Frontostriatal Disorder of Cerebral Metabolism in Never-Medicated Schizophrenics Buchsbaum, MD; Richard J. Haier, PhD; Steven G. Potkin, MD; Keith Nuechterlein, PhD; Bracha, MD; Mark Katz, MD; James Lohr, MD; Joseph Wu, MD; Stephen Lottenberg, MD; Paul A. Jerabek, PhD; Mignone Trenary, MA; Richard Tafalla, PhD; Chandra Reynolds; William E. Bunney, Jr, Monte S. H. Stefan
scanned 18 patients with schizophrenia who had received neuroleptic medication and 20 age- and sex-matched controls by positron emission tomography with 18-F-fluorodeoxyglucose (fludeoxyglucose F 18) as a tracer of glucose metabolism. Subjects performed the Continuous Performance Test during 18-F-fluorodeoxyglucose uptake. Scan results were converted to metabolic rates, and computer algorithms were used to identify cortical regions. Pervious reports of relative hypofrontality in schizophrenia were confirmed, indicating that this finding is not an artifact of previous treatment. Significantly reduced ratios of inferior and medial frontal regions to occipital cortex were found, together with diminished metabolism in the basal ganglia. This suggests the presence of a combined frontostriatal dysfunction in schizophrenia. \s=b\ We
never
(Arch Gen Psychiatry. 1992;49:935-942)
disorders of attention, volition, and future-oriented Theplanning abnormalities1 associated be that diminished function
with frontal-lobe in this area might found in schizophrenia. Since the initial regional cerebral blood flow studies of Ingvar and Franzen,2 a series of studies with both cerebral blood flow3,4 and positron emission tomography (PET)46 have found that there is di¬ minished functional activity in the frontal lobes in many schizophrenics, especially when activity is measured rela¬ tive to that in posterior regions of the brain. While most studies have confirmed this effect, the magnitude of the hypofrontal tendency has varied greatly. Task activity during the functional study, subject age, and illness sever¬ ity all seem to be important factors. Short-term treatment with neuroleptic medication does not appear to have a clear effect on frontal-lobe metabolism,7 but the possibility that
suggest
Accepted for publication November 27, 1991. From the Department of Psychiatry and Human Behavior, University
of California
at
Irvine (Drs
Buchsbaum, Haier, Potkin, Katz, Wu, Lotten-
berg, Jerabek, Tafalla, and Bunney and Mss Ternary and Reynolds); the Department of Psychiatry, University of California at Los Angeles (Dr Nuechterlein); and the Department of Psychiatry, University of Califor-
nia at San Diego (Drs Bracha and Lahr). Dr Buchsbaum is now with Mount Sinai School of Medicine, New York, NY. Reprint requests to Department of Psychiatry, Mount Sinai School of Medicine, Box 1230, 1 Gustave L. Levy PI, New York, NY 10029 (Dr
Buchsbaum).
MD
treatment affects the frontal lobes considered.5 The effect of neuroleptic medication on the basal ganglia is less equivocal. Most studies have found elevation of basal ganglia metabolism with neuro¬ leptic administration,58 and the possibility is significant
long-term neuroleptic must be
long-term changes in the basal ganglia might occur. Many studies of schizophrenics have used patients who were not currently taking medication, typically with a 2to 4-week drug-free interval. In animal studies, spontane¬ ous locomotor activity and mouthing disappear after 2 that
weeks of withdrawal, and the enhanced stereotyped response disappears after only a month; the increased dis¬ sociation constant for spiperone binding reverts to normal 2 weeks after drug withdrawal.9 Thus, the majority of im¬ aging studies appear to have avoided the artifacts of acute medication or medication withdrawal effects. However, some effects of 12 months of medication, such as the increased stimulation of striatal adenylate cyclase by dopamine, persist unchanged throughout 6 months after drug withdrawal.9 A PET study by Cleghorn et al10 on eight never-medicated schizophrenics actually found them to be hyperfrontal. Other studies have either not included never-medicated patients or have combined them with medicated ones.11 The current study presents data on a group of 18 patients, never previously treated with neuroleptics, to evaluate the metabolic activity in the frontal lobes and the basal ganglia.
