Thalamus size and outcome in schizophrenia

Schizophrenia Research 71 (2004) 473 – 484 www.elsevier.com/locate/schres Thalamus size and outcome in schizophrenia Adam M. Brickman a,b,c,*, Monte S. Buchsbaum a, Lina Shihabuddin a,d, William Byne a,d, Randall E. Newmark a, Jesse Brand a, Shabeer Ahmed a, Serge A. Mitelman a, Erin A. Hazlett a a b Department of Psychiatry, Mount Sinai School of Medicine, New York, NY, USA Department of Psychology, Queens College and The Graduate Center of the City University of New York, Flushing, NY, USA c Department of Psychiatry and Human Behavior, Brown Medical School, Providence, RI, USA d Department of Veterans Affairs, Bronx VA Medical Center, Bronx, NY, USA Received 8 October 2003; received in revised form 25 February 2004; accepted 1 March 2004 Available online 8 May 2004 Abstract The size of the thalamus was assessed in 106 patients with schizophrenia and 42 normal controls using high-resolution magnetic resonance imaging. The thalamus was traced at five axial levels proportionately spaced from dorsal to ventral directions. Patients with schizophrenia had significantly smaller thalamic areas at more ventral levels. Thalamic size was positively associated with frontal lobe and temporal lobe size. The effects were most marked in the patients with poorer clinical outcome (i.e., ‘‘Kraepelinian’’ patients). These findings are consistent with post-mortem and MRI measurement suggesting reduction in volume of the pulvinar, which occupies a large proportion of the ventral thalamus and which has prominent connections to the temporal lobe. D 2004 Elsevier B.V. All rights reserved. Keywords: MRI; Schizophrenia; Thalamus; Functional outcome 1. Introduction The thalamus has extensive and reciprocal connections to the striatum and cortex and its association nuclei, including the medial dorsal nucleus and pulvinar, are importantly involved in maintaining attention and modulation of sensory input. The * Corresponding author. Mount Sinai School of Medicine, Department of Psychiatry, Neuroscience PET Laboratory, Box 1505, One Gustave L. Levy Place, New York, NY 10029, USA. Tel.: +1-212-241-5287. E-mail address: [email protected] (A.M. Brickman). 0920-9964/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2004.03.011 disturbance of these functions in schizophrenia, together with evidence from post-mortem and neuroimaging studies of volume reduction and functional abnormalities, has implicated the thalamus as a nexus of defective circuits in schizophrenia (Jones, 1997). Both postmortem studies (reviewed in Byne et al., 2001, 2002; Danos et al., 2003) and MRI studies (see meta-analysis; Konick and Friedman, 2001) of the thalamus in schizophrenia have generally found reduced volume, but this effect has been typically small and not seen in all studies (Bridle et al., 2002; Deicken et al., 2002). In post-mortem (Byne et al., 2002; Highley et al., 2003; Young et al., 2000) and MRI studies where thalamic nuclei were indi- 474 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 vidually traced, the medial dorsal nucleus and pulvinar have appeared to be more reduced in volume than other regions (Byne et al., 2001, 2002). However, medial dorsal nucleus reduction was not found in one study (Cullen et al., 2003). Although total thalamic reduction in schizophrenia has been found in some MRI studies (Andreasen et al., 1994; Gur et al., 1998), it has not in others (Arciniegas et al., 1999; Deicken et al., 2002; Portas et al., 1998). Equivocal MRI findings of total thalamic reduction in schizophrenia could be due to a number of factors. For example, some studies could have higher functioning patients, who are perhaps more amenable to participation in MRI studies, over represented, and thus have less severe neuropathology. Given the clinical heterogeneity in the presentation of the illness, examination of subtypes of patients would help address this issue. Furthermore, as thalamic reduction is expected to be a relatively small effect, larger sample studies combined with more comprehensive examination of total thalamic size might be necessary to determine the exact nature of thalamic reduction. To address this issue, we chose to examine total thalamic size across several slice levels on the dorsal –ventral axis. Poor outcome, or ‘‘Kraepelinian,’’ patients (Keefe et al., 1987), with more severe symptoms and worse social and occupational functioning, seem especially likely to have thalamic volume loss. Indeed, although poor outcome patients are defined on the basis of clinical functional activities, several converging lines of research suggest that poor outcome patients have more severe pathology and may represent a unique schizophrenia subtype. Compared to good outcome patients, schizophrenia patients with poor outcome have more severe psychopathology (Keefe et al., 1987, 1988, 1996), are less responsive to neuroleptic treatment (Harvey et al., 1991), and have worse neuropsychological functioning (Roy et al., 2003). Poor outcome patients also have distinctive brain regional volume change. They have been shown to have larger ventricles that become progressively larger over a 5-year period (Davis et al., 1998), smaller putamens (Buchsbaum et al., 2003), smaller posterior cortical regions (Mitelman et al., 2003), and lower temporal lobe relative metabolic rates (Buchsbaum et al., 2002) compared to good outcome patients. The dichotomous classification system of good versus