Micro- and Macrostructural White Matter Integrity in Never-treated and Currently Unmedicated Patients with Schizophrenia and Effects of Short Term Antipsychotic Treatment

Abstract

Background:

Schizophrenia is associated with progressive white matter changes, but it is unclear if antipsychotic medications contribute to these. Our objective was to characterize effects of short-term treatment with risperidone on white matter diffusion indices.

Methods:

We recruited 42 schizophrenia patients (30 never-treated, 12 currently untreated) and 42 matched healthy controls in this prospective case-control neuroimaging study. Patients received a six week trial of risperidone. Using diffusion tensor imaging, we assessed microstructural (fractional anisotropy [FA], mean diffusivity [MD], and radial diffusivity [RD]) and macrostructural (radial fiber trophy) white matter integrity deficits in unmedicated patients compared to controls and change in white matter integrity in patients before and after antipsychotic treatment (mean risperidone dose at endpoint was 3.73+/−1.72mg).

Results:

At baseline, FA was decreased in the left medial temporal white matter (cluster extent: 123 voxels, peak: x= −51; y= −44; z= −7; α< .05) and MD was increased in the fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum (cluster extent 185 voxels, peak: x= −27; y= −49; z= 2; α< .04) in patients compared to controls. RD and macrostructure were not abnormal. None of the diffusion indices showed a significant change after six weeks of treatment, both with voxel-wise and whole brain white matter analyses.

Conclusions:

We demonstrate microstructural white matter integrity abnormalities in the absence of macrostructural impairment in unmedicated patients with, primarily early stage, schizophrenia. In our data, we found no significant white matter changes after short-term treatment with risperidone.

INTRODUCTION

Schizophrenia is a complex neuropsychiatric syndrome associated with subtle changes in brain structure that progress over the course of the illness (1, 2). While causes underlying this phenomenon remain poorly understood, a number of mechanisms including accelerated aging (1), glutamate related excitotoxicity (3, 4), and a variety of environmental influences are being contemplated (5). Antipsychotic medication exposure is probably the most widely debated putative environmental factor (6, 7). However, only a limited number of studies have directly investigated antipsychotic medication effects on gray matter volume, and even fewer have examined effects on white matter integrity, perhaps because imaging sequences optimized to examine white matter have only more recently become widely available.

With diffusion tensor imaging (DTI) scalar diffusion indices can be derived to describe white matter microstructural properties, and to characterize pathological processes in the microstructural composition white matter (8, 9). Fractional anisotropy (FA) measures the degree of orientational coherence of water diffusion within a voxel and is a non-specific biomarker of microstructural architecture and neuropathology, whereas mean diffusivity (MD) measures the extent to which water is able to diffuse into any direction within the voxel and is typically elevated in inflammatory states and with edema (10), whereas cell proliferation and neoplasia can decrease MD (8). Radial diffusivity (RD) quantifies diffusion perpendicular to the principal diffusion direction (i.e. the axon) and is thought to be increased in dysmyelination and demyelination (11). However, because changes in fiber diameter and tract spacing may also affect these measures, it is important to interpret microstructural alterations in the context of macrostructure (i.e. radial fiber trophy, a surrogate marker of white matter atrophy/ hypertrophy) (12).

While there is a general theme of decreased FA in schizophrenia (13, 14) and evidence of increased MD and RD (15, 16), findings in the literature are variable. Similarly, prospective studies investigating antipsychotic effects on white matter integrity have been inconsistent. Szeszko and colleagues report FA reduction and increased RD following twelve weeks of treatment with risperidone or aripiprazole in minimally treated first episode psychosis patients (17). Another group reported increased FA, decreased MD and RD, but unchanged white matter volume after twelve weeks of first or second generation antipsychotic treatment in first episode psychosis patients that were medicated at the time of enrollment (18). Wang and colleagues reported a significant decrease in absolute FA of white matter around the anterior cingulate gyrus and corona radiata in drug-naïve first episode patients compared to healthy controls after six weeks of treatment with various clinician selected antipsychotic medications (19). Similarly, Meng and colleagues found a widespread decrease in FA after six weeks of treatment with different second generation antipsychotic medications, in this study the medication dose expressed in chlorpromazine equivalents was correlated the extent of FA reduction during the treatment trial (20). Yet others found no medication related changes in FA or MD after eight weeks of treatment with different second generation antipsychotics (21), and in FA after various timeframes of treatment with a number of medications (22). The within-subject design in longitudinal studies controls for a number of subject level confounding factors, but it is possible that discrepancies in findings may be secondary to differential effects of a variety of antipsychotic drugs and to differences in data analysis and quality control techniques mitigating DTI measurement error.

