Revealing motor behavior patterns and functional improvement trajectories following spinal cord injury using unsupervised machine learning Article Swipe
YOU?
·
· 2023
· Open Access
·
· DOI: https://doi.org/10.1101/2023.10.31.564826
Spinal cord injury (SCI) disrupts motor function, with recovery involving complex adaptive motor behaviors that cannot be fully assessed with traditional manual recovery assessments and their predefined recovery criteria. To move beyond these constraints, we employed motion sequencing (MoSeq), a machine learning-driven framework that automatically segments freely moving mouse behavior into discrete, repeated behavioral “syllables”, effectively capturing entire behavioral repertoires and dynamic behavioral sequences observed throughout SCI recovery. We observed that functional improvements involved the reemergence of preinjury behaviors, emergence of adaptive behaviors, and subtle individual differences between mice. Interestingly, the temporal organization of behavioral sequences evolved consistently across injury severities, suggesting a ceiling effect in sequence reorganization. MoSeq syllables also correlated with the Basso Mouse Scale (BMS) and outperformed BMS in correlation to several injury-related histological findings. We also introduced the “Recovery Score,” a new metric derived from aggregated MoSeq data that quantifies recovery based on syllable usage patterns. The Recovery Score identified clusters of syllables with high usages, which highlights the potential of using hallmark syllables as behavioral biomarkers for assessing recovery and informing therapies. Future research should incorporate more complex experimental designs to refine this aggregation approach and advance the pursuit of a new gold standard in recovery assessment.
Related Topics To Compare & Contrast
- Type
- preprint
- Language
- en
- Landing Page
- https://doi.org/10.1101/2023.10.31.564826
- https://www.biorxiv.org/content/biorxiv/early/2023/11/05/2023.10.31.564826.full.pdf
- OA Status
- green
- Cited By
- 3
- References
- 124
- Related Works
- 10
- OpenAlex ID
- https://openalex.org/W4388211385