poor outcome patients offers one approach that may be useful in accounting for the tremendous amount of heterogeneity in functional outcome of schizophrenia patients. Poor outcome patients may be over represented in post-mortem samples that have been more uniform than MRI samples in demonstrating statistically significant volume loss, suggesting that thalamic volume loss may be a function of poor outcome. In the only direct test of this, Staal et al. (2001) assessed patient groups with good and poor outcome and showed a difference in frontal gray matter volume but not thalamic volume. Volume reduction in the cortical areas associated with the medial dorsal nucleus and pulvinar, the frontal and temporal lobes, have been more widely examined with respect to outcome. In our own study of good and poor outcome patients (Mitelman et al., 2003), superior temporal lobe volume decrease showed a stronger relationship to outcome than other cortical areas. Poor outcome associated with temporal lobe volume decrease was also shown in a sample of 56 schizophrenics (Rossi et al., 2000). Longitudinal follow-up of first-episode patients after 30 months revealed frontal and temporal volume reduction associated with clinical worsening (Gur et al., 1998). However, another longitudinal study (Ho et al., 2003) found frontal lobe white matter reduction, but no temporal lobe change associated with greater symptom severity and global gray matter volume reduction across the entire brain has also been associated with poor outcome (Cahn et al., 2002). These observed frontal and temporal lobe volume changes might be associated with parallel changes in the connection regions of the thalamus. The frontal lobe is extensively interconnected with the medial dorsal nucleus (Bachevalier et al., 1997) while the temporal lobe is more relatively more strongly interconnected with the pulvinar (Wall et al., 1982; Yeterian and Pandya, 1989, 1991). Since, in comparison to the medial dorsal nucleus, the pulvinar occupies a much greater proportion of the more ventral as compared to the more dorsal slices of the thalamus, we hypothesized that examining the size of the thalamus at different slice levels might reveal significant outcome relationships. If poor outcome patients have a defective medial dorsal nucleus/frontal lobe circuit, we might expect volume reduction at the more dorsal levels of the thalamus where the medial dorsal 475 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 nucleus occupies a greater proportion of the total thalamic volume, while if the temporal lobe/pulvinar connections were most important for outcome then volume loss would be expected to be greater at more ventral thalamic levels. Because automated methods for tracing the pulvinar and medial dorsal nucleus are not yet available, and manual nucleus-specific tracing protocols require a tremendous amount of time and effort (e.g., Kemether et al., 2003), we have applied the surrogate technique of examining thalamic shape as a preliminary study. This report presents whole thalamic volumes systematically assessed for ventrodorsal shape in a large sample of patients with schizophrenia divided into good and poor outcome. 2. Methods Schizophrenia patients (n = 106) were recruited from outpatient departments at Mount Sinai School of Medicine and the Bronx VA Medical Center and from long-term inpatient units at Pilgrim State Psychiatric Hospital. Normal control (n = 42) comparison subjects were recruited through word-ofmouth and IRB-approved advertisements. All patients met criteria for either schizophrenia (n = 95) or schizoaffective disorder (n = 11), as determined by semistructured interview with the Comprehensive Assessment of Symptom History (CASH; (Andreasen et al., 1992). Normal control subjects were screened with a modified version of the CASH and none met current or past diagnostic criteria for an Axis I psychiatric disorder. Within the schizophrenia patient group, good outcome (i.e., ‘‘non-Kraepelinian’’) and poor outcome (i.e., ‘‘Kraepelinian) status was determined by objective criteria established by Keefe et al. (1987). Specifically, Kraepelinian poor outcome patients were defined as those who met the following criteria for the previous 5 years or more: (1) continuous hospitalization, or, if living outside the hospital, complete dependence on others for food, clothing, and shelter; (2) no useful work or employment; and (3) no evidence of symptom remission (Keefe et al., 1987). Findings from a subset of participants in the current study have been included in other reports (Buchsbaum et al., 2003; Mitelman et al., 2003). Demographic and clinical data were collected at the time of MR image acquisition; these are displayed in Table 1. Schizophrenia patients and normal controls were similar in age (t = 0.46, df = 1146, p = 0.643) and Table 1 Demographic features Variable Age % Women Age neuroleptic onset PANSS positive PANSS negative PANSS general Neuroleptic exposure at scan date % None % Typical % Atypical % Both typical and atypical Neuroleptic exposure for majority of 3-year period prior to scan % None % Typical % Atypical % Both typical and atypical Data shown as mean F S.D. Normal controls (n = 42) Schizophrenia patients (n = 106) Total Non-Kraepelinian (n = 52) Kraepelinian (n = 54) 44.1 F 14.5 33.3 43.0 F 12.1 19.8 25.0 F 9.0 18.9 F 6.6 18.9 F 7.7 37.1 F 9.8 40.9 F 12.6 19.2 26.7 F 6.9 16.1 F 4.9 16.2 F 5.4 32.2 F 7.7 45.1 F 11.5 20.4 