How closed-loop interfaces can test and shape neural computation
Summary
Neural control is a forward-looking direction in our lab: we want to close the loop between neural activity, computation, and feedback, using interfaces that adapt in real time and experiments that turn correlational theories into causal tests. Our work on multi-regional signal transmission identifies structured pathways that could be targeted by feedback or perturbation, our study of state-dependent sensory encoding defines changes that a robust controller must accommodate, and our work with ultra-high-density Neuropixels probes advances the measurements needed for precise interfaces. We now ask which neural features remain stable enough for control, how decoder adaptation and user learning should be coordinated, and whether closed-loop perturbations can distinguish competing theories of neural coding and routing.
From the lab
Foundational publications from our group
Neural control is a forward-looking lab direction. These studies establish relevant principles of signal transmission, state dependence, and measurement rather than completed closed-loop BCI studies from our group.
Deciphering neuronal variability across states reveals dynamic sensory encoding
Defines state-dependent changes that a robust real-time decoder or controller must accommodate.
Ultra-high-density Neuropixels probes improve detection and identification in neuronal recordings
Advances the recording fidelity and unit identification needed for precise neural interfaces.
Multi-regional module-based signal transmission in mouse visual cortex
Identifies structured routes through which targeted feedback or perturbation could influence visual processing.
Gamma and the coordination of spiking activity in early visual cortex
Establishes how population timing relates to the coordination of spikes under different visual conditions.