Affect Regulation Based on Brain-computer Interface Towards Treatment for Depression

In this project, we will leverage on a locally developed electroencephalograph (EEG)-based BCI technology for decoding affective states of the brain, and thereby develop a closed-loop sensing and stimulation mechanism. The technology uses advanced neural signal computing on the EEG in real-time and audio feedback using a machine learning model that associates individual user's EEG characteristics in relation to music-emotion features. The system is portable and will allow emotion regulation training to be done outside of hospital setting with ease thus potentially addressing the treatment gap for MDD in the elderly.

This is a two group randomized study. One group will undergo the intervention which is 24 sessions of BCI emotion regulation training (i.e. listening to music with audio feedback to regulate emotions towards positive affect) over an 8-week period. The control group will undergo 8 weeks of music sessions, without any emotion regulation training. Both groups will undergo pre- and post- psychometric assessments looking at cognition, quality of life, functioning and emotional states. This study will also carry out functional MRI of the brain before and after 8 weeks of either training (intervention) or music music session (control) to examine changes associated with the affective BCI training. Study findings derived from psychometric assessments, EEG analysis and neuroimaging will provide evidences on efficacy and usability of this technology.

Inclusion criteria

• English-speaking
• Geriatric Depression Scale score of 4 and below
• Fit to provide informed consent
• Able to travel to study site independently

Exclusion criteria

• Presence of hearing impairment
• Presence of any known neuropsychiatric disorders (e.g. dementia, epilepsy or mental retardation)
• Contraindications for MRI such as surgical clips, cardiac pacemakers, metal implants and plates, orthodontics and claustrophobia