22.8 F 10.7 21.7 F 6.9 21.6 F 8.6 41.8 F 9.3 12.8 2.6 43.0 18.6 15.2 32.6 39.1 13.0 10.0 17.5 47.5 25.0 29.4 24.4 29.1 17.4 28.3 32.6 26.1 13.0 30.0 15.0 32.5 22.5 476 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 sex distribution (X2 = 3.04, df = 1, p = 0.080). Good and poor outcome patients were similar to each other in age (t = 1.79, df = 104, p = 0.077) and sex distribution (X2 = 0.022, df = 1, p = 0.883). Schizophrenia patients were assessed with the Positive and Negative Syndrome Scales (PANSS; Kay et al., 1987) and poor outcome patients had more severe positive symptom, negative symptom, and general subscale scores (all Fs>15.00, df = 1,97, p < 0.0001). For the schizophrenia patients, interview and clinical chart reviews were conducted to determine age of neuroleptic exposure and medication history over the 3-year period prior to scan acquisition. To examine the effects of neuroleptic exposure, we classified patient treatment at the time of MRI scan and for the predominant pattern over the previous 3 years as off medication, typical neuroleptics, atypical neuroleptics, or both typical and atypical neuroleptics. Poor outcome patients began neuroleptic treatment significantly earlier than good outcome patients (t = 2.02, df = 82, p = 0.006). The distribution of type of neuroleptic exposure was similar between the two groups at time of scan (X2 = 4.36, df = 3, p = 0.225) and for the majority of time over the three periods prior to the scan (X2 = 4.14, df = 3, p = 0.247). mm, matrix size 256  256) was used for MRI acquisition. MR images were adjusted along the anterior commissure –posterior commissure axis. 2.1. Image acquisition 2.3. Determination of frontal and temporal lobe volume The Signa 5  system (GE Medical Systems, Milwaukee, WI) with a 3D-SPGR sequence (TR = 24 ms, TE = 5 ms, flip angle = 40j and slice thickness = 1.22 2.2. Automated edge finding A semi-automated boundary-finding method based on the Sobel intensity-gradient filter was used to anatomically define thalamic edges, as has been reported previously for the thalamus (Byne et al., 2001) and caudate (Brickman et al., 2003; Buchsbaum et al., 2003). Outlining points were manually deposited with a mouse on the enhanced white matter edge with a semi-automated 3  3 local pixel maximum search (see Fig. 1). The top and the bottom of the thalamus were determined as the most dorsal axial slice showing a visible gray patch and the most ventral extent of the entire structure, respectively. Fig. 2 displays a three-dimensional representation of the five dorsal to ventral tracings. The distance between the two slices was divided by six to produce five equally spaced slices considered for analysis, as we have done similarly with caudate and putamen (Buchsbaum et al., 2003). For the absolute volumetric analysis we multiplied each slice area by the slice thickness. Gray and white matter quantification was conducted on coronal images for both frontal and Fig. 1. Top: Left: MRI of thalamus with midline. Right: Sobel gradient filter with deposited points indicating edge enhancement with image differentiation. A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 477 Fig. 2. The five slices of thalamus are displayed here in three-dimensional form with the color corresponding to MRI intensity values. Below: pixel locations included in outline for most ventral (red dots) and most dorsal (green dots) of the five slices with Talairach directions and dimensions marked. temporal lobes, described in greater detail elsewhere (Mitelman et al., 2003; Hazlett et al., 1998; Stein et al., 1998). Briefly, coronal slices were divided into 20 radial sectors in each hemisphere and Brodmann areas were assessed for gray and white pixels within each sector. For frontal lobe, we combined gray and white pixels separately for Brodmann areas 44, 45, and 46. For temporal lobe, we combined gray and white pixels separately for Brodmann areas 20, 21, and 22. Cortical size was corrected for whole brain volume by taking the ratio of each region to whole brain volume (e.g., frontal lobe white matter/whole brain volume). 2.4. Statistical analysis Repeated-measures analyses of variance (ANOVA) were used to examine size of the thalamus. As total brain size differed at a trend level ( p = 0.067) between schizophrenia patients and normal controls, both absolute and relative thalamic size were considered for analysis. Absolute size was computed in cubic millimeters and relative size as the ratio of area of ROI/ total brain size. Total brain size was calculated by summing the area of 61 contiguous coronal edges, comprised of 61 Brodmann areas. For these analyses, Diagnostic Group (2: schizophrenia versus normal 478 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 control or good outcome versus poor outcome) was a between-subjects variable, while Hemisphere (2: L, R) and Slice (5: most ventral to most dorsal) were repeated measures. Follow-up simple interactions and pairwise post hoc tests were used to identify the greatest source of variance for significant interactions involving Diagnosis. Pearson Product Moment correlational analyses were conducted between relative thalamic size at each slice level (collapsed across hemisphere) and relative frontal lobe and temporal lobe gray and white matter volumes. ventral levels, poor outcome patients had smaller thalami than good outcome patients, while at the two most dorsal levels, poor outcome patients had larger thalami than good outcome patients. Post hoc analyses between the two patient groups did not reach statistical significance. Poor outcome patients had smaller right but similar sized left thalami than good outcome patients (Diagnostic Group by Hemisphere interaction, F = 11.785, df = 1,104, p = 0.00086). No other significant effects involving Diagnostic Group were statistically significant. 