SUBJECTS AND METHODS
Subjects
male patients (mean [±SD] age, 29.6 ±7.2 years) recruited from referral through the clinical and research programs of the University of California at Los Angeles, San Diego, and Ir¬ vine served as subjects. No patient had a history of having received neuroleptic medication before scanning, and all had negative urine screen values for medications and drugs of abuse. This judgment was based on review of available clinical records and on interviews with patients and informants. One patient might have received medication for a few days in 1976 but reported discarding all capsules. The average age at onset was 24.9±7.0 years (range, 13 to 40 years). The average duration of ill¬ ness was 4.6±5.9 years. Nine patients had been ill 1 year or less and five patients has been ill 5 years or more. Three patients had never been hospitalized, and 12 were currently hospitalized. All patients were in good health on the basis of medical history,
Eighteen
Downloaded From: http://archpsyc.jamanetwork.com/pdfaccess.ashx?url=/data/journals/psych/12538/ by a University of California - San Diego User on 05/10
physical examination, and laboratory measures. Seventeen were right-handed and one was left-handed. Subjects with a history of epilepsy or head injury were excluded. No patient had an abnor¬ mal blood glucose level. Patients were diagnosed by means of DSM-III criteria by inter¬ viewers entirely independent of brain imaging data. The diag¬ nostic workup included the Present State Examination,12 modified to allow use of DSM-III criteria; the expanded version of the Brief Psychiatric Rating Scale (BPRS)13,14; and the Scale for the Assess¬ ment of Negative Symptoms.15 Assessment with the BPRS was
available for all 18 patients (mean score, 48.2±9.4). Due to the complex multicenter logistics of obtaining PET scans before treatment, not all patients received the full standardized inter¬ view and not all had been ill at least 6 months at the time of the scan. At a minimum, patients were seen for a similar clinical in¬ terview by two of us (S.G.P., J.L.) and judged acceptable for the study. Patients who had been ill for only a short interval were subsequently followed up diagnostically, with the exception of two who soon became unavailable for follow-up. Currently, of the 18 patients, 15 are schizophrenic and three schizophreniform. The control subjects (20 men with a mean age of 27.1 ±6.4 years) were screened with the same physical and laboratory examina¬ tion as the patients. All were right-handed. The controls did not have a history of psychiatric illness in themselves or in firstdegree relatives, and none of them were taking psychoactive medications by history or urine screen.
Task and Procedure
Subjects arrived in the laboratory and were briefed on the pro¬ cedure. Before PET scanning, an individually molded, thermosetting plastic head holder was made for each subject to minimize head movement. Next, two intravenous lines were inserted, one in each arm, for the administration of 18-F-fluorodeoxyglucose (fludeoxyglucose F18) and withdrawal of blood samples. Subjects were then seated in a comfortable chair in a darkened, acoustically attenuated isolation room and instructed about the procedure and the Continuous Performance Task (CPT) to be performed. A 5-minute practice session was given. Next, a dark curtain was drawn around the subject, with the subject's left arm protruding through an opening to allow blood sampling. The left arm was wrapped in a hot pack for arterialization of venous blood. The CPT was started, and then an injection of 148 to 192 MBq of 18-
F-fluorodeoxyglucose was administered. The CPT has been widely used in studies of schizophrenia and in studies of individuals with relatives who have schizophrenia.16,17 In the version of the CPT used here,18 single digits (0 to 9) were pre¬ sented for 40 milliseconds at a rate of one every 2 seconds. Subjects were told to respond with a button press, using their right hand, each time that they detected the digit 0 and that it was equally im¬ portant to respond to zeros and not to respond to digits other than zero. Targets were presented irregularly with a probability of oc¬ currence of 0.25. Stimuli were presented, silently by rear projection, on a 24x24-cm screen, with a carousel slide projector (Kodak) fitted with a shutter (Ilex No. 4 Synchro-electronic Shutter) and blurred to a degree that made digits barely recognizable (such that a 2.8diopter correction is required to refocus clearly). The subject's eyes were 1.2 m from the rear projection screen. Subjects were not spoken to during uptake, and all remained quiet and cooperative. Subjects were continuously observed to ensure that they were following in¬ structions. Small, low-level penlights were kept on for blood sam¬ pling behind the curtain. After 30 to 35 minutes of 18-Ffluorodeoxyglucose uptake, the intravenous line in the right arm was removed and the subject was allowed to void and then trans¬ ferred to the adjacent scanning room. PET
Scanning
18-F-Fluorodeoxyglucose was prepared in the cyclotron facility of the Brain Imaging Center at the University of California, Irvine. It was synthesized by means of an adaptation of the synthesis19 using aqueous fluoride F18 produced in a small-volume enriched 180 water target. All quality assurance procedures confirmed the
Precentrai
Postcentral
Middle Frontal
Superior Frontal
\J
rietal Lobule lar Gyrus Lateral Occipital 8 19 17 17
Inferior Frontal
ior Temporal iTempor Superior Temporal Posterior Temporal
Fig 1.—Key figure for cortical peel analysis. Brain is reconstructed from the Matsui and Hirano atlas2' with the use of their gyral identification.
