3D fNIRS Simulator Generates Unlimited Synthetic Brain Recordings
Open-source simulator tackles fNIRS data scarcity with high-fidelity Monte Carlo modeling.
Deep Dive
A new high-fidelity 3D fNIRS simulator uses mesh-based Monte Carlo simulations to generate unlimited labeled synthetic brain recordings. It combines anatomically accurate sensitivity profiles with parameterized models of hemodynamic responses, systemic physiology, and nonsystematic artifacts. Validated on open-source finger-tapping, pain-assessment, and surgical-skill datasets, this open-source tool enables testing denoising algorithms, data augmentation, mechanistic modeling, and in silico experimentation.
Key Points
- Uses mesh-based Monte Carlo simulations for high-fidelity synthetic fNIRS data generation.
- Models hemodynamic responses, systemic physiology, and nonsystematic artifacts for realism.
- Validated on three open-source datasets: finger-tapping, pain-assessment, and surgical-skill.
Why It Matters
Provides abundant labeled synthetic data to train AI models for fNIRS analysis, overcoming dataset scarcity.