PhD defence of Shibam Debbarma – Wearable Flexible Biopotential Measurement Systems

Abstract

Smart wearables can provide clinical standard health monitoring services in our day-to-day life. One such wearable application is smart headband/eye-mask based biopotential monitoring systems, which employ electroencephalography (EEG) and electrooculography (EOG) techniques for sleep assessment. However, most of the commercial biopotential wearables are rigid, bulky, and uncomfortable to wear.

In this research, we propose a wearable for EOG measurement which is implemented on flexible Polyimide PCB with integrated printed gold contact electrodes and the biopotential acquisition. The wearable can be easily integrated with headbands/eye-masks. The system performance is validated using a MATLAB based algorithm for the detection of different eye activities such as blinks, winks, and eye movements. But this prototype requires electrode gel for better EOG detection and is sensitive to random motion artifacts. To overcome these limitations, the EOG wearable design is redesigned with parallel non-contact (or capacitive) electrode pairs, which have better sensitivity and do not require gel for EOG detection. The parallel electrode pairs are configured differentially for sensing and reducing motion artifacts during EOG measurement. The proposed wearable is then validated for acquiring EOG signals in the presence/absence of motion. However, forehead/eye-masks based wearables are still uncomfortable to wear and prone to displacements due to movements.

In recent studies, EEG signals are successfully acquired intra-orally, with the help of oral appliances like mandibular advancement devices (MADs). Here, we propose a smart MAD which integrates flexible EEG electrodes, accelerometer, and the measurement unit. The system can measure intra-oral EEG and motion signals (such as tongue movements, teeth grinding, and gulping) simultaneously. An EMD-ICA based algorithm is also developed in MATLAB to identify the motion corrupted EEG segments using the accelerometer data and denoise them accordingly. The smart MAD and the proposed algorithm are also validated for detecting ‘eye open’ and ‘eye close’ activities from the acquired intra-oral EEG spectrums, in the presence/absence of intra-oral motions. This smart MAD system for intra-oral EEG can be a potential alternative solution to headband/eye-masks based wearables.