Directory UMM :Data Elmu:jurnal:B:Biological Psichatry:Vol47.Issue11.2000:

Reduced Number of Mediodorsal and Anterior
Thalamic Neurons in Schizophrenia
Keith A. Young, Kebreten F. Manaye, Chang-Lin Liang, Paul B. Hicks, and
Dwight C. German
Background: The thalamus is a brain region of interest in
the study of schizophrenia because it provides critical
input to brain regions such as the prefrontal, cingulate,
and temporal cortices, where abnormalities have been
repeatedly observed in patients with schizophrenia. Postmortem anatomic studies have rarely investigated the
thalamus in this population.
Methods: Postmortem tissue was obtained from the left
hemisphere of eight male schizophrenic patients and eight
male age-matched control subjects. The optical dissector
stereologic procedure was used to count neurons in the
mediodorsal (MD) and anteroventral/anteromedial (AV/
AM) nuclei of the thalamus.
Results: The number of neurons and volume of the MD
were significantly reduced by 35% and 24%, respectively.
The MD cell number reduction was a consistent finding;
every control subject had more and every schizophrenic
subject had fewer than 3.5 million neurons. Neuron

number was also significantly reduced (16%) in the
AV/AM nuclei.
Conclusions: The present data indicate that schizophrenia is associated with robust reductions in nerve cell
numbers in nuclei that communicate with the prefrontal
cortex and limbic system. These thalamic anatomic deficits
may be responsible, in part, for previous reports of such
prefrontal cortical abnormalities as reduced synaptic
density, reduced volume, and metabolic hypofunction.
Biol Psychiatry 2000;47:944 –953 © 2000 Society of Biological Psychiatry
Key Words: Schizophrenia, postmortem brain, thalamus,
mediodorsal nucleus, anterior nucleus, stereology

From the Department of Psychiatry and Behavioral Science, Neuropsychiatry
Research Program, Texas A&M University System Health Science Center and
Central Texas Veterans Health Care System, Temple (KAY), the Department
of Psychiatry, University of Texas Southwestern Medical School, Dallas
(KFM, C-LL, DCG), and the Department of Psychiatry, Scott and White
Memorial Hospital, Temple (PBH), Texas.
Address reprint requests to Keith A. Young, Ph.D., Central Texas Veterans Health
Care System, Neuroscience Laboratory, Neuropsychiatry Research Program

(151T), 1901 S. 1st Street, Temple TX 76504.
Received June 23, 1999; revised October 25, 1999; revised January 28, 2000;
accepted February 8, 2000.

© 2000 Society of Biological Psychiatry

Introduction

S

chizophrenia is a major mental illness that affects
approximately one in every 100 people (Andreasen
and Carpenter 1993). Individuals with schizophrenia exhibit a lack of motivation, abnormal thought content and
process (e.g., delusions, paranoia, hallucinations and discontinuity of associations), and disturbance of emotion
(i.e., inappropriate affect; Andreasen 1997a). Many
schizophrenic patients also perform poorly on neuropsychologic tasks involving the frontal lobes and exhibit
abnormal metabolic activity patterns in the prefrontal
cortex (Berman et al 1986; Fletcher et al 1998; Weinberger and Gallhofer 1997) and anterior cingulate cortex
(Haznadar et al 1997). Schizophrenic individuals who
exhibit large numbers of errors of perseveration on the

Wisconsin Card Sorting Test or who perform poorly on
verbal working memory or sustained attention tests lack
the normal pattern of metabolic activation in the dorsolateral prefrontal cortex (Goldberg et al 1994; Stevens et al
1998). Morphologic abnormalities in the dorsolateral prefrontal and anterior cingulate cortex have been identified
as a possible anatomic basis for the behavioral and
metabolic abnormalities of schizophrenia (Arnold and
Trojanowski 1996; Buchanan et al 1998; Goldstein et al
1999; Rajkowska et al 1998; Selemon et al 1998).
The thalamus, which is an important source of subcortical input to the cortex, has become a recent focus of
research in schizophrenia (Andreasen 1997b; Jones 1997a;
Scheibel 1997). The majority of thalamic studies have
utilized magnetic resonance imaging (MRI) (Andreasen et
al 1994; Buchsbaum et al 1996; Gur et al 1998; Staal et al
1998). The data indicate that the thalamus of schizophrenic subjects is smaller than that of control subjects,
although the magnitude of the volume reductions (10%) is
relatively small. Quantitative postmortem morphologic
studies, on the other hand, have found more substantial
thalamic abnormalities in schizophrenic subjects. Using
stereologic cell counting procedures, Pakkenberg (1990)
reported a 40% reduction in the number of mediodorsal

(MD) thalamic neurons in schizophrenic subjects, a finding apparently not related to neuroleptic treatment (Pak0006-3223/00/$20.00
PII S0006-3223(00)00826-X

The Thalamus in Schizophrenia

BIOL PSYCHIATRY
2000;47:944 –953

945

Table 1. Specimen Characteristics
Agea

ID

Schizophrenic Group
A
76
B
75

C
65
D
63
E
68
F
71
G
65
H
40
Control Group
I
94
J
45
K
59
L

60
M
60
N
81
O
72
P
48

TIF

PMI

Brain weightb

Cause of death

12
84

72
48
168
4
84
22

6
9
14
11
19
4
30
36

1,440
1,300
1,635
1,230

1,260
1,135
1,270
1,600

Congestive heart failure
Cancer
Aortic aneurysm
Congestive heart failure secondary to pneumonia
Congestive heart failure secondary to pneumonia
Pneumonia secondary to emphysema
Myocardial infarction
Accident

7
48
12
168
168
8

8
24

8
11
14
6
8
28
30
18

1,100
1,400
1,490
1,500
1,430
1,210
1,620
1,660

Congestive heart failure
Heart failure
Myocardial infarction secondary to lymphoma
Heart failure
Myocardial infarction
Congestive heart failure secondary to emphysema
Heart failure
Heart failure

TIF, time in formalin (months); PMI, postmortem interval (hours).
a
Age in years.
b
Brain weight in grams.

kenberg 1992). Considering the prominent frontal cortical
deficits of schizophrenic individuals, an MD neuron number reduction is particularly intriguing because the MD is
a major source of subcortical input to the frontal cortex
(Giguere and Goldman-Rakic 1988; Ray and Price 1993).

