PHASOR: An Open-Source Pipeline for Hidden Markov Model Analysis of Stimulus-Locked Retinal Multielectrode-Array Population Activity

PHASOR decodes stimulus-locked retinal population states with near one-to-one correspondence to seven generative states in synthetic data, selecting model orders from seven to nine (mean 7.9).

• Why it matters: Linking recurring population activity to stimulus timing and single-unit response classes requires multiple coordinated steps, and existing methods are not packaged into a reusable workflow.
• What they did: The pipeline aligns neural activity to the stimulus clock, classifies responses via Fourier-based methods, fits hidden Markov models with Chow-Liu tree emissions, and selects model order, then visualizes states.
• The result: In ten synthetic datasets, decoded states matched the generative count exactly in three cases and showed a single extra low-occupancy state; applied to ten ex vivo recordings (including rd1 and L-AP4) it produced stimulus-phase/contrast-dependent response classes and state occupancy/dwell differences.

AI summary of the abstract · may contain errors · not medical advice · read the paper.