3.2. Absolute thalamic size 3. Results 3.1.1. Schizophrenia patients versus normal controls There were no significant interactions or a main effect involving Diagnostic Group. 3.2.1. Schizophrenia patients versus normal controls Schizophrenia patients had smaller thalami at the two most ventral levels and larger thalami at the two most dorsal levels than normal controls (Diagnostic Group by Slice interaction, F = 3.47, df = 4,584, p = 0.0082; see Fig. 3). 3.1.2. Good outcome versus poor outcome patients An interaction between Diagnostic Group and Slice Level ( F = 7.377, df = 4,416, p = 0.00001) indicated a double-dissociation pattern: at the two most 3.2.2. Good outcome versus poor outcome patients Poor outcome patients had significantly smaller thalami than good outcome patients (main effect of Diagnostic Group, F = 6.22, df = 1,104, p = 0.014), 3.1. Relative thalamic size Fig. 3. Absolute thalamus slice volume in patients with schizophrenia and normal controls. A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 particularly at the three most ventral levels (Diagnostic Group by Slice interaction, F = 9.66, df = 4,416, p = 0.00001) and in the right hemisphere (Diagnostic Group by Hemisphere interaction, F = 12.13, df = 1,104, p = 0.0007) Fig. 4. Neuroleptic treatment at the time of scan (none, typical, atypical, both) did not significantly affect thalamic size (effect of size by neuroleptic status, F = 0.46, df = 3,78, p = 0.70; neuroleptic status not available on some patients). Similarly, predominant neuroleptic for 3 years before the scan had no significant effects (main effect of Neuroleptic Status F = 0.33, df = 3,78, p = 0.81; Diagnostic Group by Neuroleptic Status, F = 2.08, df = 3,78, p = 0.11). 3.2.3. Good outcome patients versus poor outcome patients versus normal controls When the three groups were considered together, there was a significant Diagnostic Group by Slice interaction ( F = 7.32, df = 8,580, p < 0.0001). As can be seen in Fig. 5, the most robust thalamic reductions were in ventral slices of thalamus in poor outcome patients compared to good outcome patients and normal controls. Dorsal differences were not as robust; poor outcome patients had similar dorsal aspects of thalamus as normal controls, whereas good outcome patients had reduction compared to the other two groups. 479 3.3. Clinical correlates of thalamic volume We explored the correlations between dorsal and ventral thalamic volume and PANSS positive, negative, and general scores, sex, age at which the patient was first treated, and duration of neuroleptic treatment in all schizophrenia patients grouped together. For these analyses, ventral thalamus volume was averaged across the three most ventral levels and dorsal was averaged across the two most dorsal levels. For ventral thalamus, significant correlations were obtained for PANSS positive (r = 0.25, df = 77, p < 0.05), PANSS General (r = 0.24, p < 0.05, df =77, r < 0.22) scores, and age at which the patient was first treated (r = 0.25, df = 77, p < 0.05). For dorsal thalamus, there was only a significant correlation found for PANSS Positive scores (r = 0.22, df =77, p < 0.05). There was no significant correlation with age, sex, or duration of treatment. When considered separately, correlations between dorsal thalamus volume and PANSS positive scores (r = 0.34, df = 43, p < 0.05) in the good outcome group and between dorsal thalamus volume and PANSS negative scores (r = 0.33, df = 36, p < 0.05) in the poor outcome patients reached statistical significance. Other comparisons were not statistically significant, although the effect sizes were similar to the overall group correlations. Fig. 4. Absolute thalamus slice volumes in patients with good and poor outcome. 480 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 Fig. 5. Absolute thalamus slice volumes in both schizophrenia patient groups and normal volunteers together. As the largest group differences in thalamus size were seen in right ventral thalamus, we explored correlations between dorsal and ventral thalamus on the right side and the clinical variables described above in the poor outcome patients only. There was a significant positive association between right ventral volume and PANSS negative scores (r =0.34, df = 36, p < 0.05). Further, larger right ventral thalamus was associated with a later onset of treatment (r = 0.33, df = 36, p < 0.05) and with a shorter duration of Table 2 Correlations coefficients between relative thalamic size at each level, collapsed across hemisphere, and sum of frontal and thalamic grey and white matter in all schizophrenia patients together, good outcome patients, and poor outcome patients Patient group Thalamus slice All Schizophrenia patients Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Slice Good outcome patients only Poor outcome patients only * p < 0.0500. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 (ventral) (dorsal) (ventral) (dorsal) (ventral) (dorsal) Sum relative frontal white 0.14 0.08 0.15 0.13 0.01 0.18 0.05 0.17 0.10 0.14 0.12 0.12 0.16 0.19 0.00 Sum relative frontal grey 0.03 0.16 0.21* 0.25* 0.21* 0.26 0.15 0.09 0.03 0.10 0.19 0.32* 0.37* 0.39* 0.35* Sum relative temporal white 0.08 0.09 0.15 0.15 0.06 0.01 0.05 0.01 0.02 0.06 0.16 0.19 0.28* 0.27* 0.19 Sum relative temporal grey 0.18 0.20* 0.26* 0.26* 0.17 0.13 0.16 0.22 0.16 0.12 0.20 0.23 0.32* 0.32* 0.20 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 treatment (r = 0.33, df = 36, p < 0.05). Correlations between right dorsal thalamus and clinical measures did not reach statistical significance. There were no significant associations between dorsal or ventral thalamus volume and type of neuroleptic exposure at the time of scan or for the majority of time over the 3-year period prior to the scan for either schizophrenia group alone or when considered together. 3.4. Correlations with other brain regions There was a tendency for larger frontal gray matter to be associated with larger dorsal thalamic areas while larger temporal gray matter is associated with larger ventral thalamic areas, at least for slices 2 – 4 (see Table 2). 4. Discussion Because of marked attentional and sensory-perceptual disturbances characteristic of schizophrenia, the thalamus has been implicated as a central structure in its underlying pathophysiology. This theory is in line with the conceptual role of the thalamus as a gatekeeper of information flow to and from relevant areas of the cortex (Jones, 1997). However, as several lines of investigation have implicated other brain regions, including frontal lobe (e.g., Buchsbaum et al., 1982; Goldman-Rakic and Selemon, 1997; Ingvar, 1974), temporal lobe (e.g., Nopoulous et al., 1997; Shenton et al., 2001), and striatum (e.g., Shihabuddin et al., 1998; 2001), it is most likely that the thalamus is but one of several structures implicated in dysfunctional schizophrenia-associated subcortical –cortical circuitry. In fact, the interconnetions among the implicated regions suggests that the function of one must be considered together with the function of the others. Some studies have shown reduction in size in the thalamus in schizophrenia patients compared to matched normal control comparison subjects (Andreasen et al., 1994; Gur et al., 1998), whereas others (Arciniegas et al., 1999; Deicken et al., 2002; Hazlett et al., 1999; Portas et al., 1998) have not. The results from the current study found that schizophrenia patients do not have overall reduced absolute or relative thalamic size compared to normal comparison 481 subjects (i.e., no main effect of Diagnostic Group). Reported negative findings of thalamic size reduction in schizophrenia could be due to specific level or nucleus dysfunction, which is not evident when the entire thalamus is considered as a whole. That is, there may be insufficient power and insufficient sensitivity to detect a small expected effect when only a single thalamic level or whole thalamus is traced (Konick and Friedman, 2001). This was supported by our finding of a significant Diagnostic Group by Slice Level interaction, which showed that schizophrenia patients do have reduced thalamic size at more ventral levels of thalamus. The reduction of more ventral aspects of the thalamus in schizophrenia implicates thalamo-temporal dysfunction. Indeed, correlational analyses demonstrated an association of larger frontal gray matter with larger dorsal thalamic areas while larger temporal gray matter was associated with more ventral thalamic areas, consistent with significant differences in the extent of reciprocal interconnection between the two areas. The assessment of volume in the thalamus and the cortex were done two entirely independent ways and on geometrically different (axial and coronal) slices and thus could not be related to a tracer bias and are difficult to assign to systematic signal intensity differences along an axis. An alternative explanation to the finding of reduced ventral aspects of thalamus in poor outcome patients and in all patients compared to controls is the possibility of thalamic shape differences among the groups. Schizophrenia patients as a group had slightly larger thalami at more dorsal levels than normal volunteers (see Fig. 3) and the effect appeared to be mostly driven by poor outcome patients (see Fig. 4). Thus, there is the possibility of nucleus reorganization, such that more ventral aspects of thalamus have shifted dorsally in poor outcome. This type of abnormal thalamic organization is consistent with the idea of faulty neurodevelopment in specific brain regions (Innocenti et al., 2003) that might be specific to poorer outcome patients. When the three groups were considered together, however, most of the group differences were accounted for by dramatic reduction in ventral portions of the thalamus in the poor outcome patients and differences in dorsal thalamus among the three groups were not as robust (see Fig. 5). Therefore, a more likely explanation of our findings is that 482 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 the loss of volume in poor outcome patients appears to result in reduced tissue mass in the ventral portion of the thalamus. Poor outcome patients had significantly smaller absolute thalami and significantly smaller absolute and relative ventral aspects of the thalamus than good outcome patients. The findings are consistent with our previous report on a subset of these patients (Mitelman et al., 2003), which demonstrated smaller