Journal of Clinical and Experimental Neuropsychology

ISSN: 0168-8634 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/ncen19

The subcortical dysfunction hypothesis of memory

deficits in depression: Neuropsychological
validation in a subgroup of patients

Paul J. Massman , Dean C. Delis , Nelson Butters , Renee M. Dupont & J.

Christian Gillin

To cite this article: Paul J. Massman , Dean C. Delis , Nelson Butters , Renee M. Dupont & J.
Christian Gillin (1992) The subcortical dysfunction hypothesis of memory deficits in depression:
Neuropsychological validation in a subgroup of patients, Journal of Clinical and Experimental
Neuropsychology, 14:5, 687-706, DOI: 10.1080/01688639208402856

To link to this article: http://dx.doi.org/10.1080/01688639208402856

Published online: 04 Jan 2008.

Submit your article to this journal

Article views: 62

View related articles

Citing articles: 66 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ncen20

Download by: [Central Michigan University] Date: 02 October 2015, At: 18:50
 Journal of Clinical and Experimental NmpsycholOgY 0168-8634/9U1405-M87$3.~
. NO.5, p ~a.7-706
1992, V O ~14, @ Swets & &itlinger

The Subcortical Dysfunction Hypothesis of Memory

Deficits in Depression:
Neuropsychological Validation
in a Subgroup of Patients*

Paul J. Massman, Dean C. Delis, Nelson Butters

Renee M. Dupont, J. Christian Gillin
Downloaded by [Central Michigan University] at 18:50 02 October 2015

San Diego Veterans Affairs Medical Center

University of California, San Diego, School of Medicine

ABSTRACT

The subcortical dysfunction hypothesis of verbal learning and memory deficits in

depression was evaluated by comparing the memory test profiles of unipolar
depressives (n=40) and bipolar depressives (n=9) with those of patients with a
prototypical subcortical dementia (Huntington’s disease, HD), patients with a
prototypical cortical dementia (Alzheimer’s disease, AD), and normal controls. In
a discriminant function analysis that well-differentiated the HD, AD, and normal
subjects, it was found that 28.6% of the depressed patients were classified as HD
patients (DEP-HD subjects), 49.0% were classified as normals (DEP-N subjects),
none were classified as AD patients, and 22.4% were not well-classified. The DEP-
HD group closely resembled the HD group on additional indices of verbal learning
and memory, and differed from the DEP-N group, which strongly resembled the
normal control group. DEP-N patients also performed significantly better than
DEP-HD patients on a number of other neuropsychological tests (e.g., WAIS-R
Digit Symbol, category fluency, Trail Making Test Part B). The findings provide
support for the subcortical dysfunction hypothesis, but only for a subgroup of
depressed patients. Implications for differentiating depressive “pseudodementia”
from AD are discussed.

* This study was supported by Mental Health Clinical Research Center Grant MH 30914,
grant MH 42575, and fellowship training grant MH 18399 from the National Institute of
Mental Health; by NIA grant AG-05131 to the University of California at San Diego; and
by the Department of Veterans Affairs Medical Research Service. We wish to thank
David Salmon for contributing subjects’ test results.
Paul J. Massman is now at the University of Houston, Department of Psychology and
Baylor College of Medicine, Department of Neurology.
Address correspondence to: Paul J. Massman, Ph.D., Department of Psychology, Univer-
sity of Houston, 4800 Calhoun Road, Houston, TX 77204-5341 USA.
Accepted for publication: December 10, 1991.
 688 PAUL J. MASSMAN ET AL.

Depressed patients have been reported to be impaired on numerous verbal memory

tasks (Danion et al., 1991; Hart, Kwentus, Taylor, & Harkins, 1987; King, Caine,
Conwell, & Cox, 1991 ;Kopelman, 1986; Weingartner, Cohen, Murphy, Martello,
& Gerdt, 1981; Weingarmer & Silberman, 1982; Wolfe, Granholm, Butters,
Saunders, & Janowsky, 1987), but their deficiencies do not resemble the storage
and encoding deficits typical of amnesic patients. Although depressives perform
poorly on verbal recall tasks, they often attain normal levels of performance
when recognition measures are employed (Dunbar & Lishman, 1984; Frith et al.,
1983; King et al., 1991; Lamer. 1977; Miller & Lewis, 1977; Silberman,
Weingartner, Laraia, Bymes, & Post, 1983; Watts, Moms, & MacLeod, 1987;
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Wolfe et al., 1987). Also, unlike amnesics and patients with Alzheimer’s disease
(AD), depressed patients make few intrusion errors and retain over long delay
periods most of the information they had managed to learn initially (Breslaw,
Kocsis, & Belkin, 1980; Kopelman, 1986; Massman & Bigler, 1992; Weingartner
& Silberman, 1982; Wolfe et al., 1987).
This pattern of memory test performance closely parallels that displayed by
patients with so-called “subcortical” or basal ganglia dementias. For example,
patients with Huntington’s disease (HD), a prototypical subcortical dementia,
often exhibit verbal free recall impairment as severe as that shown by amnesics
or AD patients, but generate fewer intrusion errors, exhibit far better retention of
information, and perform significantly better on recognition memory tests (But-
ters et al., 1988; Delis et al., 1991). Their memory test profiles suggest that
depressives and subcortical dementiapatients have difficulties initiating systematic
search (retrieval) of short- and long-term memory.
Besides the noted neuropsychologicalparallels, it is also of interest that there
is a markedly increased incidence of depression in subcortical dementia patients
(Caine & Shoulson, 1983; Mayeux, Stern, Rosen, & Leventhal, 1981). and de-
pressive symptoms frequently precede the onset of motor disability. Additionally,
a number of neuroradiological and neuropathological studies have provided evi-
dence that subcortical abnormalitiesmay play a role in affective disorders. Baxter
et al. (1985) found that the ratio of caudate nucleus metabolism to mean hemispheric
metabolic rate was significantly lower in unipolar depressives than in normals.
Brown et al. (1986) performed postmortem neuropathological assessments, and
reported that the brains of male patients who had diagnosesof “affectivepsychosis”
had smaller caudate nuclei than Alzheimer’s specimens (but their caudate nuclei
were larger than HD brain specimens). Using SPECT images, O’Connell et al.
(1989) found that depressives had lower regional cerebral blood flow (rCBF)
ratios in the caudate than did schizophrenics, manics, and patients with atypical
psychoses. Finally, recent MRI studies have consistently reported a high frequency
of white matter abnormalities (e.g., leukoencephalopathy, infarction) in patients
with major affective disorder (Coffey et al., 1988; Coffey, Figiel, Djang, Saunders,
& Weiner, 1989; Dupont et al., 1990; Figiel et al., 1991; Krishnan et al., 1988;
Lesser et al., 1991; Zubenko et al., 1990).
When one closely inspects the results reported in the neuropsychological and
 SUBCORTICAL DYSFUNCTION HYPOTHESIS 689

neuroradiological studies, it is clear that there is markedly greater variability

within depressed groups than within normal groups, and that there is usually
considerable overlap between groups. Thus, it is possible that only a subgroup of
depressives displays memory dysfunction and neurophysiological changes sug-
gestive of subcortical involvement. The importance of considering the heterogeneity
among patients with affective disorder was illustrated in an MRI study by Dupont
et al. (1990). It was found that 9 of 19 bipolar patients had subcortical signal
hyperintensities and that none of the 10 normal controls had such hyperintensities.
The bipolar patients with abnormal MRIs had a history of more hospitalizations
than those patients with normal MRIs, and performed more poorly on free recall,
Downloaded by [Central Michigan University] at 18:50 02 October 2015

fluency, and timed symbol manipulation tasks.

