Right-sided brain lesions predominate among patients with lesional mania: evidence from a systematic review and pooled lesion analysis

Abstract

Despite claims that lesional mania is associated with right-hemisphere lesions, supporting evidence is scarce, and association with specific brain areas has not been demonstrated. Here, we aimed to test whether focal brain lesions in lesional mania are more often right- than left-sided, and if lesions converge on areas relevant to mood regulation. We thus performed a systematic literature search (PROSPERO registration CRD42016053675) on PubMed and Web-Of-Science, using terms that reflect diagnoses and structures of interest, as well as lesional mechanisms. Two researchers reviewed the articles separately according to PRISMA Guidelines, selecting reports of adult-onset hypomania, mania or mixed state following a focal brain lesion, for pooled-analyses of individual patient data. Eligible lesion images were manually traced onto the corresponding MNI space slices, and lesion topography analyzed using standard brain atlases. Using this approach, data from 211 lesional mania patients was extracted from 114 reports. Among 201 cases with focal lesions, more patients had lesions involving exclusively the right (60.7%) than exclusively the left (11.4%) hemisphere. In further analyses of 56 eligible lesion images, while findings should be considered cautiously given the potential for selection bias of published lesion images, right-sided predominance of lesions was confirmed across multiple brain regions, including the temporal lobe, fusiform gyrus and thalamus. These, and several frontal lobe areas, were also identified as preferential lesion sites in comparisons with control lesions. Such pooled-analyses, based on the most comprehensive dataset of lesional mania available to date, confirm a preferential association with right-hemisphere lesions, while suggesting that several brain areas/circuits, relevant to mood regulation, are most frequently affected.

Fig. 1. Article selection flowchart.

Article selection ws performed according to PRISMA Statement.

Fig. 2. Lesion distribution by major brain areas.

a Lesion distribution for all cases in the literature review. b Lesion distribution for cases with vascular lesions. c Lesion distribution for cases with non-vascular lesions. d Comparison of lesion distribution for tumor cases identified in this literature review (n = 19) with tumor distribution described by Ostrom et al. for a large patient database (n = 169934). In what brain tumors are concerned, regions were defined according to the International Classification of Diseases for Oncology (ICD-O), without considering tumours originating from the meninges (n = 14), ventricles (n = 2), cranial nerves (n = 1), or of unspecified origin (n = 4). “Other Brain” refers to lesions spanning multiple areas (C71.8: “neoplasm involving two or more sites, corpus callosum and tapetum”) or when areas were insufficiently specified (C71.9: “intracranial site, cranial fossa not otherwise specified, anterior cranial fossa, middle cranial fossa, posterior cranial fossa and suprasellar”). “Cerebrum” refers to multiple subcortical structures (C71.0: “basal ganglia, central white matter, unspecified cerebral cortex, cerebral hemisphere, cerebral white matter, corpus striatum, globus pallidus, hypothalamus, insula, internal capsule, island of Reil, operculum, pallium, putamen, rhinencephalon, supratentorial brain not otherwise specified and thalamus”). ^aDoes not include the following case series, which did not provide enough information on individual lesion etiology: Carran 2003, Robinson 1988, Starkstein 1987 and Starkstein 1991 (See Supplementary Material for complete references – Supplementary Table S3). *p value < 0.05.

Fig. 3. Distribution of brain lesions associated with secondary mania in 56 patients with eligible lesion images.

a Comparison between right vs. left-sided lesions. Each lesion was traced manually onto a common brain atlas (MNI) and projected on the closest depicted slice. Numbers above slices indicate z-coordinates in MNI space. The color code indicates maximal number of lesions overlapping on a single voxel. b Subtraction plot contrasting 29 right-sided stroke patients with secondary bipolar disorder (red-yellow) versus 439 unselected right hemisphere stroke patients (blue-green). In this plot, a value of, for example, 30, reflects that the voxel is damaged 30% more frequently in patients with secondary bipolar disorder than in unselected patients (for more details on the method see Rorden and Karnath). To improve visualization, lesions of mania patients were projected onto the closest depicted slice before plot generation.