Abstract
One of the most exciting areas of rehabilitation research is brain-controlled prostheses, which translate electroencephalography (EEG) signals into control commands that operate prostheses. However, the existing brain-control methods have an obstacle between the selection of brain computer interface (BCI) and its performance. In this paper, a novel BCI system based on a facial expression paradigm is proposed to control prostheses that uses the characteristics of theta and alpha rhythms of the prefrontal and motor cortices. A portable brain-controlled prosthesis system was constructed to validate the feasibility of the facial-expression-based BCI (FE-BCI) system. Four types of facial expressions were used in this study. An effective filtering algorithm based on noise-assisted multivariate empirical mode decomposition (NA-MEMD) and sample entropy (SampEn) was used to remove electromyography (EMG) artifacts. A wavelet transform (WT) was applied to calculate the feature set, and a back propagation neural network (BPNN) was employed as a classifier. To prove the effectiveness of the FE-BCI system for prosthesis control, 18 subjects were involved in both offline and online experiments. The grand average accuracy over 18 subjects was 81.31 ± 5.82% during the online experiment. The experimental results indicated that the proposed FE-BCI system achieved good performance and can be efficiently applied for prosthesis control.
Highlights
With the increase in the number of disabled persons with amputations or spinal cord injuries, many studies have focused on the development of prosthetic technology to restore lost motion function (Ziegler-Graham et al, 2008)
To verify the performance of EMG artifacts removed by noise-assisted multivariate empirical mode decomposition (NA-MEMD) combined with sample entropy (SampEn), the comparisons between the original EEG signals and the artifact-attenuated EEG signals were conducted associated with the grand average Furrowing Brow in FC5 and FC6
It is observed that the frequency bands higher than 30 Hz were mainly located in the components of Intrinsic mode functions (IMFs) 4 - IMF 5 and their SampEn values were significantly higher than 0.45
Summary
With the increase in the number of disabled persons with amputations or spinal cord injuries, many studies have focused on the development of prosthetic technology to restore lost motion function (Ziegler-Graham et al, 2008). Several types of prosthesis have been developed, ranging from passive cosmetic prostheses to body-powered limbs, from EMG-based prostheses to EEG-based prostheses (Lee et al, 2014). The earliest prostheses were passive cosmetic devices, which can only help a person seem less awkward in social situations but not change posture (Cordella et al, 2016). Body-powered prostheses gradually replaced the passive cosmetic prostheses due to their simple design and effectiveness. The shortcoming of these types of prosthesis is that they can only control one joint at a time by mechanical linkage (Kistenberg Robert, 2014).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