temporal lobe areas in the poor outcome cohort and, again, suggest trophic effects related to outcome. The results of the current study are consistent with our in vivo and postmortem structural data demonstrating volume loss in the medial dorsal nucleus and pulvinar and with our postmortem data suggesting volume and neuronal loss restricted to the medial pulvinar (Byne et al., 2001, 2002; Kemether et al., 2003). These data are also consistent with the ventral lateral posterior nucleus volume decreases reported (Danos et al., 2002) and our results might reflect both pulvinar and ventral lateral posterior nucleus volume loss. It should be noted that in our MRI studies, where both medial dorsal and pulvinar were traced, the whole thalamic volume was not significantly reduced nor was whole thalamic volume minus medial dorsal and pulvinar volume (Byne et al., 2001). Whole thalamic volume measures are probably only an indirect indicator of more marked association nuclei and regional loss. Our finding of greater right hemisphere volume reduction in poor outcome patients is consistent with the recent report (Sullivan et al., 2003) of greater right hemisphere than left hemisphere volume reduction in inpatients. The right posterior region, generally consisting of the pulvinar, was also confirmed as smaller in our earlier study (Buchsbaum et al., 1996) consistent with the current results and with the report of Sullivan and colleagues. Voxel-based morphometry (Hulshoff Pol et al., 2001) found focal gray matter density decreases in the medial dorsal region (Talairach 3, 19, 5) and these were larger on the right. However, recent volumetric voxel-based morphometry found left hemisphere and dorsal decreases not extending into the pulvinar or right hemisphere (Ananth et al., 2002). The findings from the current study also raise the question of whether poor outcome is more associated with bilateral pathology, instead of the unilateral left hemisphere pathology often reported in studies of schizophrenia groups in general (e.g., Crow, 2000). We did not replicate the finding of smaller thalami in patients taking atypical neuroleptics (Sullivan et al., 2003). Shifts from earlier treatment with conventional neuroleptics, noncompliance in outpatients, and differences in duration of treatment make replication of chronic medication effects difficult and issues of nonresponsiveness to typical drugs further confound these analyses. A significant relationship between type of neuroleptic treatment at the time of scan or type during the 3-year period prior to scanning and dorsal or ventral thalamic volume was not found in either patient groups. This finding was somewhat inconsistent with other reports of a significant positive association between antipsychotic treatment response and volumetric expansion of the thalamus (Strungas et al., 2003). However, unlike in previous studies (Strungas et al., 2003), patients in the current study had been chronically treated with neuroleptics and were symptomatically stable. Follow-up studies of never-previously medicated patients are necessary to resolve this question. Finally, some interesting correlates of thalamic size emerged from the current study. Thalamic size was positively associated with measures of positive psychopathology in dorsal and ventral areas and with general psychopathology in ventral levels. The finding has been reported by other investigators (Portas et al., 1998), and is consistent with the notion of thalamus as central to sensory-gating and information flow to cortex (Jones, 1997). Our findings of a significant positive association between dorsal and ventral thalamic size in poor outcome patients and severity of negative symptoms were somewhat idiosyncratic and require further exploration. Significant associations between increased thalamic volume and both shorter duration of illness and older age of onset in the poor outcome group only provide some preliminary evidence of thalamic degeneration. Taken together with other recent reports, these data provide further evidence of thalamic volumetric deficits in schizophrenia and suggest that poorer outcome may be associated with more ventral thalamic volume loss. They suggest that variability in the reports of thalamic volume loss may be related to the volume loss restricted to pulvinar, medial dorsal and ventrolateral posterior regions. Detailed examination of thalamic regions in larger samples of patients will A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 be helpful in relating thalamic loss to patterns of disease outcome and regional cortical change. Acknowledgements This work was supported by a VA Merit Award (2571-005) and by grants from the National Institute of Mental Health (MH60023, MH56489, MH60384S). References Ananth, H., Popescu, I., Critchley, H.D., Good, C.D., Frackowiak, R.S., Dolan, R.J., 2002. Cortical and subcortical gray matter abnormalities in schizophrenia determined through structural magnetic resonance imaging with optimized volumetric voxelbased morphometry. Am. J. Psychiatry 159 (9), 1497 – 1505. Andreasen, N., Flaum, M., Arndt, S., 1992. The comprehensive assessment of symptoms and history (CASH): an instrument for assessing diagnosis and psychopathology. Arch. Gen. Psychiatry 49, 615 – 623. Andreasen, N.C., Arndt, S., Swayze II, V., Cizadlo, T., Flaum, M., O’Leary, D., Ehrhardt, J.C., Yuh, W.T., 1994. Thalamic abnormalities in schizophrenia visualized through