Meta-analysis of diffusion tensor imaging studies in schizophrenia

Abstract

The objective of the study was to identify whether there are consistent regional white matter changes in schizophrenia. A systematic search was conducted for voxel-based diffusion tensor imaging fractional anisotropy studies of patients with schizophrenia (or related disorders) in relation to comparison groups. The authors carried out meta-analysis of the co-ordinates of fractional anisotropy differences. For the meta-analysis they used the Activation Likelihood Estimation (ALE) method hybridized with the rank approach used in Genome Scan Meta-Analysis (GSMA). This system detects three-dimensional conjunctions of co-ordinates from multiple studies and permits the weighting of studies in relation to sample size. Fifteen articles were identified for inclusion in the meta-analysis, including a total of 407 patients with schizophrenia and 383 comparison subjects. The studies reported fractional anisotropy reductions at 112 co-ordinates in schizophrenia and no fractional anisotropy increases. Over all studies, significant reductions were present in two regions: the left frontal deep white matter and the left temporal deep white matter. The first region, in the left frontal lobe, is traversed by white matter tracts interconnecting the frontal lobe, thalamus and cingulate gyrus. The second region, in the temporal lobe, is traversed by white matter tracts interconnecting the frontal lobe, insula, hippocampus–amygdala, temporal and occipital lobe. This suggests that two networks of white matter tracts may be affected in schizophrenia, with the potential for ‘disconnection’ of the gray matter regions which they link.

Introduction

Brain changes in schizophrenia may involve abnormalities in a network of gray and white matter regions (Csernansky and Cronenwett, 2008). However, the architecture of these changes has been more precisely mapped in gray matter than in white matter. Gray matter reductions have been located in limbic, paralimbic and frontal cortical regions and thalamus (Ellison-Wright et al., 2008a, Glahn et al., 2008, Wright et al., 2000, Shenton et al., 2001). However, the distribution of white matter changes remains uncertain (Kubicki et al., 2007, Kanaan et al., 2005). Two broad theories have been proposed to describe the pattern of white matter changes: the global and macro-circuit theories (Buchsbaum et al., 2006a).

According to the global theory, white matter reductions occur uniformly throughout the brain, possibly as a result of genetic abnormalities in the protein pathways controlling myelination (Konrad and Winterer, 2008). The alternative macro-circuit theory proposes that specific white matter tracts are disrupted in schizophrenia, either as a cause or a consequence of a disorder in the gray matter regions they connect (Konrad and Winterer, 2008).

At a functional level, considerable evidence has accrued for the abnormal integration of neural systems activated in patients with schizophrenia performing cognitive tasks. The evidence for this ‘functional dysconnectivity’ derives from a wide range of neurophysiologic and neuroimaging studies (Friston, 2005). Such abnormalities in functional connectivity could be due to abnormalities of axonal connectivity between regions (Bullmore et al., 1997), or they could be attributed to abnormal synapse formation and plasticity (Stephan et al., 2006).

The objective of this meta-analysis was to test the different implications of these theoretical accounts of white matter changes in schizophrenia. We focused on neuro-imaging studies using diffusion tensor imaging (DTI). This is a magnetic resonance method which measures the diffusion properties of water molecules (Assaf and Pasternak, 2008). In general, the diffusion of water molecules is increased in white matter where it can occur parallel to the fibers. This property enables DTI to map both the distribution and integrity of white matter within the brain. One DTI measurement is called fractional anisotropy (FA). In the brain FA is high in white matter, low in gray matter and close to zero in cerebro-spinal fluid. We have selected studies employing voxel-based analyses of FA, since these analyse white matter throughout the brain, rather than region of interest studies which pre-select limited parts of the brain for analysis. Voxel-based analyses generally report the three-dimensional co-ordinates where there are maximal FA differences between patients and controls and these provide the data for meta-analysis.

Although most DTI studies of schizophrenia have identified FA reductions, with a few exceptions (Foong et al., 2002), diverse white matter regions have been implicated (Kubicki et al., 2007, Kanaan et al., 2005, Buchsbaum et al., 2006a, Friedman et al., 2008). Evidence that there is functional dysconnectivity in schizophrenia has recently been interpreted in terms of changes in micro-circuit synaptic plasticity rather than macro-circuit white matter abnormality (Stephan et al., 2006). An important part of the evidence cited against the macro-circuit theory was that ‘diffusion weighted imaging studies have delivered negative results or widely varying findings’.

This meta-analysis tests whether there are consistently located white matter deficits (possibly super-imposed on global changes) whose detection in individual studies depends on statistical variation (i.e. the macro-circuit theory). Combining the results from multiple studies increases the power to detect these deficits. White matter regions which are preferentially affected may identify specific tract deficits contributing to functional disconnection of the regions they link (Burns et al., 2003). Our null hypothesis was that the coordinates of FA reductions (if present) followed a uniform random distribution (i.e. consistent with the global model of axonal disruption).