In this study, the subcortical dysfunction hypothesis of memory deficits in
depression was evaluated with the presumption that there is substantial heteroge-
neity in the memory performances of depressives. Thus, rather than examining
group means only, the memory performances of individual patients were exam-
ined. Using key variables from a standardized test of verbal learning and memory,
samples of mildly demented HD (subcortical dementia) patients, mildly demented
AD (cortical dementia) patients, and normal controls were differentiated in a
discriminant function analysis (DFA). Based on the results of past studies, it was
predicted that the classification accuracy achieved in the DFA should be high
(e.g., Butters et al., 1988; Butters, Wolfe, Granholm, & Martone, 1986; Delis et
al., 1991; Kramer et al., 1988, Kramer, Levin, Brandt, & Delis, 1989). The same
discriminant function was then applied to samples of unipolar and bipolar patients.
It was hypothesized that a substantial proportion of these patients would be
classified as HD (subcortical dementia) patients, a sizable percentage would be
classified as normals, and that few or none would be classified as AD (cortical
dementia) patients.

METHOD

Subjects
A total of 129 subjects were tested in this study: 40 unipolar depressed patients, 9 bipolar
depressed patients, 20 HD patients, 20 AD patients, and 40 normal controls (NC). The
unipolar and bipolar patients met DSM-III criteria for major affective disorder based on
information obtained from the Schedule for Affective Disorders and Schizophrenia -
Lifetime version (SADS-L) interview (Spitzer, Endicott, & Robins, 1978). This interview
was conducted by a well-trained psychiatry research fellow, and the DSM-III diagnoses
were made in consultation with a board-certified psychiatrist. All bipolar patients had
documentation of at least one prior manic episode requiring treatment. Subjects also had
elevated scores on the Hamilton Depression Scale (Hamilton, 1960) at the time of
neuropsychological testing (see Table 1). Results of thorough physical diagnostic work-
ups (including CBC and tests of thyroid function, liver function, and syphilis) were
negative. Exclusion criteria included a history of cerebrovascular disease, head trauma, or
other serious neurologic disease.
Twenty-one of the unipolar patients and four of the bipolar patients met Research
Diagnostic Criteria (RDC; Spitzer et al., 1978) for a diagnosis of alcoholism. All of the
subjects were tested while alcohol-free, and none exhibited signs of alcohol withdrawal or
 690 PAUL J. MASSMAN ET AL.

Table 1. Means and Standard Deviations of Group Characteristics.

Age Years of Dementia Rating Hamilton

Group (Years) Education Scale Score Score
NC 46.1 13.3 -
(n=40) (13.4) (1.8)
HD 45.2 13.2 120.1 -
(n=20) (10.8) (2.2) (8.0)
AD 68.9* 13.4 118.4
Downloaded by [Central Michigan University] at 18:50 02 October 2015

(n=20) (6.1) (3.2) (6.0)

Unipolar 43.8 13.1 - 17.3

(n=40) (10.2) (2.7) (8.2)

Bipolar 42.4 13.9 - 21.1

(n=9) (10.4) (1.6) (7.6)
* Significantly different from all other groups at p < .001.

significant liver dysfunction. In the unipolar group, 33 patients were not taking any
psychotropic medications at the time of testing (they were undergoing diagnostic evaluation
and had not yet begun their medication regimen), 4 patients were taking antidepressant
medication, 2 were taking lithium and an antipsychotic medication. and 1 was taking an
antipsychotic medication. In the bipolar group, 5 patients were medication-free, 2 were on
lithium, 1 was on lithium and an antidepressant, and 1 was on an antidepressant. None of
the unipolar or bipolar patients had undergone ECT.
The normal subjects were either spouses of the HD or AD patients or were paid
participants obtained through newspaper advertisements. Volunteers and respondents with
a history of alcoholism, learning disabilities, serious neurologic, systemic, or psychiatric
illness were excluded. The NC group was composed so that it would match the HD and
AD groups on years of education, and would match the depressed groups on age and years
of education.
The HD patients were evaluated by a psychologist, psychiatrist, and neurologist and
were diagnosed on the basis of a positive family history of HD, the presence of involun-
tary choreiform movements, and the presence of dementia. They had been experiencing
motor symptoms for a mean of approximately 5.0 years. HD subjects were not assessed
with the Hamilton Depression Scale because their HD-induced akinesia and other
“neurovegetative” symptoms would have resulted in artificially inflated scores, even in
patients whom did not experience depressed mood or depressive thoughts. The team of
professionals who evaluated the HD patients conferred to decide if patients met DSM-III
criteria for major affective disorder. None of the HD patients in the present study were
given this diagnosis, and none were taking antidepressant medication.
The AD patients were diagnosed by two senior staff neurologists according to the
criteria for probable AD developed by the National Institute of Neurological and Com-
municative Disorders and Stroke (NINCDS) and the Alzheimer’s Disease and Related
Disorders Association (ADRDA) (McKhann et al.. 1984). A number of laboratory tests
were performed to rule out various viral, metabolic, or traumatic causes of dementia.
Patients with a history of severe head injury, alcoholism, or serious and prolonged psychi-
 SUBCORTICAL DYSFUNCTION HYPOTHESIS 691

atric disturbance were excluded. To reduce the possibility of including multi-infarct

dementia, patients with a score of five or greater on the Hachinski scale (Hachinski et al.,
1975) were also excluded.
Because the differentiation of subcortical- and cortical-type memory deficits should
be clearest in the earliest stages of the diseases, the HD and AD subjects were selected
from a larger pool of subjects based on having relatively high scores on the Dementia
Rating Scale (DRS; Mattis, 1976). Also, the HD and AD groups’ mean scores on the DRS
were matched (see Table 1). so that differences between these groups could be attributed
to type rather than to seventy of dementia.
The distributions of men and women in the NC group (24 men. 16 women), HD group
(15 men, 5 women), and AD group (1 1 men, 9 women) did not differ significantly ( y > .30
for all chi-squares). The unipolar and bipolar patients were recruited primarily from a
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Veterans Affairs Medical Center, so almost all were men (there were 3 women in the
unipolar group, and 1 woman in the bipolar group).
Additional characteristics of the groups are shown in Table 1. As expected, the AD
subjects were significantly older than subjects in all other groups (p < .001 for all t tests),
but there were no other significant group differences in age ( p > .40for all I tests). To control
for differences in age and sex distributions across the groups, standardized scores from
the California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober, 1987) which
are stratified by age and sex, were used in the analyses. The Hamilton Depression Scale
scores of the unipolar and bipolar groups did not differ significantly (p > .20).Finally, the
HD and AD groups’ mean DRS scores did not differ (p > .45).

