Imagined Speech Classification Accuracy and the Signal Acquisition Procedure

Mohammad Khani, Gholam Reza; Asghari Bejestani, Mohammad Reza

doi:10.22034/ijiems.2023.399363.1061

Imagined Speech Classification Accuracy and the Signal Acquisition Procedure

Document Type : Original Article

Authors

Gholam Reza Mohammad Khani ¹

Mohammad Reza Asghari Bejestani ²

¹ Electrical & IT department, Iranian Research Organization for Science and Technology (IROST), Tehran, Iran

² Electrical & IT department, Iranian Research Organization for Science and Technology (IROST), Tehran, Iran

10.22034/ijiems.2023.399363.1061

Abstract

Imagined speech recognition is one of the most interesting approaches to BCI development. A lot of works have been done in this area. Many different experiments have been designed and hundreds of combinations of feature extraction methods and classifiers have been examined. Reported classification accuracies range from the chance level to more than 90%. Based on non-stationary nature of brain signals, we have introduced 3 classification modes according to time difference in inter and intra-class samples. The modes can explain the diversity of reported results and predict the range of expected classification accuracies from the brain signal accusation procedure. In this paper, a few samples are illustrated by inspecting results of some previous works. It has been shown that, in current state of art researches on imagined word classification, if signal accusation schema falls in the mixed time mode, the accuracy can reach to more than 90 percent's, but for more realistic short and long time modes, it's hard to attain good results.

Keywords

EEG

Imagined word

classification