The goal of this study was to characterize the effects of short term treatment with a single antipsychotic medication on white matter diffusion indices. We scanned a group of patients with schizophrenia and schizoaffective disorder (SZ) before and after six weeks of treatment with risperidone and scanned a group of matched healthy controls (HC) twice, approximately six weeks apart. We examined if (1) FA would be decreased, but MD and RD would be increased in never-treated and currently unmedicated SZ compared to HC, (2) if macrostructural white matter integrity would differ between HC and never-treated and currently unmedicated SZ, and (3) if six weeks of treatment with risperidone would affect micro- and macrostrcutural white matter integrity. In addition, we tested the hypothesis that greater duration of untreated psychosis (DUP) would be associated with poorer microstructural white matter integrity.

METHODS AND MATERIALS

Participants and study design

SZ were recruited from outpatient clinics, inpatient units, and the emergency room at the University of Alabama at Birmingham (UAB). HC matched on age, sex, and parental occupation were recruited by advertisements. Written informed consent was obtained prior to enrollment in this UAB Institutional Review Board approved study.

Diagnoses were established by review of medical records, the Diagnostic Interview for Genetic Studies (DIGS), and consensus of two board certified psychiatrists (ACL and NVK). The Brief Psychiatric Rating Scale (BPRS)(23) and Repeatable Battery for The Assessment of Neuropsychological Status (RBANS)(24) were used to assess symptom severity and cognition, respectively. Participants were excluded if they had major neurological or medical conditions, history of head trauma with loss of consciousness, substance use disorders (excluding nicotine) within six months of imaging, were prescribed medications known to affect brain function, were pregnant or breastfeeding, or had MRI contraindications. Controls with a personal or family history in a first-degree relative of an Axis I disorder were excluded.

We enrolled 42 SZ who were medication naïve or off antipsychotic medications for at least two weeks (determined by self report; patients were not taken off medications to meet this criterion) in a six-week trial of oral risperidone using a flexible dosing regimen. We chose this medication because it is commonly prescribed, now available as generic medication and thus one of the more affordable second generation antipsychotic medications in the US, and is considered first line treatment in schizophrenia, specifically first episode patients (25). Scans were obtained prior to treatment and after six weeks of treatment. Nine SZ dropped out before study completion. Medication was managed by ACL and NVK. Risperidone was started at 1–3 milligrams and titrated in 1–2 milligram increments; dosing was based on therapeutic and side effects. Use of concomitant medications was permitted as clinically indicated. Fourteen patients were prescribed benztropine, two trazodone, two fluoxetine, one sertraline, one desvenlafaxine, one mirtazapine, one amitriptyline, and one clonazepam. Compliance was monitored with pill counts at each visit. We also enrolled 42 HC, 38 of those were scanned twice, on average 51.5 +/− 19.8 days apart.

Data from 5 HC (one time point), and 14 SZ baseline as well as 5 SZ endpoint scans were included in a prior publication (16). No longitudinal DTI data has been included in prior publications.

Data acquisition

Imaging was performed on a 3 T head-only MRI scanner (Siemens Magnetom Allegra, Erlangen, Germany) using a circularly polarized transmit/receive head coil. A T1 weighted scan was acquired for anatomical reference (MPRAGE, TR/TE/TI= 2300/3.93/1100ms, flip angle 12°, phase encoding direction A>P, field of view= 256× 256mm, 1mm isotropic voxel size). Two diffusion-weighted runs were acquired, each non-collinearly distributed along 30 directions [b= 1000 s/mm², TR/TE=9200/96ms, field of view=246× 246mm, matrix= 112× 112, 60 slices, interleaved acquisition, 2.2mm isotropic voxel size]. Both runs were concatenated to increase the signal-to-noise ratio (for three SZ datasets only one run was acquired). Five images with no diffusion gradients (b0; b= 0 s/mm²) were also acquired.

