Cortical-inspired Architectures
Hierarchical, recurrent patterns for efficient representation learning.
Research that illuminates the mind
Our mission is to decode neural dynamics and build AI that learns with human-like efficiency, transparency, and adaptability.
Neural Signal Decoding
Transforming fMRI, EEG, and MEG into interpretable cognitive maps.
Causal Inference in Brains
Perturbation-driven insights for mechanism-level understanding.
Neuro-symbolic Systems
Blending continuous and discrete reasoning with biological priors.
Responsible NeuroAI
Privacy-preserving learning and governance for neural data.
Neuroadaptive Interfaces
Closed-loop systems that respond to cognitive states in real-time.
Publications
-
2025 NeuroAI
Self-supervised Cortical Maps for Vision-Language Alignment
Demonstrates emergent topography consistent with primate ventral stream.
PDF -
2024 ICLR
Causal Transformers for Neural Perturbation
Predictive modeling of stimulation effects with uncertainty bounds.
PDF -
2023 NeurIPS
Privacy-Preserving Federated fMRI
Cross-institution learning with differential privacy guarantees.
PDF
