Brain organoid MEA electrophysiology is a rapidly growing field with no established analysis standards. Different groups use different metrics, thresholds, and algorithms — making cross-study comparison difficult. We present Axon (v6.3), an open-source Python pipeline for end-to-end analysis of NWB-formatted brain organoid MEA recordings with direct DANDI Archive integration. Axon computes per-unit firing rate, inter-spike interval statistics, coefficient of variation, Spike Time Tiling Coefficient (STTC) cross-unit synchrony matrices, adaptive network burst detection, and graph-theoretic network topology via a single command-line interface. Version 6.3 adds full SpikeInterface integration for spike sorting of raw high-density MEA data, supporting four algorithms (mountainsort5, spykingcircus2, tridesclous2, simple) with DC-corrected RMS-based channel selection and per-unit quality metrics (SNR, ISI violation ratio, firing rate). We validate the pipeline on four public DANDI datasets spanning 29 subjects: DANDI:001603 (human and mouse brain organoids and ex vivo mouse cortex), DANDI:001132 (human hippocampal organoids, HD-MEA raw signal), DANDI:000774 (in vitro MEA), and DANDI:001611 (chronic dissociated cortical culture). All analysis modules are dataset-agnostic. Source code is available at github.com/metin/axon.
Metin (Mon,) studied this question.