.5 46.5

38 0 38 48 32 16 54 33 21 86

44.2

M

SD

Range

63.3

14.2

22–84

58.6

76.4

0.95–373.6

16.7

6.1

5–35

26.5

3.5

13–30

13.5 144.9 19.8 36.3

5.2 31.1 6.6 7.1

0–21 15–168 2–30 6–40

55.8

61.1 38.9

81 58.0 80 16.3 74 78 80 83 31

36.0

Abbreviations: CES-D, Center for Epidemiological Studies-Depression Scale; M, mean; MMSE, Mini Mental State Examination. *Numbers in parentheses indicate corresponding educational levels according to revised classification system of Verhage46 (1964). yNumbers in parentheses indicate range of possible score.

WML were present in 31 patients. After 1 year poststroke, patients had a mean Mini Mental State Examination score of 26.5 (SD 3.5). A score lower than 24, suggestive of probable dementia,68 was found in 16.3% of the patients. Mean scores and SD on the Token Test, BNT, JLO, and the Letter Cancellation Task are shown in Table 1. CES-D scores were normally distributed (Med 5 17). Of the patients, 58% obtained a score of 16 or higher, indicating presence of psychological distress. Mean score on the CES-D was 16.7 (SD 6.1).

Memory Performance and Lesion Characteristics Table 2 presents the mean raw scores on the RAVLT and Doors Test for the total patient group and for subgroups with different lesion characteristics. On the subtest of verbal immediate recall, patients (N 5 76) showed a mean score of total correct answers over 5 trials of 28.8 (SD 12.2). After the 20-minute delay, patients recalled a mean of 5.4 (SD 3.3) words. Of the patients, 43% scored below the 10th percentile of immediate recall, and 29%

STROKE LESION AND VERBAL AND VISUAL MEMORY DEFICITS

133

Table 2. Mean performance of patients on the Rey Auditory-Verbal Learning Test and the Doors Test RAVLT

All patients Lesion side LH RH Lesion level Mainly subcortical Mainly cortical Lesion location Anterior Posterior

Doors Test

N

Immediate recall

Delayed recall

Recognition

N

Recognition

76

28.8 (12.2)

5.4 (3.3)

25.9 (3.8)

83

9.1 (2.1)

37 39

26.3 (12.8) 31.3 (11.3)

4.7 (3.5) 6.0 (3.1)

25.3 (4.0) 26.4 (3.5)

44 39

9.3 (1.8) 8.8 (2.4)

45 31

26.6 (11.5) 32.1 (12.6)

4.7 (3.0) 6.2 (3.5)

25.3 (4.2) 26.5 (3.1)

46 37

9.0 (2.3) 9.2 (1.9)

26 18

31.0 (13.2) 28.6 (10.6)

6.4 (3.4) 5.4 (3.1)

25.5 (4.3) 26.6 (3.1)

31 20

9.0 (1.7) 9.1 (2.5)

Abbreviations: LH, left hemisphere; RAVLT, Rey Auditory-Verbal Learning Test; RH, right hemisphere. Raw scores; SD in parentheses. Patients subdivided according to lesion side (left and right hemisphere), lesion level (subcortical and cortical), and cortical lesion location (anterior and posterior lesions).

below the 10th percentile of delayed recall, both indicating a disordered memory performance.10 On the verbal recognition subtest, a mean of 25.9 (SD 3.8) items was correctly recognized of 30 items. Of the patients, 44% obtained a recognition score below 27, indicating impaired performance. On the Doors Test, the patients (N 5 83) had a mean raw score of 9.1 (SD 2.1). Patients obtained a mean scaled score of 8.0 (SD 3.5), whereas 25% of the patients had a scaled score of 5 or less, indicating impaired visual recognition memory.64 Univariate analyses of variance with age, sex, and educational level as covariates showed that patients with a left hemispheric lesion performed significantly poorer on the trials of verbal immediate recall (F 5 6.7, df 5 1,75, P 5 .01) and the delayed recall trial (F 5 6.4, df 5 1,75, P 5 .01) than patients with a right hemispheric lesion. No significance (P 5 .09) was observed for lesion side on verbal recognition. No significant differences in verbal immediate and delayed recall and recognition were found between patients with cortical and subcortical lesions, or between patients with anterior and posterior lesions. On the Doors Test, differences between subgroups of patients were not significant.

