Upregulating excitability of corticospinal pathways in stroke patients using TMS neurofeedback; A pilot study

Abstract

Upper limb weakness following a stroke affects 80% of survivors and is a key factor in preventing their return to independence. State-of-the art approaches to rehabilitation often require that the patient can generate some activity in the paretic limb, which is not possible for many patients in the early period following stroke. Approaches that enable more patients to engage with upper limb therapy earlier are urgently needed. Motor imagery has shown promise as a potential means to maintain activity in the brain's motor network, when the patient is incapable of generating functional movement. However, as imagery is a hidden mental process, it is impossible for individuals to gauge what impact this is having upon their neural activity. Here we used a novel brain-computer interface (BCI) approach allowing patients to gain an insight into the effect of motor imagery on their brain-muscle pathways, in real-time. Seven patients 2-26 weeks post stroke were provided with neurofeedback (NF) of their corticospinal excitability measured by the size of motor evoked potentials (MEP) in response to transcranial magnetic stimulation (TMS). The aim was to train patients to use motor imagery to increase the size of MEPs, using the BCI with a computer game displaying neurofeedback. Patients training finger muscles learned to elevate MEP amplitudes above their resting baseline values for the first dorsal interosseous (FDI) and abductor digiti minimi (ADM) muscles. By day 3 for ADM and day 4 for FDI, MEP amplitudes were sustained above baseline in all three NF blocks. Here we have described the first clinical implementation of TMS NF in a population of sub-acute stroke patients. The results show that in the context of severe upper limb paralysis, patients are capable of using neurofeedback to elevate corticospinal excitability in the affected muscles. This may provide a new training modality for early intervention following stroke.

Keywords: Brain-computer interface; Motor imagery; Neurofeedback; Stroke; TMS; Upper limb.

Conflict of interest statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Copyright © 2020 The Author(s). Published by Elsevier Inc. All rights reserved.

Figures

Fig. 1

Experimental setup, neurofeedback display and EMG traces. Panel A shows the general experimental setup with patient seated in front of a monitor viewing feedback of MEP amplitude (adapted from Cretu et al., 2019). Panel B shows the visual sequence of events for one trial. Each trial commenced with a ‘traffic lights’ display (i) indicating background muscle activity in non-involved muscles, followed by a video (ii) showing a wrist or finger movement, during which the TMS pulse was delivered. Following TMS, a rectangular bar representing the MEP amplitude from the paretic muscles was displayed (iii). The colour of the bar (green or red) and a positive or negative soundbyte indicated to the patient whether their MEP had exceeded (or not) the amplitude of their baseline values recorded prior to training. Panel C shows EMG traces from one patient (NF6) during their final session (Day 4), with exemplar trials shown at i) baseline, ii) during the final block of TMS neurofeedback and iii) at rest post neurofeedback training. TMS delivery is denoted by a black dashed line. A period of 105 ms background EMG prior to TMS is shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Percentage change in MEP amplitudes from resting baseline in stroke affected limb. Boxplots with superimposed individual datapoints show MEP amplitudes from stroke affected finger muscles expressed as percentage of resting baseline for 6 patients over 4 days of TMS NF training of averaged FDI/ADM. Data for all 3 blocks of neurofeedback are shown separately. Values above 0 (dotted line) indicate increased MEP amplitudes.

Fig. 3

Percentage change in MEP amplitudes from baseline collapsed across training days. Data for each day of training were averaged across the 3 training blocks. Individual patients’ data are plotted as separate data points. Black trendline shows median values across the 6 patients for FDI (Panel A) and ADM (Panel B).

Fig. 4

Case study of patient engaging in TMS NF of wrist extensor. Data shown are raw MEP amplitudes recorded from Extensor Carpi Radialis (ECR) from one patient with very poor residual finger or wrist function. Red points indicate baseline measurements on each separate day, followed by three blocks of TMS NF from ECR. The shaded area in Day 1 indicates where resistance was applied by a therapist in order to evoke MEPs, as none were present at rest. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)