Image copyright: Nature
Simple behavioral analysis (SimBA) as a platform for explainable machine learning in behavioral neuroscience
Published in Nat Neuroscience (2024)
Authors
Nastacia L. Goodwin, Jia J. Choong, Sophia Hwang, Kayla Pitts, Liana Bloom, Aasiya Islam, Yizhe Y. Zhang, Eric R. Szelenyi, Xiaoyu Tong, Emily L. Newman, Klaus Miczek, Hayden R. Wright, Ryan J. McLaughlin, Zane C. Norville, Neir Eshel, Mitra Heshmati, Simon R. O. Nilsson & Sam A. Golden
Paper presented by Dr. Rajtarun Madangopal and selected by the NIDA TDI Paper of the Month Committee
Publication Brief Description
Simple Behavioral Analysis (SimBA) is an open-source machine learning platform designed for the automated detection of animal behaviors. It leverages markerless key-point pose tracking from open-source programs like SLEAP and DeepLabCut to transform tracking data into features that describe relationships between body parts over time. Researchers can train classifiers within SimBA to detect specific behaviors based on these features. A key feature of SimBA is its integration with Shapley Additive Explanations (SHAP) which provides users with quantitative insights into which features are critical to their behavioral predictions. The explainability of the outcome facilitates the standardization and sharing of behavioral definitions across labs, making complex behavioral analysis accessible to non-specialists. Further, by computing and sharing these explainability metrics, behavioral classifiers can be reconceptualized as shareable reagents akin to the commonly used Research Reagent Identifiers (RRIDs) system for wet lab reagents, enhancing reproducibility and interpretability between research groups.
Simple Behavioral Analysis (SimBA) as a platform for explainable machine learning in behavioral neuroscience Journal Article
In: Nat Neurosci, 2024, ISSN: 1546-1726.
