Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

A longitudinal study investigating sub-threshold symptoms and white matter changes in individuals with an ‘at risk mental state’ (ARMS)

Schizophrenia Research, 1-3, 162, pages 7 - 13



Evidence supports disruption in white matter (WM) connectivity in established schizophrenia, however, it is unclear when these abnormalities occur during the course of illness and if they are progressive. Here we investigated whether WM abnormalities predate illness onset by examining a group of individuals with an ‘at risk mental state’ (ARMS) and assess whether there is evidence of progressive change. We hypothesized that WM abnormalities are associated with symptom change.


Sixteen healthy controls and 41 ARMS subjects at baseline underwent Diffusion Tensor Imaging (DTI). Sub-threshold positive symptoms were measured using the Scale of Prodromal Symptoms (SOPS). Imaging and symptoms were re-administered in the ARMS group after one year (52 weeks). Fractional anisotropy (FA) value differences between ARMS and control groups at baseline were localized using the method of Tract-Based Spatial Statistics (TBSS).


At baseline, FA was significantly reduced in a sub-region of the corpus callosum (CC) in the ARMS group as a whole compared to controls. This reduction was also found in the 34 individuals who did not transition (ARMS-N) during the one-year follow-up. However, the ARMS-N group showed a significant improvement in sub-threshold positive symptoms at follow-up, which was correlated with an increase in FA in the same CC region (r = − 0.664, p < 0.001).


There was a significant FA reduction in the CC in individuals at high risk for psychosis regardless of transition status at one year. This suggests that WM abnormalities in the CC may represent a biological vulnerability to psychosis. Improvement in sub-threshold positive symptoms was associated with improvement in measures of WM integrity in the CC. This may suggest that neurobiological ‘resilience’ is associated with improved outcomes, although this notion requires future study.

Keywords: At risk mental state, Corpus callosum, Diffusion tensor imaging, Prodrome, Psychosis, White matter integrity.

1. Introduction

Many studies have revealed that multiple brain regions are implicated in schizophrenia ( Shenton et al., 2001 ). It is postulated that these abnormalities reflect dysfunction of neural connectivity consistent with the disconnection hypothesis of schizophrenia ( Friston, 1998 ). Recently, it has been possible to directly examine the disconnection hypothesis by assessing white matter (WM) connectivity using diffusion tensor imaging (DTI) ( Zalesky et al., 2011 ). A meta-analysis of DTI studies in schizophrenia by Bora et al. (2011) identified significant fractional anisotropy value (FA) reductions in the genu of the corpus callosum, anterior cingulate/medial frontal WM and the anterior limb of the internal capsule and the external capsule/corona radiata. The regions presented in this meta-analysis were largely consistent with abnormalities proposed to be involved in schizophrenia (Crow, 1998 and Andreasen, 1999). DTI studies have also identified relationships between FA and various psychiatric symptoms, including positive symptoms, negative symptoms and hallucinations as well as cognitive deficits (for review: Walterfang et al., 2011 ).

However, whether these widespread WM impairments identified with DTI predated the onset of psychotic symptoms or developed over the course of the illness progressively remains unclear (Yung and McGorry, 1996 and Pantelis et al, 2003), although this is suggested by progressive WM volumetric changes in UHR individuals converting to psychosis ( Walterfang et al., 2008 ). In a longitudinal DTI study of subjects with ‘at risk mental states’ (ARMS), Peters et al. (2010) found no baseline FA differences in any brain region between those ARMS patients developing psychosis (ARMS-P) compared with those not developing psychosis (ARMS-N). In contrast, Bloemen et al. (2010) reported that ARMS-P had lower baseline FA in the putamen and in the superior temporal gyrus compared to the ARMS-N. Further, in this study, positive symptom scores in ARMS-P subjects were negatively correlated with FA in the middle temporal lobe, while for the whole ARMS group (ARMS-W) positive symptoms negatively correlated with FA in the right superior temporal lobe. In a previous study, we found a negative correlation between change in FA in the genu of the corpus callosum and change in negative symptoms in ARMS-W during the 52-week study period ( Saito et al., in submission ). These reports suggest that relationships between symptoms and FA may be relevant to all patients with ARMS, rather than only apparent in those who transition to psychosis.

While transition to psychosis is defined by the expression of prominent positive symptoms, sub-threshold psychotic symptoms gradually develop before onset of psychosis and this increasing severity may be associated with progressive changes in multiple brain regions ( Pantelis et al., 2005 ). To date, no longitudinal DTI studies have examined the relationship of sub-threshold positive symptom changes to WM changes. Further, although, the positive symptoms of almost 80% of ARMS subjects do not exceed the threshold for psychosis and usually recover during the follow-up period, no DTI studies have examined the relationship between recovery of sub-threshold symptoms and the WM changes of these so-called “false positives”.

We hypothesized that the severity of sub-threshold positive symptoms that emerge in ARMS subjects is associated with reduction of WM integrity. To elucidate the relationship between sub-threshold positive symptoms and WM integrity in ARMS subjects, we investigated the relationship between cross sectional and longitudinal FA and sub-threshold positive symptoms in ARMS subjects followed for one year.

