ReviewSystematic review and voxel-based meta-analysis of diffusion tensor imaging studies in bipolar disorder
Introduction
Structural (Emsell and Mcdonald, 2009, Kempton et al., 2008) and functional (Cerullo et al., 2009) imaging studies have revealed regional brain differences in bipolar disorder compared to healthy controls. Abnormal interaction between prefrontal and limbic areas (Phillips et al., 2008) may account for the characteristic emotional dysregulation and executive deficits in bipolar disorder (BD). Subtle microstructural abnormalities of white matter revealed by diffusion tensor imaging (DTI) studies have been hypothesised to be responsible for a functional disconnection between cortical, limbic and other brain regions (Mahon et al., 2010).
The first DTI investigations of white matter in bipolar disorder were region-of-interest (ROI) studies which focussed on frontal and limbic regions, placing ROI's in frontal white matter (Adler et al., 2004, Adler et al., 2006, Haznedar et al., 2005, Beyer et al., 2005, Frazier et al., 2007), in specific limbic or frontal white matter tracts (Haznedar et al., 2005, Wang et al., 2008a, Pavuluri et al., 2009, Gonenc et al., 2010), or using limbic or frontal tracts as reconstructed by tractography (Houenou et al., 2007, Mcintosh et al., 2008, Wang et al., 2009, Lin et al., 2011, Benedetti et al., 2011a). Other white matter regions are less well represented, but include internal capsule (Haznedar et al., 2005, Pavuluri et al., 2009), posterior white matter (Adler et al., 2006), periventricular white matter (Macritchie et al., 2010), and corpus callosum (Yurgelun-Todd et al., 2007, Wang et al., 2008b, Pavuluri et al., 2009, Macritchie et al., 2010).
ROI studies have the advantage of being driven by a priori hypotheses, but are thus also subject to bias. Whole-brain studies are less presumptive, though more statistically complex as each white matter voxel is analysed independently. Two whole-brain analysis methods are commonly used. Voxel-based analysis (VBA) analyses all white matter voxels, correcting for multiple comparisons and noise by reporting only contiguous clusters of significant voxels. Tract-based spatial statistics (TBSS) isolate the central core of white matter tracts with the highest FA and report significant clusters within that white matter skeleton (Smith et al., 2006). Results from whole-brain studies suggest widespread white matter abnormalities in bipolar disorder, but not all data are consistent.
Attempts to summarise the literature have also been inconsistent. In a review of DTI studies in bipolar disorder, Heng et al. (2010) noted that white matter changes are widespread but consistent, with decreased FA in frontal and prefrontal lobes. In their review, Brambilla et al. (2009) concluded that the evidence from diffusion imaging studies suggested impairments in fronto-limbic circuitry (cingulum bundle and uncinate fasciculus), inter-hemispheric connectivity (corpus callosum) and fronto-parieto-temporal connections (superior longitudinal fasciculus). In a meta-analysis of affective disorders (including major depression), Sexton et al. (2009) reported reduced FA in frontal and temporal regions and tracts in bipolar and unipolar depression patients relative to controls.
Vederine et al. (2011) performed a meta-analysis of whole-brain voxel-based DTI studies in BD using anatomical likelihood estimation (ALE), a widely used meta-analytical tool for voxel-based neuroimaging studies (Turkeltaub et al., 2002), to locate clusters of most consistently altered fractional anisotropy (FA) in BD. They analysed the peak voxel co-ordinates of clusters of decreased FA from ten whole-brain DTI studies in BD, omitting clusters of increased FA. Two significant clusters of decreased FA were identified, both on the right side. The first cluster was located close to the right parahippocampal gyrus posteriorly, with tractography from that cluster implicating various long association tracts. The second cluster was close to the right anterior and subgenual cingulate cortex, implicating the inferior fronto-occipital and uncinate fasciculi as well as forceps minor.
There has not yet been a systematic review examining whether certain tracts are more consistently and more robustly affected than other tracts in BD. Indeed, it is unclear whether certain classes of tracts—association, projection or commissural—are preferentially affected. Knowledge of the regional distribution of WM changes, particularly if tract specific, may provide clues to aetiology and pathogenesis. In the first part of this review, we aim to identify which white-matter tracts, if any, are most consistently implicated by changes in anisotropy in bipolar disorder.
