Implicit Behavioral Decoding from Next-Step Spike Forecasts at Population Scale

TL;DR

This paper introduces Mamba, a neural network model that predicts upcoming neural activity and behavioral states from spike data, enabling efficient decoding in brain-computer interfaces with improved accuracy over traditional methods.

Contribution

Mamba demonstrates that a single model trained on next-step spike forecasts can decode behavior and stimulus information more accurately than raw spike count classifiers.

Findings

Mamba achieves 75.7% trial vote accuracy in mouse choice decoding.

Mamba outperforms linear decoders on raw spike counts by 4-6 percentage points.

A session calibration improves decoding accuracy to near asymptote within 100-150 trials.

Abstract

Closed-loop brain-computer interfaces often require both a forecast of upcoming neural population activity and a readout of the animal's behavioral state. A single Mamba forecaster, trained only on next-step spike counts at Neuropixels scale, can deliver both in one forward pass. A lightweight per-session linear head reading the model's predicted rates decodes behavior better than the same linear classifier reading the raw spike counts, under matched temporal context. We test on the Steinmetz visual-discrimination benchmark, which spans 39 sessions, roughly 27,000 neurons, and 1,994 held-out trials. Across three training seeds, Mamba's predicted rates decode mouse choice at 75.7$\pm$0.2% trial vote, roughly 2.3 times chance level, and stimulus side at 66.1$\pm$0.6%, about twice chance. Compared to a matched 500 ms-context linear decoder on the raw spike counts, Mamba wins at trial vote by 4-6 pp on response and 4-6 pp on stimulus side. A session-start calibration block of about 100-150 trials brings the readout within 1-2 pp of asymptote, and the full pipeline fits inside the 50 ms bin budget on workstation-class GPUs typical of tethered chronic Neuropixels recordings.