There have been few morphologic studies of the thalamus
in subjects with schizophrenia. We expected our study to
replicate the finding of Pakkenberg (1990), which described
a reduction in the number of MD neurons, and also extend
the examination of the thalamus by studying another thalamic
region that projects to the limbic cortex, the anterior thalamic
nuclei. The optical disector stereologic cell counting procedure was used to estimate the total number of neurons in
control and schizophrenic postmortem brains. Neurons were
counted in the MD nucleus, which projects to the dorsolateral
prefrontal and orbitofrontal cortex, and the anteroventralanteromedial (AV/AM) nuclei, which project to the cingulate
and entorhinal cortex. We found robust reductions in the
number of neurons in both the MD and AV/AM thalamus in
schizophrenic subjects.

Methods and Materials
Brains
The left thalamus was obtained from eight male control
subjects with no history of psychiatric disturbance and eight
male schizophrenic subjects diagnosed according to DSM-IV
criteria (Table 1). The Diagnostic Evaluation after Death was
used to obtain postmortem diagnosis of schizophrenic subjects. The brains were obtained from Terrell State Hospital,
the Waco VA Medical Center, and the Stanley Foundation
Neuropathology Consortium. All brains were grossly normal

at autopsy and there was no evidence of ischemic brain
damage in the thalamus of any of the subjects. Neuropathologic examinations were available on 13 of the brains. There
was no histopathologic evidence of neurodegenerative
changes in the cortex or hippocampus of any of these brains.
The schizophrenic patients were chronically treated with
neuroleptic drugs until near the time of death. Schizophrenic
and control groups did not differ significantly in age (schizophrenic group mean 6 SD, 65.4 6 11.3 years; control group,
64.9 6 16.7 years), postmortem interval (schizophrenic
group, 16.1 6 11.5 hours; control group, 15.3 6 9.2 hours), or
time in formalin fixative (schizophrenic group, 61.7 6 53
months; control group, 55.3 6 71 months). One potential diagnostic confound in the present study was lack of
information about the alcohol or drug abuse history in many
of the subjects, particularly during adolescence and early
adulthood.

Histology
Before sectioning the thalamus, the tissue block was immersed in a
solution of 20% sucrose and 10% formalin for 2–3 weeks. A single
block containing the entire left thalamus was serially sectioned in
the coronal plane on a freezing microtome. Although the microtome
thickness was set at 60 mm, the actual average distance advanced for
each cut was measured to be 56 mm. After Nissl staining every 20th
section with cresyl violet, the regional boundaries of the two
thalamic nuclei (Jones 1997b) were delineated at low magnification
with a 2.53 objective (Figure 1). The MD begins just caudal and
ventral to the AV and ends at the level of the rostral pulvinar. Its
medial border is near the lateral wall of the third ventricle except
where it merges into the massa intermedia. It is bounded laterally by
the fasciculus mammillo-thalamicus anteriorly and by the internal
medullary lamina and centromedian nucleus posteriorly. The large

946

K.A. Young et al

BIOL PSYCHIATRY
2000;47:944 –953

Figure 1. Photomicrographs of Nisslstained coronal sections of the thalamus
in control case C. Frozen sections were
cut at 60-mm thickness on a freezing
microtome. After staining and coverslipping, the section thickness measured
22–25 mm. A dashed line delineates the
boundaries of the anteroventral/anteromedial nuclei of the thalamus (AV/AM)
at several different rostro-caudal levels.
The AV/AM begins in section 141 and
ends at the level of section 281. The
mediodorsal nucleus of the thalamus
(MD) begins at the level of section 241
and ends at the level of section 444.
Also illustrated are the laterodorsal nucleus (LD) and the habenula (H). Four
levels of the thalamus are illustrated
spanning a distance of 18.18 mm, from
rostral (section 141) to caudal (section
444). Marker bars, 1.0 mm.

cells of the centrolateral nucleus clearly delineate the lateral boundary of the MD for most of its rostral-caudal extent. The densocellular portion of the MD was included. The lateral ventricle, fornix,
anterior nuclei, and laterodorsal nucleus lie on its superior surface
and are clearly recognizable. Most posteriorly, the inferior limit
merges with the nucleus parafascicularis and centromedian nuclei.
The external borders of the AV/AM nuclei are clearly demarcated.
The nuclei are encapsulated by the white matter track of the internal
medullary lamina; however, the border between the two nuclei often
is difficult to identify in Nissl stained coronal sections. We therefore
combined the AV nucleus with the AM nucleus to form a single
region of interest (AV/AM) for cell counting purposes.
The tissue processing and staining procedures produced approximately 50% tissue shrinkage. The final Z-axis thickness in
all specimens measured between 22 and 25 mm. Thus, we used
a 5-mm upper guard zone, a 7–10-mm lower guard zone and a
10-mm-thick counting frame. The sampling coefficient of error
(CE; Scheaffer et al 1996) was estimated with the StereoInvestigator software (version 1.3; MicroBrightField, Colchester, VT)
providing a measure of the variance of the sampling procedure.
The biological CE provides an estimate of the variability in
neuron counts between specimens (Table 2). The CE estimates in
the present study indicate that sampling variances for the MD
and AV/AM regions were lower in magnitude than the inherent
biological variance, indicating that the stereologic sampling
parameters used in the study were sufficient to detect relatively
small group differences between schizophrenic and control
subjects.

