Decision strategies in neuropsychology III: The relationship among lateralized dysfunction, etiology and depression

ofClinical

Archives Neuropsychology, Vol. 5. pp. 341-358, Printed in the USA. All rights reserved.

1990 Copyright

0 1990 National

0887-6177190 $3.00 + .OO Academy of Neuropsychology

Decision Strategies in Neuropsychology III: The Relationship Among Lateralized Dysfunction, Etiology and Depression Leslie K. Ross and Charles J. Long University

of Tennessee-Memphis

State University,

Neuropsychology

Lab

The present study examines depression as a function of lateralized dysfunction and etiology. The proposed hypothesis is: patients with left hemisphere lesions would be depressed, whereas patients with right hemisphere lesions would not. A total of 61 patients (26 right hemisphere, 35 left hemisphere) having either (I stroke, tumor, or penetrating head wound, completed the MMPI and were given a neuropsychological evaluation. Lesion localization was based on electroencephalography and neuroradiological procedures. A regression formula derived from the MMPI assessed depression. E-vo analyses were performed between: (i) right and left lesioned groups of mixed etiology and a control group, and (ii) right and left cerebrovascular lesioned groups and a control group. Potentially confounding factors, such as, diagnosis, demographic factors, degree of neuropsychological impairment, presence or absence of aphasia, and lesion localization were examined. Regardless of etiology, left and right lesioned patients did not significantly differ in their depression score, nor was either group considered depressed based on the criteria used in this study to measure depression.

While it is generally accepted that neurological damage can have both cognitive and emotional effects, this relationship remains poorly understood. Part of the controversy in this area pertains to whether the concommitant emotional effects represent a psychological reaction by the individual to their deficits or whether they are based on pathophysiological changes. This study will examine the later aspect, in terms of site of lesion and its relative contribution to alterations in affect.

Requests for reprints Memphis, TN 38152.

should

be sent to Charles

347

J. Long Ph.D.,

Memphis

State University,

348

L. K. Ross and C. J Long

Many studies have looked at the issue of changes in emotional characteristics as a function of lesion location and have arrived at conflicting conclusions. Considerable research asserts that left hemisphere (LH) damage as opposed right hemisphere (RH) damage leads to a depressive state (Black, 1975; Gainotti, 1972; Gasparrini, Satz, Heilman, & Coolidge, 1978; Robinson, 1986; Robinson, Kubos, Starr, Rao, & Price, 1984; Robinson & Price, 1982; Robinson, Starr, Kubos, & Price, 1983). Some researchers consider intrahemispheric lesion location an important factor (Robinson, 1986; Robinson, Kubos, Starr, Rao, & Price, 1984; Robinson & Szetela, 1981; Sinyor, Jacques, Kaloupek, Becker, Goldenberg, & Coopersmith, 1986). Robinson and his associates conclude the severity of depression is greater for left frontal than for left occipital strokes. Sinyor et aI. (1986) found a trend for the severity of depression to be greater with increasing proximity of the lesion to the frontal pole of either hemisphere. In addition, they asserted that posterior lesions of the RH were separately associated with higher depression scores. Lastly, others have failed to observe interhemispheric differences in depression (Dikeman & Reitan, 1974; Gass & Russell, 1986; Williams, Little, & Klein, 1986). For example, Finklestein et al. (1982), did not find interhemispheric differences when measuring depression by both behavioral methods and the dexamethasone suppression test. Several methodological factors along with variations in sampling characteristics, such as, age and educational level of the patients, presence or absence of aphasia, etiology, level of cognitive impairment, and/or differing measures used for the assessment of depression, might be contributing to the inconclusiveness of research findings in this area. Premorbid factors, such as age and education, play a role in a patient’s psychological adjustment to brain damage and its functional consequences; therefore, they need to be taken into consideration when investigating emotional sequelae following brain damage (Gass & Russell, 1987; Lezak, 1983). It has been demonstrated, by correlation analysis, significant relationships exist between cognitive impairment, functional physical impairment, quality of social support, and age in their contribution to depression (Robinson, Bolla-Wilson, Kaplan, Lipsey, & Price, 1986; Robinson, Starr, Kubos, & Price, 1983). Many of the studies cited in this paper do not adequately control for these variables. It can be concluded that the discrepancies among the studies in this area are probably multifactorial, interacting in a manner that compounds differences in the subject population exponentially rather than linearly. An important consideration when studying the effects of lesion location on depression, is the measure used to assess depression. A variety of techniques have been employed ranging from behavioral observation and verbal expressions, to one of the many psychometric measures of depression. Frequently used scales are the MMPI and Zung Self-Rating Depression Scale.

