Motor performance and functional connectivity between the posterior cingulate cortex and supplementary motor cortex in bipolar and unipolar depression

Abstract

Although implicated in unsuccessful treatment, psychomotor deficits and their neurobiological underpinnings in bipolar (BD) and unipolar (UD) depression remain poorly investigated. Here, we hypothesized that motor performance deficits in depressed patients would relate to basal functional coupling of the hand primary motor cortex (M1) and the posterior cingulate cortex (PCC) with the supplementary motor area (SMA). We performed a longitudinal, naturalistic study in BD, UD and matched healthy controls comprising of two resting-state functional MRI measurements five weeks apart and accompanying assessments of motor performance using a finger tapping task (FTT). A subject-specific seed-based analysis describing functional connectivity between PCC-SMA as well as M1-SMA was conducted. The basal relationships with motor performance were investigated using linear regression models and all measures were compared across groups. Performance in FTT was impaired in BD in comparison to HC in both sessions. Behavioral performance across groups correlated significantly with resting state functional coupling of PCC-SMA, but not of M1-SMA regions. This relationship was partially reflected in a reduced PCC-SMA connectivity in BD vs HC in the second session. Exploratory evaluation of large-scale networks coupling (SMN-DMN) exhibited no correlation to motor performance. Our results shed new light on the association between the degree of disruption in the SMA-PCC anticorrelation and the level of motor impairment in BD.

Keywords: Bipolar disorder; Finger tapping; Major depressive disorder; Posterior cingulate cortex; Resting state functional magnetic resonance imaging; Supplementary motor area.

Fig. 1

Methods: (A) Connectivity model between regions of interest: supplementary motor area (SMA) and right hemisphere hand primary motor cortex (M1) as part of the sensory motor network (red), posterior cingulate cortex (PCC) as part of the default mode network (blue), (B) Study design: unipolar (UD), bipolar (BD) depressed and healthy control subjects (HC) performed the finger-tapping-task (FTT) on both days of both visits, FTTa and FTT b versions as described in methods section. The data of the second days was not used for the analysis of motor performance. The clinical assessment and the MRI sessions took place on one of the days. The blood sample was taken as close as possible before the MRI measurement, (C) Seed definition technique: exemplification of the overlapping area (yellow) based on FreeSurfer individual segmentation of grey matter (green) and a spherical seed (red) at the MNI coordinates x = 40, y = − 20, z = 54 of functional coordinates for right hemisphere hand motor cortex (M1)⁶⁶ (D) and at x = 0 y = − 53, z = 2 for the posterior cingulate cortex¹⁴

Fig. 2

Motor performance scores: Marginal means (mean number of correctly tapped sequences per 30 seconds) and confidence intervals from repeated-measures ANOVA of motor performance scores for healthy controls (HC, blue), unipolar (UD, pink) and bipolar (BD, green) patients. * = p < 0,05; The values displayed in the graph are HC visit 1 M ± SD = 12.7 ± 6.1, UD visit 1 M ± SD = 9.8 ± 4.2, BD visit 1 M ± SD = 8.6 ± 4.6, HC visit 2 M ± SD = 14.4 ± 5.8, UD visit 2 M ± SD = 12.1 ± 5.0, BD visit 2 M ± SD = 10.8 ± 4.7. Significant differences are seen only between HC and BD (ANOVA factor “group” F(2,76) = 4.122; p = 0.020) with BD showing a significantly reduced number of correctly performed finger tapping trials compared to the HC (Post-hoc-test p = 0.023)

Fig. 3

PCC connectivity differences: Resting state functional connectivity differences of the posterior cingulate cortex (PCC) ANOVA “main effect” contrast and pairwise differences in healthy controls, bipolar and unipolar depressed at the second visit (HC2, BD2, UD2, p FWEc < 0.05) as seem in an SPM12 full factorial model. HC2 showed higher connectivity measured by anticorrelation between the PCC and the bilateral SMA (red circle) than BD2. UD2 showed higher connectivity measured by anticorrelation between the PCC and the bilateral SMA (red circle) than BD2. Colorbar represents t-values

Fig. 4

PCC connectivity correlated with motor performance Resting state functional connectivity of the posterior cingulate cortex (PCC) linearly correlated with the motor performance in the finger-tapping-task (p FWEc < 0.05). The motor performance of all participants is correlated with the connectivity (anticorrelation) strength between the PCC and the left supplementary motor area (SMA, red circle) as seen in an SPM12 linear regression model. Colorbar represents t-values

Fig. 5

PCC connectivity overlaps in the SMA: Resting state functional connectivity differences of the posterior cingulate cortex (PCC) between groups and connectivity of the PCC correlated with motor performance overlap in the supplementary motor area (SMA) at an axial plane (z_MNI = 56, p FWEc < 0.05). A Healthy control subjects (HC2) show higher connectivity than the bipolar depressive subjects (BD2) at visit 2. B Unipolar depressive subjects (UD2) present higher connectivity than the bipolar depressive subjects (BD2) at visit 2. C The connectivity of the PCC to a part of the left SMA is correlated with the motor performance in the finger-tapping-task. D Connectivity clusters seen in (A, red), (B, blue) and (C, green) overlap in different extends of the SMA, as depicted in pink (A⋂B), located at medially in both hemispheres, in yellow (A⋂C), located anterior of the white area in the left hemisphere, in turquoise (B⋂C), located posterior to the white area in the left hemisphere, and in white (A⋂B⋂C), located in the left hemisphere between pink, yellow, and turquoise

Fig. 6

M1 connectivity differences: Resting state functional connectivity differences of the right hemisphere hand motor area (M1) ANOVA “main effect” contrast and pairwise differences between bipolar and unipolar depressed at first and second visits (BD, UD, p FWEc < 0.05) as seem in an SPM12 full factorial model. BD showed stronger connectivity than UD between M1 and the bilateral SMA (red circle) in visit 1 than in visit 2. No regions were evidenced in the UD>BD contrast. Colorbar represents t-values