Stereology
One of us (KFM), who was blinded to the identification of the
brains, performed all stereologic procedures. Before the start of
the experiment, a pilot study was performed to determine the

The Thalamus in Schizophrenia

BIOL PSYCHIATRY
2000;47:944 –953

947

Table 2. Stereologic Sampling Parameters

3

Sampling grid size (mm 3 10 )
Mean # sections counted (6 SD)a
Mean # of neurons counted (6 SD)a
Sampling CE (Scheaffer et al 1996),
neuron numberb
Sampling CE (Gunderson and Jensen
1987), neuron numberc
Biological CE, neuron numberd

MD

AV/AM

2.0 3 1.5
11.1 (1.0)
345 (101)
0.049

1.2 3 1.0
9.5 (1.4)
204 (33)
0.064

0.05

0.07

0.064

0.054

MD, mediodorsal nucleus of the thalamus; AV/AM, anteroventral/anteromedial
nuclei of the thalamus.
a
For schizophrenic and control groups combined (n 5 16).
b
Estimated coefficient of error for the variance in sampling, given for a
representative brain (Scheaffer et al 1996).
c
Estimated coefficient of error for the variance in sampling, nugget corrected
estimator, given for a representative brain (Gunderson and Jensen 1987).
d
Coefficient of error for the between-subject variance of the control group 5
SEM/mean (n 5 8).

optimal counting procedures. Using a “counting frame” that
measured 140 3 80 3 10 mm, we varied the number of counting
frames (Figure 2) to identify from 100 to 600 neurons in several
replicate counts of a set of nine sections through the MD in one
case. The results indicated that counting 200 – 600 neurons
produced similar estimates of total neuron number. We adjusted
the number of counting frame placements used to count cells in
the two nuclei based on this result. Outlines of each region of
interest were digitized using StereoInvestigator software.
Nine to 11 sections were analyzed in each thalamic region.
Using more than 100 counting frames systematically placed
throughout the entire rostral-caudal extent of the nucleus, the
average density of neurons in the counting frames was determined using a 253 oil immersion objective (N.A. 0.75). Neurons
were identified as Nissl positive cell bodies containing a nucleolus that was clearly in focus within the counting frame (Figure
3). Thalamic nuclear volumes were calculated with Cavalieri’s
estimate from areas calculated by the StereoInvestigator software
using the outlined borders in each tissue section. The total
number of neurons in each region was estimated by multiplying
the average counting frame neuronal density and the calculated
volume of the thalamic region of interest (cell number/mm3 3
total nucleus volume [mm3] 5 estimated total cell number).

Statistics
A Student’s t test was used to compare cell number, nuclear
volume, and neuron density between the schizophrenic group and
control group. The p values reflect Bonferroni correction for two
multiple comparisons. Analysis of covariance for age, postmortem interval, and time-in-formalin was performed with neuron
number as the dependent variable.

Results
Total Cell Number
Anatomic changes were readily apparent in the MD and
AV/AM of the schizophrenic group, the most prominent

Figure 2. Stereologic cell counting requires the use of a “counting frame” and a “sampling grid.” (A) The dimensions of the
counting frame used in our study and the dimensions of the
sampling grid used to count cells in the mediodorsal nucleus of
the thalamus (MD). (B) Neurons whose nucleoli fall within the
counting frames are counted and used to calculate an average
neuronal density for the nucleus, which is multiplied by the
volume of the entire nucleus to determine an estimate of total
neuron number. AV, anteroventral nucleus.

change being a marked reduction in the total number of
neurons (Table 3). There was a mean of 4.15 million
neurons in the MD from control subjects and a mean of
2.70 million neurons in the MD of schizophrenic
subjects (t 5 5.14, p , .001), a 35% reduction. In every
control subject, the MD nucleus contained more than
3.5 million neurons; in every schizophrenic subject, it
contained less (Figure 4). Values obtained in the current
study for control MD neuron numbers are in agreement
with two previously published values (Harding et al
1994; Xuereb et al 1991). Analysis of covariance was
used to test for effects of several covariates on MD neuron
number estimates. These included age [F(1,11) 5 0.75,
p . .4], time in formalin fixative [F(1,11) 5 0.13, p . .7],
and postmortem interval [F(1,11) 5 0.4, p . .5]. The main
effect of diagnosis remained significant after accounting
for these covariates [F(1,11) 5 24.7, p , .0004]. Neuron
numbers were also reduced by an average of 16% in the
AV/AM of schizophrenic subjects, from 807,000 to
674,0000 (t 5 2.91, p , .012). Unlike the MD, in the
AV/AM there was overlap in the range of neurons for
schizophrenic subjects and control subjects.

Nucleus Volume and Cell Density
We also tested for differences in nuclear volume and
neuron density between schizophrenic and control subjects. In addition to a reduction in neuron number, there
was a significant 24% reduction in the volume of the
MD in schizophrenics (t 5 3.73, p 5 .002; Table 3 and
Figure 5). Although there was a 17% reduction in the
volume of the AV/AM in the schizophrenic brains, the

948

K.A. Young et al

BIOL PSYCHIATRY
2000;47:944 –953

Figure 3. Examples of cells counted in the
anterior nucleus of the thalamus in control
case C. (A, B) Single cells, each with a
darkly stained nucleolus, in the center of a
“counting frame” (80 3 140 mm). Only
cells with a clearly visible nucleolus that
could be focused within a 10-mm Z depth
were counted. These photomicrographs
were taken with the 253-oil objective used
for cell counting in the present study.

difference between the schizophrenic group and the
control group was not statistically significant (t 5 1.92,
p 5 .077). The MD was found in the same number of
sections along the rostro-caudal length in schizophrenic
group (11.25 6 1.2 sections) and the control group
(10.88 6 0.9 sections; t 5 0.693, p 5 .49). Similarly,
the AV/AM was found in the same number of sections
along the rostro-caudal length in schizophrenic brains
(9.37 6 1.8 sections) and control brains (9.62 6 1.3

sections; t 5 0.283, p 5 .78). Neuronal densities in
both the MD and AV/AM nuclei were not significantly different in the two groups (Table 3); however,
six of the eight schizophrenic subjects had lower MD
neuron densities than each of the control subjects. There
was no evidence for a neuron density reduction in the
AV/AM because both subject groups had the same
range of density values and virtually identical mean
densities.

