Anodic Transcutaneous Spinal Cord Stimulation for Eliciting Upper Limb Motion: An Alternative Hypothesis

Introduction: Cervical transcutaneous spinal cord stimulation (tSCS) and brain-computer interface (BCI) have been used to improve upper limb motor recovery in patients with neurological impairments.1 However, the real-time integration between BCI and tSCS requires further investigation. Our study aimed to develop a BCI-controlled tSCS system and investigate its potential advantages in inducing neuroplasticity of upper limb corticospinal excitability.

Methods: A non-disabled participant was recruited to perform BCI-tSCS intervention within a single 30-min session following the experimental protocol shown in Figure 1A. The BCI system was used to detect pinch motor imagery in real-time from a single electroencephalography (EEG) electrode showing the strongest event-related desynchronization (ERD) activity during the BCI calibration prior to the intervention.2 During the BCI-tSCS intervention, if the BCI system detected ERD activity within 15 s during the “Go” epoch, continuous tSCS was triggered for 5 s (Figure 1A). If the BCI system did not detect ERD activity within 15 s, tSCS was not triggered, and the subsequent trial was initiated. When tSCS was triggered, 1 ms long biphasic rectangular pulses were applied continuously with the highest tolerable intensity of 25 mA at a frequency of standard 30 Hz. The tSCS cathode electrode was placed over C7-T1 cervical spine, and two anode electrodes were placed over iliac crests (Figure 1A). The neurophysiological assessment was conducted to investigate corticospinal excitability via single-pulse transcranial magnetic stimulation (TMS) applied on the primary motor cortex to elicit ten motor-evoked potentials (MEP) at first dorsal interosseous muscle before (Pre), immediately (Post0), and 30 min (Post30) after the intervention. The Friedman test and the Wilcoxon signed-rank post hoc comparison were examined to compare the peak-to-peak amplitude of MEP responses across three assessment time points.

Results: During the 30-min intervention, the success rate of the BCI-tSCS system was 86.1% resulting in 105 successful tSCS stimulation out of 122 trials. After the BCI-tSCS intervention, MEP responses were facilitated at the Post0 time point compared with before the intervention (Pre) (Figure 1B).

Discussion and Conclusion: We developed a real-time BCI-tSCS system with a success rate comparable with previous BCI-based neuroplasticity applications.2 The BCI-tSCS intervention was effective in inducing short-duration corticospinal excitability facilitation. The effectiveness of the BCI-tSCS was likely attributed to the involvement of Hebbian-type learning,2 where the use of BCI synchronizes the descending voluntary commands with the sensory nerve activation through tSCS. Therefore, the real-time BCI-controlled tSCS system may benefit for rehabilitation applications to improve the motor recovery of patients with neurological impairments.