The Mouse Action Recognition System (MARS): multi-worker behavior annotation data

Creators: Segalin, Cristina¹; Williams, Jalani¹; Karigo, Tomomi¹; Hui, May¹; Zelikowski, Moriel²; Sun, Jennifer J.¹; Perona, Pietro¹; Anderson, David J.¹; Kennedy, Ann³

1. California Institute of Technology
2. University of Utah
3. Northwestern

Description

This dataset was collected to quantify inter-annotator variability in manual labeling of social behaviors in freely interacting mice. We provided a group of graduate students, postdocs, and technicians in the Anderson lab with written descriptions of close investigation, mounting, and attack behaviors (included below), and instructed them to score a set of ten resident-intruder videos, all taken from unoperated mice. Annotators were given front- and top-view video of social interactions, and scored behavior using either Bento or the Caltech Behavior Annotator, both of which support simultaneous display of front- and top-view video and frame-by-frame browsing and scoring. All but one annotator (Human 4) had previous experience scoring mouse behavior videos; Human 4 had previous experience scoring similar social behaviors in flies.

Files

annotations.zip

Files (38.4 GB)

Name	Size	Actions
annotation_parsers.py md5:0d731e8427a697cfb5476d2a942c1129	14.8 kB	Download
annotations.zip md5:c2329a85e3632ff0a2d153ba1fbe1473	126.3 kB	Preview Download
loadAnnotSheet.m md5:954ad6db17f2533d0531e1f20b036ea6	1.9 kB	Download
videos_seq.zip md5:0e4a99269c9050a0739cbe12aa6108bd	38.4 GB	Preview Download

Additional details

DOI

Resource type: Dataset
Publisher: CaltechDATA
Languages: English
Formats: This dataset includes behavior videos (.seq files) from 10 standard resident-intruder interactions, as well as frame-by-frame manual annotations by eight individuals, in the Bento .annot format. The accompanying script loadAnnotSheet.m can be used to load a .annot file into Matlab, or use annotation_parsers.parse_annotations() in Python.