2. Method

2.1. Participants

The ARMS individuals were recruited at the Toho University Omori Medical Center (baseline). Experienced psychiatrists diagnosed participants as having ARMS using the Structured Interview for Prodromal Syndrome (SIPS) ( Miller et al., 2003 ) at the time of the first consultation. Individuals with clearly diagnosed brain disease or substance dependence were excluded. The ARMS-W group was treated at the “Youth Clinic” and at the “Il Bosco” day-care center that was established for persons with early psychosis (Mizuno et al, 2009 and Nemoto et al, 2012). Participants showing severe deterioration of clinical symptoms received antipsychotics even in the absence of apparent positive symptoms ( Yung et al., 2007 ). Transition to psychosis during the follow-up period was defined using the SIPS. After a 52-week follow-up period, the ARMS-W group was divided into ARMS-P and ARMS-N. To investigate the influence of antipsychotics on the ARMS-N group, we further divided ARMS-N subjects into those who were not prescribed antipsychotics (ARMS-NN) and those treated with antipsychotics (ARMS-NA) ( Fig. 1 ). The mean dosage of antipsychotics at baseline and 52 weeks was expressed as milligram equivalents of chlorpromazine ( Woods, 2003 ) and dosage was log10transformed to reduce skewness (CPZeq-log).


Fig. 1 Flow chart outlining of participants. ARMS-W: overall group of ARMS subjects; ARMS-P: ARMS subjects who developed psychosis during the 1-year follow-up period; ARMS-N:ARMS subjects who did not develop psychosis during the 1-year follow-up period; ARMS-NN: ARMS subjects who did not develop psychosis and were not prescribed antipsychotics during the follow-up period; ARMS-NA: ARMS subjects who did not develop psychosis but were prescribed antipsychotics during the follow-up period.

Healthy control subjects were recruited from independent sources in the community and were interviewed in detail by experienced psychiatrists. These individuals had no current psychiatric disorder and no history of psychiatric illness, head trauma, neurological illness, serious medical illness or substance dependence. Written informed consent was obtained from all the participants after the study had been explained in full. This study was approved by the Ethics Committee of the Toho University School of Medicine.

2.2. Scaling of the severity of sub-threshold positive symptoms

The Scale of Prodromal Symptoms (SOPS) ( Miller et al., 2003 ) is contained within the SIPS. The SOPS has been designed to define and diagnose the psychosis prodrome, to characterize the symptoms and their severity and, importantly, to assess change longitudinally. The SOPS is a 19-item scale designed to measure the severity of prodromal symptoms and changes over time. The SOPS contains four subscales for Positive, Negative, Disorganization, and General symptoms. In this study, sub-threshold positive symptom severity of ARMS subjects was calculated as the sum of the five SOPS positive items (POS) at baseline and at 52 weeks. For longitudinal analysis, the one-year change in POS (ΔPOS) was calculated by subtracting POS score at baseline from the score at 52 weeks.

2.3. Image acquisition

The ARMS subjects and controls underwent MRI scans (EXCELART Vantage, XGV 1.5T; Toshiba Medical Systems, Tokyo, Japan) with a five-channel head coil at baseline; DTI images were acquired by using a single-shot spin-echo echo-planar sequence in 30 axial sections. The whole brain diffusion-weighted images were recorded along 6 gradient directions using ab-value of 1000 s/mm2together with unweighted (b = 0) images. For each image, we used the following parameters: field of view = 260 × 260 mm2; matrix size 128 × 128; voxel resolution 1.02 × 1.02 × 5 mm3; TE = 100 ms; TR = 7,668 ms; number of signal average 3. After 52 weeks from baseline scan, ARMS subjects underwent a second scan using the same instrument and same method.

Although some of subjects underwent SIPS and MRI repeatedly during the follow-up period, to analyze strictly, we adopted the SIPS and MRI data at 52 weeks for each subject.

2.4. Image processing

In the previous study, we measured the FA values, using Tract Specific Analysis (TSA), of each of the three CC sub-regions that were manually segmented based on brain anatomy ( Saito et al., in submission ). However, in this study, using Tract-Based Spatial Statistics analysis (TBSS), we analyzed the precise difference in FA values of cerebral white matter between ARMS-W and controls ( Smith et al., 2006 ). Briefly, correction for the effects of head movement and eddy currents was performed. A brain mask was created by using one of the b = 0 (no diffusion weighted) images. A diffusion tensor was then fitted to each voxel comprising the mask and FA images were computed. Each image was aligned to a common space (FMRIB58-FA standard-space image) using the nonlinear registration tool FNIRT which uses ab-spline representation of the registration warp field ( Rueckert et al., 1999 ). Next, the mean FA image was created and thinned to create a mean FA skeleton, which represents the centers of all tracts common to the group. Each subject's aligned FA data were then projected onto this skeleton.

Group comparisons between ARMS-W and controls were carried out using t-tests. These were computed using the randomize function implemented in FSL, and implementing the recently developed threshold-free cluster-enhancement method for proper statistical inference of spatially distributed patterns. A p-value of < 0.05 was regarded as significant. Family-wise error was used to correct for multiple comparisons. In this way, significant regions of interest (ROIs) were identified and delineated for further analysis, as below.

2.5. Quantification of FA of significant ROIs

To analyze the longitudinal FA differences between the various subgroups at baseline and 52 weeks, we specifically investigated FA in the ROIs identified in the previous step. Using a custom MATLAB ( www.mathworks.com.jp ) script, the average of the FA across all the voxels comprising the significant regions identified with TBSS was computed. For longitudinal analysis, the one-year changes in FA (ΔFA) of each group were calculated by subtracting the FA values at baseline from the values at 52 weeks.

2.6. Statistical analysis

Data were analyzed using SPSS version 20 ( www.spss.com ). We first compared differences in baseline POS scores and ROI FA values between controls, ARMS-NN, ARMS-NA and ARMS-P using ANOVA. Post-hoc tests were examined using Tukey HSD. We also examined the longitudinal changes in FA and POS scores respectively between ARMS-NN and ARMS-NA between baseline and one year, using repeated measures ANOVA that included a between-group factor (ARMS-N subgroups) and a within-subject factor (baseline and 52 weeks).

To investigate the longitudinal relationship between FA changes and POS changes during follow-up, we analyzed the correlation between ΔFA and ΔPOS in the ARMS subgroups.

