Saina Charkas is a PhD candidate at the University of Twente, working within the Biomedical Signals and Systems (BSS) group and the Stress in Action consortium. Her research focuses on developing interpretable and generalizable methods for stress monitoring in daily life using multimodal physiological, behavioral, and contextual data. She aims to bridge controlled laboratory findings and real-world applications through advanced signal processing and artificial intelligence. Her work includes designing validation pipelines for physiological stress features and building white-box and hybrid models to capture stress dynamics. With a background in Biomedical Engineering, she aims to advance wearable-based health monitoring and personalized interventions.

Research question
“How can interpretable and generalizable multimodal models be developed to reliably detect and explain acute stress in real-life settings using physiological, behavioral, and contextual data?”

Abstract
Stress monitoring in daily life requires methods that move beyond controlled laboratory settings toward robust, real-world applicability. This research focuses on developing a standardized pipeline for validating physiological stress features across individuals and contexts, ensuring interpretability and generalizability. It explores interpretable (white-box) and hybrid (gray-box) modeling approaches to uncover the dynamic physiological mechanisms underlying acute stress. By integrating multimodal data from wearable and smartphone-based sensors, including physiological, behavioral, and contextual signals, the project aims to enable continuous, passive stress detection and support personalized, real-time interventions in everyday environments.

PhD student,
University of Twente