Table 3. Thalamic Neuron Number, Nuclear Volume, and Neuronal Density
Neuron number
MD
AV/AM
Number 106 6 SD (n)
Control group
Schizophrenic group
t test

4.15 6 0.53 (8)
2.70 6 0.59 (8)
p , .001

0.807 6 0.101 (8)
0.674 6 0.071 (7)
p 5 .012

Volume
MD

AV/AM
mm3 6 SD (n)

248 6 35 (8)
188 6 29 (8)
p 5 .002

52.3 6 7.9 (8)
43.3 6 10.1 (7)
p 5 .077

MD, mediodorsal nucleus of the thalamus; AV/AM, anteroventral/anteromedial nuclei of the thalamus.

Neuronal density
MD
AV/AM
Neurons/mm3 6 SD (n)
1,696 6 256 (8)
1,426 6 416 (8)
p 5 .33

1,558 6 300 (8)
1,568 6 330 (7)
p 5 .79

The Thalamus in Schizophrenia

Figure 4. Cell number in schizophrenic and control thalamic
nuclei. Thalamic neuron numbers are reduced in schizophrenia.
The mean 6 SD is depicted in the figure. There are fewer
neurons in the mediodorsal (MD) and anterior (AV/AM) thalamic nuclei of schizophrenic subjects compared with normal
control subjects. The identification of individual subjects (letters
A–P) corresponds to the identifications given in Table 1. The
number of neurons in the schizophrenic group is 35% lower in
the MD (p , .001) and 16% lower in the AV/AM (p , .012).

Discussion
Thalamic Neuron Number Reductions in
Schizophrenia
Our study clearly indicates that there is a marked reduction
in the number of MD neurons in the postmortem schizophrenic brain. It represents the first replication of an initial
report of a reduction in the number of MD neurons in
schizophrenic subjects (Pakkenberg 1990). A striking
observation of both our study and the earlier report
(Pakkenberg 1990) is that there is virtually no overlap in
the ranges of the MD neuron numbers between normal
control subjects and schizophrenic subjects. In our experiment, each of the eight control subjects had more neurons
than each of the 8 schizophrenic subjects, whereas in the
Pakkenberg study, 11 of the 12 control subjects had more
neurons than each of the 12 schizophrenic brains. These
findings are also remarkable because of the magnitude of
the MD neuron number reductions: on average more than
one third of the normal complement of MD neurons is
missing from the schizophrenic brain. The large magnitude of the deficit and the consistency of the reduction
from patient to patient are uncommon in schizophrenia
research, in which anatomic deficits are not found consistently in every patient (Arnold and Trojanowski 1996).
The marked reduction in neuron number in the MD is
arguably the most robust neuroanatomic deficit in schizophrenia research reported to date.
The number of MD neurons in normal subjects has been
estimated to be approximately 4 million in two previous

BIOL PSYCHIATRY
2000;47:944 –953

949

Figure 5. Volume of schizophrenic and control thalamic nuclei.
The volume of the mediodorsal nucleus of the thalamus (MD) is
reduced in schizophrenia. The mean 6 SD for the two groups are
illustrated. The identification of individual subjects (letters A–P)
corresponds to the identifications given in Table 1. There was a
significant 24% reduction in the volume of the MD in schizophrenic subjects (p , .002), but no significant reduction in the
volume of the anterior thalamic nuclei (AV/AM; p . .077).

studies (Harding et al 1994; Xuereb et al 1991) and in our
study; however, Pakkenberg (1990) reported that control
brains have approximately 2 million MD neurons. This
disparity is most likely due to differences in defining the
borders of the MD (B. Pakkenberg, personal communication,
August 1999). Thus, although both Pakkenberg (1990) and
Xuereb et al (1991) used similar paraffin embedding tissue
processing procedures, MD volumes in the Pakkenberg study
were 50% smaller than in the Xuereb et al study.
Our study is the first to quantify the number of neurons
in the anterior nucleus of the thalamus in the schizophrenic
brain. Neuron numbers in the AV/AM were reduced by
16% on average. Unlike the MD, however, there was
overlap between normal and control neuron numbers.
These data suggest that AV/AM neuron number reductions may be present in only a subset of schizophrenic
individuals, whereas MD reductions are more consistently
found in the great majority of patients. It also has been
reported that there is a reduction in parvalbumin-positive
neurons in the AV nucleus of schizophrenic subjects
(Danos et al 1998). It is interesting that the AV/AM is
interconnected with medial temporal lobe (hippocampal)
circuits of the limbic system, where many morphologic
and functional abnormalities have been observed in
schizophrenia (see Arnold and Trojanowski 1996 for
review).
The thalamic neuron reductions so far revealed are
present in nuclei providing input to prefrontal and temporolimbic cortex, in which anatomic deficits have been
observed in postmortem and MRI studies. In a recent
report, for example, Goldstein et al (1999) used MRI to

950

BIOL PSYCHIATRY
2000;47:944 –953

measure the volume of 48 separate regions of the cerebral
cortex. The greatest volumetric reductions were observed
in the middle frontal gyrus, frontoorbital cortex, the insula
and the anterior cingulate cortex. The first three of these
brain regions have strong connectivity with the MD
nucleus (Giguere and Goldman-Rakic 1988; Ray and Price
1993), whereas the anterior cingulate is a primary cortical
target of the AV/AM nuclei (Vogt et al 1987). Our data
suggest that thalamic neuron number reductions may
contribute to cortical volume deficits. Further quantitative
study of all thalamic nuclei is underway to determine the
full extent of thalamic abnormalities in schizophrenia.

