Mouse Lockbox Dataset: Behavior Recognition of Mice Solving Mechanical Puzzles

Patrik Reiske, Marcus Nicolaas Boon, Niek Andresen, Sole Traverso, Katharina Hohlbaum, Lars Lewejohann, Christa Thöne-Reineke, Olaf Hellwich, Henning Sprekeler

Keywords: dataset, video, action classification, behavior, mice, machine learning, computer vision, neuroscience, computational neuroscience

Abstract: Machine learning and computer vision have a major impact on the study of natural animal behavior, as they enable automated action classification of large bodies of videos. Mice are the standard mammalian model system in many fields of research, but the open datasets that are currently available to refine machine learning methods mostly focus on either simple or social behaviors. In this work, we present a large video dataset of individual mice solving complex mechanical puzzles, so-called lockboxes. The dataset consists of a total of well over 110 hours of animal behavior, recorded with three cameras from different perspectives. As a benchmark for frame-level action classification methods, we provide human-annotated labels for all videos of two different mice, that equal 13% of our dataset. The used keypoint (pose) tracking-based action classification framework illustrates the challenges of automated labeling of fine-grained behaviors, such as the manipulation of objects. We hope that our work will help accelerate the advancement of automated action and behavior classification in the computational neuroscience community. An anonymized preview of our dataset is available for the reviewers of this manuscript at https://www.dropbox.com/scl/fo/h7nkai8574h23qfq9m1b2/AP4gNZOpDJJ7z0yGtbWQiOc?rlkey=w36jzxqjkghg0j0xva5zsxy2v&st=5r9msqjw&dl=0

Primary Area: datasets and benchmarks

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 10803