Chronic sensing-enabled devices for deep brain stimulation (DBS) permit long-term intracranial neural recordings in naturalistic, ambulatory settings. However, interpreting chronic neural signals requires concurrent behavioral context from wearables or other sources. The objective reported in this preprint was to develop a reproducible end-to-end framework for continuous acquisition and synchronization of wearable-derived behavioral data alongside chronic DBS recordings to enable longitudinal neurobehavioral research.
The authors describe a publicly available software framework that automates several steps required for producing analysis-ready multimodal datasets. Key components include automated wearable data ingestion, handling of neural recording artifacts, epoch-based temporal synchronization across data streams, and generation of datasets prepared for downstream analysis. The framework is presented as open-source to lower technical barriers to combining behavioral and neural recordings over extended periods.
To test the framework, the team applied it to simultaneous recordings collected from three participants diagnosed with obsessive-compulsive disorder (OCD). Chronic intracranial neural recordings were acquired using sensing-enabled DBS devices, and wearable-derived behavioral data were recorded with an Oura Ring. The combined dataset incorporated device settings and parameters from the implanted stimulators alongside wearable measures of sleep-wake state, physical activity, and autonomic physiology.
Using the framework, the authors synchronized 6,384 hours of intracranial neural recordings with wearable-derived behavioral data. The resulting neurobehavioral datasets spanned months of ambulatory monitoring and integrated multiple modalities: chronic neural signals, sleep–wake annotation, physical activity metrics, autonomic physiology, and DBS parameter records. These integrated datasets are intended to support longitudinal analysis and potential biomarker discovery.
The team performed benchtop testing to quantify temporal alignment accuracy. They report that a Medtronic Percept device exhibited 12 seconds of clock drift relative to network time over a one-week period. The authors note that this observed temporal error is substantially smaller than the 10-minute sampling interval used for the chronic neural recordings in their datasets, indicating that clock drift is unlikely to compromise reliable alignment with wearable-derived data at that sampling cadence.
All data supporting the study are available online at the project repository identified in the source (https://dabi.loni.usc.edu/projects/M7FD26LM28KI). The research received ethical approval from the IRB of Baylor College of Medicine (H-43144). The authors confirm that necessary participant consent was obtained and institutional forms archived. The preprint explicitly states that the article has not been certified by peer review and should not be used to guide clinical practice.
The authors position this framework as a reproducible, practical foundation for continuous acquisition, synchronization, and analysis of wearable behavioral data alongside chronic DBS sensing. By providing tools for automated ingestion, artifact handling, and epoch-based synchronization, the framework aims to enable scalable longitudinal neurobehavioral studies and facilitate behaviorally annotated neural datasets for biomarker discovery. Limitations inherent in the report include that the work is presented as a preprint not yet peer-reviewed; additional validation and broader deployment are implied but not detailed in the source.
Declared funding sources include the National Institute of Mental Health (R01MH139889), the National Institute of Neurological Disorders and Stroke (UH3NS136631), the Brain & Behavior Research Foundation, and the Robert and Janice McNair Foundation. Several authors disclosed competing interests: E.A.S. reported multiple research funding sources, training income, past consulting, stock options under $5,000, and royalties; W.K.G. received donated devices from Medtronic and has consulting agreements; S.A.S. has consulting agreements with several device companies and is co-founder of a neurotechnology firm. The source lists the full competing interest statements and author declarations.