Thalamic Volume and Density Measurements
In the MD of schizophrenic subjects, a significant reduction in tissue volume was observed. These data support
previous MRI studies that have identified modest thalamic
volume reductions in schizophrenic individuals (Andreasen et al 1990; Flaum et al 1995; Gur et al 1998; Staal
et al 1998). In one study (Buchsbaum et al 1996), for
example, the volume of the left anterior thalamus was
selectively reduced by 15%. Because there is a reduction
in the volume of the thalamus in never-medicated schizophrenic individuals (Buchsbaum et al 1996; Gur et al
1998; Pakkenberg 1992), it appears unlikely that neuroleptic effects are responsible for such volume reductions.
Increased volumes of the lateral ventricles accompany
smaller thalamic volumes in schizophrenic subjects (Portas et al 1998). These MRI findings raise the possibility
that thalamic volume reductions may be associated with
the enlargement of the lateral ventricles observed in
schizophrenia (Cannon et al 1998; see review by McCarley et al 1999). The modest reduction in whole thalamic
volume (,10%) reported in some MRI studies of schizophrenic individuals (Andreasen et al 1990; Flaum et al
1995) and reports of normal thalamic size in other studies
(Wolkin et al 1998; for review, see McCarley et al 1999)
suggest that thalamic volume changes in schizophrenia
may be limited to select regions of the thalamus such as
the AV/AM and MD.
In our study, the density of neurons in the MD and
AV/AM nuclei of schizophrenic subjects was normal even
though total neuron numbers were significantly reduced in
these nuclei. The 15% lower mean density of neurons in
the MD was not significantly different from control
subjects, although it is possible that a larger sample size
would have identified this change as significant. Our data
suggest that there is a large effect of nuclear volume
reduction on neuron number in the MD of schizophrenics.
Nonetheless, the findings do not exclude the possibility
that neuron number reductions in this region are the
product of both a density and a volume reduction. The

K.A. Young et al

findings in the MD contrast with the pattern observed in
the AV/AM, in which neuronal density was virtually
identical in both groups. This latter observation replicates
a recent study (Danos et al 1998) that identified differences in the density of a select population of thalamocortical neurons in the AV but did not find a significant
change in the density of AV neurons.
Historically, anatomic studies have relied on measurement of neuronal density and regional volume, measurements that in our study provided a less robust indication of
thalamic neuropathology than total neuron counts. Because postmortem tissue preservation and tissue shrinkage
influence volume and density measurements, there is often
greater variance associated with these measures compared
with neuron number estimates. Variable tissue shrinkage
from subject to subject has no net effect on total neuron
counts when stereologic procedures are used (West 1994).
The observation that neuronal density and nuclear volumes are reduced only modestly or are unaffected in the
thalamus of schizophrenic individuals provides one explanation for why the thalamus has not been previously
recognized as a locus of robust anatomic change in
schizophrenia.

Relationship between Thalamic Deficits in
Schizophrenia and Cortical Anatomy and Function
Because of the intimate connections between the thalamus
and cortex, thalamic neuron number reductions in schizophrenia may have profound influences on cortical anatomy
and function. There are reciprocal connections between
the MD and dorsolateral prefrontal, orbitofrontal, frontal
eye fields, and insular cortices (Giguere and GoldmanRakic 1988; Mufson and Mesulam 1984), and the AV/AM
region with the anterior cingulate cortex (Vogt et al 1987).
Thalamic neuron number reductions in MD and
AV/AM may contribute to the reported anatomic changes
in the frontal and cingulate cortex of schizophrenic subjects. In particular, the findings are relevant to the “reduced neuropil” hypothesis (Selemon and Goldman-Rakic
1999). This hypothesis is based on the observation that
there are increased densities of neurons in the middle
lamina of the dorsolateral prefrontal cortex of schizophrenic individuals, but no reductions in neuron number,
suggesting that there is a decreased amount of cortical
neuropil in certain lamina (Akbarian et al 1995; Daviss
and Lewis 1995; Selemon et al 1998; Selemon and
Goldman-Rakic 1999). Other prefrontal anatomic deficits
include decreased dendritic spine density for middle lamina pyramidal neurons (Garey et al 1998), reductions in
synaptic protein levels (Glantz and Lewis 1997), and
abnormalities in the distribution of serotonin-1A/2A receptors (Burnet et al 1996). In the anterior cingulate

The Thalamus in Schizophrenia

cortex, there are changes in lamina II and III GABA
receptors (Benes 1993) and other subtle findings (Benes
1993; Heckers 1997; Lahti et al 1998). Schizophrenic
individuals have abnormalities in the distribution of
“guidepost” interstitial white matter neurons (Akbarian et
al 1996) in the frontal cortex. Because thalamocortical
axons project into lamina III and IV in the frontal cortex
(Giguere and Goldman-Rakic 1988), a reduction in the
number of thalamocortical axons innervating the frontal
lobes could contribute to the observed reductions in
neuropil and spine density. Our findings of neuron number
reductions in the thalamus of schizophrenic subjects suggests that perturbation of thalamocortical connections may
contribute to reduced neuropil and other cortical deficits in
schizophrenia.
Thalamic cell number reductions in schizophrenia may
also contribute to metabolic activation deficits observed in
the frontal cortex. Thalamocortical projections provide
excitatory input to the frontal cortex, which influences the
level of cortical metabolism and blood flow. For instance,
animal studies have shown that inhibition of MD neuron
activity selectively depresses metabolic activity in the
frontal cortex (Young et al 1994); in humans, thalamic
vascular accidents often induce frontal lobe metabolic
hypoactivity (Casselli et al 1991; Szelies et al 1991). In
schizophrenic individuals, a reduction in the number of
thalamic neurons available to activate or sustain frontal
cortex neural activity may impair metabolic activation
during performance of verbal working memory and sustained attention tests, resulting in frontal cortical metabolic
hypoactivity (Berman et al 1986; Fletcher et al 1998;
Goldberg et al 1994; Stevens et al 1998; Weinberger and
Gallhofer 1997).
Reductions in MD and AV/AM neuron number, through
interactions with frontal circuits, have the potential to
influence behaviors relevant to schizophrenia. MD lesions
in rats produce increased perseverative behaviors (Hunt
and Aggleton 1998), and perseverative behaviors are
observed in schizophrenic subjects performing the Wisconsin Card Sorting Task (Berman et al 1986; Goldberg et
al 1994). Strokes that selectively affect the MD and
AV/AM can induce a schizophrenialike thought disorder
characterized by bizarre, disconnected, and incoherent
speech patterns (Chatterjee et al 1997). Accidental thalamic lesions that affect the MD and anterior nuclei are
particularly prone to create disturbances in verbal memory
and language (Casselli et al 1991; Squire and Moore 1979;
Szelies et al 1991). Finally, the thalamus is metabolically
activated during episodic hallucinations (Silbersweig et al
1995), and elevated thalamic, frontal, and cingulate metabolic activity is associated with delusions and thought
disorder (Erkwoh et al 1997; Sabri et al 1997). Because of
the important role played by the thalamus in support of

