ABEL: an active-learning behavior estimation and labeling platform
- 1University of North Carolina at Chapel Hill
Abstract
Detailed behavior analysis is essential for thorough characterization of ethologically relevant behaviors in model organisms, yet manual annotation of the full behavioral repertoire remains subjective, time intensive, and susceptible to observer error. Advances in machine learning have enabled high-throughput pose estimation on recorded video, but tools for behavior classification from pose and video data are still developing. Instead of hand-scoring every frame of video, experimenters can instead label a small subset of video frames and software trained through machine learning makes predictions on the rest. Here, we present an Active-learning Behavior Estimation and Labeling (ABEL) platform that uses clip-level active learning (i.e., human labeling of short video snippets) with multimodal features (pose, video, context/ROI) to train robust behavior classifiers. We rigorously validated ABEL-derived behavior predictions against expert human observers and field-standard automated software, across diverse rodent behavioral assays. Across eight assays and 45 behaviors, model training required 19.5 hours of human annotation in total, with the reviewer scoring ~8% of available video. Models trained in ABEL achieved a mean precision-recall area under the curve (PR-AUC - a 0-1 score of how well a model balances missed detections against false alarms, with 1 being perfect) of 0.90 (SD 0.09, range 0.60-0.99), with no association between performance and behavior prevalence (r = 0.20). This was aided by custom tools, Essence Extractor and UMAP Interactive Selection, for targeted discovery of high probability clips which reduce the clip review needed to find a rare behavior 6-fold relative to random sampling and 10-fold relative to labeling whole videos. As a biological validation, we assessed how ABEL-derived behaviors relate to underlying neuronal calcium dynamics. Behavior labels were tightly synced with neuronal signatures distinct from ambiguous behavior and randomly chosen, behavior-unrelated time windows (shuffle control). Together, these data indicate that ABEL provides an efficient platform for frame-precise classification of distinct ethologically relevant behaviors.
Related
- Toddlers’ Active Gaze Behavior Supports Self‐Supervised Object Learning
- Machine learning-assisted Repli-Histo labeling reveals distinct transcription-dependent constraints on chromatin motion in living cells
- B-REAL: A keypoint-less behavior platform for offline classification and real-time detection
- Deep-learning-based design of an orthogonal self-labeling protein from K-Ras(G12C)
- A programmable platform for the study of active wrinkle patterns and cell migration directed by dynamic curvotaxis
BibTeX
@misc{ritchie2026abel,
title = {{ABEL: an active-learning behavior estimation and labeling platform}},
author = {J. L. Ritchie and B. E. George and A. V. Roland and C. G. Krieman and B. N. Bender and M. R. Eberle and W. C. Der and L. S. Kooyman and A. M. Lawes and R. T. Scott and T. K. O'Buckley and M. M. R. McLean and C. J. Gallagher and J. Besheer and T. L. Kash},
howpublished = {bioRxiv},
year = {2026},
month = aug,
doi = {10.64898/2026.08.30.748115},
url = {https://www.biorxiv.org/content/10.64898/2026.08.30.748115v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - ABEL: an active-learning behavior estimation and labeling platform
AU - J. L. Ritchie
AU - B. E. George
AU - A. V. Roland
AU - C. G. Krieman
AU - B. N. Bender
AU - M. R. Eberle
AU - W. C. Der
AU - L. S. Kooyman
AU - A. M. Lawes
AU - R. T. Scott
AU - T. K. O'Buckley
AU - M. M. R. McLean
AU - C. J. Gallagher
AU - J. Besheer
AU - T. L. Kash
PY - 2026
DA - 2026/08/30
PB - bioRxiv
DO - 10.64898/2026.08.30.748115
UR - https://www.biorxiv.org/content/10.64898/2026.08.30.748115v1
AB - Detailed behavior analysis is essential for thorough characterization of ethologically relevant behaviors in model organisms, yet manual annotation of the full behavioral repertoire remains subjective, time intensive, and susceptible to observer error. Advances in machine learning have enabled high-throughput pose estimation on recorded video, but tools for behavior classification from pose and video data are still developing. Instead of hand-scoring every frame of video, experimenters can instead label a small subset of video frames and software trained through machine learning makes predictions on the rest. Here, we present an Active-learning Behavior Estimation and Labeling (ABEL) platform that uses clip-level active learning (i.e., human labeling of short video snippets) with multimodal features (pose, video, context/ROI) to train robust behavior classifiers. We rigorously validated ABEL-derived behavior predictions against expert human observers and field-standard automated software, across diverse rodent behavioral assays. Across eight assays and 45 behaviors, model training required 19.5 hours of human annotation in total, with the reviewer scoring ~8% of available video. Models trained in ABEL achieved a mean precision-recall area under the curve (PR-AUC - a 0-1 score of how well a model balances missed detections against false alarms, with 1 being perfect) of 0.90 (SD 0.09, range 0.60-0.99), with no association between performance and behavior prevalence (r = 0.20). This was aided by custom tools, Essence Extractor and UMAP Interactive Selection, for targeted discovery of high probability clips which reduce the clip review needed to find a rare behavior 6-fold relative to random sampling and 10-fold relative to labeling whole videos. As a biological validation, we assessed how ABEL-derived behaviors relate to underlying neuronal calcium dynamics. Behavior labels were tightly synced with neuronal signatures distinct from ambiguous behavior and randomly chosen, behavior-unrelated time windows (shuffle control). Together, these data indicate that ABEL provides an efficient platform for frame-precise classification of distinct ethologically relevant behaviors.
ER -