Regression analysis was used to investigate factors that influenced POS changes over the follow-up in the ARMS-N subgroups. The dependent factor was ΔPOS. The regressor of interest was ΔFA. For ARMS-NN, age, sex and ΔFA were included as regressors. For ARMS-NA, age, sex, ΔFA and the mean dosage of antipsychotics during the one-year follow-up period were included as regressors. The mean dosage of antipsychotics during the one-year follow-up period was calculated as the mean value of antipsychotic dosage (milligram equivalents of chlorpromazine) at the two time points (baseline and 52 weeks) and mean dosage for each person was log10transformed to reduce skewness (CPZeq-mean-log).

3. Results

3.1. Demographic and characteristics of the groups

Participants were sixteen controls and 41 ARMS-W. Seven of 41 (17.1%) subjects transitioned to psychosis during the one-year follow-up. Among these 7 ARMS-P subjects 2 subjects did not undergo the SIPS and 2 subjects did not undergo MRI at follow-up (i.e. only four underwent both SIPS and MRI at both baseline and 52 weeks). Thus, this group was not included in the longitudinal analysis (i.e. analyses were limited to the ARMS-N subgroups).

Among 34 ARMS-N, 23 subjects had been prescribed antipsychotics (ARMS-NA) during the follow-up. Of 23 ARMS-NA subjects followed up at 52 weeks, 5 subjects did not undergo SIPS and 1 subject did not undergo MRI, and among 11 ARMS-NN subjects followed up, 2 subjects did not complete SIPS. The demographic characteristics of the groups are shown in Table 1 .

Table 1 Demographic data, scores for sub-threshold positive symptoms (POS), and fractional anisotropy values (FA).

  Baseline 52 weeks
Participants (male/female) n = 16 (8/8) n = 11 (3/8) n = 23 (6/17) n = 7 (1/6) p = 0.273 n = 11 (3/8) n = 23 (6/17) n = 7 (1/6) p = 0.789
Age at baseline and 52 weeks (years) 23.19 (2.86) 24.18 (7.88) 23.35 (6.49) 20.71 (5.53) p = 0.673 25.09 (7.80) 24.48 (6.45) 21.20 (5.26) p = 0.549
Subject underwent SOPS (male/female) n = 11 (3/8) n = 21 (4/17) n = 7 (1/6) p = 0.778 n = 9 (2/7) n = 18 (4/14) n = 5 (1/4) p = 0.994
POS score at baseline and 52 weeks 17.36 (3.26) 19.05 (4.18) 19.14 (4.67) p = 0.499 15.22 (5.09) 13.22 (4.953) 16.80 (4.09) p = 0.301
Subject underwent DTI (male/female) n = 16 (8/8) n = 11 (3/8) n = 23 (6/17) n = 7 (1/6) p = 0.273 n = 11 (3/8) n = 22 (6/16) n = 5 (1/4) p = 0.942
FA value in ROI at baseline and 52 weeks 0.47 (0.04) 0.39 (0.05) 0.40 (0.07) 0.43 (0.06) p = 0.002 lowastlowast 0.38 (0.04) 0.40 (0.04) 0.43 (0.06) p = 0.128
Subject underwent DTI and SOPS at baseline and 52 weeks (male/female) n = 9 (2/7) n = 17 (4/13) n = 4 (1/3) p = 0.994 n = 9 (2/7) n = 17 (4/13) n = 4 (1/3) p = 0.994
Dose of CPZeq-log 0 1.75 (0.62) 2.06 (0.39) 0 1.80 (0.64) 2.37 (0.32)

lowastlowast p < 0.01.

ARMS-NN: ARMS subjects who did not develop psychosis and were not prescribed antipsychotics during the 1-year follow-up period; ARMS-NA: ARMS subjects who did not develop psychosis but were prescribed antipsychotics during the follow-up period; ARMS-P:ARMS subjects who developed psychosis during the follow-up period.

lowastp < 0.05.

There were no significant group differences in dose of medication between ARMS-NA and ARMS-P at baseline (t(28) = − 1.361, p = 0.18). Within ARMS-NA, there was no significant difference in dose of medication between baseline and 52 weeks (t(22) = − 0.727, p = 0.48). All ARMS-P subjects received antipsychotics at baseline because they had shown evidence of clinical deterioration, even though positive symptoms were not prominent. Further, while these subjects converted to psychosis during the one-year follow-up period with increased dose of antipsychotics during the acute phase, all patients had been treated at the “Youth Clinic” and the “Il Bosco” and reached remission before 52 weeks. Consequently, the mean dose of antipsychotics was reduced again by 52 weeks.

3.2. POS score of ARMS subgroups

Assessment at baseline did not identify any significant differences in POS between ARMS-NN, ARMS-NA and ARMS-P at baseline (F(2,36) = 0.709, p = 0.50). Repeated measures analysis examining changes over time (excluding the ARMS-P because of low numbers) indicated that there was no significant main effect of group (F(1,25) = 0.064, p = 0.80) between ARMS-NN and ARMS-NA. However, there was a significant effect of time for POS score (F(1,25) = 19.001, p < 0.001) and a significant time × group interaction (F(1,25) = 4.985 p = 0.035), with greater POS reduction observed in the ARMS-NA group.

3.3. FA differences between the ARMS group and controls at baseline

Comparing FA between groups, significant differences were found between the ARMS-W and controls at baseline using TBSS, in one cluster involving part of the genu and body of the corpus callosum (CC) (confirmed with JHU ICBM-DTI-81 WM Labels) ( Fig. 2 ).


Fig. 2 Significant FA differences between ARMS-W group and controls. ARMS-W: overall group of ARMS subjects; Image is radiologically oriented (participant’s left is to the right) ; p-value < 0.05 was regarded as significant (after family-wise error correction for multiple comparison).