Correlational Patterns among Predictors and Dependent Variables Bivariate analyses showed weak to strong associations of demographic, clinical, and lesion characteristics and the memory test scores (r 5 .23-.85). Older age was associated with more WML (r 5 .46) and a lower educational level (r 5 –.27). Educational level was positively associated with language comprehension and naming. Lesions of right hemispheric patients were larger and these patients showed a poorer performance on naming. Patients with pre-existent WML had smaller lesions.

Scores on the verbal immediate recall, delayed recall, and recognition subtests of the RAVLT were strongly interrelated (r 5 .68-.85). Correlations of the RAVLT measures with scores on the Doors Test were low but positive. In summarizing the correlation patterns between background variables and the dependent variables, poorer verbal immediate and delayed recall and recognition were significantly correlated (P , .05) with older age, lower education, language comprehension, and naming. In addition, men performed worse on immediate recall compared with women. Poorer performance on the Doors Test correlated significantly (P ,.05) with lower education.

Predictive Models for Memory Performances Tables 3 and 4 show the results from hierarchic multiple regression analyses on the RAVLT scores and the Doors Test. The analyses were done in 3 steps. For the RAVLT, in the first step age, sex, and educational level were significant independent predictors of verbal immediate, delayed recall, and recognition performance (b 5 0.17– 0.44). In all 3 memory processes, age was the strongest predictor. Together, these variables explained 41%, 38%, and 21% of the scores on verbal immediate recall, delayed recall, and recognition performance, respectively (Table 4). In the second step, adding the clinical variables to the model of the first step, the Token Test was a predictor of verbal immediate recall and a weak predictor for verbal recognition. Performance on the BNT predicted verbal delayed recall. The clinical characteristics of WML and CES-D had no predictive value for any of the 3 verbal memory indices. Together these variables explained an additional 21%, 18%, and 24% of total variance in, respectively, immediate recall, delayed recall, and recognition scores. In the third model, with the addition of the lesion characteristics to the earlier steps, lesion side, level, and

E.A. SCHOUTEN ET AL.

134

Table 3. Hierarchic regression analyses for prediction of patient performance on the Rey Auditory-Verbal Learning Test (N 5 76) and Doors Test (N 5 83) in 3 steps: demographic variables, clinical characteristics, and lesion characteristics RAVLT Immediate recall RAVLT Delayed recall RAVLT Recognition Doors Test Recognition

I. Demographics Age Sex Educational level II. Clinical characteristics Token Test Boston Naming Test Letter Cancellation JLO CES-D WML III. Lesion characteristics Lesion side Lesion level Lesion volume (log)

b

P

b

P

b

P

b

P

–.44 .28 .27

,.001 .006 .01

–.42 .17 .30

,.001 .07 .005

–.34 .18 .17

.005 .10 .13

.04 .01 .53

.68 .90 ,.001

.35 .18 .02 –.24 –.06 –.14

.009 .22 .81 .05 .46 .15

.13 .41 –.02 –.24 –.01 –.03

.33 .008 .84 .06 .97 .76

.24 .28 .09 –.04 –.18 .09

.10 .09 .44 .75 .08 .43

–.05 .20 .15 .30 –.22 .09

.73 .23 .22 .04 .04 .40

.23 .31 –.30

.01 .001 .004

.28 .35 –.25

.004 ,.001 .01

.23 .31 –.33

.04 .006 .01

.07 .07 –.23

.57 .53 .08

Abbreviations: CES-D, Center for Epidemiological Studies-Depression Scale; JLO, Judgment of Line Orientation; log, natural logarithm; RAVLT, Rey Auditory-Verbal Learning Test; WML, white matter lesions. Coding dichotomous variables: sex: 0 5 male, 1 5 female; lesion side: 0 5 left, 1 5 right; lesion level: 0 5 mainly subcortical, 1 5 mainly cortical.

