The mesencephalic locomotor region shapes decision-making during movement - Dataset and Code for Shin et al 2026

Abstract

Shin et al. 2026 v2.0 — Revised release Data and analysis code accompanying the revised manuscript: The mesencephalic locomotor region shapes decision-making during movementWooyeon Shin, Juri Kim, Dajung Jung, Yunjae Kim, Se-Bum Paik, Jeongjin Kim. Nature Communications (under revision). AbstractDecision-making during movement requires the brain to integrate internal motor states with external cues effectively. However, the neural mechanisms by which ongoing motor signals dynamically update decision-making remain unclear. The mesencephalic locomotor region (MLR), an evolutionarily conserved locomotion center, is a candidate region where motor and cognitive signals may converge. To explore this, we established an auditory decision-making task for actively moving mice and combined it with extracellular recordings in the MLR. We found that movement speed (fast vs. slow) shifts decision behavior, reducing go-rate and, for reward cues, prolonging response latency. Analysis of neuronal activities in the MLR using general linear models revealed that MLR neurons encode both movement and decision-related variables, demonstrating multiplexing at the single-cell level. To further investigate how these multiplexing neurons contribute to the decision-making process, we applied activity-based clustering and identified multiplexing MLR clusters engaged during decision-making. We confirmed that these neuronal clusters successfully predicted decision outcomes. By using a drift-diffusion model to simulate decision-making over time, we clarified the relationship between these MLR clusters and drift rate. Finally, we tested the causal role of MLR in decision-making by labeling decision-period-active neurons with Cal-Light and showing that their optogenetic inhibition modulated decision performance. Collectively, these results demonstrate that the MLR contributes to evaluation during decision-making, suggesting that this locomotor hub also contributes to state-dependent decision-making. ContentsThis repository contains the behavioral and electrophysiological data (behavior sessions, MLR single-unit recordings, Cal-Light behavior sessions and task-naive MLR Neuropixels recordings) and the MATLAB and Python code used for the behavioral analyses, generalized linear encoding models, ROC and SVM decoding, and drift-diffusion modeling (HDDM and PyDDM) reported in the paper. See README.md for the folder structure, system requirements, installation, instructions for use and the scripts that generate each figure. Changes from v1.0- Code reorganized into numbered scripts per analysis, with paths relative to the repository root- Added analyses performed during revision: event-window GLM, task-naive recordings, PyDDM time-limited models and the preparatory-licking filter LicenseCode: MIT License. Data: Creative Commons Attribution 4.0 International (CC BY 4.0). CitationIf you use these data or code, please cite the associated article (preprint: https://doi.org/10.21203/rs.3.rs-9445221/v1).