In the second part, we update and extend the previous meta-analysis using effect-size signed differential mapping (ES-SDM) (Radua and Mataix-Cols, 2009), a voxel-based meta-analytic technique similar to ALE which has been used to study white matter FA and volume changes during development as well as in several psychiatric disorders (Radua et al., 2010, Bora et al., 2011, Peters et al., 2012).
Section snippets
Paper selection
Pubmed/MEDLINE and EMBASE databases were systematically searched using the strategy ((“bipolar” OR “mania” OR “depression”) AND (“DTI” OR “diffusion tensor”)) for papers published up to December 2012. In addition, reference lists of included studies were hand searched for suitable papers. Papers had to meet the following inclusion criteria: (1) be published in English in a peer-reviewed journal, (2) contain original data from subjects 18–65 years old, (3) compare a bipolar disorder group with a
Study selection
Systematic search of databases revealed abstracts of 244 English papers on 30 December 2012, which were reduced to 15 whole-brain studies using DTI to compare FA between bipolar patients and healthy controls in adults (Fig. 1). Ten VBA studies (252 patients and 256 controls) and five TBSS studies (138 patients and 98 controls) met inclusion criteria.
Subject and scan characteristics
Table 1 shows the clinical features of the subject groups. High heterogeneity of subject characteristics in terms of diagnosis, chronicity,
Discussion
The tract count indicates that all major classes of white matter tracts are implicated in bipolar disorder. Although both VBA and TBSS implicate all classes of tract, the long association tracts are more implicated in VBA analyses than TBSS. VBA implicates the anterior more than the posterior projection fibres, with the reverse being true for TBSS.
In general, larger tracts were implicated more frequently than smaller tracts. There may be a correspondence between how often tracts are implicated
Limitations
Limitations in our review relate to tract counting, ES-SDM sensitivity, study heterogeneity and consideration of only fractional anisotropy. Our tract counting method is systematic and quantitative but lacks statistical rigour. The important confounder of tract size and architecture is likely to influence labelling by virtue of ease of achieving threshold cluster size, and relative prominence and consistency in atlases leading to labelling bias. In addition, the accurate identification of
Conclusions
Despite these limitations, the findings here suggest that all major classes of white matter are affected in bipolar disorder, with larger tracts implicated more frequently. ES-SDM meta-analysis reveals that anterior limbic and posterior temporoparietal regions are both affected. Most findings are of decreased FA. The anterior findings are consistent with current models of emotion regulation. The posterior findings may be related to cognitive deficits in bipolar disorder, but require further
Role of funding source
This review was performed as part of residency training in psychiatry and no extra funding was required.
Conflict of interest
None of the authors have any conflicts of interest to declare.
Acknowledgements
None of the authors have any acknowledgements to make. There were no other financial or intellectual contributors to the paper.
References (73)
- et al.
Global prefrontal and fronto-amygdala dysconnectivity in bipolar I disorder with psychosis history
Biological Psychiatry
(2013) - et al.
Disruption of white matter integrity in bipolar depression as a possible structural marker of illness
Biological Psychiatry
(2011) - et al.
Neuroanatomical abnormalities in schizophrenia: a multimodal voxelwise meta-analysis and meta-regression analysis
Schizophrenia Research
(2011) - et al.
A diffusion tensor imaging tractography atlas for virtual in vivo dissections
Cortex
(2008) - et al.
Voxel based morphometric and diffusion tensor imaging analysis in male bipolar patients with first-episode mania
Progress in Neuropsychopharmacology and Biological Psychiatry
(2012) - et al.
Sustained attention-deficit confirmed in euthymic bipolar disorder but not in first-degree relatives of bipolar patients or euthymic unipolar depression
Biological Psychiatry
(2005) - et al.
Assessment of white matter abnormalities in paranoid schizophrenia and bipolar mania patients
Psychiatry Research
(2011) - et al.
Age-related changes in grey and white matter structure throughout adulthood
Neuroimage
(2010) - et al.
Combined diffusion tensor imaging and transverse relaxometry in early-onset bipolar disorder
Journal of the American Academy of Child and Adolescent Psychiatry
(2010) - et al.
Similarities and differences of white matter connectivity and water diffusivity in bipolar I and II disorder
Neuroscience Letters
(2011)