Motor imagery-based neuro-feedback system using neuronal excitation of the active synapses.

Citation metadata

From: Annales des Telecommunications(Vol. 76, Issue 7-8)
Publisher: Springer
Document Type: Report; Brief article
Length: 260 words

Document controls

Main content

Abstract :

Keywords: Electroencephalogram (EEG); Embedded system; Neuronal excitation; Neuro-feedback; Reactive frequency band (RFB); Reverse filtration (RF) and short-time cepstrum Abstract Neuronal excitation enables identifying the features of an electroencephalogram (EEG) signal for motor imagery detection. We propose a novel feature extraction algorithm supported by short-term cepstrum-based inverse filtration of neuronal excitation of the active synapse. The maximum power of the estimated neuronal excitation is subjected to a two-class Bayesian probabilistic classifier. The feature extraction algorithm with the Bayesian probabilistic classifier significantly improves the brain--computer interface performance compared with that of other conventional methods of EEG signal processing such as wavelet with a Bayesian classifier, autocorrelation and CSP filter with a naïve Bayes classifier over the BCI competition II and IV datasets. Consequently, this neuronal excitation feature allows the authors to develop a motor imagery neuro-feedback system the performance of which achieves 87.2% average classification accuracy, which is 14% greater than that of the wavelet-based algorithm and 6.2% greater than that of the TRSP-based algorithm, with 53 ms of processing time allotted for each instruction in a real-time experiment. However, brain signal variation across different subjects and sessions significantly impairs decision accuracy. Our neuronal excitation base feature extraction algorithm minimizes these variations in classification accuracy. Author Affiliation: (1) BIT Mesra, Ranchi, India (2) Cambridge Institute of Technology, Ranchi, India (3) The PNG University of Technology, Lae, Morobe, Papua New Guinea (4) Federal University of Piauí (UFPI), Teresina, Piauí, Brazil (5) Instituto de Telecomunicações, Aveiro, Portugal (c) Article History: Registration Date: 11/13/2019 Received Date: 05/25/2019 Accepted Date: 11/13/2019 Online Date: 12/12/2019 Byline:

Source Citation

Source Citation   

Gale Document Number: GALE|A667937594