3.4. Group differences in FA at baseline

The region identified in Fig. 2 , where the significant FA reduction was detected in the ARMS-W compared to controls was defined as the ROI.

At baseline the FA in the ROI differed significantly across groups (controls, 0.47 ± 0.04; ARMS-NN, 0.39 ± 0.05; ARMS-NA, 0.40 ± 0.07; ARMS-P, 0.43 ± 0.06); (F(3,53) = 5.701, p = 0.002). Post-hoc tests revealed significant differences in FA between controls and ARMS-NN (p = 0.004) and between controls and ARMS-NA (p = 0.005) ( Fig. 3 ).


Fig. 3 Mean FA value of controls and ARMS subgroups at baseline. ARMS-NN: ARMS subjects who did not develop psychosis and were not prescribed antipsychotics during the follow-up period; ARMS-NA: ARMS subjects who did not develop psychosis but were prescribed antipsychotics during the follow-up period; ARMS-P: ARMS subjects who developed psychosis during the follow-up period. *P < 0.05; **P < 0.01.

3.5. Longitudinal relationship between FA changes and changes in POS in ARMS-N subgroups

For comparison of FA between ARMS-NN and ARMS-NA, there was no significant effect of group (F (1,31) = 0.517, p = 0.48), time (F (1,31) = 0.339 p = 0.567) or time × group interaction (F (1,31) = 1.000, p = 0.33).

Nine ARMS-NN, 17 ARMS-NA and 4 ARMS-P subjects underwent both MRI and SOPS at baseline and 52 weeks. There was a significant correlation between ΔFA and ΔPOS between baseline and one year for each of the ARMS subgroups (ARMS-W, n = 30, r = − 0.568, p = 0.001; ARMS-N, n = 26, r = − 0.664, p < 0.001; ARMS-NN, n = 9, r = − 0.776, p = 0.014; ARMS-NA, n = 17, r = − 0.586, p = 0.013). However, the correlation was not significant in the ARMS-P group (n = 4, r = 0.139, p = 0.86) ( Fig. 4 ).


Fig. 4 Correlations between ∆FA and ∆POS for ARMS subgroups. Green triangles: scatter of ARMS-NN; Blue diamonds: scatter of ARMS-NA; Red X: scatter of ARMS-P. Red broken line: regression line of ARMS-P; Black solid line: regression line of ARMS-N. ∆POS: one-year change in POS; ∆FA: one-year change in FA; ARMS-NN: ARMS subjects who did not develop psychosis and were not prescribed antipsychotics during the follow-up period; ARMS-NA: ARMS subjects who did not develop psychosis but were prescribed antipsychotics during the follow-up period; ARMS-P: ARMS subjects who developed psychosis during the follow-up period.

The determinants of ΔPOS were examined further using regression to assess the effects of FA, age, gender and mean dosage of antipsychotics during the one-year follow-up. Only ΔFA significantly correlated with ΔPOS for both subgroups (ARMS-NN p = 0.03; ARMS-NA: p = 0.004). Mean antipsychotic dose was not associated with ΔPOS in the ARMS-NA group (ARMS-NA: p = 0.96) ( Table 2 ).

Table 2 Results of regression analyses for ΔPOS of ARMS-N subgroups.

Regressor of ΔPOS β t p (95% CI) β t p (95% CI)
age − 0.248 − 0.932 0.394 − 0.882 to 0.384 − 0.452 − 1.889 0.083 − 0.596 to 0.042
sex − 0.441 − 1.756 0.139 − 15.509 to 2.918 0.431 1.802 0.097 − 0.782 to 8.252
ΔFA − 0.714 − 3.012 0.030 − 245.837 to − 19.433 lowast − 0.791 − 3.548 0.004 − 160.759 to − 38.430 lowastlowast
CPZeq-mean-log   0.010 0.046 0.964 − 4.718 to 4.922

lowast p < 0.05.

lowastlowast p < 0.01.

ARMS-N:ARMS subjects who did not develop psychosis during the 1-year follow-up period; ARMS-NN: ARMS subjects who did not develop psychosis and were not prescribed antipsychotics during the follow-up period; ARMS-NA: ARMS subjects who did not develop psychosis but were prescribed antipsychotics during the follow-up period; ∆POS: one-year change in POS; ∆FA: one-year change in FA.

4. Discussion

In this study we examined WM integrity in prepsychotic individuals considered at-risk for psychosis.

In the cross sectional analysis, FA values in the CC for the ARMS-P group was intermediate relative to controls and ARMS-N group ( Fig. 3 ). While this result is not in accordance with our hypothesis that severity of symptoms would be associated with reduction of WM integrity or previous findings that reported FA reductions in the left CC of ARMS-P group ( Carletti et al., 2012 ), our sample of converters was small. However, some recent studies suggest that FA reductions in the brain of schizophrenia patients were inhomogeneous across different sub-regions of the CC. Thus, Schneiderman et al., (2009) reported that although the DTI index was lower in most regions of CC, it was higher in the right body and left anterior portion of the genu in male schizophrenia patients compared to controls.

In contrast, in the ARMS-N group there was a significant FA reduction in the CC compared to healthy controls at baseline. These results raise the possibility that sub-threshold psychotic symptoms observed in ARMS-N are associated with potentially deleterious biological differences affecting white matter integrity.

In the longitudinal analysis, only seven individuals converted to psychosis (ARMS-P) during the follow-up period and because only 4 subjects underwent both MRI and SOPS at follow-up, our main results relate to those who did not convert to psychosis. Interestingly, this ARMS-N group showed a significant improvement in sub-threshold positive symptoms (POS) during one-year follow-up and this improvement was associated with increased FA. This finding was not explained by potential confounders including age and gender. Further, analysis of the ARMS-N subjects (including medicated and antipsychotic-naïve) identified the same relationship between POS and FA. While these findings suggest that this change was not explained by use of antipsychotics, further study is needed to increase the number of unmedicated subjects. These results indicate that improvement in the severity of sub-threshold psychotic symptoms was associated with improved indices of white matter integrity in a sub-region of the left CC.

