Show a desktop notification when the AI TA finishes replying
Notify me when classmates post messages in the forum
Play an alert sound whenever there is a new notification
A self-built multimodal physiological sensing device developed in stages: fNIRS and PPG are complete, and EEG integration is in progress. It explores real-time monitoring of cognitive load, attention and learning states.
Prototype preview · simulated signals
Uedu Brain is developed in stages: the fNIRS and PPG modules are complete, and EEG module integration is in progress. The three complementary sensing technologies capture learners' brain and physiological activity from different angles.
Use near-infrared light to penetrate the scalp and measure haemoglobin oxygenation changes in the prefrontal cortex, indirectly reflecting brain-region activation and cognitive workload.
Optical sensors detect changes in blood vessel volume to calculate heart rate and heart rate variability (HRV), which reflect autonomic nervous system activity.
Measure brain potential activity via scalp electrodes and capture brain waves in different frequency bands such as α, β, θ and δ, reflecting levels of attention, cognitive load and relaxation.
Using consumer wearables in real university classes, we streamed beat-to-beat intervals (BBI), computed HRV metrics afterwards and analyzed them at the individual level. HRV responses to classroom events varied greatly between individuals, and group averages can mask individual patterns. The Uedu Brain PPG module follows the same approach: store only BBI, compute HRV afterwards, and analyze person by person.
Proposes a framework that brings the structure of clinical reasoning to physiological signal processing: when AI interprets physiological signals, each step maps to mechanisms validated in the medical literature, and the reasoning path can be reviewed by teachers, students and researchers. Uedu Brain will adopt this as its interpretation principle, keeping readings traceable and reviewable.
This study explores deploying an EEG sleep staging model on edge devices, combined with the reasoning capabilities of large language models (LLMs), to enable real-time, privacy-preserving sleep quality assessment. Uedu Brain's EEG module integration is in progress, and this study lays the groundwork for its edge computing and local inference.