magnetic resonance image averaging. Science 266, 221. Arciniegas, D., Rojas, D.C., Teale, P., Sheeder, J., Sandberg, E., Reite, M., 1999. The thalamus and the schizophrenia phenotype: failure to replicate reduced volume. Biol. Psychiatry 45 (10), 1329 – 1335. Bachevalier, J., Meunier, M., Lu, M.X., Ungerleider, L.G., 1997. Thalamic and temporal cortex input to medial prefrontal cortex in rhesus monkeys. Exp. Brain Res. 115 (3), 430 – 444. Brickman, A.M., Buchsbaum, M.S., Shihabuddin, L., Hazlett, E.A., Borod, J.C., Mohs, R.C., 2003. Striatal size, glucose metabolic rate, and verbal learning in normal aging. Brain Res. Cogn. Brain Res. 17 (1), 106 – 116. Bridle, N., Pantelis, C., Wood, S.J., Coppola, R., Velakoulis, D., McStephen, M., Tierney, P., Le, T.L., Torrey, E.F., Weinberger, D., 2002. Thalamic and caudate volumes in monozygotic twins discordant for schizophrenia. Aust. N. Z. J. Psychiatry 36 (3), 347 – 354. Buchsbaum, M.S., Ingvar, D.H., Kessler, R., Waters, R.N., Cappelletti, J., van Kammen, D.P., King, A.C., Johnson, J.L., Manning, R.G., Flynn, R.W., Mann, L.S., Bunney Jr., W.E., Sokoloff, L., 1982. Cerebral glucography with positron tomography. Use in normal subjects and in patients with schizophrenia. Arch. Gen. Psychiatry 39, 251 – 259. Buchsbaum, M.S., Someya, T., Teng, C.Y., Abel, L., Chin, S., Najafi, A., Haier, R.J., Wu, J., Bunney, W.E., 1996. PET and MRI of the Thalamus in Never-Medicated Patients with Schizophrenia. Am. J. Psychiatry 153, 191 – 199. Buchsbaum, M.S., Shihabuddin, L., Hazlett, E.A., Schroder, J., Haznedar, M.M., Powchick, P., Spiegel-Cohen, J., Wei, T., Singer, 483 M.B., Davis, K.L., 2002. Kraepelinian and non-Kraepelinian schizophrenia subgroup differences in cerebral metabolic rate. Schizophr. Res. 55 (1 – 2), 25 – 40. Buchsbaum, M.S., Shihabuddin, L., Brickman, A.M., Miozzo, R., Prikryl, R., Shaw, R., Davis, K., 2003. Caudate and putamen volumes in good and poor outcome patients with schizophrenia. Schizophr. Res. 64 (1), 53 – 62. Byne, W., Buchsbaum, M.S., Kemether, E., Hazlett, E.A., Shinwari, A., Mitropoulou, V., Siever, L.J., 2001. Magnetic resonance imaging of the thalamic mediodorsal nucleus and pulvinar in schizophrenia and schizotypal personality disorder. Arch. Gen. Psychiatry 58 (2), 133 – 140. Byne, W., Buchsbaum, M.S., Mattiace, L.A., Hazlett, E.A., Kemether, E., Elhakem, S.L., Purohit, D.P., Haroutunian, V., Jones, L., 2002. Postmortem assessment of thalamic nuclear volumes in subjects with schizophrenia. Am. J. Psychiatry 159 (1), 59 – 65. Cahn, W., Pol, H.E., Lems, E.B., van Haren, N.E., Schnack, H.G., van der Linden, J.A., Schothorst, P.F., van Engeland, H., Kahn, R.S., 2002. Brain volume changes in first-episode schizophrenia: a 1-year follow-up study. Arch. Gen. Psychiatry 59 (11), 1002 – 1010. Crow, T.J., 2000. Schizophrenia as the price that homo sapiens pay for language: A resolution of the central paradox in the origin of the species. Brain Res. Brain Res. Rev. 31 (2 – 3), 118 – 129. Cullen, T.J., Walker, M.A., Parkinson, N., Craven, R., Crow, T.J., Esiri, M.M., Harrison, P.J., 2003. A postmortem study of the mediodorsal nucleus of the thalamus in schizophrenia. Schizophr. Res. 60 (2 – 3), 157 – 166. Danos, P., Baumann, B., Bernstein, H.G., Stauch, R., Krell, D., Falkai, P., Bogerts, B., 2002. The ventral lateral posterior nucleus of the thalamus in schizophrenia: a post-mortem study. Psychiatry Res. 114 (1), 1 – 9. Danos, P., Baumann, B., Kramer, A., Bernstein, H.G., Stauch, R., Krell, D., Falkai, P., Bogerts, B., 2003. Volumes of association thalamic nuclei in schizophrenia: a postmortem study. Schizophr. Res. 60 (2 – 3), 141 – 155. Davis, K.L., Buchsbaum, M.S., Shihabuddin, L., Spiegel-Cohen, J., Metzger, M., Frecska, E., Keefe, R.S., Powchick, P., 1998. Ventricular enlargement in poor-outcome schizophrenia. Biol. Psychiatry 43, 783 – 793. Deicken, R.F., Eliaz, Y., Chosiad, L., Feiwell, R., Rogers, L., 2002. Magnetic resonance imaging of the thalamus in male patients with schizophrenia. Schizophr. Res. 58 (2 – 3), 135 – 144. Goldman-Rakic, P.S., Selemon, L.D., 1997. Functional and anatomical aspects of prefrontal pathology in schizophrenia. Schizophr. Bull. 23 (3), 437 – 458. Gur, R.E., Cowell, P., Turetsky, B.I., Gallacher, F., Cannon, T., Bilker, W., Gur, R.C., 1998. A follow-up magnetic resonance imaging study of schizophrenia. Relationship of neuroanatomical changes to clinical and neurobehavioral measures. Arch. Gen. Psychiatry 55 (2), 145 – 152. Harvey, P.D., Putman, K.M., Davidson, M., Kahn, R.S., Powchik, P., McQueeney, R., Keefe, R.S., Davis, K.L., 1991. Brief neuroleptic discontinuation and clinical symptoms in Kraepelinian and non-Kraepelinian chronic schizophrenic patients. Psychiatry Res. 38, 285 – 292. 