Data processing

In preparation for DTI registration in TORTOISE (version 2.5.2)(26), the T1 weighted images were skull stripped and processed through FATCAT in AFNI (27, 28) to create a volume with an approximate T2-weighted contrast. This “imitation T2” image was spatially aligned with the DTI images and used solely for providing an anatomical reference with requisite contrast similar to the DTI b0 volume (29).

After visual inspection of raw images, we calculated bulk motion indices. The root-mean-square (RMS) of the 6 motion parameters (translations and rotations) was calculated both for absolute (RMS_abs; average of absolute displacement of each image and the first b0 image) and relative (RMS_rel; average of absolute displacement between adjacent images, which reduces the likelihood of a few large movements biasing the outcome of motion estimates (30)) movement(31). Datasets with RMS_abs of greater than half the voxel edge length (1.1mm) or RMS_rel of >0.05mm per run were excluded from further analyses. One SZ baseline and one SZ week six datasets were rejected based on these criteria.

Artifacts due to between-volume bulk body motion, eddy currents, and EPI susceptibility-induced geometric distortions were removed using DIFF_PREP with a single interpolation step (32). For each dataset, the first b0 image was selected as the reference for registration; subsequent b0 images were affine registered to the initial b0 image. Prior to registration, diffusion weighted and structural images were upsampled at a factor two using a bicubic algorithm and smoothed with a Perona-Malik anisotropic edge favoring gradient based filter (33) to compute the transformations from moving to fixed images. After computation of transformations, original images were used to create the registered images. Bspline correction was done using the imitation T2 image. Diffusion images were resampled to 1.5mm isotropic voxels. Gradient tables were rotated along with motion correction (34, 35). Tensors were computed with DIFF_CALC using a linear fitting algorithm. We computed maps for FA, MD, and RD. Resulting maps were visually inspected for anomalies in scalar diffusion parameters. After quality control, a total of ten datasets were excluded from further analysis (one HC and three SZ baseline as well as two HC and four SZ week 6 scans); in addition five datasets were processed without imitation T2 images.

To spatially normalize diffusion images to the Illinois Institute of Technology atlas (IIT2) space (36), we implemented an optimized non-linear image registration procedure using a modified version of 3dQwarp in AFNI (37). The warping optimization implements an iterative refinement, where an input image is repeatedly processed through an optimizer in smaller and smaller patches, incorporating convergence criteria at each level to better resolve artifacts, with a final patch size of 5× 5× 5 millimeters. Resulting deformation maps were then used together with diffusion images to calculate radial fiber trophy, a surrogate marker of macrostructural white matter change (12). Briefly, stretch of the deformation field in the plane perpendicular to the predominant direction of diffusion is computed using slight modifications of traditional tensor based morphometry, a method for identifying macroscopic differences in brain structure based on the Jacobian determinant of the deformation matrix (38, 39). Before statistical analyses, a 5mm median filter was applied with 3dLocalstat to improve the signal to noise ratio while maintaining edges (40).

