Projects
A living lab of experiments in computational neuroscience and neurotechnology. Each project addresses a real clinical problem with testable technical approaches.
AbsenceLens
AI-Assisted EEG for Absence Epilepsy
Problem
Many children with staring spells are misdiagnosed as inattentive or having ADHD when they may have absence seizures. Traditional EEG review is time-intensive, and subtle patterns can be missed.
Approach
- →Use open EEG datasets to build pattern recognition models
- →Detect generalized spike-and-wave patterns (especially 3 Hz)
- →Highlight suspicious segments for neurologists to review
- →Build explainable visualizations to show why the model flagged each segment
Tech Stack
ICU Seizure Detection
Continuous EEG Monitoring for Critical Care
Problem
Non-convulsive seizures in ICU patients are common but often missed because continuous EEG monitoring requires constant expert review. Delayed detection worsens outcomes.
Approach
- →Detect rhythmic patterns and evolving waveforms in continuous EEG
- →Alert clinicians to potential seizure activity in real-time
- →Reduce false positives through multi-stage filtering
- →Support integration with existing ICU monitoring systems
Tech Stack
Med Tuning Assistant
Balancing Seizure Control and Cognition
Problem
Anti-seizure medications often have cognitive side effects. Finding the right dose that controls seizures without impairing memory, attention, or mood is challenging and requires careful tracking.
Approach
- →Track seizure frequency, medication changes, and cognitive metrics over time
- →Use simple models to suggest dosage adjustments
- →Visualize trends to help patients and doctors make informed decisions
- →Privacy-first, patient-controlled data
Tech Stack
Note: These are research prototypes, not clinical products. All work follows ethical guidelines and respects patient privacy.