I develop cortical brain-computer interfaces
How can we leverage the statistical structure in cortical activity to build neural interfaces with higher performance and reliability? My work uses novel computational methods both to develop neurotechnology, and add to our scientific understanding of the principles that underlie activity in motor cortex.
See my Google Scholar Page for a full list of publications.
Recent lines of work:
Decoding from latent population dynamics
We demonstrated excellent online BMI control of a cycling task using latent neural features. This novel approach has implications for generalized online control across a broad range of movements.
Schroeder KE*, Perkins SM*, Wang Q, and Churchland MM. Robust neural control of virtual locomotion enabled by a novel decoding strategy. J Neurosci https://doi.org/10.1523/JNEUROSCI.2687-20.2021
Decoding sensorimotor finger signals for dextrous control
Previous BMI studies had primarily focused on predicting and controlling whole-arm movements. Here, we investigated the continuous decoding of precise finger movements in rhesus macaques using a novel task in which they are required to bend the fingers according to a visual cue.
We also quantified tactile sensory responses within the same M1 cortical populations used for motor BMI control, showing that M1 neurons are tuned to specific tactile fingertip inputs in both nonhuman primates and humans. These findings help to explain challenges encountered by state-of-the-art decoders and suggest avenues for improvement.
Schroeder, K.E., Irwin, Z.T., Bullard, A.J., et al. Robust tactile sensory responses in finger area of primate motor cortex relevant to prosthetic control. Journal of Neural Engineering
Irwin, Z.T., Schroeder, K.E., Vu P.P., et al. Neural control of finger movement via intracortical brain–machine interface. Journal of Neural Engineering
Understanding sensorimotor integration
The neural mechanisms of anesthetic-induced unconsciousness have yet to be fully elucidated, in part because of the diverse molecular targets of anesthetic agents. We demonstrated, using intracortical recordings in macaque monkeys, that information transfer between structurally connected cortical regions is disrupted during anesthesia, despite preserved primary sensory representation.
Schroeder, KE, Irwin, ZT, Gaidica, M, et al. (2016) Disruption of corticocortical information transfer during ketamine anesthesia in the primate brain. NeuroImage
Willsey M, Nu CS, Nason SR, Schroeder KE, …, Chestek CA. Neural Dynamics in Primate Cortex During Exposure to Subanesthetic Concentrations of Nitrous Oxide, eNeuro