Statistical analyses

HC were considered negative controls, facilitating descriptive characterization of abnormalities in SZ and assessing temporal stability of diffusion indices. To examine whole brain voxel-wise group differences in scalar diffusion indices and atrophy measures within a IIT2 based white matter mask we used AFNI’s 3dttest++; a priori defined co-variates included age and gender (15, 41). Because baseline head motion differed between groups (RMSabs baseline: HC 0.56+/− 0.10mm; SZ 0.62+/− 0.14mm; t= −2.18; p= 0.03; RMS_abs week 6: HC 0.56+/− 0.10mm; SZ 0.62+/− 0.16mm; t= −1.83; p= .09; RMS_rel baseline: HC 0.14+/− 0.002mm; SZ 0.17+/− 0.005mm; t= −3.24; p= .002; RMS_rel week 6: HC 0.15+/− 0.03mm; SZ 0.16+/− 0.05mm; t= −1.33; p= .19) we included RMSrel as additional covariate in analyses. To examine change over time in voxelwise analyses, we used AFNI’s 3dttest++ to compare diffusion indices at baseline and after six weeks in each group separately. Clustsim, an algorithm that uses randomization/permutation simulation to produce 10,000 iterations of noise only generated t-tests and determine global cluster-level threshold values, was implemented to control for the false positive rate (voxelwise threshold Bonferroni corrected p= .0008 [accounting for 12 comparisons at p_uncorrected= .01]; cluster threshold α= 0.05) (42). As a post hoc analysis, we extracted whole brain white matter scalar diffusion indices for each participant and compared baseline group differences and changes over time in each group using linear models, with age, gender and RMS_rel included as covariates. We also examined the relationship between DUP (in patients who were antipsychotic medication-naïve) and voxelwise white matter integrity in the subgroup of medication-naïve patients with 3dttest++ including the log transformed DUP as regressor. In an exploratory fashion we also examined partial correlations between clinical measures in SZ and whole brain white matter scalar diffusion indices as well as diffusion indices in areas of abnormal white matter integrity, controlling for age, gender and RMS_rel.

A schematic of our data analysis pipeline can be found in Figure 1.

Figure 1:

Schematic of the diffusion data processing pipeline. Major steps in preprocessing, image registration and statistical analyses are depicted here. The blue color signifies software programs used/ data processing steps applied, the green color signifies quality control steps. Abbreviations: DWI, diffusion weighted image; RMS_abs and RMS_rel, absolute and relative head motion; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity

Final analyses included good quality datasets from 41 HC and 38 SZ at baseline, and 36 HC and 28 SZ at week 6. A total of 36 HC and 27 SZ had good quality datasets for both baseline and week 6 scans.

RESULTS

HC and SZ did not differ in gender, age, or parental socioeconomic status (Table 1). BPRS scores significantly decreased after six weeks of treatment.

Table 1:

Demographics and clinical measures^a

	SZ (n=42)	HC (n=42)	t/X2/F	p value
Gender (% male)	61.9	61.9	0.00	1.0
Age	26.62 (9.00)	27.88 (9.43)	0.63	.53
Parental Occupationb	7.16 (6.13)	5.45 (4.18)	16.28	.30
Smoking (packs per day)	0.38 (0.53)	0.23 (0.43)	−1.47	.15
Diagnosis
Schizophrenia	34
Schizoaffective Disorder	5
Brief Psychotic Disorder	3
Illness Duration (years; mean, SD, median, range)c	15.00 (8.67; 18.00; 1–25)
Prior antipsychotic treatment
Antipsychotic naive (yes/ no)	30/ 12
Antipsychotic free interval (months; mean, SD, median, range)c	17.33; (36.91; 3.25; 0.5–120)
Risperidone dose at week 6 (in mg; mean, SD, median, range)	3.73 (1.72; 4.0; 1–8)
BPRSd
Total
Baseline	50.26 (9.30)
Week 6e	32.36 (9.60)g
Positive
Baseline	10.93 (3.46)
Week 6e	5.24 (2.55)h
Negative
Baseline	7.19 (3.36)
Week 6e	5.73 (2.41)i
RBANSf
Total index	71.82 (14.37)	89.67 (12.73)	5.89	< .01
Immediate memory	77.42 (16.35)	97.07 (15.33)	5.55	< .01
Visuospatial	73.24 (16.52)	82.48 (15.41)	2.59	.01
Language	85.50 (12.82)	96.64 (14.04)	3.76	< .01
Attention span	80.39 (19.67)	94.07 (18.45)	3.21	< .01
Delayed memory	71.79 (20.57)	91.38 (10.13)	5.32	< .01

SZ, schizophrenia; HC, healthy control

^aMean (SD) unless indicated otherwise

^bRanks determined from Diagnostic Interview for Genetic Studies (1 – 18 scale); higher rank (lower numerical value) corresponds to higher socioeconomic status