Procedure
The following tests were utilized in the study: CVLT; Controlled Oral Word Association
Test (generation of words beginning with the letters “F”,“A”, and “S” in 1-min periods;
Benton. 1968); category fluency test (generation of animal names, fruits, and vegetables
in 1-min periods); WAIS-R Digit Symbol subtest, with incidentalrecallof symbols following
completion of the third line of test stimuli (Kaplan, Fein, Moms, & Delis, 1991; Wechsler,
1981); Trail Making Test Part B (Reitan, 1958); a 15-item version of the Benton Visual
Retention Test-multiple-choice recognition format with a 15-s delay between stimulus
presentation and recognition (Benton, 1974); and a 28-item version of the Boston Naming
Test (Kaplan, Goodglass, & Weintraub, 1983). All tests were administered by trained
psychological technicians under standardized conditions.
Because the CVLT is the integral measure in the study, this test will be described in
detail. The CVLT involves the oral presentation of a “shopping” list of 16 items (List A)
over five immediate-recall trials. The items on the list are presented in the same order on
all five trials and examinees are asked to recall the items in any order they please, and to
recall all the items they can, including those they reported on previous trials. The list
consists of four items from each of four categories (fruits, spices, clothing, and tools), but
examinees are not informed of this. Adjacent words on the list are from different catego-
ries, which affords an assessment of the degree to which an examinee uses the active,
effective learning strategy of recalling the words in semantic clusters. Also, the items are
not prototypical exemplars of their categories (e.g.. ‘tangerines’ is in the fruit category but
‘apples’ is not), so if examinees start guessing prototypical category members, these
responses are correctly labeled as intrusion errors.
Following the five learning trials, a second, interference list (List B) is presented for
one trial. Following recall of List B, free and category-cued recall of List A is tested. After
a 20-min interval in which nonverbal testing occurs, free recall, category-cued recall, and
recognition of List A are assessed. The yes/no recognition test contains the 16 List A
target items and 28 distracters (List B items, items phonemically-similar to the target
words, items prototypical of the List A categories, and items unrelated to the target
words).
 692 PAUL J. MASSMAN ET AL.

Table 2. Description of the CVLT Indices.

Variables Description
List A Total Trials 1-5 Total number of List A words recalled across Trials 1-5
Semantic Clustering Ratio Ratio of List A words from the same category recalled
together over chanceexpected semantic clustering given
number of words and categories recalled
Serial Clustering Ratio Ratio of List A words recalled in the same order as they
were presented over chance-expected serial clustering given
the number of words recalled
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Percent Primacy Recall Percentage of total words recalled on Trials 1-5 that are
from the primacy region of List A (first four words)
Percent Recency Recall Percentage of total words recalled on Trials 1-5 that are
from the recency region of List A (last four words)
Recall Consistency Percentage of List A words d e d on one of the first four
trials that are also recalled on the very next aial (i.e.. Trials
2-5)
Learning Slope The slope of a least-squares regression line calculated to
fit changes in correct response scores across Trials 1-5
List B Recall Number of List B words recalled on the one immediate
recall trial
List B vs. List A Trial 1 Change in recall of List B relative to Trial 1 of List A;
reflects vulnerability to proactive interference
Short-Delay Free Recall Number of List A words recalled immediately after the
List B trial without re-presentation of List A
Short-Delay Cued Recall Number of List A words recalled when category names are
provided
Long-Delay Free Recall Number of List A words recalled after a 20-min delay of
nonverbal testing
Long-Delay Cued Recall Number of List A words recalled when category names are
provided
Free Recall Intrusions Total number of nontarget items reported on all free recall
trials of Lists A and B
Cued Recall Instrusions Total number of nontarget items reported on the two cued
recall trials of List A
Recognition Hits Number of List A words identified on a yes/no recognition
test that includes 28 distracter items
False Positives Number of distracter items identified as List A items on
the recognition test
Recognition Discriminability Signal detection measure of the ability to discriminate hits
(signal) from false positives (noise) on the recognition test
Discriminability vs. Trial 5 Standardized score on Recognition Discriminability minus
standardized score on List A Trial 5; a measure of im-
provement on recognition relative to free recall
 SUBCORTICAL DYSFUN<JIION -1.5 693

The paper-and-pencil protocols were scored using the CVLT scoring software (Fridlmd
& Delis, 1987). Extensive normative data on the various CVLT indices made it possible to
convert raw scores to standard scores within different age groups and by sex, controlling
for the influence of these two important variables. Descriptions of the CVLT variables
analyzed in this study are provided in Table 2.

RESULTS

Based on the results of previous investigations (Butters et al., 1985,1988; Delis

et al., 1991; Kramer et al., 1988, 1989), three indices were selected from the
Downloaded by [Central Michigan University] at 18:50 02 October 2015

CVLT which would likely prove successful in differentiating the NC, HD, and
AD patients. The first variable was the standard score on the measure of total
recall over the five learning trials, which has been found to distinguish HD and
AD patients from normals. The second variable was the standard score on the
measure of number of intrusions occurring on the two cued recall trials. This
measure has been shown to differentiate AD patients (who tend to generate many
intrusion errors when provided with categorical cues) from HD patients and
normals. The third variable was the differencebetween the standard scores obtained
on recognition discriminability and free recall on the fifth (final) learning trial.
Because the recognition test occurs after a delay period, this difference score is
tapping both retention of the list items over time, and the ability to profit from
the aid to retrieval provided by a recognition format. HD patients typically dis-
play supranormal improvement on recognition testing relative to free recall (so
this measure should help discriminate them from the NC subjects) and adequate
retention, whereas AD patients do not benefit from a recognition format and have
severe retention deficits (so the measure should aid the differentiation of HD
from AD patients).
Figure 1 shows the scores of the groups on the three variables described
above (the DEP-N and DEP-HD groups in the figure will be referred to later). A
mean standard score of -1.00 would indicate that subjects in the group scored an
average of one standard deviation below persons of their age group and sex in the
CVLT normative sample. As predicted, the HD patients exhibited List A recall
impairment as severe as that of the AD patients, but these patients showed great
improvement on the delayed recognition test, whereas the AD subjects became
even more impaired on recognition testing (relative to their peers in the CVLT
normative sample). The group difference on the discriminability - Trial 5 meas-
ure was highly significant (p < .001 for the 1 test). Also, as predicted, the AD
patients obtained higher (more impaired) standard scores on the cued recall intrusion
measure than the HD patients (p < .OOl). The NC group differed significantly
from the HD group on all three variables (p < .001 for the immediate recall and
discriminability - Trial 5 scores; p < .01 for the cued intrusion score). The NC
group also differed significantly from the AD group on all three measures (p<.OOl
for the immediate recall and cued intrusion variables;pc.01 for the discriminability
- Trial 5 variable).
 694 PAUL J. MASSMAN ET AL..

Normal
Downloaded by [Central Michigan University] at 18:50 02 October 2015

C
m la DEP-N
il DEP-HD
ra HD
Ab

List A Total Cued Intrusions Discrim - Trial 5

Fig. 1. CVLT standard scores on the DFA measures in the Normal, DEP-N, DEP-HD,
HD, and AD groups.

