My-ndful buddy: an interactive, personalised mental wellness aide

• Main Author: Low, Calvin Soo Yee
• Other Authors: Guan Cuntai
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering
• Online Access: https://hdl.handle.net/10356/165984
• Institution: Nanyang Technological University

People across the world, including Singapore, experience high levels of mental stress and anxiety every day. Prolonged exposure to such conditions can be detrimental to their mental and physical health. Different mindfulness practices and mobile applications aim to alleviate stress and improve mental well-being. Still, they cannot meet the unique and diverse needs of individuals. Furthermore, mindfulness techniques can be hard to cultivate, and existing applications often lack personalisation. To fill this gap, we incorporate real-time EEG neurofeedback and personalised music to enhance the efficacy of the relaxation technique. Our online survey results showed most people are not familiar with mindfulness practices, though aware of stressful conditions, but listen to music every day. We hypothesise that listening to personalised music while performing mindfulness activities can improve relaxation levels. Our application will adapt the music and meditation guidance according to the user’s cognitive and relaxation states, enabling tailored anti-stress remedies to their specific needs. We designed an experiment comprising attention, affect and arousal calibration tasks with five music listening sessions with varying music features according to the valence-arousal complexity model. Using our proposed experiment protocol, we collected multimodal data from 32 subjects: EEG, eye gaze, behavioural and physiological data. Initial evaluations and participants’ feedback show that most participants feel more relaxed after the experiment. With these initial promising results, we are currently analysing those multimodal data using shallow and deep learning methods with inferential statistics to validate our research hypotheses.