Neural Networks architectures and it's applications in EEG signal classification for BCI

Authors

DOI:

https://doi.org/10.5335/rbca.v14i1.13070

Keywords:

Activation Functions, Artificial Neural Networks, Brain Computer Interface, EEG

Abstract

A lot of people around the world suffer from some kind of motor illness that interferes with their daily lifes. One of the ways to help to improve the life of such people is the so called Brain Computer Interface. However, this method has some room to improve in its accuracy.
This Article seeks to explore and to compare the large amount of Neural Networks Architectures for classification of EEG signals for Brain Computer Interfaces, using even the less-known Complex-Valued Networks, and try with new activation functions . The methodology of this work involves preprocessing of labeled-EEG signals, splitting the frequency components of the signal in bands of frequency based on the brainwaves frequencies, defined as delta (0.5-4HZ), theta (4-8HZ), alpha (8-13HZ), and beta (above 13HZ) The resulting time frames will then be used to feed the several evaluated-to-be architectures.

Downloads

Download data is not yet available.

Published

2022-04-11

Issue

Section

Original Paper

How to Cite

[1]
2022. Neural Networks architectures and it’s applications in EEG signal classification for BCI. Brazilian Journal of Applied Computing. 14, 1 (Apr. 2022), 55–69. DOI:https://doi.org/10.5335/rbca.v14i1.13070.