As predicted, a DFA using these indices fared well in differentiating the NC,
HD, and AD groups. With 80 subjects and 3 variables included in the analysis,
the subjectsto variables ratio was satisfactorilylarge to yield reliable DF coefficients
(Stevens, 1986). However, this sample size was not large enough to permit in-
dependent-sample cross-validation. A jackknife procedure thus was used in the
classification of subjects, in which each subject was classified based on the DF
derived from the remaining (n-1)subjects (Lachenbruch, 1967).
In the DFA, all three variables contributed to significant reductions in Wilks’
Lambda, and the final value of Lambda was .11, p c .001. The immediate recall
measure correlated highly with the first canonical DF ( T = 37). and the cued in-
trusion measure was negatively associated with scores on the first canonical DF
(T = -.49). This DF appeared to achieve good separation between the three groups.
The discriminability - Trial 5 score was correlated strongly with the second
canonical DF (T = .92), and this function enhanced the differentiation of the HD
group from the other groups. As shown in Table 3, jackknife DFs correctly
classified 90.0% (72/80) of the subjects.
Next, because the DF did discriminate well between the NC, HD, and AD
subjects, it was applied to the depressed patient groups. The classification rates
yielded by the DF are presented in Table 3. Although the proportion of patients
classified as HD subjects appears somewhat larger in the bipolar group than in
the unipolar group, the relative frequencies in the two groups did not differ
) p > .30. Of course, given the small number of subjects
significantly, ~ ~ (=11.14,
 SUBCORTICAL DYSFUNCTION HYPOTHESIS 695

Table 3. DF ClassificationResults.
Predicted Group
Actual Group NC HD AD

NC 36 1 3
HD 1 18 1
AD 0 2 18
Unipolar 28 12 0
Bipolar 4 5 0
Downloaded by [Central Michigan University] at 18:50 02 October 2015

in the bipolar group, this is not a powerful test of potential group differences.
Combining the results of the two depressive groups, 65.3% (32/49) of the pa-
tients were classified as NC subjects, 34.7%(17/49) were classified as HD patients,
and none were classified as AD patients.
Fifteen of the 32 depressives classified as normals had an RDC diagnosis of
alcoholism and 10 of the 17 depressives classified as HD patients had this diag-
nosis. These relative frequencies did not differ significantly, ( ~ ~ ( =1 .25,p
) > .60),
indicating that alcoholism was not related to classification as an HD patient.
Furthermore, two-way ANOVAs, with NC versus HD classification as one factor
and presence versus absence of alcoholism as a second factor, revealed that the
Classification X Alcoholism interaction was not significant for any of the three
CVLT variables in the DFA (p > .20 for all interaction effects). In general, an
RDC diagnosis of alcoholism was not associated with significant memory or
cognitive impairment; depressed patients with a positive alcohol history did not
differ significantly from depressed patients with a negative alcohol history on
the three CVLT variables included in the DFA ( p > .65 for all t tests) or any of the
eight additional neuropsychological measures listed in Table 7 (p > .10 for all r
tests). Six of the depressives classified as normal were taking psychotropic
medication and five of the depressives classified as HD patients were taking
medication. These relative frequencies were not significantly different, (2( 1) =
.24, p > .60).
The foregoing results support the hypothesis that a subgroup of depressives
displays a profile of memory performance similar to patients with subcortical
dementia. However, it is somewhat problematic to ‘force’ these patients to be
classified as NC, HD, or AD subjects, because some of them might not fit into
any of the three groups very well. Thus, it was decided to classify only those
patients who obtained posterior probabilities (i.e., the likelihood of group mem-
bership given a particular DF score) exceeding .70 (with three groups, the lowest
possible posterior probability for favored group membership is .34). Using this
criterion, the patients who only marginally resembled NC or HD subjects were
identified and excluded from further analyses.
 6% PAUL J. MASSMAN ET AL.

Table 4. DF Classificationwith Posterior Probability Criterion of .70.

Predicted Group
Actual Group NC HD AD None
Unipolar 21 9 0 10
Bipolar 3 5 0 1
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Table 4 reveals the results produced by this strict classification procedure.

Combining the unipolar and bipolar groups, 49.0% of the patients were classified
as NC subjects, 28.6% were classified as HD patients, 0.0% were classified as
AD patients, and 22.4% were unclassified. In this unclassified group, 8 of the
patients most closely resembled the NC subjects and 3 most closely resembled
HD patients. The relative frequencies of NC and HD classifications were not
significantly different in the unipolar and bipolar groups, 2(1) = 1.64, p > .20.
As in the original analysis, alcoholism was not related to classification status
v(1) = .11, p > .70),nor was medication status (&) = .21, p > .60).Also, two-
way ANOVAs again revealed that the Classification X Alcoholism interaction
was not significant for any of the three CVLT variables included in the DFA (p >
.18 for all interaction effects).
The classification rates indicated that even when a stringent criterion for
group membership was employed, substantial proportions of depressed patients
were classified as normal subjects or subcortical dementia patients. Given these
sizable percentages, it was appropriate to compare those depressives classified as
normal subjects (DEP-N, n = 24) with those classified as HD patients @EP-HD,
n = 14) on demographic variables, psychiatric characteristics, additional CVLT
measures, and other neuropsychological test scores. Because a number of com-
parisons between the groups were performed, it was decided to adopt a more
conservative criterion of statistical significance that would not also severely
undermine statistical power: a group effect was considered significant at a p -
level of .01.Effects with p-levels between .01 and .05 were considered nonsig-
nificant trends.
As shown in Table 5, the DEP-N and DEP-HD groups did not differ signifi-
cantly in age (p = .08), years of education, duration of illness, number of previ-
ous hospitalizations, rate of hospitalization over the duration of illness, or Ham-
ilton scores (p > .30 for all t tests). There was, however, a trend toward the DEP-
N patients having later ages of onset than the DEP-HD patients (p = .04).In the
DEP-N group, 54.2% (1 3/24) patients had an onset older than age 32, whereas in
the DEP-HD group, only 21.4% (3/14)of the patients had an onset this late.
Not surprisingly, there were many significant group differences on CVLT
variables. In addition to contrasting the DEP-N and DEP-HD groups’ CVLT
scores, the DEP-N group was compared with the NC group, and the DEP-HD
 SUBCORTICAL DYSFUNCTION HYPOTHESIS 697

Table 5, Demographic and Psychiatric Characteristics of Depressive Subgroups.

DEP-N DEP-HD
M SD M SD
Age 46.4 10.8 40.1 9.9
Education 12.9 1.9 12.8 1.6
Age of Onset 33.7+ 12.3 25.9 7.8
Duration of 12.8 10.3 14.1 10.0
Disorder
Hospitalizations 1.2 1.8 1.9 3.3
Hospitalization 1.2 2.6 1.0 1.2
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Rate (per 10 years)

Hamilton Score 18.9 8.8 18.4 7.4

+ Nonsignificant trend toward group difference (.01 < p < .05).

group was compared with the HD group. The DEP-N and NC groups did not
differ in age or years of education (p > .85 for both t tests), nor did the DEP-HD
and HD groups (p > .15 for age, p > .60 for years of education). The DEP-N and
DEP-HD groups were compared with the AD group on only several key CVLT
measures. The DEP-N and DEP-HD patients were, of course, significantly younger
than the AD patients (p < .001 for both t tests), but there were no group differ-
ences in education (p > S O for both t tests).
Figure 1 illustrates that on the variables included in the DFA, the DEP-N and
DEP-HD groups differed substantially on List A total recall and on the recogni-
tion discriminability - Trial 5 difference score (p < .001 for both t tests). There
was only a nonsignificant trend toward higher (worse) cued intrusion scores in
the DEP-HD group than in the DEP-N group (p = .03). Furthermore, there were
no significant differences between the DEP-N and NC groups on these three
variables (p > .10 for all t tests), nor were there differences between the DEP-HD
and HD groups (p > S O for all t tests). The DEP-N subjects differed significantly
from the AD patients on all three measures (p < -001for all r tests). The DEP-HD
group did not differ from the AD group on List A total recall (p > .70),but ob-
tained much higher discriminability- Trial 5 difference scores ( p < .OOl) and lower
(better) cued intrusion standard scores (p < .Ol).
Next, the performances of the DEP-N, DEP-HD, NC, and HD groups on
CVLT measures not included in the DFA were examined (see Table 6). These
variables are not independent from the variables included in the DFA, so these
analyses are not intended to provide validation of the memory profile subgroups.
Rather, examination of the additional CVLT indices contributes a more detailed
description of how the subgroups differed. The variables were grouped in ac-
cordance with the six-factor structure of the CVLT (for age-residualized scores)
reported by Delis, Freeland, Kramer, and Kaplan (1988). MANOVAs were con-
ducted using these variable groupings, and it was found that the overall group
 698 PAUL J. MASSMAN ET AL.

