Skip to main content
2024 47th International Conference on Telecommunications and Signal Processing (TSP)

Full Program »
Video (.mp4)
View File

Imagined Speech Recognition and the Role of Brain Areas Based on Topographical Maps of EEG Signal

Nowadays, brain-computer interface (BCI) technologies are oriented toward developing an intuitive and effective system for decoding speech-related processes from brain activity data. Electroencephalography (EEG) is the most often used method for recording the electrical waves of the human brain from the outer scalp. However, there are several challenges related to the nature of EEG signals, such as non-linearity, non-stationary and low signal-to-noise ratio (SNR). Hence, decoding imagined speech EEG signals and classifying several imagined words is challenging. In this paper, we propose an approach based on topographic maps from the raw EEG data of imagined speech to distinguish between several imagined English words using CNN deep learning models. Furthermore, we investigated the role of brain hemispheres and lobes in distinguishing between imagined words based on topographic images created from EEG data. The experiment results show the effectiveness of using EEG topographic images to identify imagined speech based on subject-dependent approaches with an average accuracy of 76%. Moreover, the results indicate that the left hemisphere of the brain is primarily engaged when right-handed individuals imagine words. Also, the brain’s frontal lobe contributes to a high level of success in discriminating between imagined speech based on EEG signals.

Yasser F. Alharbi
King Saud University
Saudi Arabia

Yousef Ajami Alotaibi
King Saud University
Saudi Arabia


Privacy Policy

Powered by OpenConf®
Copyright ©2002-2024 Zakon Group LLC