484 A.M. Brickman et al. / Schizophrenia Research 71 (2004) 473–484 Hazlett, E.A., Buchsbaum, M.S., Haznedar, M.M., Singer, M.B., Schnurr, D.B., Jimenez, E.A., Buchsbaum, B.R., Troyer, B.T., 1998. Prefrontal cortex glucose metabolism and startle eyeblink modification abnormalities in unmedicated schizophrenia patients. Psychophysiology 35, 186 – 198. Hazlett, E.A., Buchsbaum, M.S., Byne, W., Wei, C.T., SpiegelCohen, J., Geneve, C., Kinderlehrer, R., Haznedar, M., Shihabuddin, L., Siever, L.J., 1999. Three-dimensional analysis with MRI and PET of the shape, size, and function of the thalamus in the schizophrenia spectrum. Am. J. Psychiatry 156, 1190 – 1199. Highley, J.R., Walker, M.A., Crow, T.J., Esiri, M.M., Harrison, P.J., 2003. Low medial and lateral right pulvinar volumes in schizophrenia: a postmortem study. Am. J. Psychiatry 160 (6), 1177 – 1179. Ho, B.C., Andreasen, N.C., Nopoulos, P., Arndt, S., Magnotta, V., Flaum, M., 2003. Progressive structural brain abnormalities and their relationship to clinical outcome: a longitudinal magnetic resonance imaging study early in schizophrenia. Arch. Gen. Psychiatry 60 (6), 585 – 594. Hulshoff Pol, H.E., Schnack, H.G., Mandl, R.C., van Haren, N.E., Koning, H., Collins, D.L., Evans, A.C., Kahn, R.S., 2001. Focal gray matter density changes in schizophrenia. Arch. Gen. Psychiatry 58 (12), 1118 – 1125. Ingvar, D.H., 1974. Regional cerebral blood flow in organic dementia and in chronic schizophrenia. Triangle 13 (1), 17 – 23. Innocenti, G.M., Ansermet, F., Parnas, J., 2003. Schizophrenia, neurodevelopment, and corpus callosum. Mol. Psychiatry 8 (3), 261 – 274. Jones, E., 1997. Cortical development and thalamic pathology in schizophrenia. Schizophr. Bull. 23, 483 – 501. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13 (2), 261 – 276. Keefe, R.S., Mohs, R.C., Losonczy, M.F., Davidson, M., Silverman, J.M., Kendler, K.S., Horvath, T.B., Nora, R., Davis, K.L., 1987. Characteristics of very poor outcome schizophrenia. Am. J. Psychiatry 144 (7), 889 – 895. Keefe, R.S., Mohs, R.C., Davidson, M., Losonczy, M.F., Silverman, J.M., Lesser, J.C., Horvath, T.B., Davis, K.L., 1988. Kraepelinian schizophrenia: a subgroup of schizophrenia? Psychopharmacol. Bull. 24, 56 – 61. Keefe, R.S., Frescka, E., Apter, S.H., Davidson, M., Macaluso, J.M., Hirschowitz, J., Davis, K.L., 1996. Clinical characteristics of Kraepelinian schizophrenia: replication and extension of previous findings. Am. J. Psychiatry 153, 806 – 811. Kemether, E., Buchsbaum, M.S., Byne, W., Hazlett, E.A., Haznedar, M., Brickman, A.M., Platholi, J., Bloom, R., 2003. Magnetic resonance imaging of mediodorsal, pulvinar, and centromedian nuclei of the thalamus in patients with schizophrenia. Arch. Gen. Psychiatry 60 (10), 983 – 991. Konick, L.C., Friedman, L., 2001. Meta-analysis of thalamic size in schizophrenia. Biol. Psychiatry 49 (1), 28 – 38. Mitelman, S.A., Shihabuddin, L., Brickman, A.M., Hazett, E.A., Buchsbaum, M.S., 2003. MRI assessment of gray and white matter distribution in Brodmann’s areas of the cortex in good and poor outcome schizophrenia. Am. J. Psychiatry 160 (12), 2154 – 2168. Portas, C.M., Goldstein, J.M., Shenton, M.E., Hokama, H.H., Wible, C.G., Fischer, I., Kikinis, R., Donnino, R., Jolesz, F.A., McCarley, R.W., 1998. Volumetric evaluation of the thalamus in schizophrenic male patients using magnetic resonance imaging. Biol. Psychiatry 43 (9), 649 – 659. Rossi, A., Bustini, M., Prosperini, P., Marinangeli, M.G., Splendiani, A., Daneluzzo, E., Stratta, P., 2000. Neuromorphological abnormalities in schizophrenic patients with good and poor outcome. Acta Psychiatr. Scand. 101 (2), 161 – 166. Roy, M., Lehoux, C., Emond, C., Laplante, L., Bouchard, R., Everett, J., Merette, C., Maziade, M., 2003. A pilot neuropsychological study of Kraepelinian and non-Kraepelinian schizophrenia. Schizophr. Res. 62, 155 – 163. Shihabuddin, L., Buchsbaum, M.S., Hazlett, E.A., Haznedar, M.M., Harvey, P.D., Newman, A., Schnur, D.B., Spiegel-Cohen, J., Wei, T., Machac, J., Knesaurek, K., Vallabhajosula, S., Biren, M.A., Ciaravolo, T.M., Luu-Hsia, C., 1998. Dorsal striatal size, shape, and metabolic rate in never-medicated schizophrenics performing a verbal learning task. Arch. Gen. Psychiatry 55 (3), 235 – 243. Shihabuddin, L., Buchsbaum, M.S., Hazlett, E.A., Silverman, J., New, A., Brickman, A.M., Mitropoulou, V., Nunn, M., Fleischman, M.B., Tang, C., Siever, L.J., 2001. Striatal size and relative glucose metabolic rate in schizotypal personality disorder and schizophrenia. Arch. Gen. Psychiatry 58 (9), 877 – 884. Staal, W.G., Hulshoff Pol, H.E., Schnack, H.G., van Haren, N.E., Seifert, N., Kahn, R.S., 2001. Structural brain abnormalities in chronic schizophrenia at the extremes of the outcome spectrum. Am. J. Psychiatry 158 (7), 1140 – 1142. Stein, D.J., Buchsbaum, M.S., Hof, P.R., Siegel Jr., B.V., Shihabuddin, L., 1998. Greater metabolic rate decreases in hippocampal formation and proisocortex than in neocortex in Alzheimer’s disease. Neuropsychobiology 37, 10 – 19. Strungas, S., Christensen, J.D., Holcomb, J.M., Garver, D.L., 2003. State-related thalamic changes during antipsychotic treatment in schizophrenia: Preliminary observations. Psychiatry Res. 124 (2), 121 – 124. Sullivan, E.V., Rosenbloom, M.J., Serventi, K.L., Deshmukh, A., Pfefferbaum, A., 2003. Effects of alcohol dependence comorbidity and antipsychotic medication on volumes of the thalamus and pons in schizophrenia. Am. J. Psychiatry 160 (6), 1110 – 1116. Wall, J.T., Symonds, L.L., Kaas, J.H., 1982. Cortical and subcortical projections of the middle temporal area (MT) and adjacent cortex in galagos. J. Comp. Neurol. 211 (2), 193 – 214. Yeterian, E.H., Pandya, D.N., 1989. Thalamic connections of the cortex of the superior temporal sulcus in the rhesus monkey. J. Comp. Neurol. 282 (1), 80 – 97. Yeterian, E.H., Pandya, D.N., 1991. Corticothalamic connections of the superior temporal sulcus in rhesus monkeys. Exp. Brain Res. 83 (2), 268 – 284. Young, K.A., Manaye, K.F., Liang, C., Hicks, P.B., German, D.C., 2000. Reduced number of mediodorsal and anterior thalamic neurons in schizophrenia. Biol. Psychiatry 47 (11), 944 – 953.