Table 6. Group Standard Scores on CVLT Measures Not Included in the DFA.

Normal DEP-N DEP-HD HD

M SD M SD M SD M SD
~ ~

General Verbal Learning

Short-Delay Free Rec. -.05 .99 -.33 1.40 -3.00. 1.04 -3.20 1.11
Short-Delay CuedRec. -.13 .99 -.42 1.06 -2.64 1.01 -3.05 1.28
Long-Delay Free Rec. -.30 .99 -.38 .88 -2.57 1.16 -2.95 1.32
Long-Delay CuedRec. -.lo 1.01 -30 .89 -2.71 1.38 -3.40 1.31
Consistency .08 .89 -.42 .78 -2.29 1.33 -2.10 1.52
Recognition Hits -.20 .91 -.46 1.35 -2.50 2.14 -1.35 1.79
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Response Discrimination
Free Recall Intrusions -.05 1.04 .00 1.06 .36 .84 .65 1.57
Recognition False Pos. 20 .91 .21 .51 1.07 1.14 1.65 1.63

Proactive Effect
List B Recall -.05 1.06 -.58 1.28 -1.64 1.28 -1.80 1.11
List B vs. Trial 1 .30 1.20 -.25 1.26 .07 1.39 .40 1.05
Learning Strategy
SernanticClustering -.08 .92 -.25 .90 -1.14 .77 -1.35 .88
Serial Clustering .03 1.12 .08 .78 SO 1.29 .oo 1.12

Serial Position Effect

Percent Primacy Rec. -.20 .72 .25 .99 .21 1.76 -.95 2.11
Percent Recency Rec. -23 .77 -.29 .91 SO 1.29 2.50 1.85

Acquisition Rate
Learning Slope .30 1.1 1 .04 1.04 -1.29 .83 -1.15 1.09

Note. See text for MANOVA and follow-up t test results.

effect was significant for General Verbal Learning, Response Discrimination,

Proactive Effect, Learning Strategy, and Serial Position Effect (Wilks’ Lambda
significant at p < .001 for all analyses). The groups also differed on the Acquisi-
tion Rate factor, which was represented by a single variable, Learning Slope
(F(3, 94) = 13.53, p < .001).
Three 2-group comparisons were of most interest: DEP-N versus DEP-HD,
DEP-N versus NC, and DEP-HD versus HD. Therefore, MANOVAs were com-
puted for each of these comparisons, using the variable groupings outlined in
Table 6. A Bonferroni correction was applied to each set of three comparisons,
so a multivariate result was considered significant at a p-level of .017 (.05/3).
Follow-up t tests were conducted where appropriate, using Bonferroni correc-
tions for significance.
MANOVAs revealed that the DEP-N and DEP-HD groups differed signifi-
cantly on the General Verbal Learning variables (p < .001), the Response Dis-
crimination variables (p c .Ol), the Proactive Effect variables (p < .001), and the
 SUBCORTICAL DYSFUNmON H Y P O ~ I s 699

Learning Strategy variables (p < .016). These groups did not differ on the Serial
Position Effect variables (p > .09). Follow-up t tests showed that the DEP-N
patients obtained significantly higher standard scores than the DEP-HD patients
on all of the General Verbal Learning measures (p c .001 for all t tests). On Re-
sponse Discrimination, the DEP-N patients had lower (better) False Positive
scores than the DEP-HD patients 0,< .004), but the two groups did not differ in
their Free Recall Intrusion scores (p > .25). Regarding the Proactive Effect vari-
ables, the DEP-N group had significantly higher List B recall standard scores
than the DEP-HD group (p < .02), but the groups’ List B - List A Trial 1 scores
did not differ (p > .45). On the Learning Strategy variables, the DEP-N patients
Downloaded by [Central Michigan University] at 18:50 02 October 2015

made better use of the semantic clustering strategy (p < .005),but the two groups’
serial clustering scores did not differ (p > .20). Lastly, the DEP-N patients had
significantly better Acquisition Rates (i.e., Learning Slope standard scores) than
the DEP-HD patients (p < .001).
In contrast to the many group differences found between the DEP-N and
DEP-HD groups, the performances of the DEP-N and NC groups, and the DEP-
HD and HD groups, were quite similar. MANOVAs revealed that the DEP-N and
NC groups failed to differ significantly on the General Verbal Learning meas-
ures @ > .07), the Response Discrimination measures (p > .95),the Proactive Effect
measures 0,> .IS), the Learning Strategy measures (p > .75), and the Serial Po-
sition Effect measures (p > .lo). The ANOVA result on the Acquisition Rate
measure also was not significant (p > .35). The DEP-HD and HD groups did not
differ on General Verbal Learning (p > .15), Response Discrimination (p > S O ) ,
Proactive Effect (p > .40),Learning Strategy (p > .30),or Acquisition Rate (p >
.70). The lone set of variables on which these groups did differ significantly was
Serial Position Effect (p < .008 for the MANOVA). Follow-up t tests showed that
the HD patients had significantly higher (more abnormal) Percent Recency scores
than the DEP-HD patients (p < .OOl), but that the groups’ Percent Primacy scores
were not significantly different (p > .lo).
The performances of the DEP-N and DEP-HD groups on a small set of addi-
tional neuropsychological measures are presented in Table 7. The DEP-N patients
obtained significantly higher WAIS-R Digit Symbol age-scaled scores than the
DEP-HD patients @ < .OOl), and performed better on the subsequent test of
incidental recall (p < .003), producing more of the symbols correctly matched
with their corresponding numbers. Also, the DEP-N subjects completed the Trail
Making Test Part B more quickly than the DEP-HD patients. There was substan-
tial heterogeneity of variance between groups in Trail Making Part B time scores,
and a nonparametric Mann-Whitney test revealed that the distributions of times
in the groups differed significantly (p < .W).The groups performed equivalently
on the Benton Visual Retention Test with a delayed four-choice recognition
format (r, > .90). There was a trend toward the DEP-HD patients spontaneously
generating fewer correct responses on the Boston Naming Test ( p < .02), but they
performed as well as the DEP-N patients when phonemic cues were provided ( p
> .30).
 700 PAUL J. MASSMAN ET AL.

Table 7. Neuropsychological Test Performances of Depressive Subgroups.

DEP-N DEP-HD
M SD M SD
Digit Symbol 9.47** 1.47 7.07 1.77
Age-Scaled Score
Digit Symbol 7.13* 2.17 4.29 2.92
Incidental Recall
Trail-Making B (sec.) 77.79* 23.45 116.57 44.07
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Benton VRT 12.58 2.17 12.38 1.71

Boston Naming
#Correct
I 26.13+ 2.05 24.36 2.24
With Phonemic Cues 27.04 1.16 26.64 1.34

Fluency
FAS 34.67 10.21 33.50 10.08
Category 43.13+ 8.14 34.71 11.14
~ ~~ ~~~

** SigniFicantly differunt from DEP-HD at p < .001.

* Significantlydifferent from DEP-HD at p < .01.
+ Nonsignificant trend toward group difference (.01 c p c .05).