For each atlas slice, the perimeter was identified and the percentage lo¬ cation of each sulcal division recorded. For each positron emission to-
the perimeter was identified and the percentage loca¬ the perimeter used to calculate the mean glucose level within the sector of the cortical peel. Next, an area-weighted average of each cortical region was computed.
mographic scan, tion
on
18-F-fluorodeoxyglucose to be within specifications of pharma¬ ceutical quality as recommended by the US Pharmacopeia. Nine planes (CTI NeuroECAT, Computer Technology and Im¬ aging, Knoxville, Tenn) at 10-mm increments and parallel to the canthomeatal (CM) line were acquired 45 to 100 minutes after 18F-fluorodeoxyglucose injection. Scans were performed with both shadow and septa shields in a configuration with measured inplane resolution of 7.6 and 10.9 mm in the z-dimension (axial). A calculated attenuation correction and a smoothing filter were used. The scanner was calibrated each scan day, with a cylindric phantom, and compared with well counter data. Scan Slice Processing and Selection Scans were transformed to glucose metabolic rate as described elsewhere.2" Each slice was then neuroanatomically matched to a standard atlas21 by an observer (CR.) without reference to diag¬ nostic group. We then outlined the cortex and averaged metabolic rates in a strip along the slice perimeter as described elsewhere.20 A weighted average of these strip segments was then calculated across slices to obtain estimates of metabolic rate in the major gyri of the brain (Fig 1). This method, which we developed in 1982, has been used in a number of studies to examine the cerebral cortex, and its advantages over rectangular stereotaxis and discrete regions of interest were reviewed by Harris et al.22 To match our earlier methods, subcortical structures were as¬ sessed by means of stereotaxic coordinates derived from the neuroanatomic atlas.21 The caudate and putamen were measured with the use of the same atlas and a stereotaxic atlas method as described elsewhere.8,20,23 Values were also expressed as relative metabolic rate for the cor¬ tex as metabolic rate/whole-brain metabolic rate, and for subcor¬ tical regions of interest as metabolic rate / whole-slice metabolic rate. To validate both stability of head positioning and the stereotaxic coordinates, we obtained magnetic resonance (MR) images in five subjects scanned twice 4 weeks apart and 20 subjects scanned once (repetition time, 3000 milliseconds; echo time, 33 milliseconds; four echoes) at 10-mm intervals, with a slice thickness of 7.5 mm. The subjects' heads were positioned in the MR imager with the use of the same thermoplastic mask used with the PET and an identical plastic jig to rest the head and posterior flanges of the mask. The CM line was drawn on the mask. A right triangle was drawn with one leg on the CM line and one leg perpendicular to the CM line. A right triangle of plastic tubing filled with petroleum jelly was attached to the mask. Slices at different heights above the CM line have varying distances between the two points. We measured the
Downloaded From: http://archpsyc.jamanetwork.com/pdfaccess.ashx?url=/data/journals/psych/12538/ by a University of California - San Diego User on 05/10
Schizophrenics (n=18)
Normal Controls (n=20)
Region
Left Frontal
Superior Middle Inferior Precentrai *Values are mean±SD. tP