volume were predictors of verbal immediate recall (b 5 .23-.31). Patients with a left hemispheric lesion, subcortical lesion, and larger lesion were associated with lower scores on the measure of verbal immediate recall. These variables explained an additional 9% of total variance in verbal immediate recall scores (P 5 .002). Lesion side, level, and volume were predictors of verbal delayed recall. These variables explained an additional 11% of variance in verbal delayed recall scores (P 5 .001). Lesion side, level, and volume also predicted verbal recognition performance. Together, lesion side, level, and volume explained 10% of total variance in verbal recognition (P 5 .01) scores in addition to the variance explained by demographic and clinical characteristics. In total, 71%, 69%, and 56% of variance in verbal immediate recall, delayed recall, and recognition scores, respectively, was explained after the third step. For the Doors Test, the first model (demographic variables) only yielded 27% explained variance, with educational level as the single predictor (b 5 .53). After addition of the clinical variables, the percentage explained variance was raised by 17%, with JLO and CES-D as contributors (b 5 .30 and 5.22, respectively). The final model, adding the lesion characteristics, did not significantly enlarge the amount of explained variance. The total percentage was 46%.

Discussion Our study investigated the relationship between lesion characteristics (lesion side, level, and volume) and

episodic memory performance of patients with ischemic stroke in the chronic phase after stroke. Our results show that 1 year poststroke, compared with norms, a large number of patients were impaired on verbal immediate and delayed recall as well as on verbal recognition. Furthermore, one quarter of the patients had impairments of recognizing learned visual information. Studies1,9 of administering the RAVLT at 3 months and 2 years after stroke, respectively, also showed data consistent with the notion of impaired verbal memory performance, at least in 40% of the patients.

Lesion Side In this study we expected left hemispheric patients to be more hampered with memory dysfunctions (in particular verbal memory) than right hemispheric patients, complying with the classic notion of verbal memory dysfunction with lesions of the left hemisphere and nonverbal memory dysfunction with lesions of the right hemisphere.6,7 Patients with left hemispheric lesions indeed showed a poorer performance on the verbal memory subtests than patients with right hemispheric lesions, but the difference was not as large as expected. However, these findings are consistent with results of Hochstenbach et al1 showing that left hemispheric lesion side predicts poorer performance on the RAVLT subtests of immediate recall, delayed recall, and recognition at 3 months after stroke than right hemispheric lesion side. In our study, first the influence of demographic and clinical characteristics was removed. Of the latter variables,

STROKE LESION AND VERBAL AND VISUAL MEMORY DEFICITS

135

Table 4. Increase of explained variance and significance level after each step in the hierarchic regression analyses of Table 3

I. Demographics II. Clinical characteristics III. Lesion characteristics

RAVLTImmediate recall

RAVLTDelayed recall

RAVLTRecognition

Doors TestRecognition

R2 change (P)

R2 change (P)

R2 change (P)

R2 change (P)

.41 (,.001) .21 (,.001) .09 (.002)

.38 (,.001) .18 (.002) .11 (.001)

.21 (.002) .24 (.001) .10 (.01)

.27 (,.001) .17 (.01) .03 (.39)

Abbreviations: RAVLT, Rey Auditory-Verbal Learning Test. See text for total explained variances.

in particular, naming ability and language comprehension may be of influence. Yet, lesion side was still a predictor of verbal immediate and delayed recall, and verbal recognition. The univariate analysis of variance controlling for age, sex, and education, however, showed a significant difference between left and right hemispheric patients on immediate and delayed recall. No such result was found for recognition of the RAVLT. An explanation for this might be that the active process of verbal recall is more dependent on verbal processes mediated by the left hemisphere than the more passive process of verbal recognition. On the other hand, the distribution of the patient recognition scores was skewed, although one cannot speak of ceiling effects, because not more than 17% of the patients had the maximum score. Nevertheless, one cannot rule out some influence from this fact. In our study, lesion side was not a predictor of visual recognition memory. This contradicts the classic association of verbal processes in the left hemisphere and visuospatial functioning in the right. However, it is consistent with recent results of Van Zandvoort et al,12 who found no significant effect of lesion side on visual recognition performance on the Doors Test in the first 3 weeks and at 1 to 2 years after stroke, and neither was an association found in the acute phase of stroke by Nys et al.20 It may not be excluded this was due to the limited range of scores on the Doors Test (especially in our study, using only set A), although the sensitivity of the test has clearly been established.62-64 Alternatively, it has been proposed that in addition to verbal function, the dominant left hemisphere mediates generalized cognitive activities other than language, thereby playing a role in visuospatial memory as well.2,3,12 Yet another explanation might be that the Doors Test supposes an unintended assessment of verbal (memory) function as well, by means of unaware verbal encoding or verbal mediation of the pictures. Although the doors are not usefully nameable, it is suggested that certain distinctive features may be encoded verbally as well. From these results and ours it can be assumed that visual memory performance is sensitive for both left and right hemispheric damage, whereas verbal memory function is mainly sensitive for left hemispheric lesions.5,6,14