Concerning fluency test performances, Table 7 shows that the DEP-N and
DEP-HD groups had similar mean scores on the FAS test (p > .85), but that there
was a trend toward higher category fluency scores in the DEP-N group (p -c -02).
The animal fluency performancesof the DEP-N and DEP-HD groups were examined
in more detail by tabulating the number of animal’categories which the subjects’
accessed (e.g., household pets, farm animals, jungle animals, fish, birds), and the
quantity of members generated in each category. If a subject produced members
of one category, switched to another grouping, then returned to the original
category, this category was counted twice. Interestingly, there was a trend (p = .05)
toward the DEP-HD patients accessing more total categories than the DEP-N
patients, but they generated an average of only 1.8 members per category, whereas
the DEP-N patients produced a mean of 2.8 members per category, a significant
group difference (p c .001).These findings suggest that the DEP-HD patients
were conducting a somewhat chaotic, ‘shallow’search through semantic memory.
The DEP-HD subjects accessed an average of 4.7 categories in which they gener-
ated only a single member, while the DEP-N subjects accessed a mean of 2.4
such categories, a significant group effect (p < .Ol). Finally, in their ‘best’ cat-
egory, the DEP-HD patients produced an average of only 3.6 members, which
was significantly poorer (p c .OOl) than the DEP-N patients’ average of 6.1
members.
 SUBCORTICALDYSFUNCTION HYPOTHESIS 701

DISCUSSION

The major finding of this study was that 65% of unipolar and bipolar depressed
patients were classified as normals, 35% were classified as HD patients, and
none were classified as AD patients, using three key verbal learning and memory
variables. Similar results were obtained when a stringent criterion of group
membership was employed: nearly 50% of unipolar and bipolar depressed pa-
tients closely resembled normals, nearly 30% closely resembled HD patients,
none resembled AD patients, and the remaining 20% weakly resembled normal
or HD patients. The findings are consonant with the proposition that there is
Downloaded by [Central Michigan University] at 18:50 02 October 2015

greater than normal heterogeneity in the learning and memory abilities of depressed
patients, and indicate that ‘lumping’ together the performances of all patients
obscures important differences between them. In the current study, if merely the
group means of the normal and depressed groups had been contrasted, there
would have been significant group differences on most of the CVLT variables,
despite the fact that at least half of the depressed patients had normal memory
profiles. This analysis of group means would also have underestimated the learning
and memory deficits displayed by the sizable subgroup of DEP-HD patients.
Based on the current findings, it appears that the subcortical dysfunction
hypothesis of memory deficits in depression is viable, but only for a subgroup of
patients. Although a subgroup of our depressed patients showed a memory pro-
file highly consistent with a subcortical dysfunction hypothesis, these behavioral
data should not be interpreted directly as indicating that these patients have
subcortical brain pathology. These results do, however, provide neuropsychological
corroboration for previous neuroradiological and neuropathological findings
suggestive of subcortical abnormalities in depressives (Baxter et al., 1985; Brown
et al., 1986; Coffey et al., 1989; O’Connell et al., 1989; Zubenko et al., 1990).
In addition to their close similarities to the HD patients on the measures
included in the DFA, the DEP-HD patients strongly resembled the HD patients
on nearly all other CVLT variables as well (and differed from the DEP-N patients,
who closely resembled the NC subjects). The DEP-HD subjects, like the HD
subjects, displayed severely impaired performanceson all of the free recall measures
(including the interference word list); inconsistent item recall from trial to trial;
inadequate use of the active, efficient semantic clustering learning strategy; and
deficient rate of improvement across learning trials. These impairments suggest
that the DEP-HD and HD patients did not efficiently utilize systematic organi-
zational strategies when learning the material.
Like the HD patients, the DEP-HD subjects showed dramatic improvement
on recognition testing, suggesting that retrieval defects contributed strongly to
their strikingly impaired free recall performances. However, because their per-
formance did not normalize completely on recognition testing (hit rate and false
positive rate were both mildly impaired relative to the DEP-N patients), it appears
that they did have at least mild encoding difficulties. Similar conclusions have
been reached in several previous investigations of the memory disorders of HD
 702 PAUL I.MASSMAN ET AL..

patients (Butters et al., 1986, 1988). Unlike AD patients, the DEP-HD patients
displayed good retention of the information they had learned and did not gener-
ate many intrusion errors. This indicates that the DEP-HD subjects had intact
memory storage abilities and could discriminateadequatelyamong stored materials.
The results of the study appear to have implications for the differentiation of
“pseudodementia” from AD, a diagnostic question encountered frequently by
neuropsychologists,psychiatrists,and neurologists (Caine, 1981;Folstein & Rabins,
1991). Since our DEP-HD patients exhibited immediate free recall impairment
as severe as that shown by the mildly demented AD patients, these patients
would seem to be candidates for the diagnosis of pseudodementia. However, two
Downloaded by [Central Michigan University] at 18:50 02 October 2015

features of their CVLT performances clearly distinguished them from AD patients:

(1) their cued intrusion error scores were dramatically lower (better) than the AD
patients’ scores; and (2) compared with their poor immediate free recall, the
DEP-HD patients improved substantiallyon the delayed recognition test, whereas
the AD patients’ recognition scores were even more impaired than their immedi-
ate free recall scores. None of the 49 depressed subjects in the study were classified
as an AD patient based on these features of memory performance, suggesting
that the ‘pseud0dementia’-ADdifferential diagnosis can be made very accurately
when the appropriate memory measures are utilized.
Besides their distinctive CVLT performances, the DEP-N and DEP-HD sub-
jects differed significantly on a number of additional neuropsychological tests.
These differences support the validity of the distinction between the subgroups,
which was based on memory test performance alone. The DEP-HD patients
performed more poorly than DEP-N patients on Digit Symbol, Trail Making Test
Part B, and category fluency, suggesting that they had difficulty with speeded
information processing tasks requiring rapid retrieval and/or mental manipulation
of symbols. These types of difficulties have also been documented in subcortical
dementia patients (Butters, Granholm, Salmon, Grant, & Wolfe, 1987; Caine,
Bamford, Schiffer, Shoulson, & Levy, 1986; Josiassen, Cuny, & Mancall, 1983;
Starkstein et al., 1989; Troster, Salmon, McCullough, & Butters, 1989). On the
category fluency task, the DEP-HD patients actually accessed more categories
than the DEP-N patients, but generated fewer members per category. HD patients
have also been found to produce few exemplars per category on fluency tests
(Troster et al., 1989). The DEP-HD subjects’ somewhat haphazard category
fluency performance suggests that they had difficulty organizingsystematicretrieval
strategies, and is analogous to their poorly organized free recall performances on
the CVLT (i.e., inconsistent recall and deficient semantic clustering).
The DEP-HD group displayed mild difficulty spontaneously generating cor-
rect responses on the Boston Naming Test, which may reflect mild impairment in
retrieving low-frequency object names (e.g., trellis, abacus). When provided
with phonemic cues to aid their retrieval, the DEP-HD patients performed as well
as the DEP-N patients, suggesting that their fundamental semantic abilities were
intact. The DEP-HD and DEP-N groups did not differ on the Benton Visual
Retention Test with a delayed recognition format or on the FAS fluency test. The
 SUBCORTICALDYSFUNCTION HYPOTHESIS 703