scanned 18 patients with schizophrenia who had received neuroleptic medication and 20 age- and sex-matched controls by positron emission tomography with 18-F-fluorodeoxyglucose (fludeoxyglucose F 18) as a tracer of glucose metabolism. Subjects performed the Continuous Performance Test during 18-F-fluorodeoxyglucose uptake. Scan results were converted to metabolic rates, and computer algorithms were used to identify cortical regions. Pervious reports of relative hypofrontality in schizophrenia were confirmed, indicating that this finding is not an artifact of previous treatment. Significantly reduced ratios of inferior and medial frontal regions to occipital cortex were found, together with diminished metabolism in the basal ganglia. This suggests the presence of a combined frontostriatal dysfunction in schizophrenia. \s=b\ We
never
(Arch Gen Psychiatry. 1992;49:935-942)
disorders of attention, volition, and future-oriented Theplanning abnormalities1 associated be that diminished function
with frontal-lobe in this area might found in schizophrenia. Since the initial regional cerebral blood flow studies of Ingvar and Franzen,2 a series of studies with both cerebral blood flow3,4 and positron emission tomography (PET)46 have found that there is di¬ minished functional activity in the frontal lobes in many schizophrenics, especially when activity is measured rela¬ tive to that in posterior regions of the brain. While most studies have confirmed this effect, the magnitude of the hypofrontal tendency has varied greatly. Task activity during the functional study, subject age, and illness sever¬ ity all seem to be important factors. Short-term treatment with neuroleptic medication does not appear to have a clear effect on frontal-lobe metabolism,7 but the possibility that
suggest
Accepted for publication November 27, 1991. From the Department of Psychiatry and Human Behavior, University
of California
at
Irvine (Drs
Buchsbaum, Haier, Potkin, Katz, Wu, Lotten-
berg, Jerabek, Tafalla, and Bunney and Mss Ternary and Reynolds); the Department of Psychiatry, University of California at Los Angeles (Dr Nuechterlein); and the Department of Psychiatry, University of Califor-
nia at San Diego (Drs Bracha and Lahr). Dr Buchsbaum is now with Mount Sinai School of Medicine, New York, NY. Reprint requests to Department of Psychiatry, Mount Sinai School of Medicine, Box 1230, 1 Gustave L. Levy PI, New York, NY 10029 (Dr
Buchsbaum).
MD
treatment affects the frontal lobes considered.5 The effect of neuroleptic medication on the basal ganglia is less equivocal. Most studies have found elevation of basal ganglia metabolism with neuro¬ leptic administration,58 and the possibility is significant
long-term neuroleptic must be
long-term changes in the basal ganglia might occur. Many studies of schizophrenics have used patients who were not currently taking medication, typically with a 2to 4-week drug-free interval. In animal studies, spontane¬ ous locomotor activity and mouthing disappear after 2 that
weeks of withdrawal, and the enhanced stereotyped response disappears after only a month; the increased dis¬ sociation constant for spiperone binding reverts to normal 2 weeks after drug withdrawal.9 Thus, the majority of im¬ aging studies appear to have avoided the artifacts of acute medication or medication withdrawal effects. However, some effects of 12 months of medication, such as the increased stimulation of striatal adenylate cyclase by dopamine, persist unchanged throughout 6 months after drug withdrawal.9 A PET study by Cleghorn et al10 on eight never-medicated schizophrenics actually found them to be hyperfrontal. Other studies have either not included never-medicated patients or have combined them with medicated ones.11 The current study presents data on a group of 18 patients, never previously treated with neuroleptics, to evaluate the metabolic activity in the frontal lobes and the basal ganglia.
