Pseudo-Online BMI Based on EEG to Detect the Appearance of Sudden Obstacles during Walking

Abstract

The aim of this paper is to describe new methods for detecting the appearance of unexpected obstacles during normal gait from EEG signals, improving the accuracy and reducing the false positive rate obtained in previous studies. This way, an exoskeleton for rehabilitation or assistance of people with motor limitations commanded by a Brain-Machine Interface (BMI) could be stopped in case that an obstacle suddenly appears during walking. The EEG data of nine healthy subjects were collected during their normal gait while an obstacle appearance was simulated by the projection of a laser line in a random pattern. Different approaches were considered for selecting the parameters of the BMI: subsets of electrodes, time windows and classifier probabilities, which were based on a linear discriminant analysis (LDA). The pseudo-online results of the BMI for detecting the appearance of obstacles, with an average percentage of 63.9% of accuracy and 2.6 false positives per minute, showed a significant improvement over previous studies.

Analysis of Human Gait Using Hybrid EEG-fNIRS-Based BCI System: A Review

“…In many online BCI systems, IMUs are used to measure the real movement to ensure smooth operation. Elvira et al ( 2019 ) placed 7 IMUs on lumber, thigh, shin, and foot to measure the actual movement to detect stop intention. IMUs serve as feedback to ensure the real physical stop.…”

“…Due to LDA's simplicity and effectiveness, it is widely used in the classification of EEG and fNIRS signals for gait disorders (Bulea et al, 2014 ; Rea et al, 2014 ; Salazar-Varas et al, 2015 ; Naseer et al, 2016 ; Gui et al, 2017 ; Khan R. A. et al, 2018 ; Costa-Garciacutea et al, 2019 ; Elvira et al, 2019 ). Fazli et al ( 2012 ) used LDA to classify MI tasks and found that simultaneous EEG-fNIRS measurement helped increase the classification accuracy by 5%.…”

“…A pseudo-online BCI system to detect the unexpected obstacle was developed with an average accuracy of 63.9%, which can help its implementation in real-time BCI systems. Future works can help increase the accuracy of such a BCI system to make them more feasible for real-time applications (Elvira et al, 2019 ). A similar EEG-based study was performed to detect the sudden appearance of obstacles for the lower limb exoskeleton during walking with an average accuracy achieved 79.5% (Salazar-Varas et al, 2015 ).…”

Assistance Robotics and Biosensors 2019

“…In upper-limb rehabilitation, the physiological interaction of the patient with the robotic device is critical. In [7], two different physiological control methods based on EEG and electrooculography (EOG) are evaluated, while measuring stress levels from skin conductance level (SCL) and heart rate variability (HRV). This study shows that EEG control is associated with a higher level of stress and mental workload when compared to EOG control.…”

Vibration-Based Structural Health Monitoring Using Piezoelectric Transducers and Parametric t-SNE

“…Dimensionality reduction is very important, because it alleviates undesired properties of high-dimensional spaces, such as "the curse of dimensionality" [5]. In the literature, various dimensionality reduction methods have been proposed: (i) linear methods, such as principal component analysis (PCA) [6,7] and linear discriminant analysis (LDA) [8,9], and (ii) nonlinear methods, such as isometric mapping (ISOMAP) [10,11] and the non-parametric version of t-distributed stochastic neighbor embedding (t-SNE) [12].…”