We found that improvements in WM measures were accompanied by improvement in such symptoms, suggesting a dynamic process that may be relevant to the onset of disorder. While our sample of ARMS-P was inadequate to examine this further, an intriguing hypothesis is that improvement in white matter over time reduces the risk for developing psychosis. This may imply neurobiological resilience in a subgroup of young people deemed at risk for psychosis (see: Pantelis and Bartholomeusz, 2014 ).

The CC is the largest commissure in the human brain, connecting neocortical regions of the two hemispheres. This includes the left and right frontal areas, associated with specific higher cognitive functions ( de Lacoste et al., 1985 ). Further, the normal human brain is characterized by a pattern of gross anatomical asymmetry, which has been associated with the processing of language ( Cook, 2002 ). Individuals with schizophrenia show a loss of the normal asymmetry in regions such as superior temporal gyrus relevant to language (e.g., Crow et al., 2007 ). Asymmetry in schizophrenia is also observed in anterior cingulate region close to the CC ( Yücel et al., 2002 ), which has been associated with impairments in executive function ( Fornito et al., 2006 ). Crow (1998) has proposed a translocation misconnection syndrome based on such evidence together with evidence for asymmetry in the CC ( Highley et al., 1999 ), a finding that is consistent with our findings implicating left CC. Impairment of CC in schizophrenia had been consistently reported in neuroimaging studies (e.g. Walterfang et al., 2006 ), including the meta-analysis of schizophrenia DTI studies finding FA reductions in this region ( Patel et al., 2011 ). Similar to our study, Carletti et al. (2012) reported reduction of DTI indices in the left CC of ARMS individuals, which were intermediate between first episode schizophrenia and controls. Additionally, Knochel et al. (2012) reported volume and FA reductions of CC in schizophrenia as well as first-degree relatives suggesting a role for genetic influences. This may also suggest that such measures represent potential endophenotypic markers for schizophrenia.

Our findings indicate that the increase of FA in CC was associated with improvement or recovery of POS rather than antipsychotics in ARMS-N group. There is already some evidence of relationships between psychiatric symptoms and the impairment of CC in schizophrenia (Bleich-Cohen et al, 2012, Nakamura et al, 2012, and Serpa et al, 2012). These reports support our findings in ARMS, suggesting that changes in psychiatric symptom severity are associated with changes of CC WM integrity.

Taken together with the consistent reports of impairment of frontal CC in the ARMS as well as schizophrenia, our results of FA reduction in anterior CC in the ARMS group are in accordance with previous studies. However, to our knowledge, this is the first study that reports FA reduction in ARMS that have not made the transition to psychosis, so-called “false positive” cases; and the first study to report FA changes over time in such a group. These results raise the possibility that the “false positives” do not simply express sub-threshold psychotic symptoms, but also manifest neurobiological risk for developing psychosis and that our findings of improvement in FA over a 12-month period relate to ‘protective’ factors or ‘neurobiological resilience’ ( Pantelis and Bartholomeusz, 2014 ).

On the other hand, theoretically, true positives (those making the transition) will be outnumbered by “false positives”. However, there is the possibility of “false false positives”, that is those who would have made the transition to psychotic disorder had it not been for some intervention or change in circumstances ( Yung et al., 2007 ). This is particularly the case in our study, as all ARMS patients underwent various therapies including a proportion receiving medication. However, we have demonstrated that a positive change in FA was associated with fewer symptoms, suggesting that neurobiological vulnerability can be modified.

Takeuchi et al. (2010) reported that working memory training correlated with increased FA in the WM regions adjacent to the intraparietal sulcus and the body of the CC after training in healthy people. Penades et al. (2013) reported that schizophrenia patients who underwent cognitive remediation therapy showed increase of FA in the genu of the CC. Such studies suggest that changes at the earliest stages of psychosis and pre-psychosis are dynamic and changeable, potentially reflecting plasticity. Importantly, our findings are indicative of positive neurobiological outcomes that may be associated with amelioration of symptoms and possibly illness prevention. Further work should explore ways to reduce such neurobiological vulnerability, including measures of WM integrity.

We acknowledge some limitations to our study. First, the risk of Type II errors should be considered in view of the relatively small sample sizes, particularly for the ARMS-P. Second, while TBSS has been widely used to identify focal between-group differences in FA, this approach is associated with several limitations, including disregard for potential effects outside of the ‘skeleton’ and evidence showing the tract-specificity of the skeleton projection step may be low for some fiber geometries (Zalesky, 2011 and Keihaninejad et al, 2012). Despite these methodological considerations, TBSS is currently still considered the leading technique for voxel-wise diffusion imaging analysis as many alternative approaches are far less reproducible and may have similar problems ( Bach et al., 2014 ).

Third, although the effect of gender differences was not significant in our study, we did not specifically design the study to examine the role of sex differences ( Crow et al., 2013 ). Savadjiev et al., (2014) , in their study of schizophrenia patients, reported a significant diagnosis by sex interaction of DTI indices in the left anterior CC. Further, they reported a difference in the correlation between the severity of patients' negative symptoms and DTI indices in the left anterior CC in both sexes. These results coincide with previous studies suggesting that the structural change underlying the continuum of psychosis relates to the interaction of laterality and sex (Bora et al, 2012 and Lagopoulos et al, 2013). Further work is required to clarify the effects of gender, laterality or clinical subtypes on FA change on CC.

Role of funding source

Public Grant from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) KAKENHI Grant Number 21591533 to Masafumi Mizuno.