^cincludes only patients who are not antipsychotic naïve (n=12), illness duration since first diagnosis

^dBrief Psychiatric Rating Scale (1 – 7 scale); positive (conceptual disorganization, hallucinatory behavior, and unusual thought content); negative (emotional withdrawal, motor retardation, and blunted affect)

^en= 33

^fRepeatable Battery for the Assessment of Neuropsychological Status

^gt= 11.19; p< .01

^ht= 9.45; p< .01

ⁱt= 2.98; p< .01

White matter integrity in medication-naïve and unmedicated SZ compared to HC

At baseline, FA was reduced in the left medial temporal white matter and MD was increased in the left fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum in SZ compared to HC (Figure 2). RD and radial fiber trophy did not differ between groups, both when the entire SZ group was considered and when only antipsychotic medicationnaïve subjects were considered. Post hoc analysis did not show group differences in whole brain white matter scalar diffusion indices (Figure 3). We found no associations between DUP and FA, MD, or RD.

Figure 2:

White matter microstructural integrity abnormalities in never-treated and currently unmedicated patients with schizophrenia (SZ) compared to healthy controls (HC). Fractional anisotropy is decreased in the left medial temporal white matter (cluster extent: 123 voxels, peak: x= −51; y= −44; z= −7; α< .05) and mean diffusivity is increased in the fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum (cluster extent 185 voxels, peak: x= −27; y= −49; z= 2; α< .04) in SZ compared to HC. Clusters are projected on the IIT2 white matter atlas template. Numbers adjacent to slices indicate x, y, and z coordinates in Talairach convention. Left= Right. Color bar indicates z scores.

Figure 3:

White matter scalar diffusion indices in healthy controls (HC; n= 41) and never treated and currently unmedicated patients with schizophrenia (SZ; n=38). We observed no significant group differences in fractional anisotropy, mean diffusivity, or radial diffusivity. Dots represent individual measurements, and the horizontal lines represent the group mean.

In a secondary analysis, we repeated the same computations but included only patients who were antipsychotic naïve and their matched controls. Findings in this subgroup were largely consistent with those in the larger group, where FA was is decreased in the left medial temporal white matter (cluster extent: 80 voxels, peak: x= −51; y= −36; z= −14; α< .10) and MD was increased in the fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum (cluster extent 199 voxels, peak: x= −28; y= −49; z= 2; α< .04) in antipsychotic-naïve SZ compared to HC. No group differences were seen in RD.

Changes in white matter integrity after six weeks

In HC, none of the white matter measures changed over time. We also observed no changes in micro- or macrostructural white matter integrity after six weeks of treatment in SZ. This was true even when lowering the statistical threshold to p_uncorrected< .001. Post hoc analyses did not show significant change in whole brain white matter scalar diffusion indices after six weeks in either group (Table 2, Figure 4). When we examined change in diffusion indices over time in each group separately with paired-sample t-tests, we again did not find significant changes over time in HC (FA: t= −1.00; p= .32 MD: t= −0.96; p= .34 ; RD: t= − 0.01; p= .99) and after six weeks of treatment in SZ (FA: t= 0.21; p= .76; MD: 1.45; p= .16; RD: t= 0.14; p= .89).

Table 2:

Whole brain white matter microstructural integrity measurements^a

	SZ (n= 39)	HC (n= 41)	F	p value
Baselineb
FA	0.38 ± 0.02	0.39 ± 0.02	0.23	0.63
MD	0.83 ± 0.02	0.82 ± 0.03	1.72	0.19
RD	0.67 ± 0.04	0.65 ± 0.05	0.99	0.32
Change from Baseline to Week 6c
FA	0.001 ± 0.018	−0.002 ± 0.010	0.32	0.57
MD	0.004 ± 0.015	−0.002 ± 0.013	2.99	0.09
RD	0.001 ± 0.030	−0.000 ± 0.027	0.03	0.86

SZ, schizophrenia; HC, healthy control; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity.

^aMean (SD)

^bAll linear models controlled for age, gender, and relative movement of the participant at baseline.