SUBJECTS AND METHODS
Subjects
male patients (mean [±SD] age, 29.6 ±7.2 years) recruited from referral through the clinical and research programs of the University of California at Los Angeles, San Diego, and Ir¬ vine served as subjects. No patient had a history of having received neuroleptic medication before scanning, and all had negative urine screen values for medications and drugs of abuse. This judgment was based on review of available clinical records and on interviews with patients and informants. One patient might have received medication for a few days in 1976 but reported discarding all capsules. The average age at onset was 24.9±7.0 years (range, 13 to 40 years). The average duration of ill¬ ness was 4.6±5.9 years. Nine patients had been ill 1 year or less and five patients has been ill 5 years or more. Three patients had never been hospitalized, and 12 were currently hospitalized. All patients were in good health on the basis of medical history,
Eighteen
Downloaded From: http://archpsyc.jamanetwork.com/pdfaccess.ashx?url=/data/journals/psych/12538/ by a University of California - San Diego User on 05/10
physical examination, and laboratory measures. Seventeen were right-handed and one was left-handed. Subjects with a history of epilepsy or head injury were excluded. No patient had an abnor¬ mal blood glucose level. Patients were diagnosed by means of DSM-III criteria by inter¬ viewers entirely independent of brain imaging data. The diag¬ nostic workup included the Present State Examination,12 modified to allow use of DSM-III criteria; the expanded version of the Brief Psychiatric Rating Scale (BPRS)13,14; and the Scale for the Assess¬ ment of Negative Symptoms.15 Assessment with the BPRS was
available for all 18 patients (mean score, 48.2±9.4). Due to the complex multicenter logistics of obtaining PET scans before treatment, not all patients received the full standardized inter¬ view and not all had been ill at least 6 months at the time of the scan. At a minimum, patients were seen for a similar clinical in¬ terview by two of us (S.G.P., J.L.) and judged acceptable for the study. Patients who had been ill for only a short interval were subsequently followed up diagnostically, with the exception of two who soon became unavailable for follow-up. Currently, of the 18 patients, 15 are schizophrenic and three schizophreniform. The control subjects (20 men with a mean age of 27.1 ±6.4 years) were screened with the same physical and laboratory examina¬ tion as the patients. All were right-handed. The controls did not have a history of psychiatric illness in themselves or in firstdegree relatives, and none of them were taking psychoactive medications by history or urine screen.
Task and Procedure
Subjects arrived in the laboratory and were briefed on the pro¬ cedure. Before PET scanning, an individually molded, thermosetting plastic head holder was made for each subject to minimize head movement. Next, two intravenous lines were inserted, one in each arm, for the administration of 18-F-fluorodeoxyglucose (fludeoxyglucose F18) and withdrawal of blood samples. Subjects were then seated in a comfortable chair in a darkened, acoustically attenuated isolation room and instructed about the procedure and the Continuous Performance Task (CPT) to be performed. A 5-minute practice session was given. Next, a dark curtain was drawn around the subject, with the subject's left arm protruding through an opening to allow blood sampling. The left arm was wrapped in a hot pack for arterialization of venous blood. The CPT was started, and then an injection of 148 to 192 MBq of 18-
F-fluorodeoxyglucose was administered. The CPT has been widely used in studies of schizophrenia and in studies of individuals with relatives who have schizophrenia.16,17 In the version of the CPT used here,18 single digits (0 to 9) were pre¬ sented for 40 milliseconds at a rate of one every 2 seconds. Subjects were told to respond with a button press, using their right hand, each time that they detected the digit 0 and that it was equally im¬ portant to respond to zeros and not to respond to digits other than zero. Targets were presented irregularly with a probability of oc¬ currence of 0.25. Stimuli were presented, silently by rear projection, on a 24x24-cm screen, with a carousel slide projector (Kodak) fitted with a shutter (Ilex No. 4 Synchro-electronic Shutter) and blurred to a degree that made digits barely recognizable (such that a 2.8diopter correction is required to refocus clearly). The subject's eyes were 1.2 m from the rear projection screen. Subjects were not spoken to during uptake, and all remained quiet and cooperative. Subjects were continuously observed to ensure that they were following in¬ structions. Small, low-level penlights were kept on for blood sam¬ pling behind the curtain. After 30 to 35 minutes of 18-Ffluorodeoxyglucose uptake, the intravenous line in the right arm was removed and the subject was allowed to void and then trans¬ ferred to the adjacent scanning room. PET
Scanning
18-F-Fluorodeoxyglucose was prepared in the cyclotron facility of the Brain Imaging Center at the University of California, Irvine. It was synthesized by means of an adaptation of the synthesis19 using aqueous fluoride F18 produced in a small-volume enriched 180 water target. All quality assurance procedures confirmed the
Precentrai
Postcentral
Middle Frontal
Superior Frontal
\J
rietal Lobule lar Gyrus Lateral Occipital 8 19 17 17
Inferior Frontal
ior Temporal iTempor Superior Temporal Posterior Temporal
Fig 1.—Key figure for cortical peel analysis. Brain is reconstructed from the Matsui and Hirano atlas2' with the use of their gyral identification.