Lesion Level It was found that lesion level predicted verbal memory function. It was a moderate predictor of both recall measures and recognition. We found that patients with subcortical lesions performed less well on verbal memory subtests than patients with cortical lesions. These results seem contradictory to common sense at first glance. This is in contrast with the results of Nys,13 who found more cognitive dysfunctions after cortical rather than subcortical stroke, but it is in line with the study by Wagner and Cushman,69 who found that patients with subcortical vascular lesions have more verbal memory impairments than patients with cortical lesions. Previous studies2,14,32 have proposed that cognitive impairments after subcortical infarction could be explained on the basis of frontal lobe dysfunction. They assumed that small subcortical infarctions may disrupt corticocortical association pathways, thalamocortical pathways, or corticostriatal pathways, producing frontal cortical disconnection. Memory functions are mediated by various neural networks of widely distributed cortical and subcortical areas and their multiple, reciprocal interconnections. A strategically positioned subcortical lesion may disrupt an entire neuronal circuit and its memory processing, provided that other structures cannot take over the function of the damaged area. If it is assumed that more widespread functional zones of the cortex have more options to restore cognitive processing through neural plasticity; a subcortical lesion might have a larger impact on memory function than a cortical lesion. Nevertheless, the finding that subcortical patients perform less well on the RAVLT than cortical patients could make sense, if one respects the assumption that cortical lesions produce more aphasia than subcortical lesions. Following the second step in the regression analysis, we controlled for the influence of aphasia (Table 3). One possible explanation is that additional variance in verbal memory performance could be predicted by left hemispheric subcortical rather than cortical lesions. No significant effect of lesion level was found in our study with regard to visual memory performance. This result is consistent with findings of Lunge et al70 showing no differences in visual memory performance between

E.A. SCHOUTEN ET AL.

136

patients with cortical and subcortical lesions. Unfortunately, the more precise location of lesions within subcortical areas (e.g., hippocampus and thalamus) could not be incorporated as a variable in our study, due to small sample size. Likewise, lesion location within the cortex was not entered into the regression model. However, Nys13 demonstrated that cortical lesions located in the temporal lobes hampered recovery of verbal memory in the first year of stroke, whereas lesions of the temporal and temporo-occipital lobes hampered recovery of visual memory. The study of visual memory in our study is limited in that the Doors Test assesses visuospatial recognition but not recall. Further research is needed to clarify the influence of specific cortical and subcortical sites on verbal and visual memory functioning in patients with stroke.

Lesion Volume Lesion volume was a predictor of verbal memory but not of visuospatial memory. Larger lesions produced poorer verbal memory performance. Previous studies examining the effect of lesion volume on cognitive functioning showed different results. Corbett et al32 also found a correlation (although weak) between lesion volume and verbal recognition performance. However, both Hochstenbach et al34 in their study on stroke in the basal ganglia and Nys et al20 in their study on acute cognitive impairment after stroke failed to demonstrate a significant effect of lesion volume on memory. Further research is needed into the influence of lesion volume on poststroke memory impairment. One of the strengths of our study was the relatively large sample size, compared with other (follow-up) studies investigating the long-term outcomes of cognitive function in patients with stroke. Also, we corrected for the influence of poststroke language and visuospatial deficits. A limitation of our study was that the influence of other possible determinants of poststroke memory performance, such as lesion location within cortical and subcortical areas, could not be investigated due to reduction of sample size. The influence of other poststroke cognitive impairments on memory performance was not studied either, such as deficits of attention and visual perception. Furthermore, our results might be limited by the fact that the Doors Test used to assess (visual) memory performance, assesses the visuospatial recognition process only. Thus, learning, active, immediate, and delayed recall of visual material were not assessed. Because different structures are involved in encoding rather than in retrieval, results of this study with regard to visual recognition cannot be generalized to processes of visual recall. Further research should include tests that measure more active visual recall processes in addition to passive recognition of learned visual material. In conclusion, our study demonstrated that the lesion characteristics of hemispheric lesion side, cortical and subcortical level, and lesion volume are weak to moderate

predictors of verbal memory performance 1 year after stroke. Visual memory performance could not be predicted from these lesion characteristics. It is recommended that further research with larger samples, more clinical variables, and newer imaging techniques to visualize real-time neurologic activity during memory testing be performed so that understanding of the neural mechanisms underlying poststroke memory dysfunction can be improved. Acknowledgment: We would like to thank Theo D. Witkamp, MD, neuroradiologist, for his help in analyzing the MRI scans.

References 1. Hochstenbach J, Mulder T, Van Limbeck J, et al. Cognitive decline following stroke: A comprehensive study of cognitive decline following stroke. J Clin Exp Neuropsychol 1998;20:503-517. 2. Madureira S, Guerreiro M, Ferro JM. Dementia and cognitive impairment three months after stroke. Eur J Neurol 2001;8:621-627. 3. Desmond DW, Moroney JT, Sano M, et al. Recovery of cognitive function after stroke. Stroke 1996;27:1798-1803. 4. Hochstenbach J, Donders R, Mulder T, et al. Long-term outcome after stroke: A disability-orientated approach. Int J Rehabil Res 1996;19:189-200. 5. Brown GG. Eyler Zorrilla LT. Neuropsychological aspects of stroke. In: Waldstein SR, Elias MF, eds. Neuropsychology of cardiovascular disease. London: Lawrence Erlbaum Associates, 2001:301-324. 6. Schnider A, Landis T. Memory loss. In: Caplan L, ed. Stroke syndromes. New York: Cambridge University Press, 1995:145-150. 7. Lezak MD, Howieson DB, Loring DW. Neuropsychological assessment. New York: Oxford University Press, 2004. 8. Hochstenbach J. Rehabilitation is more than functional recovery. Disabil Rehabil 2000;22:201-204. 9. Hochstenbach JB, Den Otter R, Mulder TW. Cognitive recovery after stroke: A 2-year follow-up. Arch Phys Med Rehabil 2003;84:1499-1504. 10. Rasquin SMC, Lodder J, Ponds RWHM, et al. Cognitive functioning after stroke: A one-year follow-up study. Dement Geriatr Cogn Disord 2004;18:138-144. 11. Sachdev PS, Brodaty H, Valenzuela MJ, et al. Progression of cognitive impairment in stroke patients. Neurology 2004;63:1618-1623. 12. Van Zandvoort MJE, Kessels RPC, Nys GMS, et al. Early neuropsychological evaluation in patients with ischemic stroke provides valid information. Clin Neurol Neurosurg 2005;107:385-392. 13. Nys GMS. The neuropsychology of acute stroke: Characterization and prognostic implications. Utrecht: Utrecht University, 2005. 14. Ferro JM. Neurobehavioral aspects of deep hemisphere stroke. In: Caplan L, ed. Stroke syndromes. New York: Cambridge University Press, 1995:151-161. 15. Kelley WM, Miezin FM, McDermott KB, et al. Hemispheric specialization in human dorsal frontal cortex and medial temporal lobe for verbal and nonverbal memory encoding. Neuron 1998;20:927-936. 16. Golby AJ, Poldrack RA, Brewer JB, et al. Material-specific lateralization in the medial temporal lobe and prefrontal

STROKE LESION AND VERBAL AND VISUAL MEMORY DEFICITS

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33. 34.

35.