Naoyuki Katagiri and Masafumi Mizuno designed the study and wrote the protocol. Junichi Saito, Naohisa Tsujino, Taiju Yamaguchi, Shinya Ito, Nobuyuki Shiraga and Shigeki Aoki were involved at the conceptualization level of the project. Keiko Morita and Naoyuki Katagiri collected the data. Masaaki Hori, Keigo Shimoji, Issei Fukunaga, contribute to image processing. Naoyuki Katagiri, Takahiro Nemoto, Andrew Zalesky, Dominic Dwyer and Christos Pantelis analyzed the data. Naoyuki Katagiri wrote the draft of this manuscript. Masafumi Mizuno, Christos Pantelis and Takahiro Nemoto contributed to the writing and revision of the manuscript. All authors contributed to and have approved the final manuscript.

Conflict of interest

All authors declare that they have no conflicts of interest.


Image processing of this research was supported by a Grant-in-Aid for Scientific Research on Innovative Areas (Comprehensive Brain Science Network) from the Ministry of Education, Science, Sports and Culture of Japan (221S0003). Christos Pantelis analyzed the data and contributed to the writing and revision of the manuscript. Christos Pantelis was supported by a National Health and Medical Research Council (NHMRC) Senior Principal Research Fellowship (ID: 628386) and NARSAD Distinguished Investigator Award (USA, ID: 18722). Andrew Zalesky was supported by an Australian National Health and Medical Research Council (ARC) Career Development Fellowship (ID: GNT1047648).


  • Andreasen, 1999 N.C. Andreasen. A unitary model of schizophrenia: Bleuler's “fragmented phrene” as schizencephaly. Arch. Gen. Psychiatry. 1999;56(9):781-787 Crossref
  • Bach et al., 2014 M. Bach, F.B. Laun, A. Leemans, C.M. Tax, G.J. Biessels, B. Stieltjes, K.H. Maier-Hein. Methodological considerations on tract-based spatial statistics (TBSS). Neuroimage. 2014;100:358-369
  • Bleich-Cohen et al., 2012 M. Bleich-Cohen, M. Kupchik, M. Gruberger, M. Kotler, T. Hendler. Never resting region–mPFC in schizophrenia. Schizophr. Res.. 2012;140(1–3):155-158 Crossref
  • Bloemen et al., 2010 O.J. Bloemen, M.B. de Koning, N. Schmitz, D.H. Nieman, H.E. Becker, L. de Haan, P. Dingemans, D.H. Linszen, T.A. van Amelsvoort. White-matter markers for psychosis in a prospective ultra-high-risk cohort. Psychol. Med.. 2010;40(8):1297-1304 Crossref
  • Bora et al., 2011 E. Bora, A. Fornito, J. Radua, M. Walterfang, M. Seal, S.J. Wood, M. Yucel, D. Velakoulis, C. Pantelis. Neuroanatomical abnormalities in schizophrenia: a multimodal voxelwise meta-analysis and meta-regression analysis. Schizophr. Res.. 2011;127(1–3):46-57 Crossref
  • Bora et al., 2012 E. Bora, A. Fornito, M. Yücel, C. Pantelis. The effects of gender on grey matter abnormalities in major psychoses: a comparative voxelwise meta-analysis of schizophrenia and bipolar disorder. Psychol. Med.. 2012;42(2):295-307 Crossref
  • Carletti et al., 2012 F. Carletti, J.B. Woolley, S. Bhattacharyya, R. Perez-Iglesias, P. Fusar Poli, L. Valmaggia, M.R. Broome, E. Bramon, L. Johns, V. Giampietro, S.C. Williams, G.J. Barker, P.K. McGuire. Alterations in white matter evident before the onset of psychosis. Schizophr. Bull.. 2012;38(6):1170-1179 Crossref
  • Cook, 2002 N.D. Cook. The Speciation of Modern Homo Sapiens. T.J. Crow (Ed.) (Oxford University Press, 2002) 169-194
  • Crow, 1998 T.J. Crow. Schizophrenia as a transcallosal misconnection syndrome. Schizophr. Res.. 1998;30(2):111-114 Crossref
  • Crow et al., 2007 T.J. Crow, P. Paez, S.A. Chance. Callosal misconnectivity and the sex difference in psychosis. Int. Rev. Psychiatry. 2007;19(4):449-457 Crossref
  • Crow et al., 2013 T.J. Crow, S.A. Chance, T.H. Priddle, J. Radua, A.C. James. Laterality interacts with sex across the schizophrenia/bipolarity continuum: an interpretation of meta-analyses of structural MRI. Psychiatry Res.. 2013;210(3):1232-1244 Crossref
  • de Lacoste et al., 1985 M.C. de Lacoste, J.B. Kirkpatrick, E.D. Ross. Topography of the human corpus callosum. J. Neuropathol. Exp. Neurol.. 1985;44(6):578-591 Crossref
  • Fornito et al., 2006 A. Fornito, M. Yücel, S.J. Wood, T. Proffitt, P.D. McGorry, D. Velakoulis, C. Pantelis. Morphology of the paracingulate sulcus and executive cognition in schizophrenia. Schizophr. Res.. 2006;88(1–3):192-197 Crossref
  • Friston, 1998 K.J. Friston. The disconnection hypothesis. Schizophr. Res.. 1998;30(2):115-125 Crossref
  • Highley et al., 1999 J.R. Highley, M.M. Esiri, B. McDonald, M. Cortina-Borja, B.M. Herron, T.J. Crow. The size and fibre composition of the corpus callosum with respect to gender and schizophrenia: a post-mortem study. Brain. 1999;122(1):99-110 Crossref
  • Keihaninejad et al., 2012 S. Keihaninejad, N.S. Ryan, I.B. Malone, M. Modat, D. Cash, G.R. Ridgway, H. Zhang, N.C. Fox, S. Ourselin. The importance of group-wise registration in tract based spatial statistics study of neurodegeneration: a simulation study in Alzheimer's disease. PLoS ONE. 2012;7(11):e45996 Crossref
  • Knochel et al., 2012 C. Knochel, V. Oertel-Knochel, R. Schonmeyer, A. Rotarska-Jagiela, V. van de Ven, D. Prvulovic, C. Haenschel, P. Uhlhaas, J. Pantel, H. Hampel, D.E. Linden. Interhemispheric hypoconnectivity in schizophrenia: fiber integrity and volume differences of the corpus callosum in patients and unaffected relatives. NeuroImage. 2012;59(2):926-934
  • Lagopoulos et al., 2013 J. Lagopoulos, D.F. Hermens, S.N. Hatton, J. Tobias-Webb, K. Griffiths, S.L. Naismith, E.M. Scott, I.B. Hickie. Microstructural white matter changes in the corpus callosum of young people with Bipolar Disorder: a diffusion tensor imaging study. PLoS ONE. 2013;8(3):e59108 Crossref
  • Miller et al., 2003 T.J. Miller, T.H. McGlashan, J.L. Rosen, K. Cadenhead, T. Cannon, J. Ventura, W. McFarlane, D.O. Perkins, G.D. Pearlson, S.W. Woods. Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: predictive validity, interrater reliability, and training to reliability. Schizophr. Bull.. 2003;29(4):703-715 Crossref
  • Mizuno et al., 2009 M. Mizuno, M. Suzuki, K. Matsumoto, M. Murakami, K. Takeshi, T. Miyakoshi, F. Ito, R. Yamazawa, H. Kobayashi, T. Nemoto, M. Kurachi. Clinical practice and research activities for early psychiatric intervention at Japanese leading centers. Early Interv. Psychiatry. 2009;3(1):5-9 Crossref
  • Nakamura et al., 2012 K. Nakamura, Y. Kawasaki, T. Takahashi, A. Furuichi, K. Noguchi, H. Seto, M. Suzuki. Reduced white matter fractional anisotropy and clinical symptoms in schizophrenia: a voxel-based diffusion tensor imaging study. Psychiatry Res.. 2012;202(3):233-238 Crossref
  • Nemoto et al., 2012 T. Nemoto, T. Funatogawa, K. Takeshi, M. Tobe, T. Yamaguchi, K. Morita, N. Katagiri, N. Tsujino, M. Mizuno. Clinical practice at a multi-dimensional treatment center for individuals with early psychosis in Japan. East Asian Arch. Psychiatry. 2012;22(3):110-113
  • Pantelis and Bartholomeusz, 2014 C. Pantelis, C.F. Bartholomeusz. Social neuroscience in psychiatry: pathways to discovering neurobiological risk and resilience. World Psychiatry. 2014;13(2):146-147 Crossref
  • Pantelis et al., 2003 C. Pantelis, D. Velakoulis, P.D. McGorry, S.J. Wood, J. Suckling, L.J. Phillips, A.R. Yung, E.T. Bullmore, W. Brewer, B. Soulsby, P. Desmond, P.K. McGuire. Neuroanatomical abnormalities before and after onset of psychosis: a cross-sectional and longitudinal MRI comparison. Lancet. 2003;361(9354):281-288 Crossref
  • Pantelis et al., 2005 C. Pantelis, M. Yücel, S.J. Wood, D. Velakoulis, D. Sun, G. Berger, G.W. Stuart, A. Yung, L. Phillips, P.D. McGorry. Structural brain imaging evidence for multiple pathological processes at different stages of brain development in schizophrenia. Schizophr. Bull.. 2005;31(3):672-696 Crossref