Depression and L&era&

349

A problem encountered when using the MMPI is exemplified by a recent study by Moehle and Fitzhugh-Bell (1988) concerning emotional disturbance and laterality of brain damage. A significant difference was not found between the RH and LH groups. The criteria employed to assess emotional functioning was based on grouped MMPI profiles in order to compare the two groups. Previous research has argued against the process of averaging different profiles because it tends to mask individual differences and affords little understanding of the personality characteristics of the individu~ patient (Long, 1981). In order to avoid the pitfalls of averaging techniques, the present study computes an index of depression for each individual patient, using a regression formula derived from the MMPI. The purpose of the study from which the formula is based was to be able to differentiate depression from other psychiatric conditions (Mezzich, Damarin, dz Erickson, 1974). The formula used in the present study was derived by Mezzich et al. for diagnostic purposes. Post and Lobitz (1980), using the depression formula to discriminate depressive from nondepressive syndromes, obtained a hit rate of 71% with respect to agreement between psychiatric diagnosis and the depression formula. The MMPI-D scale score alone accurately classified only 65 % of the patients. Post and Lobitz noted that the primary advantage of the depression formula is in eliminating ““false positive” classifications. In addition to eliminating ““false-positives” obtained by the MMPI-D scale, the depression formula discriminated depression more effectively than the Zung SelfRating Depression Scale. Using the derived depression formula of Mezzich et al. (1974), the present study examines the relationship of cerebral asymmetry and depression as a function of lesion location in a mixed etiological group (i.e., tumor, cerebrovascular, trauma), and in a second group consisting only of patients with cerebrovascular accidents. None of the patients were diagnosed as depressed according to DSM-III or the Research Diagnostic Criteria. Depression in the present patient population was based on a regression score composed of self-reported symptomatology. The second study was undertaken because studies which have consistently found differences in the degree of depression between the LH and RH examined only those patients who have experienced cerebrovascular accidents (Robinson, 1986; Robinson, Kubos, Starr, Rao, & Price, 1984; Robinson & Price, 1982; Robinson, Starr, Kubos, & Price, 1983; Robinson & Szetela, 1981). It is concluded from these prior studies that: (i) individuals with LH damage are more likely to be depressed and the depression is greater than patients with RH damage, and (ii) stroke patients are more depressed, for both severity and frequency, than traumatic brain-injured patients. The hypothesis for both studies, in terms of self-reported symptomotalo-

L. K. Ross and C. J. Long

350

gy, is that patients with left hemisphere lesions would be depressed, whereas patients with right hemisphere lesions would not. Potentially confounding factors, such as etiology, demographic factors, degree of neuropsychological impairment, presence or absence of aphasia, and lesion localization are taken into consideration by statistical methods. EXPERIMENT

1

Method Subjects. The subjects in this study were selected from larger pool of patients who were primarily noninstitutionalized medical patients referred for neuropsychological testing by neurosurgeons and other physicians over the past 17 years. Right hemisphere (RH) or left hemisphere (LH) damage was determined by nonneuropsychological criteria (e.g., both electroencephalography, and neuroradiological procedures). Using these two procedures, only patients with positive findings involving only one hemisphere were included in the experimental groups. These patients completed the MMPI and were given a modified Halstead-Reitan Neuropsychological evaluation which excludes the Category Test but includes other tests - Wechsler Memory Scale, Percent Delayed Wechsler Memory Score, Trails C (alphabet), Word Fluency. The test scores used in computing the modified impairment index are listed in Table 1. Pseudoneurologic controls were composed from referrals for neuropsychological testing who did not evidence brain damage on either the EEG and the CT Scan data. Also, subjects included in the control group did not experience head trauma or have severe emotional problems. The selection procedure yielded 35 LH-damaged subjects having a mean TABLE 1 Cut-off Scores for Modified Impairment Test and Cutoff

Index

for Impairment

Speech Perception Test > 7 Seashore Rhythm Test > 5 Finger Tapping < 51 Trails A + Trails B > 136 seconds TPT Total Time > 15.6 TPT Memory < 6 TPT Location < 5 Impairment

Index

=

No. of tests within impaired range No. of tests administered

Depression and Laterality

351

age of 42.1 years and a mean education of 12.7 years, and 26 RH subjects with a mean age of 50.5 years and a mean education of 13.4 years. The results are depicted in Table 2. Of the LH subjects there were 13 anterior lesions, 10 medial lesions, and 11 posterior lesions. The RH subjects had 12 anterior lesions, 4 medial lesions, and 10 posterior lesions. The etiology of the disorders in addition to the lesion location within the LH and RH groups are depicted in Table 3. Procedure.

