Bimanual coordination and spinal cord neuromodulation: how neural substrates of bimanual movements are altered by transcutaneous spinal cord stimulation

Abstract

Humans use their arms in complex ways that often demand two-handed coordination. Neurological conditions limit this impressive feature of the human motor system. Understanding how neuromodulatory techniques may alter neural mechanisms of bimanual coordination is a vital step towards designing efficient rehabilitation interventions. By non-invasively activating the spinal cord, transcutaneous spinal cord stimulation (tSCS) promotes recovery of motor function after spinal cord injury. A multitude of research studies have attempted to capture the underlying neural mechanisms of these effects using a variety of electrophysiological tools, but the influence of tSCS on cortical rhythms recorded via electroencephalography remains poorly understood, especially during bimanual actions. We recruited 12 neurologically intact participants to investigate the effect of cervical tSCS on sensorimotor cortical oscillations. We examined changes in the movement kinematics during the application of tSCS as well as the cortical activation level and interhemispheric connectivity during the execution of unimanual and bimanual arm reaching movements that represent activities of daily life. Behavioral assessment of the movements showed improvement of movement time and error during a bimanual common-goal movement when tSCS was delivered, but no difference was found in the performance of unimanual and bimanual dual-goal movements with the application of tSCS. In the alpha band, spectral power was modulated with tSCS in the direction of synchronization in the primary motor cortex during unimanual and bimanual dual-goal movements and in the somatosensory cortex during unimanual movements. In the beta band, tSCS significantly increased spectral power in the primary motor and somatosensory cortices during the performance of bimanual common-goal and unimanual movements. A significant increase in interhemispheric connectivity in the primary motor cortex in the alpha band was only observed during unimanual tasks in the presence of tSCS. Our observations provide, for the first time, information regarding the supra-spinal effects of tSCS as a neuromodulatory technique applied to the spinal cord during the execution of bi- and unimanual arm movements. They also corroborate the suppressive effect of tSCS at the cortical level reported in previous studies. These findings may guide the design of improved rehabilitation interventions using tSCS for the recovery of upper-limb function in the future.

Fig. 1

Experimental design. A Illustration of the KINARM exoskeleton robotic platform and experimental setup. Participants performed visually-guided reaching tasks guided by a virtual reality display. B Representation of the visually-guided tasks: top left—unimanual movement, top right—bimanual dual-goal movement, and bottom—bimanual common-goal movement. All the movements started from a home position (red circle) and ended on a target (white circle)

Fig. 2

Transcutaneous spinal cord stimulation. A Cervical tSCS were delivered through cathodic electrodes placed midline at C3–4 and C5–6 spinous processes. Two anodic electrodes were placed bilaterally over the iliac crests. B Stimulation waveform: 1ms long pulses with a carrier frequency of 10 kHz are delivered at 40 Hz

Fig. 3

Right arm raw movement traces during the execution of bimanual common-goal task in X–Y coordinates A when tSCS was off, and B when tSCS was applied to the cervical spinal cord from a representative participant. Each color is a single reaching movement, and each plot illustrates an overall of 20 repetitions per task

Fig. 4

Relationship between movement kinematics and stimulation. Values are Mean ± SE. Cervical tSCS did not significantly alter A reaction time, B movement time, and C movement error relative to the no tSCS condition. (*P < 0.05; ^#P < 0.1)

Fig. 5

Raw EEG signal and power spectral density with different filtering approaches and artifact removal from a representative participant. Single trial EEG signal recorded from the left primary motor cortex during the execution of bimanual common-goal task (A) without tSCS, (B) in the presence of tSCS. (C) Power spectrum of the concatenated EEG signal (a single trial of it is shown in A) without tSCS. D Representative welch power spectral density estimate of the EEG signal recorded from left primary motor cortex with only basic band-pass filtering of 0.1–200 Hz. E Power spectral density of the same EEG signal with the addition of 40 Hz notch filtering. F Power spectral density of the artifact-free EEG signal. A general reduction of spectral power is observed in both the alpha and beta bands when the transient high-amplitude peaks are eliminated

Fig. 6

Alpha and beta band spectral power analysis. Alpha (A, B) and beta band (C, D) spectral power during the execution of the three movement tasks with and without tSCS. An augmentation in spectral power is seen in the alpha band during both unimanual and dual-goal movements in the presence of tSCS. Additionally, elevated spectral power is observed in the beta band for unimanual movement, as well as in the beta band during common-goal movement

Fig. 7

Interhemispheric connectivity with artifact removal applied between left and right (A, C) primary motor cortex, and (B, D) primary somatosensory cortex. The alterations in interhemispheric connectivity do not exhibit consistency across frequency bands and tasks. A general elevation of interhemispheric connectivity is evident in the alpha band. However, a reduction in interhemispheric coupling is observed in the beta band, specifically over the somatosensory cortex