Temporally Synchronized Multimodal Data Capture in the Neurosurgical OR via a Novel, Open-source Software Platform

Introduction: Thalamus is an open-source, real-time software platform designed to capture and synchronize multimodal behavioral and neural data streams in clinical environments. It facilitates closed-loop control experiments by integrating high-resolution sensors with commonly used clinical hardware. This study presents the architecture of Thalamus and demonstrates its utility in a neurosurgical setting for synchronized data collection and real-time analysis.

Methods: Thalamus integrates neural, physiological, and behavioral data through a modular node-based architecture. Each node handles specific data streams, including subcortical microelectrode recordings, high-density cortical recordings, motion capture, and electrophysiology, ensuring synchronized acquisition across all modalities. The system leverages Python for high-level application management, paired with a C++ backend to manage real-time data acquisition with low latency. A core feature is its closed-loop capability, allowing real-time adjustments in response to behavioral or neural input. Synchronization is achieved using Google Remote Procedure Call (gRPC) to handle data flow with millisecond-level precision.

Results: To assess synchronization accuracy, synchronicity tests were conducted using a push-button LED circuit. Thalamus successfully synchronized voltage signals from motion capture, camera, and analog acquisition devices, with mean latencies of 6 ± 5 ms between voltage and motion capture data, and 8 ± 2.8 ms between voltage and camera feeds. This millisecond-level precision ensures accurate real-time data capture. Thalamus was further validated in a neurosurgical environment during a Parkinson’s Disease deep brain stimulation (DBS) procedure. The system synchronized subcortical microelectrode recordings, high-density cortical recordings, kinematic inertial sensors, and stretch sense sensors, allowing us to explore neural activation during hand motion.

Conclusion : Thalamus represents a significant advancement in multimodal data integration, enabling synchronized real-time data capture and closed-loop functionality in clinical environments. Its flexibility, open-source architecture, and real-time capabilities make it highly suitable for neurosurgical research and brain-computer interface development, with the potential to improve precision and personalization in neurological care.