BIOL PSYCHIATRY
2000;47:944 –953

951

cortical processing, it is clear that the substantial reduction
in the number of neurons in the MD and anterior nuclei of
schizophrenics could have profound effects on cortical
metabolic activation deficits and behavioral signs and
symptoms of schizophrenia.

Mechanisms Involved in Thalamocortical
Pathology in Schizophrenia
There are a variety of mechanisms that could potentially
account for the observed pattern of MD and AV/AM
neuron number reductions in schizophrenia. Broadly defined, the mechanisms fall into two categories: developmental and degenerative. First, fewer MD and anterior
thalamic neurons may have been generated, or they may
have died during the development and maturation of the
brain. A developmental mechanism promoting thalamic
neuron number reductions would include a failure to
develop or maintain appropriate synaptic connections with
cortical targets. Developmental processes have been
widely implicated as factors in producing anatomic deficits in schizophrenia (Andreasen 1997b; Hyde and Weinberger 1990; Jones 1997a; Raedler et al 1998; Stefen and
Murray 1997). Alternatively, a degenerative process could
be involved in MD and AV/AM nerve cell number
reductions. Thalamic neurons may develop normally but
subsequently die because of a necrotic degenerative process. This later process usually produces alterations of
glial cells (gliosis) in the regions where neuronal degeneration has occurred. As an example, following dorsolateral prefrontal cortical lesions in rats, a secondary neuronal degeneration occurs in the MD. Thalamic gliosis is
observable for at least 40 days after the lesion placement
(van Eden et al 1998); however, in schizophrenic individuals there are no increases in the number of glial cells in
the MD nucleus (Pakkenberg 1990). The lack of a gliotic
reaction in the schizophrenic brain has been interpreted as
evidence against a necrotic involvement in the disease.
Further investigation is needed to determine why there are
fewer neurons in the thalamus of patients with
schizophrenia.
In conclusion, our study identifies the MD and AV/AM
thalamus as regions of robust nerve cell number reduction
in schizophrenia. The “double hit” on the MD and anterior
thalamus in schizophrenia is consistent with other data that
indicate that the frontal cortex and limbic system are
anatomically handicapped in schizophrenia. The MD and
anterior thalamic “lesions” appear to be well positioned to
influence the signs and symptoms of schizophrenia (Andreasen 1997b) by decreasing motivation, causing failure
of mental activities, and altering emotional responses
through interactions with executive and emotional control
centers in the frontal and cingulate cortex. Historically, it

952

BIOL PSYCHIATRY
2000;47:944 –953

has been difficult to find marked abnormalities in the
schizophrenic brain that parallel those found in diseases
such as Huntington’s disease and Parkinson’s disease,
making the study of schizophrenia difficult to approach
from a mechanistic level; however, the finding of robust
neuroanatomic deficits in the thalamus provides an opportunity to better focus anatomic research in schizophrenia.
It is likely that further study of the thalamus in schizophrenia will reveal important clues about the etiology of
this debilitating mental illness.
This work was supported by NIMH (Grant No. MH55879), the Scott,
Sherwood and Brindley Foundation, Veterans Affairs VISN 17 New
Investigator Award, the Theodore and Vada Stanley Foundation, and the
John Shermerhorn Fund. The authors thank S. Bauserman, C. Conley, E.
Johnson, C. Cook, and E. Rachut for assistance in obtaining postmortem
material and Dr. E. G. Jones for assistance in validating our definition of
the mediodorsal thalamic nucleus.

References
Akbarian S, Kim JJ, Potkin SG, Hagman JO, Tafazzoli A,
Bunney WE Jr, et al (1995): Gene expression for glutamic
acid decarboxylase is reduced without loss of neurons in
prefrontal cortex of schizophrenics. Arch Gen Psychiatry
52:258 –266.
Akbarian S, Kim JJ, Potkin SG, Hetrick WP, Bunney WE Jr,
Jones EG (1996): Maldistribution of interstitial neurons in
prefrontal white matter of the brains of schizophrenic patients. Arch Gen Psychiatry 53:425– 436.
Andreasen NC (1997a): The evolving concept of schizophrenia:
From Kraepelin to the present and future. Schizophr Res
28:105–109.
Andreasen NC (1997b): The role of the thalamus in schizophrenia. Can J Psychiatry 42:27–33.
Andreasen NC, Arndt S, Swayze V, Cizadlo T, Flaum M,
O’Leary D, et al (1994): Thalamic abnormalities in schizophrenia visualized through magnetic resonance image averaging. Science 266:294 –298.
Andreasen NC, Carpenter WT (1993): Diagnosis and classification of schizophrenia. Schizophr Bull 19:199 –214.
Andreasen NC, Ehrhardt JC, Swayze VW, Alliger RJ, Yuh
WTC, Cohen G, et al (1990): Magnetic resonance imaging of
the brain in schizophrenia: The pathophysiologic significance
of structural abnormalities. Arch Gen Psychiatry 47:35– 44.
Arnold SE, Trojanowski JQ (1996): Recent advances in defining
the neuropathology of schizophrenia. Acta Neuropathol
(Berl) 92:217–231.
Benes FM (1993): Neurobiological investigations in cingulate
cortex of schizophrenic brain. Schizophr Bull 19:537–549.
Berman KF, Zec RF, Weinberger DR (1986): Physiological
dysfunction of dorsolateral prefrontal cortex in schizophrenia.
Arch Gen Psychiatry 43:126 –135.
Buchanan RW, Vladar K, Barta PE, Pearlson GD (1998):
Structural evaluation of the prefrontal cortex in schizophrenia. Am J Psychiatry 155:1049 –1055.
Buchsbaum MS, Someya T, Teng CY, Abel L, Chin S, Najafi A,

