- Email: [email protected]
The effect of neuroleptics on dysfunction in a prefrontal substrate of sustained attention in schizophrenia
Pergamon Press
Life Sciences, Vol. 43, pp. 1141-1150 Printed in the U.S.A.
THE EFFECT OF NEUROLEPTICSON DYSFUNCTION IN A PREFRONTAL SUBSTRATEOF SUSTAINEDATTENTIONIN SCHIZOPHRENIA Robert M. Cohen, William E. Semple, Michael Gross, Thomas E. Nordahl, Henry H. Holcomb, H. Susan Dowling, David Pickar* Section
on Clinical Brain Imaging, Laboratory of Cerebral Metabolism, NIMH, Bethesda, MD 20892 *Clinical Neuroscience Branch, NIHH, Bethesda, MD 20892 (Received
in final
form August
12,
1988)
Summary Regional brain metabolism was measured in patients vith schizophrenia during the performance of auditory discrimination to evaluate the effect of neuroleptic medication on the middle prefrontal cortex, an area that appears to be an important biological determinant of the sustained attention required of subjects in this task. While unmedicated patients with schizophrenia have lower than “normal” metabolic rates in this cortical area that are unrelated to performance, patients receiving neuroleptics appear to have this metabolic deficit attenuated with higher metabolism in this area associated with greater accuracy of performance. This change occurs in the context of differential neuroleptic effects on the cortical regions of the frontal cortex, increased metabolism in the basal ganglia, thalamus and temporal lobes, and no apparent effect in the parietal and occipital cortices. The findings suggest that only part of the abnormality in the prefrontal cortex determination of sustained attention in schizophrenia is sensitive to neuroleptic treatment. Recently, we reported that the positron emission tomographically determined metabolic rates of the middle prefrontal cortex of patients with schizophrenia, measured while performing an auditory discrimination task, were lower than those observed in normal controls performing the same task (1). Horeover , the metabolic rates of this region in normals, but not in unmedicated patients with schizophrenia, were directly related to their accuracy of performance. The findings were interpreted as indicating that the mid-prefrontal cortex is an important biological determinant of sustained attention and that dysfunction of this region occurs in schizophrenia. It remained to be determined whether the sustained attention problems, observed in this particular continuous performance task in schizophrenia, and those measured by other psychological procedures, arise from a specific neurotransmitter defect in the middle prefrontal cortex or from other brain structures that could affect both metabolic rate and sustained attention. To help resolve this issue, we examined middle prefrontal cortex metabolism during the performance of auditory discrimination in patients receiving clinically titrated doses of neuroleptics.
0024-3205/88
$3.00
+ .OO
1142
Neuroleptics:.Attention
in Schizophrenia
Vol.
43, No. 14,
1988
METHODS Patients met DSM III (American Psychiatric Association) criteria Subjects: for schizophrenia as determined by the consensus of two psychiatrists and were residing on research units at the time of study. Unmedicated (drug-withdrawn) patients had not received neuroleptic medications for an average of 34 days (range 13-73, s.d. = 17.5; 1 patient had never received Of the unmedicated patients, 7 were of the treatment) prior to study. undifferentiated type, 7 of the paranoid type, 1 of the residual type and 1 of The medicated patients were predominantly the disorganized hebephrenic type. of the undifferentiated (n = 2) and of the paranoid type (n - 4) with 1 patient each meeting criteria for disorganized and atypical. The medicated patients had received fluphenazine for 31.1 days prior to time of scan with a 26.1 mg/day average dose at time of scan. Demographic, state and performance characteristics of the medicated and unmedicated patients are summarized in Although there were no females in the medicated patient group, our Table I. analysis of both normal females and females with schizophrenia have not demonstrated any significant effect of gender on our normalized regional Scans of unmedicated and medicated patients were glucose metabolic rates. obtained during the same time period. The auditory continuous performance task (CPT) consisted of Behavioral Task: a random series of 500 hz tones of 1.0 second duration and 2.0 second inter-tone-interval, with an intensity of 67, 75, or 86 decibels, measured at The subject was instructed to press the hand-held the earphone-ear interface. response button when the lowest volume tone was detected. The task was pf@sented in successive 5 minute blocks for 35 minutes following [ F]-2-fluoro-2-deoxy-D-glucose (FM;) injection. A total of 220 targets and 440 distracters were presented to each subject. To minimize learning contributions to the metabolic pattern, subjects were trained to a criterion of 18 of 20 correct identifications in this task in the hour preceding Subjects received scores for correct identifications of target FDG injection. tones (hits) and incorrect identification of distracters (false alarms). Although performance of this task requires motor activity, response selection and inhibition, discrimination, memory, sensory processing, and a variety of components of attention including arousal, we believe that the task design and the PET-FDC method of measurement which averages functional brain activity over a 35 minute interval, emphasizes those aspects of the task involved in the process of sustaining attention (See Ref. 2 for a more complete discussion of these issues). Subjects, with eyes patched, began their auditory PET Scan Procedure: discr6mination task several minutes prior to the injection of a 3-5 mCi dose of [‘?]FDC and completed the task 35 min after injection. Arterialized ous blood samples were taken from the left arm for quantification of Following the 35-min uptake period, subjects were placed in ‘;@F]FDC uptake. the scanner. Seven to eight slices were obtained from each subject starting from 80% of head height and proceeding in the caudal direction. Slices were parallel to the canthomeatal line (CM) and the interslice interval was 13 mm. Scans were performed with an Ortec ECAT II scanner of 1.75 cm. full width half-maximum resolution and used a calculated attenuation correction. Raw pixel values were converted to glucose metabolic rate in Data Analysis: For the extraction of regional glucose mg/lOO g tissue/minute $4,5,6). metabolic rates, 2.9 cm boxes were placed over the regions of interest in Plane A (9.1 cm above CM), B (7.8 cm above), C (6.5 cm five standard planes; above), D (5.2 cm above), E (3.9 cm above). Two independent raters who were
Vol.
43, No. 14,
1988
Neuroleptice:
Attention
in Schizophrenia
1143
unaware of the identity and diagnosis of the individual whose scan they were evaluating selected the planes for analysis and placed the boxes through neuroanatomical matching to a standard template (7)(See Fig. 1). Anatomical structures are judged as contained within these regions on the basis of the atlas of Matsui and Mrano (8). Global glucose metabolic rates refer to the estimates of the average value for glucose metabolism obtained for all the Regional glucose metabolic rates gray matter rich areas of the brain sampled. refer to the average of standardized absolute glucose metabolic data obtained The standardization procedure from specific areas of the cortex (See Fig. 1). is designed to minimize the effects of individual variation in global glucose metabolism on regional comparisons by dividing an individual’s glucose metabolic rate in a specific region by the individual’s mean global glucose metabolic rate. The procedure is similar in principle to the “reference ratio” or “landscape method” (7,9). All statistical comparisons of regional metabolic rates use these normalized (regional/global mean) values. It is important to note that although standardization of metabolic rates was necessary in most instances to establish significance the size of the percent changes in absolute glucose metabolic rates in these same regions were always in the same direction and of the same or usually larger magnitude.
FWtwiu
Occipital
Fig.
1
Schematic representation of regions sampled in the left and right hemispheres. Regions labeled as medial, although sampled from the medial portion of the cortex, are represented as incomplete boxes on the lateral surface. The boxes outlined by dashed lines are sampled from the surface of the frontal cortex, medial to the temporal cortex. The medial and anterior frontal regions, Plane A and the medial, anterior and posterior frontal regions, Plane B are referred to in the text as superior prefrontal cortex. The medial, anterior and posterior regions, Planes C and D are referred to as middle (or mid-) prefrontal cortex. The region labeled as medial parietal is also referred to in the text as superior posterior parietal cortex.
1144
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
1988
RESULTS The patients with schizophrenia who were studied while receiving neuroleptics did not differ in age, duration of illness, number of hospitalizations, symptomatology at time of scan, or in their performance on the auditory discrimination task from the sixteen drug-withdrawn patients previously studied (See Table I). This may, however, not indicate that the patients’ symptoms and discrimination performance were not responsive to neuroleptic treatment since all patients were considered improved while they were receiving neuroleptics by their treating physicians who were unaware of their medication status or their scan results. Thus, the similar BPRS ratings may have resulted from the pre-selection criteria which were to obtain patients who were highly symptomatic, but not so ill as to preclude valid scanning. TABLE I DEMOGRAPHIC,STATE AND PERFORMANCE CHARACTERISTICS OF PATIENTS WITH SCHIZOPHRENIA DRUG-FREE (n=16) BPRS (22 item)* Total Thought Disorder Anxiety/Depression Hostility/Suspiciousness Withdrawal/Retardation
58.7 9.94 7.25 6.63 7.01
f + f f f
MEDICATED (n=8)
STATISTICAL SIGNIFICANCE
15.0 4.23 4.00 3.05 2.24
58.4 10.4 7.50 8.00 8.62
* * f f f
12.1 3.85 2.07 2.97 2.92
NS NS NS NS NS
AGE OF ONSET
18.8
f 3.43
18.4
f 3.20
NS
DURATIONOF ILLNESS (IN YEARS)
8.00
+ 6.30
10.6
* 9.59
NS
NUMBEROF HOSPITALIZATIONS
4.81
+ 4.47
4.50
f 3.92
NS
AGE AT TIME OF STUDY
26.9
f 5.83
28.8
f 9.02
NS
f 62.1 k 24.1
NS NS
SEX FEHALES HALE PERFORMANCE HITS FALSE ALARMS
*
4 12 138.3 28.5
0
8 * 66.0 * 35.0
The 22-item Brief Psychiatric Rating Scale for the period just prior to the scan.
134.8 22.0 (BPRS) (3)
is
the average
score
Differences in global glucose metabolism between the medicated and unmedicated patients were also not apparent. Glucose metabolic rates in 55 to maximize the quantitative regions of the brain (see Fig. l), selected localization of brain function, and which were the same regions examined in our earlier study of unmedicated patients with schizophrenia, were evaluated for differences.
Vol.
43, No. 14,
Neuroleptics:
1988
Attention
in Schizophrenia
1145
TABLE II SELECTEDNORMALIZED REGIONALGLUCOSEMETABOLICRATE DIFFERENCES BETWEENMEDICATEDAND MEDICATION-WITHDRAWN PATIENTS WITR SCBIZOPBRENIA FRONTALCORTEX Region
L. Posterior
L. Anterior
Medial
R. Posterior
R. Anterior
Plane A
-4.6
-7.3
-10.4**
-9.8**
-8.O*
B
-4.6
-4.7
-8.2*
-6.4**
-4.3
C
-1.5
-0.2
-3.5
-0.2
-5.3*
D
+2.6
+4.5
+3.4
+1.4
+1.7
E
-0.4
+0.3
+1.9
-1.1
-1.8
TEMPORALCORTEX Region Plane D
L. Posterior
L. Middle
+8.6
L. Anterior
+7.4
E
+9.3
R. Anterior
**
+?.5
+3.4
R. Middle
**
+5.3
R. Posterior +7.8
+5.3
L. Hippocampus * +9.2
R. Hippocampus +9.7 SUBCORTICALSTRUCTURES
L. Basal
R. Basal
Ganglia
+6.7*
Ganglia
+5.4+ L. Thalamus +6.3
R. Thalamus +3.0
The table is derived by subtracting the mean normalized (regional/global mean) regional glucose metabolic rates of medication-withdrawn patients (n=16) with schizophrenia from those of medicated patients (n=8) and multiplying by 100. Statistically significant differences based on Z-tailed t-tests are represented by +, p < 0.1; *, p < 0.05; **, p < 0.01. Values for mean normalized metabolism and standard deviation in each region for the normal control group have been reported Those for schizophrenia will be reported for a study comparing schizophrenia (2). with severe affective disorder. These data can be obtained by request.
Because our primary interest was the examination of those regions of the brain to which the ability to perform auditory discrimination localized to in normal controls, our first evaluations were of the metabolic rates of the superior posterior medial parietal cortex and the anterior cingulate gyrus (normal controls performing auditory discrimination have lower rates in these
1146
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
regions than normal controls at rest) , and of the right anterior frontal and the right posterior frontal regions (7%) of Plane C (normal controls performing auditory discrimination have higher metabolic rates in these regions than normal controls at rest) (2). These regions did not differ significantly between the drug-withdrawn and treated patient groups.
