Paving the Way for Motor Imagery-Based Tele-Rehabilitation through a Fully Wearable BCI System

Abstract

The present study introduces a brain-computer interface designed and prototyped to be wearable and usable in daily life. Eight dry electroencephalographic sensors were adopted to acquire the brain activity associated with motor imagery. Multimodal feedback in extended reality was exploited to improve the online detection of neurological phenomena. Twenty-seven healthy subjects used the proposed system in five sessions to investigate the effects of feedback on motor imagery. The sample was divided into two equal-sized groups: a "neurofeedback" group, which performed motor imagery while receiving feedback, and a "control" group, which performed motor imagery with no feedback. Questionnaires were administered to participants aiming to investigate the usability of the proposed system and an individual's ability to imagine movements. The highest mean classification accuracy across the subjects of the control group was about 62% with 3% associated type A uncertainty, and it was 69% with 3% uncertainty for the neurofeedback group. Moreover, the results in some cases were significantly higher for the neurofeedback group. The perceived usability by all participants was high. Overall, the study aimed at highlighting the advantages and the pitfalls of using a wearable brain-computer interface with dry sensors. Notably, this technology can be adopted for safe and economically viable tele-rehabilitation.

Keywords: brain–computer interface; dry sensors; electroencephalographic sensor; motor imagery; neurofeedback; tele-rehabilitation.

Figure 1

A subject using the proposed BCI system with neurofeedback in extended reality. The system involves EEG acquisition with the Helmate device, online processing, and actuators for visual–haptic feedback delivery.

Figure 2

Position on the scalp of the sensors adopted in this study. Locations are identified by the 10–20 standard system for EEG.

Figure 3

Timing of a single trial of the experimental sessions for the control group. The same timing was also used for the neurofeedback group only during the first phase of an experimental session. Notably, there was an overlap of $0.25$ $\mathrm{s}$ between the cue and the word “GO!”.

Figure 4

Timing of a single trial of the experimental sessions for phase 2 of the neurofeedback group. The same timing of the control group was used for phase 1.

Figure 5

Control group: mean classification accuracy using the best 2-second window.

Figure 6

Neurofeedback group: mean classification accuracy using the best 2-second window.

Figure 7

NASA-TLX results for both control and neurofeedback groups.

Figure 8

Time/frequency maps for a poorly performing subject from the neurofeedback group: (a) left hand imagery; (b) right hand imagery. The channels C3 and C4 are taken into account. Event-related desynchronization is depicted in red and event-related synchronization in blue.

Figure 9

Time/frequency maps associated with the best accuracy result of the same subject from Figure 8: (a) left hand imagery; (b) right hand imagery. The channels C3 and C4 are taken into account. Event-related desynchronization is depicted in red and event-related synchronization in blue.

Figure 10

Time/frequency maps associated with a subject of the control group reaching high classification accuracy: (a) left hand imagery; (b) right hand imagery. The channels C3 and C4 are taken into account. Event-related desynchronization is depicted in red and event-related synchronization in blue.