^cAll linear models controlled for age, gender, and relative movement of the participant at both baseline and Week 6.

Figure 4:

Individual changes in whole brain white matter scalar diffusion indices in healthy controls (left column; n= 36) after six weeks and patients with schizophrenia (right column, n= 27) after six weeks of treatment with risperidone. We observed no significant changes in fractional anisotropy, mean diffusivity, or radial diffusivity.

Secondary analyses in the antipsychotic naïve sample again were consistent with those in the larger group, showing no significant change in voxelwise FA, MD and RD after six weeks of treatment with risperidone. This again was true even when lowering the statistical threshold to p_uncorrected< .001.

Relationship between white matter integrity and clinical variables

Exploratory analyses revealed a modest positive correlation between MD in left fusiform/ lingual gyrus white matter and BPRS positive symptoms in SZ at baseline (r= 0.34; p= 0.04 when age and RMS_rel were included as covariates; r= 0.36; p= 0.07 when age, RMS_rel and gender were included as covariates; Figure 5), but no relationship between whole brain scalar diffusion indices at baseline and clinical symptom severity. Risperidone dose was not associated with baseline diffusion indices or change in diffusion indices (FA, MD, RD). We also did not find correlations between whole brain diffusion indices (baseline and week 6) and RBANS total scores in either HC or SZ.

Figure 5:

Relationship between Mean Diffusivity in the left fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum bundle (cluster extent: 185 voxels, Montreal Neurological Institute (MNI) peak coordinates: x= −27; y= −49; z= 2) in never-treated and currently unmedicated patients with schizophrenia and Brief Psychiatric Rating Scale (BPRS) positive symptom severity.

DISCUSSION

The purpose of this longitudinal, prospective study was to examine whole brain micro-and macrostructural white matter integrity in schizophrenia and to study effects of short-term antipsychotic treatment on these measures. We report focal decreased FA and increased MD in never-treated and currently unmedicated SZ; baseline MD abnormalities showed a modest association with positive symptom severity. No macrostructural white matter or RD alterations were observed at baseline; the latter was inconsistent with our a priori hypotheses. In addition, we did not observe a relationship between DUP and white matter microstructure in medicationnaïve patients. While six weeks of treatment with risperidone did improve psychosis severity, we did not observe significant changes associated with short-term risperidone treatment in any of the white matter indices.

Our report of decreased FA localized in the left medial temporal white matter in a group of mostly early stage schizophrenia patients replicates findings of a recent activation likelihood estimation meta-analysis showing that abnormalities in the left temporal and the right frontal lobe in first episode psychosis patients (43), and is in agreement with a recent cross-sectional study suggesting that whole-brain white matter integrity deficits are only found later in the illness (1). MD elevations in the fusiform/ lingual gyrus white matter extending to the hippocampal part of the cingulum suggest a possible inflammatory process in this area. Consistent with this, a small Positron Emission Tomography study reported evidence of hippocampal neuro-inflammation accompanied by a non-significant increase of tracer binding potential in whole-brain white matter in patients recovering from a psychotic episode (44). Furthermore, post-mortem studies in white matter revealed increased density (45) and activation (46) of microglia, the resident innate immune cells of the brain.

We did discern a modest positive relationship between positive symptom severity and MD abnormalities in the lingual/ fusiform gyrus white matter extending to the hippocampal part of the cingulum bundle), but no such association was seen with FA deficits. Similarly, Filippi et al. report an (albeit negative) relationship between MD, but not FA, and positive symptom severity in antipsychotic naïve patients with schizophrenia (47). The lack of spatial overlap between decreased FA and increased MD and the dissimilar relationships between these indices and symptom severity in our study points towards a multifactorial contribution to white matter pathology in schizophrenia. Based on findings from a twin study (48) and a study in schizophrenia patients and first degree relatives, Clark et al. argue that reduced FA may be related to genetic vulnerability for schizophrenia, while elevated MD may more directly be related to the disease processes. Pasternak and colleagues propose neuro-inflammation and axonal degeneration as two likely mechanisms of white matter alterations, with inflammation rather than degeneration as the predominant pathology at illness onset (49). This is further supported by a recent study showing widespread extracellular free water elevations, suggestive of neuro-inflammation, but only limited FA abnormalities in minimally treated first episode patients (50).

