An EEG-based brain-computer interface for gait training

Abstract

This work presents an electroencephalography (EEG)-based Brain-computer Interface (BCI) that decodes cerebral activities to control a lower-limb gait training exoskeleton. Motor imagery (MI) of flexion and extension of both legs was distinguished from the EEG correlates. We executed experiments with 5 able-bodied individuals under a realistic rehabilitation scenario. The Power Spectral Density (PSD) of the signals was extracted with sliding windows to train a linear discriminate analysis (LDA) classifier. An average classification accuracy of 0.67±0.07 and AUC of 0.77±0.06 were obtained in online recordings, which confirmed the feasibility of decoding these signals to control the gait trainer. In addition, discriminative feature analysis was conducted to show the modulations during the mental tasks. This study can be further implemented with different feedback modalities to enhance the user performance.

Effects of tDCS on Real-Time BCI Detection of Pedaling Motor Imagery

“…It should be noted that the real-time accuracies reported in this study are taking into consideration every single prediction during the experiment (every single epoch is classified). Unfortunately, the current literature on real-time BCIs is somewhat disperse in the way the results are reported [ 26 , 30 , 33 , 34 , 35 , 36 , 37 ], but whenever it is possible to compare, our results coincide rather well with those of the literature. Our results are consistent with those of Zich et al [ 33 ] which have a real-time accuracy of 55–65%, [ 26 ] with around 65% accuracy (of first 30 sessions), and [ 30 ] with around 65% accuracy.…”

“…Unfortunately, the current literature on real-time BCIs is somewhat disperse in the way the results are reported [ 26 , 30 , 33 , 34 , 35 , 36 , 37 ], but whenever it is possible to compare, our results coincide rather well with those of the literature. Our results are consistent with those of Zich et al [ 33 ] which have a real-time accuracy of 55–65%, [ 26 ] with around 65% accuracy (of first 30 sessions), and [ 30 ] with around 65% accuracy. Meanwhile, Guger et al [ 34 ] reported the results of the best time point of the best session of each of their three subjects (98%, 93% and 87%), but a careful analysis of their data shows the average real-time classification accuracy is about 80%, 65% and 65% for their three subjects respectively.…”

“…Prasad et al [ 35 ] averaged the maximum classification accuracies of each task and report them to be 60–75%. All of these studies involve upper limb or simple foot movements with the exception of the study by Liu et al [ 30 ], which involved gait. Therefore, our results actually do not stray far from those found in the literature.…”

“…Motor imagery has been detected using EEG-based BCIs in the past, but most studies focused on upper limbs or simple foot movements [ 26 , 27 , 28 , 29 ]. Much fewer studies concentrated on lower limb complex tasks such as gait or pedaling [ 30 ]. In most of these studies, BCIs have exploited in some way the fact that there is a suppression of the mu waves (8–12 Hz) and beta waves (13–30 Hz) around M1 when a motor task is being imagined [ 31 , 32 ].…”

“…In most of these studies, BCIs have exploited in some way the fact that there is a suppression of the mu waves (8–12 Hz) and beta waves (13–30 Hz) around M1 when a motor task is being imagined [ 31 , 32 ]. The literature involving real-time processing and feedback of BCI signals associated to these types of movements is scarcer, and the methods of reporting results are disperse [ 26 , 30 , 33 , 34 , 35 , 36 , 37 ]. Nevertheless, there are many relevant applications of detecting lower limb movement in real time.…”