Impacts of imagined lexical tone on Mandarin speech imagery BCI performance

Zengzhi Guo1; Hewen Zhang2; Fei Chen2

10.1109/NER52421.2023.10123911

2023

11th International IEEE EMBS Conference on Neural Engineering (IEEE/EMBS NER)

1948-3546

978-1-6654-6293-8

2023 11th International IEEE/EMBS Conference on Neural Engineering (NER)

2023-April

1-4

24-27 April 2023

Baltimore, MD, USA

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE

Speech imagery brain computer interfaces (BCIs) empower people with motor disorders to communicate their thoughts and intentions to their environment in a natural, userfriendly way. Due to its portability, non-invasiveness, safety, and low cost, functional near-infrared spectroscopy (fNIRS) is advanced as a brain imaging method for developing BCIs. To improve operators' speech imagery BCI performance, a new paradigm was studied in this work by modifying imagined lexical tone in Mandarin speech imagery BCIs motivated by the fact that speech processing with perceptual pitch variation exhibited more extensive activations in brain than that without perceptual pitch variation. For the traditional ternary yes and no Mandarin speech imagery paradigm, the subjects were asked to answer questions by covertly repeating the Mandarin syllable '(sic)' (`yes' in English, and with tone 4 in Mandarin) or '(sic)' (`no' in English, and with tone 3 in Mandarin) in mind and an unconstrained rest task was also contained. In the proposed paradigm, the subjects were asked to imagine '(sic) ' with tone 1 and '(sic)' with tone 3. Compared with the traditional paradigm, tone difference in Mandarin speech imagery of the proposed paradigm could generate more discriminative brain activities among different tasks, and the mean classification accuracy was significantly improved from 49.06% to 53.85%. These results suggest that modifying imagined lexical tone in Mandarin speech imagery could influence brain activation and is promising for improving the decoding accuracy of Mandarin speech imagery BCIs.

brain computer interface functional near-infrared spectroscopy speech imagery tone

其他

英语

CPCI-S ; EI

National Natural Science Foundation of China[61971212]

Computer Science ; Engineering ; Neurosciences & Neurology

Computer Science, Artificial Intelligence ; Engineering, Biomedical ; Neurosciences

WOS:001009053700193

20232314178134

Brain ; Brain mapping ; Chemical activation ; Image enhancement ; Infrared devices ; Near infrared spectroscopy ; Speech processing

Biomedical Engineering:461.1 ; Computer Peripheral Equipment:722.2 ; Imaging Techniques:746 ; Speech:751.5 ; Chemical Reactions:802.2 ; Chemical Products Generally:804

IEEE

1.School of Electronics and Information Engineering, Harbin Institute of Technology, Harbin, China
2.Department of Electrical and Electronic Engineering, Southern University of Science and Technology, Shenzhen, China

Zengzhi Guo,Hewen Zhang,Fei Chen. Impacts of imagined lexical tone on Mandarin speech imagery BCI performance[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE,2023:1-4.