We did not detect changes in white matter macrostructure associated with six weeks of risperidone treatment which is consistent with the only other diffusion imaging study that has examined macrostructural integrity in this context (18). Using T1 weighted imaging, Emsley et al. report no white matter volume changes after one year of treatment in a small group of antipsychotic naïve patients (51), but others suggest that white matter volume trajectories in chronic patients differ based on the treatment received (52). Interestingly, higher antipsychotic doses were associated with white matter loss and lower doses were associated with volume increase, even after accounting for illness severity (53). In a rare study of chronically ill, never-medicated patients, Xiao and colleagues found that these patients showed an accelerated and clinically relevant age-related reduction of FA in the genu of the corpus callosum compared to antipsychotic treated patients, suggesting a potential beneficial medication effect on white matter integrity (54).

Notably, we also observed no significant changes in microstructural diffusion indices after six weeks of treatment with risperidone at the voxel level and averaged across the whole brain. However, as seen in figure 3, results were variable, especially in FA. Assuming a 4% FA change (the average decrease in FA reported by Szeszko et al. (17)) to be meaningful, a subset of patients showed either an increase (four patients) or decrease (five patients) in FA after six weeks of treatment. These results could help reconcile discrepancies in findings of prior studies investigating short term antipsychotic medication effects on white matter microstructural integrity, by demonstrating heterogeneity in changes with measurements on the individual subject level. However, a definitive attribution of this heterogeneity in changes in diffusion indices to differential, possibly even dose related, medication effects rather than to divergence in disease progression between patients is impossible in human studies. Here, we did not observe any clinical correlates of change in diffusion images. It will be important for future large-scale studies to investigate the clinical correlates of this heterogeneity at the level of brain structural change in response to antipsychotic exposure in an effort to better inform our pathophysiological understanding of the illness. Another question that remains unanswered is if different antipsychotic medications affect white matter structure differentially. A recent exploratory study in a small sample of patients directly compared effects of risperidone and aripiprazole on gray matter and white matter volumes as well as neurometabolite levels and found a non-significant increase in prefrontal white matter volume in those treated with aripiprazole but not risperidone, suggesting that these drugs may have differential effects (55). However, these results need to be considered preliminary at this time, and larger scale studies making head to head comparisons of medication effects not only on macrostructural but also microstructural white matter integrity are direly needed in the field. Here, clozapine could be an especially interesting compound to compare to other drugs, as showed differential effects on FA compared to other typical antipsychotics and typical antipsychotics (56) and a small study reported a widespread increase in FA in patients that corresponded to improvement in semantic fluency after a twelve week trial of clozapine (57).

Our findings have to be seen in the context of a number of strengths and limitations. To minimize variance in data, we enrolled currently unmedicated SZ, many of them without any prior antipsychotic exposure, carefully matched groups on key demographic characteristics, and used a single antipsychotic medication in this prospective longitudinal study. We took a number of steps to assure data quality, chose a non-linear image registration algorithm, and carefully controlled for false positive rates with state of the art cluster thresholding methods (42). Ideally, we would have obtained a T2 weighted scan to aid co-registration, but were not able to do so due to time constraints. To assess for possible effects of time on diffusion indices, we included HC that were scanned twice six weeks apart. It was not possible to follow medication-free or placebo treated SZ longitudinally, as it is not ethically permissible to withhold known effective treatments from patients. Additionally, we did not have information available on prior nonresponse to risperidone in the 12 patients who had prior antipsychotic exposure, precluding us from assessing how medication history affected findings.

In summary, we demonstrate microstructural white matter integrity abnormalities in the absence of macrostructural impairment in never treated and currently unmedicated patients with schizophrenia. Our data also empirically support a multifactorial model contributing to white matter pathology in this phenotypically heterogeneous syndrome, but the underlying mechanisms remain elusive. Importantly, we found no significant white matter changes after short-term treatment with risperidone in our data.