A depression score, using the regression formula derived by Mezzich et al. (1974), was computed for each subject. The depression score is obtained from the raw scores of five scales from the MMPI: Depression = 10D - 5Pd + 4Pa - 4Sc + 2Si Figure 1 shows the mean depression score for each of the groups. A multiple regression analysis was chosen to analyze the data. Also, the effects of variables thought to be a factor in depression were taken into consideration before examining the influence of lesion location on depression. Using the SPSS/PC+ statistical package (Norusis, 1986) a regression was performed having age, education, verbal IQ (VIQ), performance IQ (PIQ), the modified impairment index, the Aphasia Screening Test (weighted scores; Russell, Neuringer, & Goldstein, 1970), and etiology as the independent variables with the depression score as the dependent variable. The default values of the SPSS/PC + were not modified. TABLE 2 Descriptive Data-Mixed Left

(n = 35) Variables

Mean

Etiologies

Right (n = 26)

SD

Mean

17.53 3.37 17.53 20.19 20.65 0.27 9.76

15 11 50.46 13.44 95.38 101.36 87.63 0.76 3.65

Control (n = 193) SD

Mean

13.79 3.02 13.79 11.09 11.35 0.24 4.01

82 111 37.82 12.90 104.60 104.58 104.00 0.32 2.13

SD

F

Sex

Male (n) Female (n) Age Education FSIQ

VIQ PIQ IIndex AST

23 12 42.12 12.65 92.64 94.35 90.29 0.65 8.60

IIndex = Impairment Index; AST = Aphasia Screening + +Chi Square (LH, RH, Controls) = 7.68*; + +Chi Square (LH, RH) = 0.71 ns; *p < 0.05;

**p < 0.01.

Test weighted

++ 13.91 2.62 13.91 13.31 12.31 0.23 2.55 score;

10X** 1.19 57.98** 7.5a** 24.98* 57.98** 30.3 1*

352

L. K. Ross and C. .J Long TABLE 3 Etiology Categories Lesion Location: Left Hemisphere and Right Hemisphere Groups

Left (n)

Right (n)

Chi-square

Tumor Vascular Trauma

13 15 I

12 10 4

0.542 ns

Anterior Medial Posterior

14 10 11

12 4 10

1,477 ns

Gr0llp

Orthogonal coding of the groups was constructed to test the hypothesis: LH-damaged patients are more depressed than RH-damaged patients. Specific comparisons between the groups were decided upon prior to the analysis, therefore, if the overall regression null hypothesis is nonsignificant, the specific comparisons can still be interpreted. ?hro orthogonal vectors were constructed: (i) comparing the LH and RH groups to the controls, and (ii) comparing the LH group to the RH group, which was entered last into the regression equation. Using a stepwise regression analysis age, education, VIQ, PIQ, the im-

200

1

Subject FIGURE 1. Mixed etiology: depression groups and the control group.

Group

Classification

ratings for the right hemisphere and the left hemisphere

353

Depression and Laterality

pairment index, the Aphasia Screening Test weighted scores, and etiology were initially entered into the equation in order to account for the variance these factors might explain in the depression score. Based on a multivariate analysis, the LH and RH groups differed significantly from the controls on age, PIQ, FSIQ, and the impairment index, but the two lateralized groups did not significantly differ from each other on these variables. The control group and the RH group significantly differed from the LH subjects on VIQ and the Aphasia Screening Test (AST). The demographic and descriptive data of the tests used in the regression analysis for the RH, LH, and control subjects is listed in Table 2. Age, explaining 2.5% (R2 = 0.025) of the variance in the depression formula, was the only variable included in the regression formula after the stepwise procedure of initially entering the seven potentially confounding variables (i.e. age, education, VIQ, PIQ, impairment index, AST, etiology) (see Table 4). The analysis of variance testing the null hypothesis of the R at this step was significant (F = 5.14, p < 0.024). Since regression analysis is based on correlation, the significance in such a small amount of explained variance is probably due to the large number of patients in the analysis. The orthogonal vector, comparing the LH and RH groups on the depression score, did not significantly increase the amount of explained variance in the depression score (F = 3.01, p < 0.084). At the last step in the regression procedure the independent variables accounted for 4.58% (R2 = 0.0458) of the variance in the depression score, which was significant (F = 3.17, p < 0.026). The t-test comparing the difference between the LH and RH group on the depression score was nonsignificant (t = 1.736, p < 0.084). Similarly, the t-test comparing the control group to the RH and LH groups was nonsignificant (t = 1.098, p < 0.273). In light of the above results, an analysis of covariance, covarying for the Results.

