Large-scale NHP recording and stimulation systems

Published:

When I arrived at CMU, the Neuro-Mechatronics Lab had no NHP behavioral assemblies; building them was my first job. The resulting systems record awake, behaving rhesus macaques with up to 256 synchronized biopotential channels — primary motor cortex and motor thalamus (stereotrode arrays) with concurrent surface HD-EMG from extrinsic arm flexors and extensors — during a wrist-based center-out task.

The same rigs support intervention: I developed stimulation/recording interfaces for dorsal-rootlet and dorsal-root-ganglion stimulation (first in mice, then monkeys), and the analysis stack that goes with them — including a temporal state-space model that forecasts slow-wave EEG power in NHPs, a step toward state-dependent, closed-loop experiments. I moved away from NHP work in 2024 as the lab’s focus shifted to human wearable interfaces.

Selected publications