For each atlas slice, the perimeter was identified and the percentage lo¬ cation of each sulcal division recorded. For each positron emission to-
the perimeter was identified and the percentage loca¬ the perimeter used to calculate the mean glucose level within the sector of the cortical peel. Next, an area-weighted average of each cortical region was computed.
mographic scan, tion
on
18-F-fluorodeoxyglucose to be within specifications of pharma¬ ceutical quality as recommended by the US Pharmacopeia. Nine planes (CTI NeuroECAT, Computer Technology and Im¬ aging, Knoxville, Tenn) at 10-mm increments and parallel to the canthomeatal (CM) line were acquired 45 to 100 minutes after 18F-fluorodeoxyglucose injection. Scans were performed with both shadow and septa shields in a configuration with measured inplane resolution of 7.6 and 10.9 mm in the z-dimension (axial). A calculated attenuation correction and a smoothing filter were used. The scanner was calibrated each scan day, with a cylindric phantom, and compared with well counter data. Scan Slice Processing and Selection Scans were transformed to glucose metabolic rate as described elsewhere.2" Each slice was then neuroanatomically matched to a standard atlas21 by an observer (CR.) without reference to diag¬ nostic group. We then outlined the cortex and averaged metabolic rates in a strip along the slice perimeter as described elsewhere.20 A weighted average of these strip segments was then calculated across slices to obtain estimates of metabolic rate in the major gyri of the brain (Fig 1). This method, which we developed in 1982, has been used in a number of studies to examine the cerebral cortex, and its advantages over rectangular stereotaxis and discrete regions of interest were reviewed by Harris et al.22 To match our earlier methods, subcortical structures were as¬ sessed by means of stereotaxic coordinates derived from the neuroanatomic atlas.21 The caudate and putamen were measured with the use of the same atlas and a stereotaxic atlas method as described elsewhere.8,20,23 Values were also expressed as relative metabolic rate for the cor¬ tex as metabolic rate/whole-brain metabolic rate, and for subcor¬ tical regions of interest as metabolic rate / whole-slice metabolic rate. To validate both stability of head positioning and the stereotaxic coordinates, we obtained magnetic resonance (MR) images in five subjects scanned twice 4 weeks apart and 20 subjects scanned once (repetition time, 3000 milliseconds; echo time, 33 milliseconds; four echoes) at 10-mm intervals, with a slice thickness of 7.5 mm. The subjects' heads were positioned in the MR imager with the use of the same thermoplastic mask used with the PET and an identical plastic jig to rest the head and posterior flanges of the mask. The CM line was drawn on the mask. A right triangle was drawn with one leg on the CM line and one leg perpendicular to the CM line. A right triangle of plastic tubing filled with petroleum jelly was attached to the mask. Slices at different heights above the CM line have varying distances between the two points. We measured the
Downloaded From: http://archpsyc.jamanetwork.com/pdfaccess.ashx?url=/data/journals/psych/12538/ by a University of California - San Diego User on 05/10
Schizophrenics (n=18)
Normal Controls (n=20)
Region
Left Frontal
Superior Middle Inferior Precentrai *Values are mean±SD. tP
