Functional and Structural Neural Plasticity Following sEMG Control of a Virtual Prosthetic Hand in an Individual with Bilateral Upper-Limb Congenital Amputation

Abstract

Here we document cortical neural connectivity changes associated with several weeks of prosthetic use in an individual with bilateral upper-limb congenital amputation. Secondly, we explore how those changes relate to prosthetic performance over time. Previous research in unilateral aplasia has shown that functional brain connectivity and activations can be disrupted in the missing hand area, and that prosthetic use can normalize those abnormalities. Functional connectivity and prosthetic use related brain changes in individuals with bilateral congenital upper limb amputations have not been defined. Here, we describe functional and structural connectivity changes measured with MRI after 10-weeks' unilateral use of an sEMG-controlled virtual prosthetic in an individual with aplasia born without either arm. We find that both functional connectivity and structural connectivity change with sEMG prosthetic use. Specifically, functional connectivity of motor regions tends to lateralize and become more hemisphere specific. Additionally, structural connectivity of motor cortico-spinal white matter projections and interhemispheric commissural projections increase after sEMG prosthetic use. These functional connectivity changes are different from those previously reported for one-handed congenital amputees, where prosthetic use normalized, not lateralized, imbalanced interhemispheric motor connectivity. Alongside these neural changes, sEMG virtual prosthetic performance both increased and decreased over time, depending on the action performed. Our results suggest that neural representations of bilateral congenital amputation and subsequent neural adaptions with unilateral prosthetic use may be distinct from those of unilateral congenital and traumatic upper-limb amputees. Consideration of condition-specific neurobiology may be critical in developing effective neuro-prostheses.Clinical Relevance- This describes the neural changes induced by sEMG prosthetic control in a congenital bilateral amputee.

Fig. 1.

Custom sleeve for myoelectric control. The sleeve has 32 embedded electrodes for recording electromyography and 3 quincunx patterns for electrocutaneous stimulation proximally on the residual limb.

Fig. 2.

Experimental setup. Participant controlled a 7-DOF virtual bionic arm using sEMG from their left residual limb.

Fig. 3.

Prosthetic performance over time. Performance improved for thumb and index finger but worsened for the little finger. Data show individual trials with grey circles and the mean and standard error of trials in blue. Linear mixed model line of fit shown in red, with marginal R2 and Bonferroni-adjusted p-value displayed.

Fig. 4.

ROI-ROI functional connectivity changes. After 10-weeks of myoelectric prosthetic use, connectivity between and within motor and sensory regions decreased. Connectivity values all greater than Fischer Z > +/− 0.3.

Fig. 5.

Seed to whole brain analysis of motor and somatosensory changes in functional connectivity. The somatosensory cortex had more pronounced lateralization. Reflecting the left-limb prosthetic training, the left, not right, hemisphere seeds showed pronounced increases in negative connectivity to posterior parietal and temporal cortex. Positive connectivity denoted by warm colors. Negative connectivity denoted by cool colors. All values greater than Fischer Z > +/− 0.3.

Fig. 6.

Seed-based motor cortex white matter tractography changes. 1000 highest-weighted motor-structural connections shown. Fiber Orientations: Purple: Superior/Inferior | Green: Anterior/Posterior | Red: