Brain game based on auditory attention detection (AAD)

• Main Author: Ngiam, Jing Xiang
• Other Authors: Smitha Kavallur Pisharath Gopi
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/149262
• Institution: Nanyang Technological University

Auditory attention decoding (AAD) is promising for use in auditory-assistive devices. Electroencephalography (EEG) is a recording method of electrical brainwave activity for numerous diagnostic and research purposes. AAD can train subjects in achieving high AAD performance which would increase the application potential. This project involves the development of a real-time neurofeedback Brain Computer Interface (BCI) Unity game designed for normal hearing subjects, using EEG signals acquired from Muse 2 headset to decode the subject’s attentional state in the game. Furthermore, this project aims to explore the feasibility of building a user-friendly and functional AAD system. This entails a high temporal resolution and an automated closed-loop neurofeedback system. The experiment conducted provided subjects with visual feedback on their ongoing performance. The data were collected from subjects and the results of the classification accuracies were evaluated to assess the performance of these normal hearing subjects. The experiment results conclude that high AAD accuracies can be achieved with a trial length of 4 seconds for single and two talker conditions. In the single talker condition experiment, most of the subjects have better decoding accuracies when attending to the left ear (88%) than the right ear (74%) at 5% significance level. Also, in the two-talker condition experiment that involves dichotic listening tasks, most ofthe subjects have better decoding accuracies when attending to the right ear (79%) than the left ear (42%) at 5% significance level. This exploratory result proves that further investigations can be done with a larger sample size population using a better ergonomic design of EEG devices.