First-author iEEG research · Speech BCI
Aaron Earle-Richardson
I build the analysis behind intracranial-EEG speech research: GPU-accelerated tensor decomposition in PyTorch, the high-gamma signal-processing pipeline that feeds it, and the decoders that read word identity back out. First author on an in-preparation paper resolving the parallel neural sub-processes that link hearing a word to saying it.
Currently seeking neural signal processing / BCI engineering roles — open to remote or the NC Research Triangle.
Flagship — first-author research
Neural Sub-Processes of Speech
First-author iEEG study resolving the parallel neural sub-processes that link hearing a word to saying it — using a GPU-accelerated PyTorch tensor decomposition (sliceTCA) over intracranial recordings from 31 patients.
Read about the researchMore projects
All projects →-
IEEG_Pipelines
Author and lead maintainer of the Cogan Lab's open-source iEEG/ECoG analysis toolkit — a PyPI package with CUDA kernels, Cython hot paths, and a parallel MATLAB API on its own CI.
-
GridQuest
A spatial-selection input game that measures achieved information transfer rate the way BCI decoders are benchmarked.
-
Visualizers
Two small from-scratch tools for building probability and signal-processing intuition by hand — exact dice-pool distributions, and a live time-domain/frequency-domain wave explorer.