Skip to main content

Achievement

Improvements to neural decoding alogrithm

Research Achievements

Improvements to neural decoding alogrithm

A second research achievement from the BCI group also involves improvements to a neural decoding algorithm. BioEngineering student Patrick Sadtler is working with Byron Yu to investigate an improvement to the usual Kalman filter algorithm used to map neuronal spike counts to control signals for a prosthetic device. Instead of running the Kalman filter on raw spike count data, they reduce the dimensionality of the dataset using a machine learning technique called Factor Analysis, then run the Kalman filter on the lower dimensional dataset. Preliminary results show a qualitative improvement from the new algorithm, with monkeys producing smoother and less erratic cursor movements.

SEE MORE: