Facilitating myoelectric-control with transcranial direct current stimulation: a preliminary study in healthy humans

Abstract

Background: Functional Electrical Stimulation (FES) can electrically activate paretic muscles to assist movement for post-stroke neurorehabilitation. Here, sensory-motor integration may be facilitated by triggering FES with residual electromyographic (EMG) activity. However, muscle activity following stroke often suffers from delays in initiation and termination which may be alleviated with an adjuvant treatment at the central nervous system (CNS) level with transcranial direct current stimulation (tDCS) thereby facilitating re-learning and retaining of normative muscle activation patterns.

Methods: This study on 12 healthy volunteers was conducted to investigate the effects of anodal tDCS of the primary motor cortex (M1) and cerebellum on latencies during isometric contraction of tibialis anterior (TA) muscle for myoelectric visual pursuit with quick initiation/termination of muscle activation i.e. 'ballistic EMG control' as well as modulation of EMG for 'proportional EMG control'.

Results: The normalized delay in initiation and termination of muscle activity during post-intervention 'ballistic EMG control' trials showed a significant main effect of the anodal tDCS target: cerebellar, M1, sham (F(2) = 2.33, p < 0.1), and interaction effect between tDCS target and step-response type: initiation/termination of muscle activation (F(2) = 62.75, p < 0.001), but no significant effect for the step-response type (F(1) = 0.03, p = 0.87). The post-intervention population marginal means during 'ballistic EMG control' showed two important findings at 95% confidence interval (critical values from Scheffe's S procedure): 1. Offline cerebellar anodal tDCS increased the delay in initiation of TA contraction while M1 anodal tDCS decreased the same when compared to sham tDCS, 2. Offline M1 anodal tDCS increased the delay in termination of TA contraction when compared to cerebellar anodal tDCS or sham tDCS. Moreover, online cerebellar anodal tDCS decreased the learning rate during 'proportional EMG control' when compared to M1 anodal and sham tDCS.

Conclusions: The preliminary results from healthy subjects showed specific, and at least partially antagonistic effects, of M1 and cerebellar anodal tDCS on motor performance during myoelectric control. These results are encouraging, but further studies are necessary to better define how tDCS over particular regions of the cerebellum may facilitate learning of myoelectric control for brain machine interfaces.

Figure 1

Schematic drawing of the cerebellar feedback-error-learning model: a feedback controller transforms sensory error in visual space into a feedback motor command which is used to train the inverse model for feedforward control.

Figure 2

Electrode montages for anodal tDCS (1 mA direct current for 15 min) of 1. primary motor cortex representation area of the right leg, where a 5 cm × 7 cm saline-soaked sponge anode was placed 1.5 cm lateral and 2 cm posterior to Cz (10–20 EEG system), 2. cerebellum of left hemisphere where the 5 cm × 7 cm saline-soaked sponge anode was placed 3 cm lateral to Inion (10–20 EEG system). The 5 cm × 7 cm saline-soaked sponge cathode was placed above the right contralateral orbit.

Figure 3

Experimental setup for myoelectric control with visual feedback. The normalized electromyogram (EMG) from tibialis anterior muscle was displayed as the TRACKING signal along with a sinusoidal TARGET signal.

Figure 4

Experimental protocols for Experiment 1 (top panel): offline anodal tDCS, and Experiment 2 (bottom panel): online anodal tDCS with myoelectric visual pursuit, each with three one-day test sessions separated by at least a week, where subjects received 15 min of anodal tDCS to M1, cerebellum, or sham tDCS.

Figure 5

The post hoc comparisons in the normalized delay in initiation (top panel) and termination (bottom panel) of muscle activity to the grouping variable, tDCS target: M1, cerebellum, sham, during 'ballistic EMG control’ for Experiment 1. The differences were considered significant if 95% confidence intervals represented by a line for the mean shown by a filled circle did not overlap.

Figure 6

Results from myoelectric the visual pursuit task (i.e., Experiment 2) for Myoelectric Training Trial# 2–6 for the tDCS groups: M1, cerebellum, sham. The top row illustrates the overall TARGET and TRACKING signals during the modulation of EMG during 'proportional EMG control’ trials, the middle row shows the effects on the normalized response latency, and the bottom row shows the effects on the mean absolute ERROR.