Behavioral and Video-Based Tracking Data from The Switchboard Test:A Flexible, Low-Cost Setup for Assessing Spatial Memory in Rats During Sequential Instrumental Tasks

This repository provides representative sample data from The Switchboard Test, a flexible, low-cost experimental paradigm developed to assess spatial memory in rats during sequential instrumental tasks. The task is designed to capture both decision-making performance and movement dynamics while remaining compatible with electrophysiological recordings and scalable behavioral testing.The experimental system consists of a microcontroller-based task execution and event-logging module that handles stimulus delivery, reward triggering, and precise timestamping of behavioral events, together with overhead video recording and pose estimation performed using DeepLabCut. The microcontroller serves as the primary temporal reference for all behavioral events, ensuring consistent and reproducible timing across experimental sessions.Due to the large size of full experimental datasets and practical constraints associated with video-linked behavioral recordings, the complete dataset is not currently hosted publicly. Instead, this repository contains representative sample files for each data modality. These samples preserve the full structural complexity of the original data, including variable definitions, timestamping schemes, session organization, and data-integration logic, thereby enabling transparency, reproducibility, and validation of analysis pipelines.Behavioral event timestamps are generated using a microcontroller-based clock and are temporally aligned with DeepLabCut-derived tracking data using a ±100 ms tolerance window. This alignment enables joint analyses of spatial movement and behavioral decisions during task performance. Multiple stages of data processing are represented, from raw event logs to analyzed performance metrics and integrated behavioral–tracking datasets.The repository includes the following data categories:Raw behavioral event logs generated during experimental sessionsAnalyzed behavioral performance data derived from raw logsProcessed DeepLabCut tracking data containing pose estimates and likelihoodsMerged datasets combining behavioral and tracking data based on temporal alignmentTask execution and data-logging code used during experimentsAll shared files correspond to real experimental sessions and are provided as representative samples rather than complete datasets. The full experimental dataset may be made publicly available in future releases as data curation and storage constraints allow. Until then, complete datasets are available from the authors upon reasonable request.