first result was expected, given that DEP-HD patients’ recognition memory should
not be affected markedly by their hypothesized subcortical system dysfunction,
particularly on a structured task in which they are given a good period of time to
encode relatively simple visual stimuli. The absence of a significant difference in
FAS scores was unexpected, considering that subcortical dementia patients typically
perform poorly on this task (Butters et al., 1986; Butters et al., 1987; Caine et al.,
1986).
The question arises as to why halfof the depressed patients in our study ex-
hibited normal learning and memory performances. The DEP-N patients did not
differ from the DEP-HD patients in age, years of education, duration of illness,
Downloaded by [Central Michigan University] at 18:50 02 October 2015

number of hospitalizations, rate of hospitalization over the duration of illness, or

in Hamilton Depression Scale scores. These results indicate that in this patient
sample, the cumulative effects of aging and repeated depressive episodes were
not the key factors in producing subcortical-type memory impairment. There was
a trend toward earlier ages of onset in the DEP-HD group, suggesting that the
DEP-HD patients may have had greater vulnerability for the development of the
disorder.
Longitudinal studies should be conducted to determine whether the learning
and memory difficulties displayed by some depressives are a state or trait phe-
nomenon. If these patients’ deficits are associated with subcortical structural
abnormalities, one would expect that they might have at least subtle memory
problems even when they are not depressed. The learning and memory abilities
of these patients may be inherently less efficient and more vulnerable to the
disruptions caused by the neurobiological changes associated with the depressed
state. Thus, relatively small differences between the “subcortical” and “normal”
patients in the non-depressed state could translate into large differences in func-
tioning when the system is taxed by depression.
In summary, it is apparent that depressives constitute a heterogeneous patient
group, some of whom display a profile of verbal learning and memory deficits
strikingly similar to brain-damaged patients with predominantly subcortical pa-
thology. Investigators studying the memory, cognitive, neuroradiological, and
neurobiological abnormalities associated with depression are advised to examine
explicitly the nature and causes of this heterogeneity, for group studies discounting
the great variability among depressives may yield misleading results.

REFERENCES

Baxtet, L.R., Phelps, M.E., Mazziotta, J.C., Schwartz, J.M., Gerner, R.H.,Selin, C.E., &
Sumida, R.M. (1985). Cerebral metabolic rates for glucose in mood disorders: Studies
with positron emission tomography and fluorodeoxyglucose F18.Archives of General
Psychiatry, 46, 243-250.
Benton, A.L. (1 968). Differential behavioral effects in frontal lobe disease.
Neuropsychologia, 6,53-60.
Benton, A.L. (1974). The Revised Visual Retention Test (4th ed.). New York: Psycho-
logical Corporation.
 704 PAUL J. MASSMAN ET AL.

Breslow, R., Kocsis, J., & Belkin, B. (1980). Memory deficits in depression: Evidence
utilizing the Wechsler Memory Scale. Perceptual and Motor Skills, 51,541-542.
Brown. R., Colter, N., Corsellis. J.A.N., Crow, T.J., Frith, C.D.,Jagoe, R.. Johnstone,
E.C.. & Marsh, L. (1986). Postmortem evidence of structural brain changes in schizo-
phrenia. Archives of General Psychiatry, 43,36-42.
Butters, N.. Granholm,E.. Salmon. D.P., Grant, I., & Wolfe. J. (1987). Episodic and
semantic memory: A comparison of amnesic and demented patients. Journal of Clini-
cal and Experimental Neuropsychology, 9,479-497.
Butters, N., Salmon, D.P.. Cullum, C.M., Cairns, P., Troster, A.I., Jacobs, D., Moss, M..
& Cermak, L.S. (1988). Differentiation of amnesic and demented patients with the
Wechsler Memory Scale-Revised. The Clinical Neuropsychologist, 2, 133-148.
Butters. N., Wolfe, J., Granholm, E., & Martone, M. (1986). An assessment of verbal
Downloaded by [Central Michigan University] at 18:50 02 October 2015

recall, recognition and fluency abilities in patients with Huntington’s disease. Cortex,
22, 11-32.
Butters, N., Wolfe. J.. Martone, M. Granholm, E., & C e d , L.S. (1985). Memory
disorders associated with Huntington’s disease: Verbal recall, verbal recognition and
procedural memory. Neuropsychologia. 23,729-743.
Caine, E.D. (1981). Pseudodementia: Current concepts and future directions. Archives of
General Psychiatry, 38, 1359-1364.
Caine, E.D., Bamford. K.A.. Schiffer, R.B., Shoulson, I., & Levy, S. (1986). A controlled
neuropsychological comparison of Huntington’sdisease and multiple sclerosis. Archives
of Neurology, 43,249-254.
Caine, E.D., & Shoulson,I. (1983). Psychiatric syndromes in Huntington’sdisease. American
Journal of Psychiatry, 140,728-733.
Coffey, C.E.. Figiel, G.S., Djang, W.T., Cress, M., Saunders. W.B., & Weiner. R.D.
(1 988). Leukoencephalopathy in elderly depressed patients referred for ECT. Biological
Psychiatry, 24, 143-161.
Coffey. C.E., Figiel. G.S., Djang, W.T., Saunders, W.B., & Weiner, R.D.(1989). White
matter hyperintensity on magnetic resonance imaging: Clinical and neuroanatomic
correlates in the depressed elderly. Journal of Neuropsychiatry and Clinical
Neurosciences, 3, 18-22.
Danion. J.M., Willard-Schroeder, D., Zimmerman, M.A., Grange, D.,Schlienger, J.L., &
Singer, L. (1991). Explicit memory and repetition priming in depression. Archives of
General Psychiatry, 48,707-71 1.
Delis, D.C., Freeland, J., Kramer. J.H., & Kaplan, E. (1988). Integrating clinical assessment
with cognitive neuroscience: Construct validation of the California Verbal Leaming
Test. Journal of Consulting and Clinical Psychology, 56, 123-130.
Delis, D.C., Kramer, J.H., Kaplan, E., & Ober, B.A. (1987). The California Verbal
Learning Test. New York: Psychological Corporation.
Delis, D.C.. Massman, P.J., Butters, N., Salmon, D.P.,Kramer, J.H., & Cermak. L. (199 1).
Profiles of demented and amnesic patients on the California Verbal Learning Test:
Implications for the assessment of memory disorders. Psychological Assessment: A
Journal of Consulting and Clinical Psychology, 3, 19-26.
Dunbar, G.C., & Lishman, W.A. (1984). Depression, recognition memory and hedonic
tone: A signal detection analysis. British Journal of Psychiatry, 144, 376-383.
Dupont. R.M., Jemigan, T.L., Butters, N., Delis, D., Hesselink, J.R., Heindel, W.,&
Gillin, J.C. (1990). Subcortical abnormalities detected in bipolar affective disorder
using magnetic resonance imaging: Clinical and neuropsychological significance. Ar-
chives of General Psychiatry. 47.55-59.
Figiel, G.S.,Krishnan, K.R.R., Rao. V.P., Doraiswamy, M., Ellinwood, E.H., Nemeroff,
C.B., Evans, D., & Boyko, 0.(1991). Subcortical hyperintensities on brain magnetic
resonance imaging: A comparison of normal and bipolar subjects. Journal of
Neuropsychiatry and Clinical Neurosciences, 3 , 18-22.
 SUBCORTICAL DYSFUNCTION HypoTHEsIS 705

Folstein, M.F., & Rabins, P.V. (1991). Replacing pseudo-demenria. Newopsychiatry,

