Neural Correlates of Mirror Visual Feedback-Induced Performance Improvements: A Resting-State fMRI Study

Abstract

Mirror visual feedback (MVF) is a promising approach to enhance motor performance without training in healthy adults as well as in patients with focal brain lesions. There is preliminary evidence that a functional modulation within and between primary motor cortices as assessed with transcranial magnetic stimulation (TMS) might be one candidate mechanism mediating the observed behavioral effects. Recently, studies using task-based functional magnetic resonance imaging (fMRI) have indicated that MVF-induced functional changes might not be restricted to the primary motor cortex (M1) but also include higher order regions responsible for perceptual-motor coordination and visual attention. However, aside from these instantaneous task-induced brain changes, little is known about learning-related neuroplasticity induced by MVF. Thus, in the present study, we assessed MVF-induced functional network plasticity with resting-state fMRI (rs-fMRI). We performed rs-fMRI of 35 right-handed, healthy adults before and after performing a complex ball-rotation task. The primary outcome measure was the performance improvement of the untrained left hand (LH) before and after right hand (RH) training with MVF (mirror group [MG], n = 17) or without MVF (control group [CG], n = 18). Behaviorally, the MG showed superior performance improvements of the untrained LH. In resting-state functional connectivity (rs-FC), an interaction analysis between groups showed changes in left visual cortex (V1, V2) revealing an increase of centrality in the MG. Within group comparisons showed further functional alterations in bilateral primary sensorimotor cortex (SM1), left V4 and left anterior intraparietal sulcus (aIP) in the MG, only. Importantly, a correlation analysis revealed a linear positive relationship between MVF-induced improvements of the untrained LH and functional alterations in left SM1. Our results suggest that MVF-induced performance improvements are associated with functional learning-related brain plasticity and have identified additional target regions for non-invasive brain stimulation techniques, a finding of potential interest for neurorehabilitation.

Keywords: mirror visual feedback (MVF); motor performance; neuroplasticity; neurorehabilitation; resting state functional connectivity.

Figure 1

Experimental setup and design. Participants participated in one experimental session. A resting-state functional magnetic resonance imaging (rs-fMRI) was acquired before (rs-fMRI_pre) and after (rs-fMRI_post) a complex ball-rotation task which was performed with two cork balls outside of the scanner. Here, participants in both groups first rotated the balls with their left hand (LH) in a counterclockwise direction for 1 min (LH_pre), followed by a 15 min training phase with the right hand (RH) in a clockwise direction (10 trials of 1 min each with 30 s breaks in between). The performing RH was covered to prevent direct view. After the training period, the performance of the LH was retested (LH_post). Only the condition used in the training phase of the ball-rotation task differed between groups: participants in the mirror group (MG) received mirror visual feedback (MVF) during the training period of the RH, whereas participants in the control group (CG) watched their resting LH. See text for details.

Figure 2

Effect of training with or without MVF on motor performance of the untrained LH. Note that there was no significant difference in baseline performance of the untrained LH between groups. (A) Absolute performance improvement of the untrained LH (ball-rotations/min). (B) Relative performance improvement of the untrained LH (%). Both groups improved their performance with the untrained LH significantly but there was a significantly higher gain in absolute as well as relative performance improvement in the MG as compared to the CG. The plots show mean values, and whiskers represent standard error (SE) values. *P < 0.05.

Figure 3

Changes in functional connectivity. (A) Significant TIME × GROUP interaction of changes in left visual cortex (V1, V2) revealing an increase of centrality in the MG. (B) Increase in centrality in MG after 15 min of training the RH with MVF in left V4 and bilateral primary sensorimotor cortex (SM1) as well as a decrease in centrality in left anterior intraparietal sulcus (aIP; Paired-t-tests). (C) Decrease in centrality in the CG in right frontopolar cortex (FPC) after RH training without MVF (Paired-t-test). P_(FWE−corr) < 0.05.

Figure 4

Association of centrality changes in rs-fMRI and relative performance improvements of the untrained LH in the MG. (A) A correlation analysis revealed a significant positive correlation between increase in centrality in left SM1 and behavioral gains of LH for MG. No such correlation could be observed for the CG. (B) Scatter plot diagram illustrating the correlation between the eigenvariate extracted from the peak voxel of the cluster in left SM1 and the performance improvements of the untrained LH in MG (model equation, y = 0.0148x − 0.3617; r = 0.842; p < 0.001).