  • Patel et al., 2011 S. Patel, K. Mahon, R. Wellington, J. Zhang, W. Chaplin, P.R. Szeszko. A meta-analysis of diffusion tensor imaging studies of the corpus callosum in schizophrenia. Schizophr. Res.. 2011;129(2–3):149-155 Crossref
  • Penades et al., 2013 R. Penades, N. Pujol, R. Catalan, G. Massana, G. Rametti, C. Garcia-Rizo, N. Bargallo, C. Gasto, M. Bernardo, C. Junque. Brain effects of cognitive remediation therapy in schizophrenia: a structural and functional neuroimaging study. Biol. Psychiatry. 2013;73(10):1015-1023 Crossref
  • Peters et al., 2010 B.D. Peters, P.M. Dingemans, N. Dekker, J. Blaas, E. Akkerman, T.A. van Amelsvoort, C.B. Majoie, G.J. den Heeten, D.H. Linszen, L. de Haan. White matter connectivity and psychosis in ultra-high-risk subjects: a diffusion tensor fiber tracking study. Psychiatry Res.. 2010;181(1):44-50 Crossref
  • Rueckert et al., 1999 D. Rueckert, L.I. Sonoda, C. Hayes, D.L. Hill, M.O. Leach, D.J. Hawkes. Nonrigid registration using free-form deformations: application to breast MR images. IEEE Trans. Med. Imaging. 1999;18(8):712-721 Crossref
  • Saito et al., in submission Saito, J., Hori, M., Nemoto, T., Katagiri, N., Shimoji, K., Ito, S., Fukunaga, I., Tsujino, N., Yamaguchi, T., Shiraga, N., Aoki, S., Mizuno, M., Abnormal white matter integrity in the corpus callosum and its correlation with psychiatric symptoms in individuals with an at-risk mental state for psychosis: A longitudinal study using Tract Specific Analysis. (in submission).
  • Savadjiev et al., 2014 P. Savadjiev, T.J. Whitford, M.E. Hough, C. Clemm von Hohenberg, S. Bouix, C.F. Westin, M.E. Shenton, T.J. Crow, A.C. James, M. Kubicki. Sexually dimorphic white matter geometry abnormalities in adolescent onset schizophrenia. Cereb. Cortex. 2014;24(5):1389-1396 Crossref
  • Schneiderman et al., 2009 J.S. Schneiderman, M.S. Buchsbaum, M.M. Haznedar, E.A. Hazlett, A.M. Brickman, L. Shihabuddin, J.G. Brand, Y. Torosjan, R.E. Newmark, E.L. Canfield, C. Tang, J. Aronowitz, R. Paul-Odouard, P.R. Hof. Age and diffusion tensor anisotropy in adolescent and adult patients with schizophrenia. Neuroimage. 2009;45(3):662-671 Crossref
  • Serpa et al., 2012 M.H. Serpa, M.S. Schaufelberger, P.G. Rosa, F.L. Duran, L.C. Santos, R.M. Muray, M. Scazufca, P.R. Menezes, G.F. Busatto. Corpus callosum volumes in recent-onset schizophrenia are correlated to positive symptom severity after 1 year of follow-up. Schizophr. Res.. 2012;137(1–3):258-259
  • Shenton et al., 2001 M.E. Shenton, C.C. Dickey, M. Frumin, R.W. McCarley. A review of MRI findings in schizophrenia. Schizophr. Res.. 2001;49(1–2):1-52 Crossref
  • Smith et al., 2006 S.M. Smith, M. Jenkinson, H. Johansen-Berg, D. Rueckert, T.E. Nichols, C.E. Mackay, K.E. Watkins, O. Ciccarelli, M.Z. Cader, P.M. Matthews, T.E. Behrens. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. NeuroImage. 2006;31(4):1487-1505 Crossref
  • Takeuchi et al., 2010 H. Takeuchi, A. Sekiguchi, Y. Taki, S. Yokoyama, Y. Yomogida, N. Komuro, T. Yamanouchi, S. Suzuki, R. Kawashima. Training of working memory impacts structural connectivity. J. Neurosci. Off. J. Soc. Neurosci.. 2010;30(9):3297-3303 Crossref
  • Walterfang et al., 2006 M. Walterfang, S.J. Wood, D. Velakoulis, C. Pantelis. Neuropathological, neurogenetic and neuroimaging evidence for white matter pathology in schizophrenia. Neurosci. Biobehav. Rev.. 2006;30(7):918-948 Crossref
  • Walterfang et al., 2008 M. Walterfang, P.K. McGuire, A.R. Yung, L.J. Phillips, D. Velakoulis, S.J. Wood, J. Suckling, E.T. Bullmore, W. Brewer, B. Soulsby, P. Desmond, P.D. McGorry, C. Pantelis. White matter volume changes in people who develop psychosis. Br. J. Psychiatry. 2008;193(3):210-215 Crossref
  • Walterfang et al., 2011 M. Walterfang, D. Velakoulis, T.J. Whitford, C. Pantelis. Understanding aberrant white matter development in schizophrenia: an avenue for therapy?. Expert. Rev. Neurother.. 2011;11(7):971-987 Crossref
  • Woods, 2003 S.W. Woods. Chlorpromazine equivalent doses for the newer atypical antipsychotics. J. Clin. Psychiatry. 2003;64(6):663-667 Crossref
  • Yücel et al., 2002 M. Yücel, G.W. Stuart, P. Maruff, S.J. Wood, G.R. Savage, D.J. Smith, S.F. Crowe, D.L. Copolov, D. Velakoulis, C. Pantelis. Paracingulate morphologic differences in males with established schizophrenia: a magnetic resonance imaging morphometric study. Biol. Psychiatry. 2002;52(1):15-23
  • Yung and McGorry, 1996 A.R. Yung, P.D. McGorry. The prodromal phase of first-episode psychosis: past and current conceptualizations. Schizophr. Bull.. 1996;22(2):353-370 Crossref
  • Yung et al., 2007 A.R. Yung, H.P. Yuen, G. Berger, S. Francey, T.C. Hung, B. Nelson, L. Phillips, P. McGorry. Declining transition rate in ultra high risk (prodromal) services: dilution or reduction of risk?. Schizophr. Bull.. 2007;33(3):673-681 Crossref
  • Zalesky, 2011 A. Zalesky. Moderating registration misalignment in voxelwise comparisons of DTI data: a performance evaluation of skeleton projection. Magn. Reson. Imaging. 2011;29(1):111-125 Crossref
  • Zalesky et al., 2011 A. Zalesky, A. Fornito, M.L. Seal, L. Cocchi, C.F. Westin, E.T. Bullmore, G.F. Egan, C. Pantelis. Disrupted axonal fiber connectivity in schizophrenia. Biol. Psychiatry. 2011;69(1):80-89 Crossref


a Department of Neuropsychiatry, Toho University School of Medicine, Tokyo, Japan

b Melbourne Neuropsychiatry Centre, Department of Psychiatry, University of Melbourne and Melbourne Health, Parkville, Melbourne, Australia

c Melbourne School of Engineering, The University of Melbourne, Melbourne, Australia

d Department of Radiology, Juntendo University Graduate School of Medicine, Tokyo, Japan

e Department of Diagnostic Radiology, Tokyo Metropolitan Geriatric Hospital, Tokyo, Japan

f Department of Social Medicine, Toho University School of Medicine, Tokyo, Japan

g Department of Radiology, Toho University School of Medicine, Tokyo, Japan

lowast Corresponding author at: Department of Neuropsychiatry, Toho University School of Medicine, 6-11-1 Omori-nishi, Ota-ku, 143-8541 Tokyo, Japan. Tel.: + 81 3 3762 4151; fax: + 81 3 5471 5774.