Neural effects of multisensory dance training in Parkinson's disease: evidence from a longitudinal neuroimaging single case study

Abstract

Dance is associated with beneficial outcomes in motor and non-motor domains in Parkinson's disease (PD) and regular participation may help delay symptom progression in mild PD. However, little is known about the neurobiological mechanisms of dance interventions for PD. The present case study explored potential neuroplastic changes in a 69-year-old male with mild PD participating in regular dance classes over 29 weeks. Functional MRI was performed at four timepoints (pre-training, 11 weeks, 18 weeks, 29 weeks), where the individual imagined a dance choreography while listening to the corresponding music. Neural activity was compared between dance-imagery and fixation blocks at each timepoint. Analysis of functionally defined regions revealed significant blood-oxygen-level-dependent (BOLD) signal activation in the supplementary motor area, right and left superior temporal gyri and left and right insula, with modulation of these regions observed over the training period except for the left insula. The results suggest the potential for dance to induce neuroplastic changes in people with PD in regions associated with motor planning and learning, auditory processing, rhythm, emotion, and multisensory integration. The findings are consistent with dance being a multimodal therapeutic activity that could provide long-term benefits for people with PD.

Keywords: Parkinson’s disease; dance; fMRI; learning; motor imagery; neuroplasticity; neurorehabilitation.

Figure 1

Regions activated during the dance-imagery task (vs. fixation) projected onto the inflated cortex of the right (RH) and left (LH) hemispheres, displayed at a statistical threshold of p < 0.0001, Bonferroni-corrected, cluster threshold k > 22.

Figure 2

(A) BOLD activation of SMA during dance-imagery shown in sagittal, coronal, and transverse view in the 3D Talairach space, displayed at a statistical threshold of p < 0.0001, Bonferroni-corrected, with cluster threshold k > 22; (B) fMRI response (average percent BOLD signal change) of SMA during the dance-imagery blocks within each of the four timepoints (dashed line indicates the start of the music played in the scanner); (C) average percent BOLD signal change of SMA between the four timepoints. Error bars represent S.E.M.

Figure 3

(A) BOLD activation of right and left STG during dance-imagery shown in sagittal, coronal, and transverse view in the 3D Talairach space, displayed at a statistical threshold of p < 0.0001, Bonferroni-corrected, cluster threshold k > 22; (B,C) average percent BOLD signal change of right and left STG during dance-imagery blocks within each of the four timepoints (dashed line indicates the start of the music played in the scanner); (D,E) average percent BOLD signal change of right and left STG between the four timepoints. Error bars represent S.E.M.

Figure 4

(A) BOLD activation of right and left insula during dance-imagery shown in in sagittal, coronal, and transverse view in the 3D Talairach space, displayed at a statistical threshold of p < 0.0001, Bonferroni-corrected, with cluster threshold k > 22; (B,C) average percent BOLD signal change of right and left insula during dance-imagery blocks within each of the four timepoints (dashed line indicates the start of the music played in the scanner); (D,E) average percent BOLD signal change of right and left insula between the four timepoints. Error bars represent S.E.M.