1988
(5%)
Significant differences were, however, observed in regions to which sustained attention did not localize to in normals. The medicated patients compared to the drug-withdrawn patients were found to have higher glucose metabolic rates throughout the temporal cortex including the hippocampus, and in the subcortical structures of the basal ganglia and thalamus, although the latter structure did not reach statistical significance. The effect of neuroleptic treatment on metabolic activity in the frontal cortex was primarily observed in the superior frontal cortex (Planes A and B) with treated patients showing significantly lower glucose metabolic rates compared to their drug-withdrawn counterparts (Table II). Statistically significant differences were not observed in the 15 sampled regions of the parietal and occipital cortices. TABLE III MID-PREFRONTAL CORTEXGLUCOSEMETABOLICRATE--PERFORMANCE ASSOCIATIONS
Brain Region Plane C medial
Schizophrenic Unmedicated
Patients Medicated
0.014
(-0.067)
-0.856**
(0.874**)
R anterior
-0.123
(0.237)
-0.742*
(0.671)
L anterior
-0.259
(0.041)
-0.608
(0.667)
Plane D medial
-0.310
(0.109)
-0.615
(0.774*)
R anterior
0.049
(0.140)
-0.870**
(0.809*)
L anterior
-0.293
(0.036)
-0.664
(0.809*)
The significant within group Pearson product-moment correlations observed between regional glucose metabolic rates of the mid-prefrontal cortex and performance are reported. The numbers outside of parentheses represent correlations to false alarms; a negative correlation indicates that higher metabolic rates in the region of interest are associated with fewer errors. The correlations within parentheses are with hits; a positive correlation indicates that higher glucose metabolic rates in the region of interest are associated with a greater number of hits. Statistically significant Additional associations are represented by *, p 5 0.05 and **, p 5 0.01. analyses using the log ratio of hits to false alarms, another measure of accuracy, did not demonstrate any significant correlations between performance and metabolic rate in these regions in the medication-withdrawn patients, but again demonstrated significant correlations for the normal controls and medicated patients (data not shown).
Vol.
43, No. 14,
Neuroleptics:
1988
Attention
1147
in Schizophrenia
whereas the drug free patients with schizophrenia do not Nevertheless, demonstrate a relationship between mid-prefrontal cortex metabolic activity the mid-prefrontal cortex in the medicated patients appears and performance, closely coupled to the patients’ performance accuracy in the direction expected on the basis of the correlations observed in the normal controls (l), the higher the metabolic rates of the mid-prefrontal cortical regions in i.e., the medicated patients the higher the number of correct responses in the task (Fig. 2 and Table III).
l* 1.20
-
1.10
-
,--a, p - NS
l
1.20
L
le
l l
:
r -
D,
P-w l
1.14
t
0.98
0 rlllltlllllll] 0
46
AUDITORY
80
120
CONTINUOUS
160
200
240
PERFORMANCE
16 SCHIZOPHRENICS
0
280
TASK
HITS
20
40
60
60
100
120
AUDIOTRY
CONTINUOUS PERFORMANCE FALSE ALARMS 16 SCHIZOPHRENICS OFF DRUG
OFF DRUG 1.10
r
%
140 TASK
r--B6 PC0001
1.06
1.66
1.04
1.02
1 .oo 0 AUDIOTRY
30
60
90
CONTINUOUS
120
160
180
210
PERFORMANCE
8 SCHIZOPHRENICS
0
240 TASK
HITS
ON DRUG
7
AUDITORY
14
21
26
25
42
49
CONTlNL.0US PERFORMANCE FALSE ALARMS 6 SCHIZOPHRENICS ON DRUG
56 TASK
Fig 2. Scatter plots of the anterior medial frontal cortex metabolic rates in Plane C and performance on the auditory discrimination task in the drug-free and This region is chosen because it is medicated patients with schizophrenia. the prefrontal cortex region with the highest associations to performance in the normal controls. The lines drawn are obtained on the basis of least squares analysis and the r values and their statistical significance are derived from the Pearson product-moment correlations. Although the drug-free patient with the highest number of false alarms appears to be an outlier, analysis without that subject does not change the absence of behavior--metabolic correlations in this group.
1148
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
1988
DISCUSSION
In a prior study, using a within subject design and somatosensory stimulation, schizophrenic patients while receiving neuroleptics appeared to have lower frontal cortex to posterior cortical metabolic rates compared to the time when these same patients were without neuroleptics (10). However this increase in relative hypofrontality, a measure of frontal cortex activity believed to reflect the brain’s general state of executive functioning, failed to reach statistical significance in the eight subjects evaluated whereas statistically significant increases in basal ganglia and temporal lobe metabolic rates were obtained. As noted in our report of dysfunction in a prefrontal substrate of both prior methods of analysis and sustained attention in schizophrenia (l), behavioral conditions studied were not ideal for examining the frontal cortex By carefully subdividing the frontal as a biological determinant of behavior. thus allowing for a sufficiently detailed regional analysis of this cortex, extremely heterogenous cortical region, and evaluating frontal cortex functional activity with respect to the specific executive function of sustained attention, we were able to improve on these design deficiencies. It was the observation of mid-prefrontal cortex involvement in the performance of a task requiring sustained attention that became the basis from which we could directly examine the functional activity of this region in unmedicated and medicated patients with schizophrenia during the performance of auditory discrimination. In this context, at least two “abnormalities” were apparent from our study of drug-withdrawn patients with schizophrenia. First, the mid-prefrontal metabolic rates of patients with schizophrenia were lower than those of normal controls, and second, whereas the accuracy of the performance of normal controls was associated with their metabolic rates in this area, the patients with schizophrenia did not display this relationship despite the greater range of performance in the patient group. In this between group design study of the effect of neuroleptics on regional brain metabolism, the previous findings, from the within subject design (lo), of substantially higher metabolic rates in the basal ganglia and temporal cortex of medicated patients compared to drug-withdrawn patients, hold true. Although this similarity together with the demographic concordance suggest that the medicated and drug-withdrawn schizophrenic groups may consist of similar types of patients, because patients in this study were not randomized to the two conditions, there is always the risk that the differences we observed between these two groups could have resulted from patient differences rather than the direct result of drug administration. Thus, in the absence of previous data from the within patient design study our findings of lowered metabolic rates in the superior frontal cortex, (lo), an area not related to task performance in normals, must be replicated before we can have confidence in this effect. Although the above effects on subcortical and temporal cortex metabolism are likely to be significant for the understanding of the molecular mechanisms of neuroleptic efficacy, we believe the most important new finding is the medication related change in the relationship between mid-prefrontal cortex metabolism and performance in the presence of continuing frontal cortex metabolic abnormality in this region of the brain to which sustained attention Two caveats, however, are in order. First, the localizes to in normals. findings could have resulted from patient differences rather than the effect of neuroleptic. Second, were the findings the result of neuroleptic uncertainty would still remain concerning whether these effects treatment, were the result of fluphenazine’s potent dopamine blocking properties or the
Vol.
43, No. 14,
1988
Neuroleptics:
Attention
in Schizophrenia
1149
Had similar results been obtained result of its other neurochemical effects. from patients treated with a variety of neuroleptics whose only common neurochemical effect were dopamine blockade this latter problem in interpretation would have been controlled for although other problems would have arisen due to the small sample size. Nonetheless, we believe the results are most parsimoniously interpreted as demonstrating, in man, an important function of the dopamine neurotransmitter pathway in the biological determination of sustained attention as mediated by Moreover, the metabolic the mid-prefrontal cortex. effects observed with fluphenazine may imply, as the earlier data on the medication-withdrawn patients did, that there may be at least two types of dysfunction in the prefrontal substrate of sustained attention in schizophrenia. The first is neuroleptic responsive, i.e. fluphenazine appears to induce coupling between the mid-prefrontal cortex and actual performance as reflected in the observed correlations (Table 3, Fig. 2). For example, untreated schizophrenic patients are engaged in unrelated (non-goal-directed) tasks that mask the changes in metabolic rate in the areas of the brain that accompany auditory discrimination; neuroleptics turn off this process revealing the expected correlation between the mid-prefrontal cortex and performance. The second dysfunction is unaffected by neuroleptic treatment, i.e. differences between drug-withdrawn and medicated patients are not found in the regional metabolic rates of the mid-prefrontal cortex in Plane C, the area demonstrating the largest increase in regional metabolism in normal controls performing auditory discrimination compared to their resting counterparts. This is the same area in the normal controls for which the regional metabolic rates are correlated with the accuracy of their performance. Whether this metabolic deficit in schizophrenia would be found to be attenuated in patients with significantly less symptomatology than the present group remains to be determined, but studies of continuous performance tasks have consistently reported deficits in patients with substantially less symptomatology than those we studied and in populations at high risk for the development of schizophrenia (11-16). Therefore, the observed mid-prefrontal cortex deficits in the drug-withdrawn psychotic patients , and the somewhat attenuated deficits observed in the medicated patients with similar symptomatology, could reflect the inability of the prefrontal cortex to direct behavior on the behalf of long range goals or “higher level” motivations and rewards in patients with schizophrenia. For example, lesions of the frontal cortex in man and in rodents can lead to heightened distractability and the expression of inappropriate behaviors (17). It is possible that dopamine dysfunction itself could lead to this type of deficit, but it is also likely that the effects of such a deficit might be expected to be attenuated or exacerbated by alterations in dopamine function as both animal and human studies have pointed to an important “gating” or “enabling” mechanism that dopamine neurotransmitter pathways play in permitting specific regions of the brain to capture behavior (18-20). Thus, neuroleptics, by blocking D -dopamine receptors, one of the proposed molecular mechanisms for the antipsycgotic efficacy of these drugs (21,22), could improve psychotic behavior by initially retarding the access of all [We assume that the large increases in brain regions to the motor system. glucose metabolism in the basal ganglia observed with prolonged neuroleptic The effectiveness treatment is the metabolic representation of this effect.] of dopamine receptor blockade on frontal cortex function might be lessened by differences in dopamine receptors in the frontal cortex compared to those in
1150
Neuroleptics:Attention in Schizophrenia
Vol. 43, No. 14, 1988
the basal ganglia, the high dopamine turnover in the frontal cortex, and the sensitivityof cortical dopamine neurons to neurolepticinduced increased turnover (23,24). Over time, the prefrontalcortex might accrue an additional advantage due to the differentialcharacteristicsof dopamine neuron adaptation that includes tolerance to the neurolepticinduced increased dopamine turnover in the striatum and limbic cortex, but not in the prefrontal cortex, at least in rodents and in non-human primates (23,25). Acknowledgments Substantialcontributionsto the conduct of this work were made by Dr. Robert Kessler, Dr. Richard Margolin, Dr. Lynn DeLisi, A. CatherineKing, John Cappelletti,and Tinya Dubyoski. REFERENCES 1.
2. 3. 4.
R.H. COHEN, W.E. SEMPLE, and M. GROSS, et al., Life Sciences 40 2031-2039(1987). R.M. COHEN, W.E. SEMPLE, M. GROSS, et al., Neuropsychiatry, Neuropsychology,and BehavioralNeurology,1 3-ZO(1988). pp. 157-169 W. GUY, ECDEU: ASSESSMENTMANUAL FOR PSYCHOPHARMACOLOGY, publicationADM 76-336 US Dept of Health, Education and Welfare (1976). L. SOKOLOFF. M. REIVICH. and C. KENNEDY. et al., J. Neurochem.28 897-916(1977).
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
R.A. BROOKS, J. Nucl. Med. 23 538-539(1982). M.E. PHELPS, S.C. HUANG, E.J.HOFFMAN, et al., Ann. Neurol. 5 371-388(1979). C. CLARK, R. CARSON, and R. KESSLER et al., J. Cerebral Blood Flow and Netab. 1 142-150(1985). T. MATSUI and A. HIRANO, AN ATLAS OF THE HUMAN BRAIN FOR COMPUTERIZED TOMOGRAPHY IGAKU SHOIN (1978). E.J. METTER, W.H. RIEGE, D.E. KUHL. and M.E. PHELPS, J. Cereb. Blood Flow and Metab. 4 l-7(1984). L.E. DELISIS H.H. HOLCOMB, and R.M. COHEN, et al., J. Cereb. Blood Flow and Metab. 5 ZOl-206(1985). A.F. MIRSKYY Ann. Rev. Psychology20 321-348(1969). C. KORNETSKY and M.H. ORZACK, J. P&h. Res. 14 69-79(1978). L.J. SEIDMAN, Psychol. Bull. 94 195-238(1983)K.H. NUECHTERLEIN,J. AbnormarPsychol. 92 4-28(1983). L. ERLENMEYER-KIMLING,Y. MARCUSE, B. COmBLATT, et al., in CHILDREN AT RISK FOR SCHIZOPHRENIA,N. F. Watt et al. (eds) pp 169-189, Cambridge University Press (1984). J.M. NEALE, K.C. WINTERS, and S. WEINTRAUB, ibid., pp. 264-278. D.T. STUSS and D.F. BENSON, THE FRONTAL LOBES, Raven Press (1986). R.D. OADES, Neurosci. and Biobehav.Rev. 2 261-282(1985). H.C. FIBIGER and A.G. PHILLIPS,Handbook of NeurophysiologyIV 647-675(1986). G.J. MOGENSON, Prog. Psychobiol.and Physiol. Psychol. 12 117-170(1987). I. CREESE, D.R. BURT, AND S.H. SNYDER, Science 192 481-a3(1976). P. SEEMAN, T. LEE, M. CHAU-WONG,and K. WONG, Nature (London)261 717-719(1976). M.J. BANNON, A.S. FREEMAN, L.A. CHIODO, B.S. BUNNEY, and R.H. ROTH, Handbook of Pharmacol.19 329-374(1987). W-H. CHANG, E-K. YEH, Y2. TSENG, M.C. CHUNG, and H-F. CHANG, Biol Psychiatry 21 422-424(1986). N.G. BACOPO%ULOS, G. BUSTOS, D.E. REDMOND, J. BAULU, and ROTH, R.H., Brain Res. 157 396-401(1978).
Life Sciences, Vol. 43, pp. 1141-1150 Printed in the U.S.A.
THE EFFECT OF NEUROLEPTICSON DYSFUNCTION IN A PREFRONTAL SUBSTRATEOF SUSTAINEDATTENTIONIN SCHIZOPHRENIA Robert M. Cohen, William E. Semple, Michael Gross, Thomas E. Nordahl, Henry H. Holcomb, H. Susan Dowling, David Pickar* Section
on Clinical Brain Imaging, Laboratory of Cerebral Metabolism, NIMH, Bethesda, MD 20892 *Clinical Neuroscience Branch, NIHH, Bethesda, MD 20892 (Received
in final
form August
12,
1988)
Summary Regional brain metabolism was measured in patients vith schizophrenia during the performance of auditory discrimination to evaluate the effect of neuroleptic medication on the middle prefrontal cortex, an area that appears to be an important biological determinant of the sustained attention required of subjects in this task. While unmedicated patients with schizophrenia have lower than “normal” metabolic rates in this cortical area that are unrelated to performance, patients receiving neuroleptics appear to have this metabolic deficit attenuated with higher metabolism in this area associated with greater accuracy of performance. This change occurs in the context of differential neuroleptic effects on the cortical regions of the frontal cortex, increased metabolism in the basal ganglia, thalamus and temporal lobes, and no apparent effect in the parietal and occipital cortices. The findings suggest that only part of the abnormality in the prefrontal cortex determination of sustained attention in schizophrenia is sensitive to neuroleptic treatment. Recently, we reported that the positron emission tomographically determined metabolic rates of the middle prefrontal cortex of patients with schizophrenia, measured while performing an auditory discrimination task, were lower than those observed in normal controls performing the same task (1). Horeover , the metabolic rates of this region in normals, but not in unmedicated patients with schizophrenia, were directly related to their accuracy of performance. The findings were interpreted as indicating that the mid-prefrontal cortex is an important biological determinant of sustained attention and that dysfunction of this region occurs in schizophrenia. It remained to be determined whether the sustained attention problems, observed in this particular continuous performance task in schizophrenia, and those measured by other psychological procedures, arise from a specific neurotransmitter defect in the middle prefrontal cortex or from other brain structures that could affect both metabolic rate and sustained attention. To help resolve this issue, we examined middle prefrontal cortex metabolism during the performance of auditory discrimination in patients receiving clinically titrated doses of neuroleptics.
0024-3205/88
$3.00
+ .OO
1142
Neuroleptics:.Attention
in Schizophrenia
Vol.
43, No. 14,
1988
METHODS Patients met DSM III (American Psychiatric Association) criteria Subjects: for schizophrenia as determined by the consensus of two psychiatrists and were residing on research units at the time of study. Unmedicated (drug-withdrawn) patients had not received neuroleptic medications for an average of 34 days (range 13-73, s.d. = 17.5; 1 patient had never received Of the unmedicated patients, 7 were of the treatment) prior to study. undifferentiated type, 7 of the paranoid type, 1 of the residual type and 1 of The medicated patients were predominantly the disorganized hebephrenic type. of the undifferentiated (n = 2) and of the paranoid type (n - 4) with 1 patient each meeting criteria for disorganized and atypical. The medicated patients had received fluphenazine for 31.1 days prior to time of scan with a 26.1 mg/day average dose at time of scan. Demographic, state and performance characteristics of the medicated and unmedicated patients are summarized in Although there were no females in the medicated patient group, our Table I. analysis of both normal females and females with schizophrenia have not demonstrated any significant effect of gender on our normalized regional Scans of unmedicated and medicated patients were glucose metabolic rates. obtained during the same time period. The auditory continuous performance task (CPT) consisted of Behavioral Task: a random series of 500 hz tones of 1.0 second duration and 2.0 second inter-tone-interval, with an intensity of 67, 75, or 86 decibels, measured at The subject was instructed to press the hand-held the earphone-ear interface. response button when the lowest volume tone was detected. The task was pf@sented in successive 5 minute blocks for 35 minutes following [ F]-2-fluoro-2-deoxy-D-glucose (FM;) injection. A total of 220 targets and 440 distracters were presented to each subject. To minimize learning contributions to the metabolic pattern, subjects were trained to a criterion of 18 of 20 correct identifications in this task in the hour preceding Subjects received scores for correct identifications of target FDG injection. tones (hits) and incorrect identification of distracters (false alarms). Although performance of this task requires motor activity, response selection and inhibition, discrimination, memory, sensory processing, and a variety of components of attention including arousal, we believe that the task design and the PET-FDC method of measurement which averages functional brain activity over a 35 minute interval, emphasizes those aspects of the task involved in the process of sustaining attention (See Ref. 2 for a more complete discussion of these issues). Subjects, with eyes patched, began their auditory PET Scan Procedure: discr6mination task several minutes prior to the injection of a 3-5 mCi dose of [‘?]FDC and completed the task 35 min after injection. Arterialized ous blood samples were taken from the left arm for quantification of Following the 35-min uptake period, subjects were placed in ‘;@F]FDC uptake. the scanner. Seven to eight slices were obtained from each subject starting from 80% of head height and proceeding in the caudal direction. Slices were parallel to the canthomeatal line (CM) and the interslice interval was 13 mm. Scans were performed with an Ortec ECAT II scanner of 1.75 cm. full width half-maximum resolution and used a calculated attenuation correction. Raw pixel values were converted to glucose metabolic rate in Data Analysis: For the extraction of regional glucose mg/lOO g tissue/minute $4,5,6). metabolic rates, 2.9 cm boxes were placed over the regions of interest in Plane A (9.1 cm above CM), B (7.8 cm above), C (6.5 cm five standard planes; above), D (5.2 cm above), E (3.9 cm above). Two independent raters who were
Vol.
43, No. 14,
1988
Neuroleptice:
Attention
in Schizophrenia
1143
unaware of the identity and diagnosis of the individual whose scan they were evaluating selected the planes for analysis and placed the boxes through neuroanatomical matching to a standard template (7)(See Fig. 1). Anatomical structures are judged as contained within these regions on the basis of the atlas of Matsui and Mrano (8). Global glucose metabolic rates refer to the estimates of the average value for glucose metabolism obtained for all the Regional glucose metabolic rates gray matter rich areas of the brain sampled. refer to the average of standardized absolute glucose metabolic data obtained The standardization procedure from specific areas of the cortex (See Fig. 1). is designed to minimize the effects of individual variation in global glucose metabolism on regional comparisons by dividing an individual’s glucose metabolic rate in a specific region by the individual’s mean global glucose metabolic rate. The procedure is similar in principle to the “reference ratio” or “landscape method” (7,9). All statistical comparisons of regional metabolic rates use these normalized (regional/global mean) values. It is important to note that although standardization of metabolic rates was necessary in most instances to establish significance the size of the percent changes in absolute glucose metabolic rates in these same regions were always in the same direction and of the same or usually larger magnitude.
FWtwiu
Occipital
Fig.
1
Schematic representation of regions sampled in the left and right hemispheres. Regions labeled as medial, although sampled from the medial portion of the cortex, are represented as incomplete boxes on the lateral surface. The boxes outlined by dashed lines are sampled from the surface of the frontal cortex, medial to the temporal cortex. The medial and anterior frontal regions, Plane A and the medial, anterior and posterior frontal regions, Plane B are referred to in the text as superior prefrontal cortex. The medial, anterior and posterior regions, Planes C and D are referred to as middle (or mid-) prefrontal cortex. The region labeled as medial parietal is also referred to in the text as superior posterior parietal cortex.
1144
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
1988
RESULTS The patients with schizophrenia who were studied while receiving neuroleptics did not differ in age, duration of illness, number of hospitalizations, symptomatology at time of scan, or in their performance on the auditory discrimination task from the sixteen drug-withdrawn patients previously studied (See Table I). This may, however, not indicate that the patients’ symptoms and discrimination performance were not responsive to neuroleptic treatment since all patients were considered improved while they were receiving neuroleptics by their treating physicians who were unaware of their medication status or their scan results. Thus, the similar BPRS ratings may have resulted from the pre-selection criteria which were to obtain patients who were highly symptomatic, but not so ill as to preclude valid scanning. TABLE I DEMOGRAPHIC,STATE AND PERFORMANCE CHARACTERISTICS OF PATIENTS WITH SCHIZOPHRENIA DRUG-FREE (n=16) BPRS (22 item)* Total Thought Disorder Anxiety/Depression Hostility/Suspiciousness Withdrawal/Retardation
58.7 9.94 7.25 6.63 7.01
f + f f f
MEDICATED (n=8)
STATISTICAL SIGNIFICANCE
15.0 4.23 4.00 3.05 2.24
58.4 10.4 7.50 8.00 8.62
* * f f f
12.1 3.85 2.07 2.97 2.92
NS NS NS NS NS
AGE OF ONSET
18.8
f 3.43
18.4
f 3.20
NS
DURATIONOF ILLNESS (IN YEARS)
8.00
+ 6.30
10.6
* 9.59
NS
NUMBEROF HOSPITALIZATIONS
4.81
+ 4.47
4.50
f 3.92
NS
AGE AT TIME OF STUDY
26.9
f 5.83
28.8
f 9.02
NS
f 62.1 k 24.1
NS NS
SEX FEHALES HALE PERFORMANCE HITS FALSE ALARMS
*
4 12 138.3 28.5
0
8 * 66.0 * 35.0
The 22-item Brief Psychiatric Rating Scale for the period just prior to the scan.
134.8 22.0 (BPRS) (3)
is
the average
score
Differences in global glucose metabolism between the medicated and unmedicated patients were also not apparent. Glucose metabolic rates in 55 to maximize the quantitative regions of the brain (see Fig. l), selected localization of brain function, and which were the same regions examined in our earlier study of unmedicated patients with schizophrenia, were evaluated for differences.
Vol.
43, No. 14,
Neuroleptics:
1988
Attention
in Schizophrenia
1145
TABLE II SELECTEDNORMALIZED REGIONALGLUCOSEMETABOLICRATE DIFFERENCES BETWEENMEDICATEDAND MEDICATION-WITHDRAWN PATIENTS WITR SCBIZOPBRENIA FRONTALCORTEX Region
L. Posterior
L. Anterior
Medial
R. Posterior
R. Anterior
Plane A
-4.6
-7.3
-10.4**
-9.8**
-8.O*
B
-4.6
-4.7
-8.2*
-6.4**
-4.3
C
-1.5
-0.2
-3.5
-0.2
-5.3*
D
+2.6
+4.5
+3.4
+1.4
+1.7
E
-0.4
+0.3
+1.9
-1.1
-1.8
TEMPORALCORTEX Region Plane D
L. Posterior
L. Middle
+8.6
L. Anterior
+7.4
E
+9.3
R. Anterior
**
+?.5
+3.4
R. Middle
**
+5.3
R. Posterior +7.8
+5.3
L. Hippocampus * +9.2
R. Hippocampus +9.7 SUBCORTICALSTRUCTURES
L. Basal
R. Basal
Ganglia
+6.7*
Ganglia
+5.4+ L. Thalamus +6.3
R. Thalamus +3.0
The table is derived by subtracting the mean normalized (regional/global mean) regional glucose metabolic rates of medication-withdrawn patients (n=16) with schizophrenia from those of medicated patients (n=8) and multiplying by 100. Statistically significant differences based on Z-tailed t-tests are represented by +, p < 0.1; *, p < 0.05; **, p < 0.01. Values for mean normalized metabolism and standard deviation in each region for the normal control group have been reported Those for schizophrenia will be reported for a study comparing schizophrenia (2). with severe affective disorder. These data can be obtained by request.
Because our primary interest was the examination of those regions of the brain to which the ability to perform auditory discrimination localized to in normal controls, our first evaluations were of the metabolic rates of the superior posterior medial parietal cortex and the anterior cingulate gyrus (normal controls performing auditory discrimination have lower rates in these
1146
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
regions than normal controls at rest) , and of the right anterior frontal and the right posterior frontal regions (7%) of Plane C (normal controls performing auditory discrimination have higher metabolic rates in these regions than normal controls at rest) (2). These regions did not differ significantly between the drug-withdrawn and treated patient groups.
1988
(5%)
Significant differences were, however, observed in regions to which sustained attention did not localize to in normals. The medicated patients compared to the drug-withdrawn patients were found to have higher glucose metabolic rates throughout the temporal cortex including the hippocampus, and in the subcortical structures of the basal ganglia and thalamus, although the latter structure did not reach statistical significance. The effect of neuroleptic treatment on metabolic activity in the frontal cortex was primarily observed in the superior frontal cortex (Planes A and B) with treated patients showing significantly lower glucose metabolic rates compared to their drug-withdrawn counterparts (Table II). Statistically significant differences were not observed in the 15 sampled regions of the parietal and occipital cortices. TABLE III MID-PREFRONTAL CORTEXGLUCOSEMETABOLICRATE--PERFORMANCE ASSOCIATIONS
Brain Region Plane C medial
Schizophrenic Unmedicated
Patients Medicated
0.014
(-0.067)
-0.856**
(0.874**)
R anterior
-0.123
(0.237)
-0.742*
(0.671)
L anterior
-0.259
(0.041)
-0.608
(0.667)
Plane D medial
-0.310
(0.109)
-0.615
(0.774*)
R anterior
0.049
(0.140)
-0.870**
(0.809*)
L anterior
-0.293
(0.036)
-0.664
(0.809*)
The significant within group Pearson product-moment correlations observed between regional glucose metabolic rates of the mid-prefrontal cortex and performance are reported. The numbers outside of parentheses represent correlations to false alarms; a negative correlation indicates that higher metabolic rates in the region of interest are associated with fewer errors. The correlations within parentheses are with hits; a positive correlation indicates that higher glucose metabolic rates in the region of interest are associated with a greater number of hits. Statistically significant Additional associations are represented by *, p 5 0.05 and **, p 5 0.01. analyses using the log ratio of hits to false alarms, another measure of accuracy, did not demonstrate any significant correlations between performance and metabolic rate in these regions in the medication-withdrawn patients, but again demonstrated significant correlations for the normal controls and medicated patients (data not shown).
Vol.
43, No. 14,
Neuroleptics:
1988
Attention
1147
in Schizophrenia
whereas the drug free patients with schizophrenia do not Nevertheless, demonstrate a relationship between mid-prefrontal cortex metabolic activity the mid-prefrontal cortex in the medicated patients appears and performance, closely coupled to the patients’ performance accuracy in the direction expected on the basis of the correlations observed in the normal controls (l), the higher the metabolic rates of the mid-prefrontal cortical regions in i.e., the medicated patients the higher the number of correct responses in the task (Fig. 2 and Table III).
l* 1.20
-
1.10
-
,--a, p - NS
l
1.20
L
le
l l
:
r -
D,
P-w l
1.14
t
0.98
0 rlllltlllllll] 0
46
AUDITORY
80
120
CONTINUOUS
160
200
240
PERFORMANCE
16 SCHIZOPHRENICS
0
280
TASK
HITS
20
40
60
60
100
120
AUDIOTRY
CONTINUOUS PERFORMANCE FALSE ALARMS 16 SCHIZOPHRENICS OFF DRUG
OFF DRUG 1.10
r
%
140 TASK
r--B6 PC0001
1.06
1.66
1.04
1.02
1 .oo 0 AUDIOTRY
30
60
90
CONTINUOUS
120
160
180
210
PERFORMANCE
8 SCHIZOPHRENICS
0
240 TASK
HITS
ON DRUG
7
AUDITORY
14
21
26
25
42
49
CONTlNL.0US PERFORMANCE FALSE ALARMS 6 SCHIZOPHRENICS ON DRUG
56 TASK
Fig 2. Scatter plots of the anterior medial frontal cortex metabolic rates in Plane C and performance on the auditory discrimination task in the drug-free and This region is chosen because it is medicated patients with schizophrenia. the prefrontal cortex region with the highest associations to performance in the normal controls. The lines drawn are obtained on the basis of least squares analysis and the r values and their statistical significance are derived from the Pearson product-moment correlations. Although the drug-free patient with the highest number of false alarms appears to be an outlier, analysis without that subject does not change the absence of behavior--metabolic correlations in this group.
1148
Neuroleptics:
Attention
in Schizophrenia
Vol.
43, No. 14,
1988
DISCUSSION
In a prior study, using a within subject design and somatosensory stimulation, schizophrenic patients while receiving neuroleptics appeared to have lower frontal cortex to posterior cortical metabolic rates compared to the time when these same patients were without neuroleptics (10). However this increase in relative hypofrontality, a measure of frontal cortex activity believed to reflect the brain’s general state of executive functioning, failed to reach statistical significance in the eight subjects evaluated whereas statistically significant increases in basal ganglia and temporal lobe metabolic rates were obtained. As noted in our report of dysfunction in a prefrontal substrate of both prior methods of analysis and sustained attention in schizophrenia (l), behavioral conditions studied were not ideal for examining the frontal cortex By carefully subdividing the frontal as a biological determinant of behavior. thus allowing for a sufficiently detailed regional analysis of this cortex, extremely heterogenous cortical region, and evaluating frontal cortex functional activity with respect to the specific executive function of sustained attention, we were able to improve on these design deficiencies. It was the observation of mid-prefrontal cortex involvement in the performance of a task requiring sustained attention that became the basis from which we could directly examine the functional activity of this region in unmedicated and medicated patients with schizophrenia during the performance of auditory discrimination. In this context, at least two “abnormalities” were apparent from our study of drug-withdrawn patients with schizophrenia. First, the mid-prefrontal metabolic rates of patients with schizophrenia were lower than those of normal controls, and second, whereas the accuracy of the performance of normal controls was associated with their metabolic rates in this area, the patients with schizophrenia did not display this relationship despite the greater range of performance in the patient group. In this between group design study of the effect of neuroleptics on regional brain metabolism, the previous findings, from the within subject design (lo), of substantially higher metabolic rates in the basal ganglia and temporal cortex of medicated patients compared to drug-withdrawn patients, hold true. Although this similarity together with the demographic concordance suggest that the medicated and drug-withdrawn schizophrenic groups may consist of similar types of patients, because patients in this study were not randomized to the two conditions, there is always the risk that the differences we observed between these two groups could have resulted from patient differences rather than the direct result of drug administration. Thus, in the absence of previous data from the within patient design study our findings of lowered metabolic rates in the superior frontal cortex, (lo), an area not related to task performance in normals, must be replicated before we can have confidence in this effect. Although the above effects on subcortical and temporal cortex metabolism are likely to be significant for the understanding of the molecular mechanisms of neuroleptic efficacy, we believe the most important new finding is the medication related change in the relationship between mid-prefrontal cortex metabolism and performance in the presence of continuing frontal cortex metabolic abnormality in this region of the brain to which sustained attention Two caveats, however, are in order. First, the localizes to in normals. findings could have resulted from patient differences rather than the effect of neuroleptic. Second, were the findings the result of neuroleptic uncertainty would still remain concerning whether these effects treatment, were the result of fluphenazine’s potent dopamine blocking properties or the
Vol.
43, No. 14,
1988
Neuroleptics:
Attention
in Schizophrenia
1149
Had similar results been obtained result of its other neurochemical effects. from patients treated with a variety of neuroleptics whose only common neurochemical effect were dopamine blockade this latter problem in interpretation would have been controlled for although other problems would have arisen due to the small sample size. Nonetheless, we believe the results are most parsimoniously interpreted as demonstrating, in man, an important function of the dopamine neurotransmitter pathway in the biological determination of sustained attention as mediated by Moreover, the metabolic the mid-prefrontal cortex. effects observed with fluphenazine may imply, as the earlier data on the medication-withdrawn patients did, that there may be at least two types of dysfunction in the prefrontal substrate of sustained attention in schizophrenia. The first is neuroleptic responsive, i.e. fluphenazine appears to induce coupling between the mid-prefrontal cortex and actual performance as reflected in the observed correlations (Table 3, Fig. 2). For example, untreated schizophrenic patients are engaged in unrelated (non-goal-directed) tasks that mask the changes in metabolic rate in the areas of the brain that accompany auditory discrimination; neuroleptics turn off this process revealing the expected correlation between the mid-prefrontal cortex and performance. The second dysfunction is unaffected by neuroleptic treatment, i.e. differences between drug-withdrawn and medicated patients are not found in the regional metabolic rates of the mid-prefrontal cortex in Plane C, the area demonstrating the largest increase in regional metabolism in normal controls performing auditory discrimination compared to their resting counterparts. This is the same area in the normal controls for which the regional metabolic rates are correlated with the accuracy of their performance. Whether this metabolic deficit in schizophrenia would be found to be attenuated in patients with significantly less symptomatology than the present group remains to be determined, but studies of continuous performance tasks have consistently reported deficits in patients with substantially less symptomatology than those we studied and in populations at high risk for the development of schizophrenia (11-16). Therefore, the observed mid-prefrontal cortex deficits in the drug-withdrawn psychotic patients , and the somewhat attenuated deficits observed in the medicated patients with similar symptomatology, could reflect the inability of the prefrontal cortex to direct behavior on the behalf of long range goals or “higher level” motivations and rewards in patients with schizophrenia. For example, lesions of the frontal cortex in man and in rodents can lead to heightened distractability and the expression of inappropriate behaviors (17). It is possible that dopamine dysfunction itself could lead to this type of deficit, but it is also likely that the effects of such a deficit might be expected to be attenuated or exacerbated by alterations in dopamine function as both animal and human studies have pointed to an important “gating” or “enabling” mechanism that dopamine neurotransmitter pathways play in permitting specific regions of the brain to capture behavior (18-20). Thus, neuroleptics, by blocking D -dopamine receptors, one of the proposed molecular mechanisms for the antipsycgotic efficacy of these drugs (21,22), could improve psychotic behavior by initially retarding the access of all [We assume that the large increases in brain regions to the motor system. glucose metabolism in the basal ganglia observed with prolonged neuroleptic The effectiveness treatment is the metabolic representation of this effect.] of dopamine receptor blockade on frontal cortex function might be lessened by differences in dopamine receptors in the frontal cortex compared to those in
1150
Neuroleptics:Attention in Schizophrenia
Vol. 43, No. 14, 1988
the basal ganglia, the high dopamine turnover in the frontal cortex, and the sensitivityof cortical dopamine neurons to neurolepticinduced increased turnover (23,24). Over time, the prefrontalcortex might accrue an additional advantage due to the differentialcharacteristicsof dopamine neuron adaptation that includes tolerance to the neurolepticinduced increased dopamine turnover in the striatum and limbic cortex, but not in the prefrontal cortex, at least in rodents and in non-human primates (23,25). Acknowledgments Substantialcontributionsto the conduct of this work were made by Dr. Robert Kessler, Dr. Richard Margolin, Dr. Lynn DeLisi, A. CatherineKing, John Cappelletti,and Tinya Dubyoski. REFERENCES 1.
2. 3. 4.
R.H. COHEN, W.E. SEMPLE, and M. GROSS, et al., Life Sciences 40 2031-2039(1987). R.M. COHEN, W.E. SEMPLE, M. GROSS, et al., Neuropsychiatry, Neuropsychology,and BehavioralNeurology,1 3-ZO(1988). pp. 157-169 W. GUY, ECDEU: ASSESSMENTMANUAL FOR PSYCHOPHARMACOLOGY, publicationADM 76-336 US Dept of Health, Education and Welfare (1976). L. SOKOLOFF. M. REIVICH. and C. KENNEDY. et al., J. Neurochem.28 897-916(1977).
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
R.A. BROOKS, J. Nucl. Med. 23 538-539(1982). M.E. PHELPS, S.C. HUANG, E.J.HOFFMAN, et al., Ann. Neurol. 5 371-388(1979). C. CLARK, R. CARSON, and R. KESSLER et al., J. Cerebral Blood Flow and Netab. 1 142-150(1985). T. MATSUI and A. HIRANO, AN ATLAS OF THE HUMAN BRAIN FOR COMPUTERIZED TOMOGRAPHY IGAKU SHOIN (1978). E.J. METTER, W.H. RIEGE, D.E. KUHL. and M.E. PHELPS, J. Cereb. Blood Flow and Metab. 4 l-7(1984). L.E. DELISIS H.H. HOLCOMB, and R.M. COHEN, et al., J. Cereb. Blood Flow and Metab. 5 ZOl-206(1985). A.F. MIRSKYY Ann. Rev. Psychology20 321-348(1969). C. KORNETSKY and M.H. ORZACK, J. P&h. Res. 14 69-79(1978). L.J. SEIDMAN, Psychol. Bull. 94 195-238(1983)K.H. NUECHTERLEIN,J. AbnormarPsychol. 92 4-28(1983). L. ERLENMEYER-KIMLING,Y. MARCUSE, B. COmBLATT, et al., in CHILDREN AT RISK FOR SCHIZOPHRENIA,N. F. Watt et al. (eds) pp 169-189, Cambridge University Press (1984). J.M. NEALE, K.C. WINTERS, and S. WEINTRAUB, ibid., pp. 264-278. D.T. STUSS and D.F. BENSON, THE FRONTAL LOBES, Raven Press (1986). R.D. OADES, Neurosci. and Biobehav.Rev. 2 261-282(1985). H.C. FIBIGER and A.G. PHILLIPS,Handbook of NeurophysiologyIV 647-675(1986). G.J. MOGENSON, Prog. Psychobiol.and Physiol. Psychol. 12 117-170(1987). I. CREESE, D.R. BURT, AND S.H. SNYDER, Science 192 481-a3(1976). P. SEEMAN, T. LEE, M. CHAU-WONG,and K. WONG, Nature (London)261 717-719(1976). M.J. BANNON, A.S. FREEMAN, L.A. CHIODO, B.S. BUNNEY, and R.H. ROTH, Handbook of Pharmacol.19 329-374(1987). W-H. CHANG, E-K. YEH, Y2. TSENG, M.C. CHUNG, and H-F. CHANG, Biol Psychiatry 21 422-424(1986). N.G. BACOPO%ULOS, G. BUSTOS, D.E. REDMOND, J. BAULU, and ROTH, R.H., Brain Res. 157 396-401(1978).