Year: 2024
Type: conference-paper
Abstract: Over 1 billion people worldwide are living with the inconvenience caused by disabilities. With the rapid development of artificial intelligence technology, human-machine interface technology can lead to a qualitative breakthrough in helping people with mobility difficulties. Although the related research aims to understand the users’ action needs through an electroencephalogram (EEG), too many electrode points are currently used, limiting this technology's application. Indeed, finding the key electrode points will ease the transformation of EEG research into a real-world application. Studying the key electrode points in AIEEG (Action Idea Electroencephalogram) is significant in achieving this goal. Hence, this study used the LSR (least squares regression) algorithm to calculate the feature coefficients of 16-channel data and conducted grouping experiments based on the characteristics of the feature coefficients. Then, used the Conv1d-GRU model to train and verify the collected data of the electrode points corresponding to each group of characteristic coefficients. The verification results demonstrate that four electrode points are required in the AIEEG to collect data and identify five action ideas. Relying on these four points provides a testing accuracy of 99%.
Author: Guoxia Zou
Institution: Guangdong Open University
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: bronze
OpenAlex ID: W4409869003