K.A. Young et al

et al (1996): PET and MRI of the thalamus in nevermedicated patients with schizophrenia. Am J Psychiatry
153:191–199.
Burnet PW, Eastwood SL, Harrison PJ (1996): 5-HT1A and
5-HT2A receptor mRNAs and binding site densities are
differentially altered in schizophrenia. Neuropsychopharmacology 15:442– 455.
Cannon TD, Van Erp GM, Huttunen M, Lonnqvist J, Salonen O,
Valanne L, et al (1998): Regional gray matter, white matter,
and cerebrospinal fluid distributions in schizophrenic patients, their siblings and controls. Arch Gen Psychiatry
55:1084 –1091.
Casselli RJ, Graff-Radford NR, Rezai K (1991): Thalamocortical
diaschesis: Single photon emission tomographic study of
cortical blood flow changes after focal thalamic infarction.
Neuropsychiatry Neuropsychol Behav Neurol 4:193–214.
Chatterjee A, Yapundich R, Mennemeier M, Mountz JM, Inampudi C, Pan JW, et al (1997): Thalamic thought disorder:
On being “a bit addled.” Cortex 33:419 – 440.
Danos P, Baumann B, Bernstein H-G, Franz M, Stauch R,
Northoff G, et al (1998): Schizophrenia and anteroventral
thalamic nucleus: Selective decrease of parvalbumin-immunoreactive thalamocortical projection neurons. Psychiatry
Res 82:1–10.
Daviss SR, Lewis DA (1995): Local circuit neurons of the
prefrontal cortex in schizophrenia: Selective increase in the
density of calbindin-immunoreactive neurons. Psychiatry Res
59:81–96.
Erkwoh R, Sabri O, Steinmeyer EM, Bull U, Sass H (1997):
Psychopathological and SPECT findings in never-treated
schizophrenia. Acta Psychiatr Scand 96:51–57.
Flaum MA, Swayze VW, O’Leary DS, Yuh WTC, Ehrhardt JC,
Arndt SV, et al (1995): Effects of diagnosis, laterality and
gender on brain morphology in schizophrenia. Am J Psychiatry 152:704 –714.
Fletcher PC, McKenna PJ, Friston KJ, Dolan RJ (1998): Brain
activations in schizophrenia during a graded memory task
studied with functional imaging. Arch Gen Psychiatry 55:
1001–1008.
Garey LJ, Ong WY, Patel TS, Kanani M, Davis A, Mortimer
AM, et al (1998): Reduced dendritic spine density on cerebral
cortical pyramidal neurons in schizophrenia. J Neurol Neurosurg Psychiatry 65:446 – 453.
Giguere M, Goldman-Rakic PS (1988): Mediodorsal nucleus:
Arial, laminar, and tangential distribution of afferents and
efferents in the frontal lobe of Rhesus monkeys. J Comp
Neurol 277:195–213.
Glantz LA, Lewis DA (1997): Reduction of synaptophysin
immunoreactivity in the prefrontal cortex of schizophrenic
subjects: Regional and diagnostic specificity. Arch Gen Psychiatry 54:660 – 669.
Goldberg TE, Torrey EF, Berman KF, Weinberger DR (1994):
Relations between neuropsychological performance and brain
morphological and physiological measures in monozygotic
twins discordant for schizophrenia. Psychiatry Res 55:51– 61.
Goldstein JM, Goodman JM, Seidman LJ, Kennedy DN, Makris
N, Lee H, et al (1999): Cortical abnormalities in schizophrenia identified by structural magnetic resonance imaging. Arch
Gen Psychiatry 56:537–547.

