Current Projects:
[Myant Inc.] Improving sleep staging using novel Heart Rate Variability features : As part of this project, I am working on improving performance over the current sleep staging algorithms by making use of newer HRV and circadian variability features along with better normalization methods and a more robust machine learning pipeline.
[Myant Inc.] Developing signal quality metrics for PPG signals: PPG analysis requires signals of high quality and a robust quality metric desirable. I am exploring both simple low complexity algorithm for online PPG analysis as well as more robust high complexity methods for PPG analysis.
[Myant Inc.] In-the-wild fatigue detection pilot using smart textiles: I ran an internal pilot for prediction of mental fatigue using a protocol which captures the effective of fatigue throughout the day when the participants are i) relax in sitting position b) performing cognitively engaging tasks. The subjective fatigue levels were then predicted using heart rate variability, activity levels and skin temperature features collected from smart underwear devices.
[Myant Inc.] Development of EEG quality metric: I am exploring quality metric for EEG analysis which could provide an idea of signal being recorded vs noise due to malfunction of device.
[2021- ] Frailty detection in geriatric population using 12-lead ECG and deep learning: This project is in collaboration with Dr. Jonathan Afilalo of Jewish General Hospital, Montreal and McGill University. We are trying to explore the possibility of frailty detection using ECG morphological features. We are exploring different 12-lead ECG input representations as well as integration of contextual information such as quality and PQRST wave points into neural network to improve learning and performance of models.
Older Projects:
[2020-2021] Remote detection of COPD and COVID using speech and heart rate data collected from Smartwatches: This project was in collaboration with Prof. Eyal de Lara, University of Toronto. I performed analysis on low-resolution smartwatch data collected from patients suffering from Chronic Obstructive Pulmonary Disease (COPD). I was able to predict the severity of COPD symptoms and exacerbation events from first few days of HRV and activity data collected from participants.
[2017-2020] Stress and anxiety monitoring of nurses in real-world conditions (TILES project): This IARPA funded project was in collaboration with Prof. Srikanth Narayanan, University of Southern California which was focused on predicting job performance and various other aspects of hospital staff (specifically nurses) using physiological, activity, social media and other data. My role in the project was working with heart rate data collected during daily shifts of hospital staff using smart-shirts and test new Heart Rate Variability features for noise robust prediction of these indices.
[2017-2021] Mental workload and stress assessment in ambulatory conditions (CRD Project): This project was in collaboration with Thales and University of Laval. The project is focused on assessing mental workload and stress in a more real-world setting where individuals take part in physical activity while completing mental workload and stress inducing tasks. As part of the project we collected three (now publicly available) datasets and derived various techniques for improving ambulatory psychological state detection.
[2015-2016] Studying trends in mental fatigue using Electroencephalography (EEG) measurements: This work was done as part of my master's thesis at IIT Kharagpur under Prof. Aurobinda Routray. I studied various EEG parameters including band energies , ratios and entropy under the influence of mental fatigue induced by simulated driving experiments. Additionally, I performed functional connectivity analysis using noise-robust motif synchronization algorithm and studied the effect of mental fatigue on various network parameters.