Akarsh Prabhakara

Keywords: ocular dynamics; non-contact, non-invasive, continuous sensing; sensing with radio
frequencies

Education
BA 2018, B.Tech in Electronics and Communication Engineering, National Institute of Technology Karnataka
PhD 2024, PhD in Electrical and Computer Engineering, Carnegie Mellon University

Research
Ocular dynamics provide deep insights into human cognitive workloads and psychological states. While glasses-based tracking relies on electrodes, blinks, micro-saccades, or gaze to indirectly infer mental states, neuroscience establishes pupillary response as the most direct metric for effort, stress, and emotional arousal. Currently, pupillary measurements are limited to episodic, short-term lab studies using benchtop or handheld IR cameras, missing critical daily fluctuations in a user’s cognitive and emotional state. A continuous tracker would be transformative for: (1) empowering users to avoid stress-inducing activities, (2) enabling preemptive clinical interventions by identifying patterns of degrading cognition, and (3) monitoring the longitudinal effects of clinical treatments. For example, dementia patients that receive anti-amyloid therapies are currently only examined episodically. Pupillometry, a key non-invasive indicator of amyloid accumulation, could monitor responses to such treatments on a fine-grained temporal scale, enabling adaptive therapies and revolutionizing cognitive care.

We propose a continuous monitoring smart glass wearable that (1) pioneers the use of lowpower 60 GHz mmWave radar for robust, light-invariant pupillometry, (2) leverages front-facing cameras on smart glasses and vision-language models to disentangle external lighting and semantic stimuli, from the raw pupillary dynamics, and extract true mental states. By estimating continuous mental states reliably, we can generate actionable, context-aware insights for personalized well-being and adaptive clinical care.

Google Scholar