Neuropsychology, and Behavioral Neurology, 4 , 3 6 4 .
Fridlund, A.J., &Delis, D.C. (1987). IBM user’s guide for the California Verbal Learning
Test. New York: The Psychological Corporation.
Frith, C.D., Stevens, M., Johnstone, E.C., Deakin, J.F.W., Lawler, P.,C Crow, T.J. (1983).
Effects of ECT and depression on various aspects of memory. British Journal of Psy-
chiatry, 142, 610-617.
Hachinski, V.C., Illiff, L.D., Zilhka, E., DuBoulay, G.H.. McAllister, V.L.,Marshall, J.,
Russell, R.W.R., & Symon, L. (1975). Cerebral blood flow in dementia. Archives of
Neurology, 32,632-637.
Hamilton, M.A. (1960). A rating scale for depression. Journal of Neurology, Neurosur-
gery, and Psychiatry, 23,56-62.
Downloaded by [Central Michigan University] at 18:50 02 October 2015

Hart, R.P., Kwentus, J.A., Taylor, J.R., & Harkins, S.W. (1987). Rate of forgetting in
dementia and depression. Journal of Consulting and Clinical Psychology, 55,101-105.
Josiassen, R.C., Curry, L.M., & Mancall, E.L. (1983). Development of neuropsychological
deficits in Huntington’s disease. Archives of Neurology, 40,791-796.
Kaplan,E.,Fein.D.,Moms,R.,&Delis,D.C. (1991). The WAIS-Rasaneuropsychological
instrument. New York: The Psychological Corporation.
Kaplan, E., Goodglass, H., & Weintraub, S. (1 983). Boston Naming Test. Philadelphia: Lea
& Febiger.
King, D.A., Caine, E.D., Conwell, Y.. & Cox, C. (1991). The neuropsychology of depression
in the elderly: A comparative study of normal aging and Alzheimer’s disease. The
Journal of Neuropsychiatry and Clinical Neurosciences, 3, 163-168.
Kopelman, M.D. (1986). Clinical tests of memory. British Journal of Psychiatry, 148,517-
525.
Kramer, J.H., Delis, D.C., Blusewicz, M.J., Brandt, I., Ober, B.A., & Strauss, M. (1988).
Verbal memory errors in Alzheimer’s and Huntington’s dementias. Developmental
Neuropsychology, 4, 1-15.
Kramer, J.H., Levin, B.E., Brandt, J., & Delis, D.C. (1989). Differentiation of Alzheimer’s,
Huntington’s, and Parkinson’s disease patients on the basis of verbal learning charac-
teristics. Neuropsychology, 3, 111-120.
Krishnan, K.R.R., Goli, V., Ellinwood, E.H.. France, R.D., Blazer, D.G.,& Nemeroff,
C.B. (1988). Leukoencephalopathy in patients diagnosed as major depressive. Biological
Psychiatry, 23, 519-522.
Lachenbruch, P.A. (1967). An almost unbiased method of obtaining confidence intervals
for the probability of misclassificationin discriminant analysis. Biomerrics, 23,639-645.
Lamer, S . (1977). Encoding in senile dementia and elderly depressives: A preliminary
study. British Journal of Social and Clinical Psychology, 16,379-390.
Lesser, I.M., Miller, B.L., Boone, K.B., Hill-Gutierrez, E., Mehringer, C.M., Wong, K., &
Mena, I. (1991). Brain injury and cognitive function in late-onset psychotic depres-
sion. Journal of Neuropsychiatry and Clinical Neurosciences. 3,33-40.
Massman, P.J., & Bigler, E.D. (1992). Memory indices differentiate reversible dementia
secondary to depression from Alzheimer’s disease. Manuscript submitted for publica-
tion.
Mattis, S. (1976). Mental status examination for organic mental syndrome in the elderly
patient. In L. Bellack & T.B. Karasu (Eds.), Geriatricpsychiatry (pp. 77-121).New York:
Grune & Stratton.
Mayeux, R., Stern, Y., Rosen, J., & Leventhal, J. ( 198 1). Depression, intellectual impairment,
and Parkinson’s disease. Neurology, 31,645-650.
McKhann, G., Drachman, D., Folstein, M., Katzman, R.. Price, R., & Stadlan. E.M.
(1984). Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA
Work Group under the auspices of Department of Health and Human Services Task
Force on Alzheimer’s disease, Neurology, 34,939-944.
 706 PAUL J. MASSMAN ET AL.

Miller, E., & Lewis, P. (1977). Recognition memory in elderly patients with depression
and dementia: A signal detection analysis. Journal ofAbnorma1Psychology. 86.84-86.
O’Connell, R.A., Van Heertum, R.L.,Billick, S.B., Holt. A.R.. Gonzalez. A., Notardonato.
H., Luck, D., & King, L.N. (1989). Single photon emission computed tomography
S P E W with [*’I]IMP in the differential diagnosis of psychiatric disorders. Journal of
Neuropsychiatry and Clinical Neurosciences, 1, 145-152.
Reitan, R.M. (1958).Validity of the Trail Making Test as an indication of organic brain
damage. Perceptual and Motor Skills. 8,271-276.
Silberman. E.K., Weingartner. H., Laraia, M.. Bymes. S.,& Post, R.M. (1983).Process-
ing of emotional properties of stimuli by depressed and normal subjects. Journal of
Nervous and Mental Disease, 171,lO-14.
Spitzer, R.L., Endicott, J., & Robins, E. (1978). Research diagnostic criteria: Rationale
Downloaded by [Central Michigan University] at 18:50 02 October 2015

and reliability. Archives of General Psychiatry, 35,773-782.

Starkstein, S.E., Preziosi, T.J.,Berthier, M.L.,Bolduc, P.L., Mayberg, H.S., &Robinson,
R.G. (1989).Depression and cognitive impairment in Parkinson’s disease. Brain, 112,
1141-1153.
Stevens, J. (1 986). Applied multivariate statistics for the social sciences. Hillsdale, NJ:
Erlbaum.
Stromgren, L.S. (1977). The influence of depression on memory. Acta Psychiatrica
Scandinavica, 56, 109-121.
Troster, A.I., Salmon, D.P., McCullough, D., & Butters, N. (1989).A comparison of the
category fluency deficits associated with Alzheimer’s and Huntington’s disease. Brain
and Language, 37.500-513.
Watts, F.N., Morris, L., & MacLeod, A.K. (1987).Recognition memory in depression.
Journal of Abnormal Psychology, 96,273-275.
Wechsler, D. (1981). Wechsler Adult Intelligence Scale- Revised. New York: Psycho-
logical Corporation.
Weingartner, H., Cohen, R.M., Murphy, D.L., Martello, J., & Gerdt, C. (1981).Cognitive
processes in depression. Archives of General Psychiatry, 38.42-47.
Weingartner, H., & Silberman, E. (1982).Models of cognitive impahment: Cognitive
changes in depression. Psychopharmacology Bulletin, 18,27-42.
Whitehead, A. (1973).Verbal learning and memory in elderly depressives. British Jour-
nal of Psychiatry, 123,203-208.
Wolfe, J.. Granholm, E., Butters, N., Saunders, E., & Janowsky, D. (1987).Verbal memory
deficits associated with major affective disorders: A comparison of unipolar and bipolar
patients. Journal of Affective Disorders, 13,83-92.
Zubenko, G.S.,Sullivan, P., Nelson, J.P., Belle, S.H., Huff, J.. &Wolf, G.L.(1990).Brain
imaging abnormalities in mental disorders of late life. Archives of Neurology, 47,1107-
1111.