cortex during memory encoding. Brain 2001;124: 1841-1854. Buckner RL, Raichle ME, Miezin FM, et al. Functional anatomic studies of memory retrieval for auditory words and visual pictures. J Neurosci 1996;16:6219-6235. Gur RC, Ragland JD, Mozley LH, et al. Lateralized changes in regional cerebral blood flow during performance of verbal and facial recognition tasks: Correlations with performance and ’effort.’ Brain Cogn 1997; 33:388-414. Herath P, Kinomura S, Roland PE. Visual recognition: Evidence for two distinctive mechanisms from a PET study. Hum Brain Map 2001;12:110-119. Nys GMS, Van Zandvoort MJE, De Kort PLM, et al. Domain-specific cognitive recovery after first-ever stroke: A follow-up study of 111 cases. J Int Neuropsychol Soc 2005;11:795-806. Daselaar SM, Veltman DJ, Witter MP. Common pathway in the medial temporal lobe for storage and recovery of words as revealed by event-related functional MRI. Hippocampus 2004;14:163-169. Kirwan CB, Stark CEL. Medial temporal lobe activation during encoding and retrieval of novel face-name pairs. Hippocampus 2004;14:919-930. Tsukiura T, Fujii T, Takahashi T, et al. Medial temporal lobe activation during context-dependent relational processes in episodic retrieval: An fMRI study. Hum Brain Mapp 2002;17:203-213. Mayes AR. Selective memory disorders. In: Craik FIM, ed. The Oxford handbook of memory. New York: Oxford University Press, 2000:427-440. McClelland JL, McNaughton BL, O’Reilly RC. Why there are complementary learning systems in the hippocampus and neocortex: Insights from the successes and failures of connectionist models of learning and memory. Psychol Rev 1995;102:419-457. Squire LR, Alvarez P. Retrograde amnesia and memory consolidation: A neurobiological perspective. Curr Opin Neurobiol 1995;5:169-177. Squire LR, Clark RE, Knowlton BJ. Retrograde amnesia. Hippocampus 2001;11:50-55. Nyberg L, Cabeza R. Brain imaging of memory. In: Craik FIM, ed. The Oxford handbook of memory. New York: Oxford University Press, 2000:501-519. Tranel D, Damasio H, Damasio AR. Amnesia caused by herpes simplex encephalitis, infarctions in basal forebrain, and anoxia/ischemia. In: Grafman J, ed. The handbook of neuropsychology. Amsterdam: Elsevier, 2000:85-110. Tulving E, Kapur S, Craik FIM, et al. Hemispheric encoding/retrieval asymmetry in episodic memory: Positron emission tomography findings. Proc Natl Acad Sci U S A 1994;91:2016-2020. Heun R, Klose U, Jessen F, et al. Functional MRI of cerebral activation during encoding and retrieval of words. Hum Brain Mapp 1999;8:157-169. Corbett A, Bennett H, Kos S. Cognitive dysfunction following subcortical infarction. Arch Neurol 1994; 51:999-1007. Exner C, Weniger G, Irle E. Implicit and explicit memory after focal thalamic lesions. Neurology 2001;57:2054-2063. Hochstenbach J, Van Spaendonck KPM, Cools AR, et al. Cognitive deficits following stroke in the basal ganglia. Clin Rehabil 1998;12:514-520. Middleton FA, Strick PL. Basal ganglia output and cognition: Evidence from anatomical, behavioral, and clinical studies. Brain Cogn 2000;42:183-200.

137

36. Van Der Werf YD, Witter MP, Uylings HBM, et al. Neuropsychology of infarctions in the thalamus: A review. Neuropsychologia 2000;38:613-627. 37. Van Der Werf YD, Tisserand DJ, Visser PJ, et al. Thalamic volume predicts performance on tests of cognitive speed and decreases in healthy aging: A magnetic resonance imaging-based volumetric analysis. Cogn Brain Res 2001;11:377-385. 38. Van Der Werf YD, Scheltens P, Lindeboom J, et al. Deficits of memory, executive functioning and attention following infarction in the thalamus; a study of 22 cases with localized lesions. Neuropsychologia 2003;41:1330-1344. 39. Schiemanck SK, Post MWM, Kwakkel G, et al. Ischemic lesion volume correlates with long-term functional outcome and quality of life of middle cerebral artery stroke survivors. Restorative Neurol Neurosci 2005;23:257-263. 40. World Health Organization. Recommendations on stroke prevention, diagnosis, and therapy: Report of the WHO task force on stroke and other cerebrovascular disorders. Stroke 1989;20:1407-1431. 41. Bamford JM, Warlow CP. Evolution and testing of the lacunar hypothesis. Stroke 1988;19:1074-1082. 42. Fisher CM. Lacunar strokes and infarcts: A review. Neurology 1982;32:871-876. 43. Schiemanck SK, Kwakkel G, Post MWM, et al. Predictive value of ischemic lesion volume assessed with magnetic resonance imaging for neurological deficits and functional outcome poststroke: A critical review of the literature. Neurorehabil Neural Repair 2006;20:492-502. 44. Brott T, Adams HP Jr, Olinger CP, et al. Measurements of acute cerebral infarction: A clinical examination scale. Stroke 1989;20:864-870. 45. Rocca B, Martin C, Viviand X, et al. Comparison of four severity scores in patients with head trauma. J Trauma 1989;29:299-305. 46. Verhage F. Intelligentie en leeftijd: Onderzoek bij Nederlanders van twaalf tot zevenenzeventig jaar [Intelligence and age: Study with Dutch people from age 12 to 77]. Assen: Van Gorcum 1964. 47. Mahoney F, Barthel D. Functional evaluation: The Barthel index. Md Med J 1965;14:61-65. 48. De Haan R, Limburg M, Schuling J, et al. Klinimetrische evaluatie van de Barthel-index, een maat voor beperkingen in het dagelijks functioneren [Clinimetric evaluation of the Barthel index, a measure of limitations in activities of daily living]. Ned Tijdschr Geneeskd 1993;137:917-921. 49. Folstein MF, Folstein SE, McHugh PR. ’Mini mental state’: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12:189-198. 50. Galasko D, Klauber MR, Hofstetter CR, et al. The minimental state examination in the early diagnosis of Alzheimer’s disease. Arch Neurol 1990;47:49-52. 51. Radloff LS. The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas 1977;1:385-401. 52. Brand AN, Jolles J, Gispen-de Wied C. Recall and recognition memory deficits in depression. J Affect Disord 1992;25:77-86. 53. Van Harskamp F, Van Dongen HR. Construction and validation of different short forms of the token test. Neuropsychologia 1977;15:467-470. 54. Kaplan E, Goodglass H, Weintraub S. The Boston naming test. Philadelphia: Lea and Febiger, 1983. 55. De Renzi E, Vignolo LA. The Token test; a sensitive test to detect receptive disturbances in aphasics. Brain 1962; 85:665-678.