Regression Entered Stepwise Method Age, educ, VIQ, PIQ, IIndex, AST, etiology Forced Entry Method Comparison 1 Comparison 2 (Constant) *p < 0.05.

TABLE 4 Analysis Summary variable

R2

R2 change

Coefficient

Age

0.025*

0.025*

0.780

LH + RH vs. controls LH vs. RH

0.031* 0.046*

0.0062 0.015

0.068 1.113 83.46

354

L. K. Ross and C. J. Long

effects of age, was performed between the three groups and their depression score. The results of this analysis was nonsignificant for the main effect of group (F = 1.58, p < 0.21). EXPERIMENT

2

Method Subjects. The same subject

selection and criteria that was used in study 1 was used for this study. The difference between the previous study and the present study is the exclusion of all etiological groups except for those subjects having a cerebrovascular accident. The pseudoneurological control group remained the same. The subject selection procedure yielded 15 LH-damaged subjects having a mean age of 52.86 years and a mean education of 13.6 years, and 10 RH subjects with a mean age of 52.0 years and a mean education of 14.8 years (see Table 5). Of the LH subjects there were 4 anterior lesions, 5 medial lesions, and 6 posterior lesions. The RH subjects had 3 anterior lesions, 2 medial lesions, and 5 posterior lesions.

Procedure. The procedure

used in this study is the same as that used in the previous analysis except the etiology variable was not entered into the regression analysis. The mean depression scores for the RH and LH cerebrovascular patients and controls are presented in Figure 2.

Descriptive

TABLE 5 Data-Stroke Right (n = 10)

Left (n = 15) Variables +Sex Male (n) Female (n) Age Education FSIQ

vIQ PIQ IIndex AST

Mean

SD

52.86 13.60 94.93 95.35 93.67 0.66 11.67

9 6 15.06 3.16 23.28 24.15 22.29 0.28 12.72

Patients

Mean

SD

52.00 14.80 95.70 1105.80 84.50 0.81 1.50

6 4 12.59 3.52 7.83 10.74 10.37 0.20 1.84

IIndex = Impairment Index; AST = Aphasia (Sex X lesion location) = 3.47 ns; *p < 0.05; **p < 0.01.

Screening

Control (n = 193) Mean

SD

F

37.82 12.90 104.60 104.58 104.00 0.32 2.13

82 111 13.91 2.62 12.44 13.31 12.31 0.23 2.55

14.49** 2.51 7.27** 4.60* 15.55** 39.62** 34.46*

Test weighted

score;

+Chi

Square

Depression and Laterality

LEFr

355

RIGHT HEMSF’HE!=IE

HEbtSPHERE Subject

Group

Classlflcatlon

FIGURE 2. Cerebrovascular subjects: depression ratings for the right hemisphere and the left hemisphere groups and the control group.

Results A multivariate analysis was performed on the demographic and descriptive variables in Table 5. The LH and RH groups differed significantly from the controls on age, PIQ, and the impairment index. The LH group, but not the RH group, differed from the controls on FSIQ. The control group and the RH group significantly differed from the LH subjects on VIQ and the AST. Age, explaining 2.6% (Rz = 0.026) of the variance in the depression formula, was the only variable, of the initial six (i.e., age, education, VIQ, PIQ, impairment index, AST), included in the regression equation after the stepwise procedure (see Table 6). The analysis of variance testing the null hypothesis of the R at this step was significant (F = 4.77, p < 0.03). As previously mentioned, since regression analysis is based on correlation, the significance in such a small amount of explained variance is probably due to the large number of patients in the analysis. The orthogonal vector, comparing the LH and RH groups on the depression score, did significantly increase the amount of explained variance in the depression score by 2.9% (F = 5.36, p c 0.02). At the last step in the regression procedure the independent variables accounted for 5.74% (RZ = 0.0574) of the variance in the depression score which was significant (F = 3.60, p < 0.015). The t-test comparing the difference between the LH and RH group on the depression