The Thalamus in Schizophrenia

Gunderson HJG, Jensen EB (1987): The efficiency of sampling
in stereology and its prediction. J Microsc 147:229 –263.
Gur RE, Maany V, Mozley PD, Swanson C, Bilker W, Gur RC
(1998): Subcortical MRI volumes in neuroleptic naive and
treated patients with schizophrenia. Am J Psychiatry 155:
1711–1717.
Harding AJ, Halliday GM, Cullen K (1994): Practical considerations for the use of the optical disector in estimating
neuronal number. J Neurosci Methods 51:83– 89.
Haznadar MM, Buchsbaum MS, Lu C, Hazlett EA, Seigel BV,
Lohr J, et al (1997): Decreased anterior cingulate gyrus
metabolic rate in schizophrenia. Am J Psychiatry 154:682–
684.
Heckers S (1997): Neuropathology of schizophrenia: Cortex,
thalamus, basal ganglia and neurotransmitter-specific projection systems. Schizophr Bull 23:401– 421.
Hunt PR, Aggleton JP (1998): Neurotoxic lesions of the dorsomedial thalamus impair the acquisition but not the performance of delayed matching to place by rats; a deficit in the
shifting response rules. J Neurosci 181:9099 –9111.
Hyde TM, Weinberger DR (1990): The brain in schizophrenia.
Semin Neurol 10:276 –286.
Jones EG (1997a): Cortical development and thalamic pathology
in schizophrenia. Schizophr Bull 23:483–501.
Jones EG (1997b): A description of the human thalamus. In:
Steriade M, Jones EG, McCormick DA, editors. Thalamus,
Vol II. New York: Elsevier, 425– 499.
Lahti RA, Roberts RC, Cochrane EV, Primus RJ, Gallager DW,
Conley RR, et al (1998): Direct determination of dopamine
D4 receptors in normal and schizophrenic postmortem brain
tissue: A [3H]NGD-94-1 study. Mol Psychiatry 6:528 –533.
McCarley RW, Wible CG, Frumin M, Hirayasu Y, Levitt JL,
Fischer IA, et al (1999): MRI anatomy of schizophrenia. Biol
Psychiatry 45:1099 –1119.
Mufson EJ, Mesulam MM (1984): Thalamic connections of the
insula in the rhesus monkey and comments on the paralimbic
connectivity of the medial pulvinar nucleus. J Comp Neurol
227:109 –120.
Pakkenberg B (1990): Pronounced reduction of total neuron
number in mediodorsal thalamic nucleus and nucleus accumbens in schizophrenics. Arch Gen Psychiatry 47:1023–1028.
Pakkenberg B (1992): The volume of the mediodorsal thalamic
nucleus in treated and untreated schizophrenics. Schizophr
Res 7:95–100.
Portas CM, Goldstein JM, Shenton ME, Hokama HH, Wible CG,
Fischer I, et al (1998): Volumetric evaluation of the thalamus
in schizophrenic male patients using magnetic resonance
imaging. Biol Psychiatry 43:649 – 659.
Raedler TJ, Knable MB, Weinberger DR (1998): Schizophrenia
as a developmental disorder. Curr Opin Neurobiol 8:157–
161.
Rajkowska G, Selemon LD, Goldman-Rakic PS (1998): Neuronal and glial somal size in the prefrontal cortex: A postmortem morphometric study of schizophrenia and Huntington’s
disease. Arch Gen Psychiatry 55:215–224.
Ray JP, Price JL (1993): The organization of projections from the
mediodorsal nucleus of the thalamus to orbital and medial

BIOL PSYCHIATRY
2000;47:944 –953

953

prefrontal cortex in macaque monkeys. J Comp Neurol
337:1–31.
Sabri O, Erkwoh R, Scheckenberger M, Owega A, Sass H, Buell
U (1997): Correlation of positive symptoms exclusively to
hyperperfusion or hypoperfusion of cerebral cortex in nevermedicated schizophrenics. Lancet 349:1735–1739.
Scheaffer RL, Mendenhall W, Ott L (1996): Elementary Survey
Sampling, 5th ed. Boston: PWS-Kent.
Scheibel AB (1997): The thalamus and neuropsychiatric illness.
J Neuropsychiatry 9:342–353.
Selemon LD, Goldman-Rakic PS (1999): The reduced neuropil
hypothesis: A circuit based model of schizophrenia. Biol
Psychiatry 45:17–25.
Selemon LD, Rajkowska G, Goldman-Rakic PS (1998): Elevated
neuronal density in prefrontal area 46 in brains from schizophrenic patients: Application of a three-dimensional, stereologic counting method. J Comp Neurol 392:402– 412.
Silbersweig DA, Stern E, Frith C, Cahill C, Holmes A,
Grootoonk S, et al (1995): A functional neuroanatomy of
hallucinations in schizophrenia. Nature 378:176 –179.
Squire LR, Moore RY (1979): Dorsal thalamic lesion in a noted
case of human memory dysfunction. Ann Neurol 6:503–506.
Staal WG, Hulshoff HE, Schnack H, Van der Schot AC, Kahn
RS (1998): Partial volume decrease of the thalamus in
relatives of patients with schizophrenia. Am J Psychiatry
155:1784 –1786.
Stefen MD, Murray RM (1997): Schizophrenia: Developmental
disturbance of brain and mind. Acta Paediatr 442:112–116.
Stevens AA, Goldman-Rakic PS, Gore JC, Fulbright RE, Wexler
BE (1998): Cortical dysfunction in schizophrenia during auditory word and tone working memory demonstrated by functional
resonance imaging. Arch Gen Psychiatry 55:1097–1103.
Szelies B, Herholz K, Pawlik G, Karbe H, Hebold I, Wolf-Dieter
H (1991): Widespread functional effects of discrete thalamic
infarction. Arch Neurol 48:178 –182.
van Eden CG, Rinkins A, Uylings HMB (1998): Retrograde
degeneration of thalamic neurons in the mediodorsal nucleus
after neonatal and adult aspiration lesions of the medial
prefrontal cortex in the rat. Implications for mechanisms of
functional recovery. Eur J Neurosci 10:1581–1589.
Vogt BA, Pandya DN, Sosene DL (1987): Cingulate cortex of
the rhesus monkey: I. Cytoarchitecture and thalamic afferents. J Comp Neurol 262:256 –270.
Weinberger DR, Gallhofer B (1997): Cognitive function in
schizophrenia. Int Clin Psychopharmacol 12(suppl 4):S29 –
S36.
West MJ (1994): New stereological methods for counting neurons. Neurobiol Aging 14:275–285.
Wolkin A, Rusinek H, Vaid G, Arena L, Lafargue T, Sanfilipo
M, et al (1998): Structural magnetic image averaging in
schizophrenia. Am J Psychiatry 155:1064 –1073.
Xuereb JH, Perry RH, Candy JM, Perry EK, Marshall E, Bonham
JR (1991): Nerve cell loss in the thalamus in Alzheimer’s
disease and Parkinson’s disease. Brain 114:1363–1379.
Young KA, Hicks PB, Randall PK, Wilcox RE (1994): Behavioral and frontal cortical metabolic effects of apomorphine
and muscimol microinjections into the mediodorsal thalamic
nucleus. J Neural Transm Gen Sect 98:119 –132.