138 56. Van Dongen HR, Van Harskamp F, Verhey Stol FW, et al. Afasie onderzoek met de Tokentest: Enige psychometrische kenmerken en evaluatie van een verkorte vorm [Study of aphasia with the Token test: Some psychometric features and evaluation of an abbreviated form]. Ned Tijdschr Psychol 1974;28:633-647. 57. Benton AL, Varney NR, Hamsher K. Visuospatial judgment: A clinical test. Arch Neurol 1978;35:364-367. 58. L’ Rey A. examen clinique en psychologie. Paris: Universitaires de France, 1964. 59. Saan RJ, Deelman BG. The 15-Words test A and B (A preliminary manual). Groningen: Department of Neuropsychology, Academic Hospital, 1986. 60. Deelman BG, Saan RJ. Memory deficits: Assessment and recovery. In: Van Zomeren AH, ed. Traumatic brain injury; clinical, social and rehabilitational aspects. Amsterdam: Swets and Zeitlinger, 1990. 61. Baddeley AD, Emslie H, Nimmo-Smith I. The doors and peoples test: A test of visual and verbal recall and recognition. Bury St Edmunds: Thames Valley Test Company 1994. 62. Morris RG, Abrahams S, Baddeley AD, et al. Doors and people: Visual and verbal memory after unilateral temporal lobectomy. Neuropsychology 1995;9:464-469.

E.A. SCHOUTEN ET AL. 63. Kemps E, Newson R. Comparison of adult age differences in verbal and visuo-spatial memory: The importance of ‘‘pure,’’ parallel and validated measures. J Clin Exp Neuropsychol 2006;28:341-356. 64. Davis C, Bradshaw CM, Szabadi E. The doors and people memory test: Validation of norms and some new correction formulae. Br J Clin Psychol 1999;38:305-314. 65. Damasio H, Damasio AR. Lesion analysis in neuropsychology. New York: Oxford University Press, 1989. 66. Van Swieten JC, Hijdra A, Koudstaal PJ, et al. Grading white matter lesions on CT and MRI: A simple scale. J Neurol Neurosurg Psychiatry 1990;53:1080-1083. 67. Cohen J. Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988. 68. Tombaugh TN, McIntyre NJ. The mini-mental state examination: A comprehensive review. J Am Geriatr Soc 1992;40:922-935. 69. Wagner MT, Cushman LA. Intellectual and memory functions after cortical and subcortical stroke. Neurorehabilitation 1992;2:45-52. 70. Lunge G, Waked W, Kirshblum S, et al. Organizational strategy influence on visual memory performance after stroke: Cortical/subcortical and left/right hemisphere contrasts. Arch Phys Med Rehabil 2000;81:89-94.