356

L. K. Ross and C. J. Long TABLE 6 Regression Analysis Summary-Stroke Entered

Stepwise Method age, educ, VIQ, PIQ, IIndex, AST Forced Entry Method comparison 1 comparison 2 (constant)

variable

Patients R2

R2 change

age

0.026*

0.026*

0.831

LH + RH vs. Controls LH vs. RH

0.029* 0.057*

0.003 0.029*

0.078 5.98 82.87

Coefficient

*p < 0.05.

score was significant (t = 2.31, p < 0.02). However, the t-test comparison of the control group versus both the RH and LH groups on the depression score was nonsignificant (t = 0.73, p < 0.467). An analysis of covariance was performed between the three groups and their depression score, covarying for the effects of age. The results of this analysis approached significance (F = 2.95, p < 0.055).

Discussion The results of this study fail to support previous findings reporting differences between LH and RH damage patients on measures of depression and lends further support to research expounding that lateralized differences do not exist for depression secondary to brain damage. It should be noted, the MMPI profiles of the three groups were not indicative of depression nor did their depression scores fall within the depressed range [depression score > 190 (Post & Lobitz, 1980)]. Therefore, the statistical comparisons examined the relative amount of symptomatology endorsed by the three groups. Within the mixed etiologies groups, 9% (n = 2) of the LH and 12.5% (n = 2) of the RH patients had scores greater than 190. There were no LH patients and only 1% (n = 1) of the RH patients within the CVA group that had a score greater than 190. In contrast, 12.4% (n = 22) control patients had scores greater than 190. The severity of depression in the present study is in direct contrast to the depression severity Robinson and his colleagues (1986; 1984; 1983; 1982; 1981) found in their cerebrovascular cases. Although diagnostic classification of depression according to DSM-III criteria and Research Diagnostic Criteria was not examined in this study, Robinson and his colleagues using this classification scheme report that 50% of their patients have significant mood disturbances. While chronicity may be a factor, Robinson (1986) report that cerebrovascular patients are vulnerable to major or minor de-

Depression and Luterality

357

pression up to two years post-stroke. He also reported that in 70% of patients diagnosed with dysthymic depression, the condition persisted for more than two years. Based on the results of the present study, it cannot be concluded that either hemisphere is more prone to a depressive condition. None of the patient populations could be considered as depressed based on the measure of depression employed in this study. Even though the regression equation itself was significant in both studies, it explained a total of 4.58% (mixed etiologies) and 5.74% (stroke etiology) of the variance in the depression score in study 1 and study 2, respectively. Since the regression procedure uses correlations, the more subjects in the analysis the greater the likelihood of achieving significance in the regression equation. Of interest is the difference in the depression scores in the two studies, suggesting etiology may be an important factor when considering hemispheric asymmetry and depression. The conclusions of this study should be interpreted with some caution. It is possible, given a brain-damage population that exhibits depression, lesion location may be an important factor. However, it is felt the patient population in this study is representative of a brain-damage population in general. Differences in the initial selection of patients may account for discrepancies between our lab and others. That is, the patients evaluated at the Neuropsychology Lab are outpatients referred by neurosurgeons and neurologists, whereas other labs might predominately evaluate in-patients. It might be that the depression effect is not as a significant factor as it may be for those patients still in the hospital or other settings. However, our lab is similar to most outpatient neuropsychological labs with neurosurgical referrals. Lastly, the experimental findings confirm our clinical impressions regarding the patients assessed at our lab, over the past 17 years, of not noticing differences in depressive behavior based upon lateralized lesions, per se. Depressive disorders in a brain damaged population are multifactorial. In the present study many factors were not considered (e.g., time since injury, lesion size, socio-economic status, etc.) which may contribute to a depressive reaction. While it is often difficult to obtain all the necessary data involved in a topic area, it is important to analyze the data in the most effective manner. The use of the regression procedure in this study examines the central issue: Are emotional characteristics in a brain damaged population a function of lesion location? Studies matching subjects on variables such as age, aphasia, cognitive impairment, and so on, do not necessarily account for the influence of these variables. The regression procedure takes into consideration these confounding variables in explaining the variance in the depression score and then examines the effect of lesion group on the depression score. The question in this analysis is: Does the inclusion of lesion group explain a significant amount of the variance in the depression score over and above that which has already been explained by the con-

358

L. K. Ross and C. .l Long

founding variables? The present results indicate lesion group does not contribute significantly to the depression score.

REFERENCES Black,

F. W. (1975). Unilateral

brain

lesions

and MMPI

performance:

A preliminary

study.

Perceptual and Motor Skills, 40, 87-93. Dikeman,

S., & Reitan,

R. M. (1974). MMPI

correlates

of localized

cerebral

lesions. perceptual

and Motor Ski& 39, 831-840. Finklestein, S., Benowitz, L. I., Baldessarini, R. J., Arana, G. W., Levine, D., Woo, W., Bear, D., Moya, K., & Stall, A. L. (1982). Mood vegetative disturbance and dexamethasone suppression test after stroke. Annals of Neurology, 12, 463-468. Gainotti, G. (1972). Emotional behavior and hemispheric side of the lesion. Cortex, 8, 41-55. Gasparrini. W. G., Satz, P., Heilman, K. M., & Coolidge, F. L. (1978). Hemispheric asymmetries of affective processing as determined by the Minnesota Multiphasic Personality Inventory. Journal of Neurology, Neurosurgery, and Psychiairy, 41,470~473. Gass, C. S., & Russell, E. W. (1986). Minnesota Multiphasic Personality Inventory correlates of lateralized cerebral lesions and aphasic deficits. Journal of Consulting and Clinical

Psychology, 54, 359-363. Gass, C. S., & Russell, E. W. (1987). MMPI correlates of performance intellectual deficits in patients with right hemisphere lesions. Journal of Clinical Psychology, 43,484-489. Lezak, M. D. (1983). Neuropsycho/ogjcu~assess~ent. New York: Oxford University Press. Long, C. J. (1981). The relationship between surgical outcome and MMPI profiles in chronic pain patients. Journal of Uinical Psychology, 37, 744-749. Mezzich, J. E., Damarin, F. L., & Erickson, J. R. (1974). Comparative validity of strategies and indices for differential diagnosis of depressive states from other psychiatric conditions using the MMPI. JournalofClinical Psychology, 42, 691-698. Moehle, K. A., & Fitzhugh-Bell, K. B. (1988). Laterality of brain damage and emotional disturbance adults. Archives of Cfinicaf ~europsycholo~y, 3, 133-144. Norusis, M. J. (1986). SFSS’PC-t. Chicago: SPSS, Inc. Post, R. D., & Lobitz, W. C. (1980). The utility of Mezzich’s MMPI regression formula as a diagnostic criterion in depression research. Journal of Consulting and Clinical Psychology,

48, 673-674. Robinson, R. G. (1986, April 15). Post-stroke mood disorders. Hospital Practice, 83-89. Robinson, R. G., Bolla-Wilson, K., Kaplan, E., Lipsey, J. R., & Price, T. R. (1986). Depression influences intellectual impairment in stroke patients. Brittrh Journal of Psychiatry, 148,

541-547. Robinson, R. G., Kubos, K. L., Starr, L. B., Rao, K., & Price, T. R. (1984). Mood disorders in stroke patients: Importance of location of lesion. Bruin, 107, 81-93. Robinson, R. G., & Price, T. R. (1982). Post-stroke depressive disorders: A follow-up study of 103 patients. Stroke, 13, 635-641. Robinson, R. G., Starr, L. B., Kubos, K. L., & Price, T. R. (1983). A two-year longitudinal study of post-stroke mood disorders: Findings during the initial evaluation. Stroke, 14, 736-741. Robinson, R. G., & Szetela, B. (1981). Mood change following left hemispheric brain injury.

Annals of Neurology, 9, 447-453. Sinyor, D., Jacques, P., Kaloupek, D. G., Becker, R., Goldenberg, M., & Coopersmith, H. (1986). Poststroke depression and lesion location: An attempted replication. Brain, 109, 537-546. Williams, J. M., Little, M. M., & Klein, K. (1986). Depression and hemispheric site of cerebra1 vascular accident. Archives of Cfinicat Neu;opsychology, 1, 393-398.