Data pipelines

Reliable analysis of neurophysiology experiments starts long before statistics: raw recordings have to be organized, synchronized, and processed the same way every time. During my doctoral work in the Cortical Plasticity Lab at KUMC, I built and maintained the data-processing pipelines — centered on the nigeLab MATLAB toolbox — that turned daily recordings from the stroke-recovery reaching study into analysis-ready datasets, with pipelines running from “the edge” (acquisition hardware) to “the cloud.”

Related tools and resources:

• Finite-state-machine (FSM) spike detector — a low-latency state machine for detecting spikes and triggering activity-dependent stimulation in real time, built by improving an open-source commercial (Intan) system: Murphy & Buccelli et al., Journal of Neural Engineering 2019.
• code.nml.wtf — code and documentation for the Neuro-Mechatronics Lab at CMU, where the current generation of these pipelines lives.
• Neuro-Mechatronics-Interfaces on GitHub — the lab’s open-source code organization.