Abnormal white matter integrity in antipsychotic-naïve first-episode psychosis patients assessed by a DTI principal component analysis

Abstract

Introduction

Diffusion tensor imaging (DTI) studies in patients with schizophrenia have shown abnormalities in the microstructure of white matter tracts. Specifically, reduced fractional anisotropy (FA) has been described across multiple white matter tracts, in studies that have mainly included patients treated with antipsychotic medications.

Objective

To compare FA in antipsychotic-naïve patients experiencing a first episode of psychosis (FEP) to FA in healthy controls to demonstrate that the variance of FA can be grouped, in a coincidental manner, in four predetermined factors in accordance with a theoretical partition of the white matter tracts, using a principal components analysis (PCA).

Methods

Thirty-five antipsychotic-naïve FEP patients and 35 age- and gender-matched healthy controls underwent DTI at 3 T. Analysis was performed using a tract-based spatial statistics (TBSS) method and exploratory PCA.

Results

DTI analysis showed extensive FA reduction in white matter tracts in FEP patients compared with the control group. The PCA grouped the white matter tracts into four factors explaining 66% of the total variance. Comparison of the FA values within each factor highlighted the differences between FEP patients and controls.

Discussion

Our study confirms extensive white matter tracts anomalies in patients with schizophrenia, more specifically, in drug-naïve FEP patients. The results also indicate that a small number of white matter tracts share common FA anomalies that relate to deficit symptoms in FEP patients. Our study adds to a growing body of literature emphasizing the need for treatments targeting white matter function and structure in FEP patients.

Introduction

Schizophrenia is a neurodevelopmental disorder in which alterations in brain connectivity have been observed (Friston and Frith, 1995, McGuire and Frith, 1996, Olabi et al., 2011), leading to ineffective communication between brain regions (Seal et al., 2008, Shi et al., 2012). These disturbances in connectivity may be caused by the interaction of biological (Weinberger, 1987, Church et al., 2002, Insel, 2010) and environmental factors during early life (Cannon et al., 2002, Tandon et al., 2008, van Haren et al., 2012), causing developmental brain alterations that predispose an individual to suffer schizophrenia (Weinberger, 1987, Singh et al., 2012).

The changes in microstructure of specific white matter tracts seem to be the anatomical substrate for the disturbances of brain connectivity in schizophrenia (Konrad and Winterer, 2008). This conclusion has been supported by postmortem and genetic studies showing myelin-related alterations (Yurgelun-Todd et al., 1996, Shergill et al., 2000, Flynn et al., 2003), as well as functional imaging studies that demonstrated connectivity abnormalities, specifically between temporo-limbic and prefrontal regions (Yu et al., 2011, Meda et al., 2012).

Diffusion tensor imaging (DTI) is a magnetic resonance imaging (MRI) technique that yields information about white matter microstructure. DTI measures the direction and degree of water diffusion that can be altered by pathologic factors such as demyelination and axonal membrane deterioration (Kubicki et al., 2005a). Fractional anisotropy (FA) is a commonly reported index reflecting white matter microstructure (Basser and Pierpaoli, 1996, Beaulieu, 2002). Several DTI studies have revealed abnormalities in the microstructure of white matter as evidenced by reduced FA in patients with schizophrenia (Kanaan et al., 2005, Kubicki et al., 2005b, Ellison-Wright and Bullmore, 2009, White et al., 2013). In a recent systematic review (Peters and Karlsgodt, 2015) conflicting results were found: some studies have shown no differences between first-episode psychosis (FEP) patients and healthy individuals (Peters et al., 2010, Kong et al., 2011, Collinson et al., 2014), while others have described white matter abnormalities (Cheung et al., 2008, Luck et al., 2011, Guo et al., 2012, Lee et al., 2012, Mandl et al., 2013, Wang et al., 2013a, Fitzsimmons et al., 2014). Some of the factors that might contribute to these results are variability in medication status, symptom history, image analysis, among others (Peters and Karlsgodt, 2015). In order to avoid some confounding variables, studies in antipsychotic-naïve FEP patients are of particular interest. Studies in patients never treated with antipsychotics have shown abnormalities in inferior longitudinal fasciculus (Cheung et al., 2008, Liu et al., 2013), superior longitudinal fasciculus (Guo et al., 2012, Mandl et al., 2013, Filippi et al., 2014), cingulum (Wang et al., 2013b), fornix (Guo et al., 2012, Filippi et al., 2014), internal capsule (Guo et al., 2012), uncinate fasciculus (Mandl et al., 2013), genu and splenium (Cheung et al., 2008, Gasparotti et al., 2009), and occipital-frontal fasciculus (Cheung et al., 2008, Liu et al., 2013).

As presented, some concordance exists in some of the white matter tracts in studies of FEP patients; however, definitive regions have not been confirmed (Wang et al., 2013a). These results could be mediated by heterogeneity of the disease, different methods of image analysis, or statistical analysis. In our study, we examine white matter microstructure in antipsychotic-naïve FEP patients, using DTI with a tract-based spatial statistics (TBSS) method (Smith et al., 2006) and principal component analysis (PCA). We chose this method because region of interest analysis is likely to have more Type II errors (false negatives) than TBSS analysis (Pasternak et al., 2012), and because TBSS also takes into account non-normal distributions of FA in certain brain regions (Smith et al., 2007).