ریزساختار بافت سفید در اسکیزوفرنی: انجمن هایی برای شناخت عصبی و علائم بالینی

Background Diffusion tensor imaging (DTI) studies in schizophrenia report widespread aberrations in brain white matter (WM). These appear related to poorer neurocognitive performance and higher levels of negative and positive symptomatology. However, identification of the most salient WM aberrations to neurocognition and clinical symptoms is limited by relatively small samples with divergent results. Methods We examined 53 well-characterized patients with schizophrenia and 62 healthy controls. All participants were administered a computerized neurocognitive battery, which evaluated performance in several domains. Patients were assessed for negative and positive symptoms. Fractional anisotropy (FA) of WM cortical regions and WM fiber tracts were compared across the groups. FA values were also used to predict neurocognitive performance and symptoms. Results We confirm widespread aberrant WM microstructure in a relatively large sample of well-characterized patients with schizophrenia in comparison to healthy participants. Moreover, we illustrate the utility of FA measures in predicting global neurocognitive performance in healthy participants and schizophrenia patients, especially for reaction time. FA was less predictive of clinical symptomatology. Conclusions Using a standardized computerized neurocognitive battery and diffusion tensor imaging we show that behavioral performance is moderated by a particular constellation of WM microstructure in healthy individuals that differs in schizophrenia.

Disruptions in brain white matter (WM) organization in schizophrenia may alter neural communication critical for sustaining neurocognitive performance and may relate to the manifestation of clinical symptoms. Diffusion tensor imaging (DTI) has facilitated in vivo study of WM integrity, as measured by fractional anisotropy (FA). Reduced FA has been documented in multiple brain regions in schizophrenia ( Kyriakopoulos et al., 2008). Studies of patients with chronic schizophrenia reported significant, widespread WM microstructural aberrations ( Kyriakopoulos et al., 2008, Thomason and Thompson, 2011 and Asami et al., 2014), while more recent investigations of early-onset psychosis ( Epstein et al., 2013 and Lee et al., 2013) and psychosis in adolescence ( White et al., 2007 and Davenport et al., 2010) have identified focal WM abnormalities. Typically, these findings are limited to major WM fiber tracts and recent evidence indicates that reductions in cortical WM microstructure are associated with cognition ( Nazeri et al., 2013). However, the specific constellation of affected brain regions varies across studies and there is likely regional specificity that relates to neurocognition or clinical symptomatology. Impaired cognition, a core feature of schizophrenia (Gur et al., 2001a), is associated with WM abnormalities (Gur et al., 2001a, Kubicki et al., 2007, Szeszko et al., 2008 and Phillips et al., 2009). Global deficits in cognition are reflective of domain-specific impairments, which are associated with aberrant WM microstructure including working memory (Sugranyes et al., 2012), executive and motor function (Perez-Iglesias et al., 2010), and verbal and visual learning abilities (Liu et al., 2013). We recently reported smaller correlations between a global measure of neurocognition, across task within-individual variability, and WM microstructure in schizophrenia as compared to healthy participants (Roalf et al., 2013). Higher within individual variability in performance speed on a computerized neurocognitive battery was associated with lower FA in the left cingulum bundle and left inferior frontal-occipital fasciculus in healthy people, but not in patients with schizophrenia. Since WM connectivity is essential for maintaining effective communication among regions, deficits in neurocognitive performance may be related, in part, to complex patterns of disrupted WM microstructure. The relation of WM findings and clinical variables has also been evaluated, including symptoms (Paillere-Martinot et al., 2001), medications (Lieberman et al., 2005) and treatment response (Marques et al., 2014). Clinical variables contribute to heterogeneity beyond demographic variables such as age and sex, which are associated with brain maturation tractography patterns (Asato et al., 2010 and Ingalhalikar et al., 2014). Clinical findings evaluating major symptom dimensions suggest that positive symptoms, such as hallucinations, are related to increases in FA (Hubl et al., 2004 and Seok et al., 2007), but see (Asami et al., 2014), while negative symptoms (Bai et al., 2009) and poor outcome (Mitelman et al., 2007) are associated with lower FA. A recent study correlated clinical symptoms with global and regional measures of FA and reported significant associations between lower FA in the left hemisphere and negative, but not positive, symptoms (Asami et al., 2014). Most of these investigations have focused on localized WM regions or tracts. Given the inconsistency in previous findings and limited sample sizes, better estimates of the relation between WM microstructure and clinical symptoms are needed. While DTI is a powerful neuroimaging technique for measuring white matter structure, methodological concerns often make group inference challenging. For example, spatial normalization, or inter-subject registration, is affected by high data dimensionality and the orientation component of the tensors (Ingalhalikar et al., 2010). Several options exist for spatial normalization (Alexander et al., 2001, Cao et al., 2006, Zhang et al., 2006a, Zhang et al., 2006b, Yang et al., 2008 and Yeo et al., 2008), including a deformable registration using orientation and intensity descriptors (DROID; Ingalhalikar et al., 2010). DROID capitalizes on the structural geometry of the diffusion tensor (Westin et al., 2002) and incorporates orientation information to improve the matching of white matter fiber tracts by accounting for the underlying fiber orientation (Ingalhalikar et al., 2010). Here, DROID is used to register all data to a common template; this method is efficient and produces robust results. The goal of this study was to 1) evaluate WM microstructural abnormalities in a large sample of patients with schizophrenia and healthy controls, and 2) relate these measures to neurocognitive performance and clinical symptoms. We hypothesized that: A) patients with schizophrenia will have lower FA values in diffuse cortical WM and along WM fiber tracts compared to healthy controls; B) prediction of performance using brain WM microstructure will result in non-overlapping networks in controls and patients; C) abnormalities in WM regions and tracts will be associated with greater symptom severity.