Unimanual Versus Bimanual Motor Imagery Classifiers for Assistive and Rehabilitative Brain Computer Interfaces

Abstract

Bimanual movements are an integral part of everyday activities and are often included in rehabilitation therapies. Yet electroencephalography (EEG) based assistive and rehabilitative brain computer interface (BCI) systems typically rely on motor imagination (MI) of one limb at the time. In this study we present a classifier which discriminates between uni-and bimanual MI. Ten able bodied participants took part in cue based motor execution (ME) and MI tasks of the left (L), right (R) and both (B) hands. A 32 channel EEG was recorded. Three linear discriminant analysis classifiers, based on MI of L-B, B-R and B-L hands were created, with features based on wide band Common Spatial Patterns (CSP) 8-30 Hz, and band specifics Common Spatial Patterns (CSPb). Event related desynchronization (ERD) was significantly stronger during bimanual compared to unimanual ME on both hemispheres. Bimanual MI resulted in bilateral parietally shifted ERD of similar intensity to unimanual MI. The average classification accuracy for CSP and CSPb was comparable for L-R task (73±9% and 75±10% respectively) and for L-B task (73±11% and 70±9% respectively). However, for R-B task (67±3% and 72±6% respectively) it was significantly higher for CSPb (p=0.0351). Six participants whose L-R classification accuracy exceeded 70% were included in an on-line task a week later, using the unmodified offline CSPb classifier, achieving 69±3% and 66±3% accuracy for the L-R and R-B tasks respectively. Combined uni and bimanual BCI could be used for restoration of motor function of highly disabled patents and for motor rehabilitation of patients with motor deficits.

Global Cortical Network Distinguishes Motor Imagination of the Left and Right Foot

“…Active training with autonomous involvement has been made possible by monitoring electromyography (EMG) signals in the affected foot. Recent interest in upper limb rehabilitation has included electroencephalography (EEG) monitoring, including the proposal of several motor imagery (MI) based rehabilitation systems [6], [7], [8]. A clinical study showed substantial improvement of arm motor function in an MI rehabilitation group, compared to a passive control group [9].…”

Decoding Continual Muscle Movements Related to Complex Hand Grasping from EEG Signals

“…In offline decoding, MAP-CNN outperformed the best conventional method with the improvement of 20.9% and 8.6% in ME and MI, respectively, for four classes including no-movement. i) CSP [31], [32]: Features and extracted by CSP on EEG data of time window 0-4 s at Hz band. LDA was used to enable the separation of the data into classes.…”

Feasibility and Safety of Bilateral Hybrid EEG/EOG Brain/Neural–Machine Interaction

“…This problem can be solved either by directly applying multiclass methods, such as naive Bayesian classifiers (Suk and Lee, 2012;Zhang et al, 2015) or multilayer perceptrons (Balakrishnan and Puthusserypady, 2005), or, as more commonly used, by decomposing the problem into several binary classifications (Lotte et al, 2018). There are different possible decomposition methods, e.g., pairwise classification (Vuckovic et al, 2018) or by hierarchical classification (Dong et al, 2017;Gundelakh et al, 2018). However, all studies have relatively low binary classification accuracies in common ranging from 50 to 70%.…”