Individual Mouse Activity Dataset for Group Environments (IMADGE)

This repository contains the auto-generated observability and behaviour labels used to model collective mouse behaviour as in Chapter 6 of my PhD Thesis “Automated Identification and Behaviour Classification for Modelling Social Dynamics in Group-Housed Mice” [1]. The video recordings are courtesy of the Mary Lyon Centre at MRC Harwell [3].

The data is released under the CC-BY 4.0 License (a copy appears as part of the repository as DATA.LICENSE).

1. Repository Structure

├── dataset         # Curated Version of the Dataset
│    ├── Adult
│    │    ├── A.df
│    │    :
│    │    └── P.df
│    └── Young
│
├── LICENSE
└── README.md

2. Dataset

This dataset contains associated features for groups of three mice in the homecage:
- The data pertains to all male cages (three mice per-cage) from the C57BL/6NTac strain at the 3-month or 12-month age group. There are in total 16 unique cages: refer to [1, §3.3] for general information about the data demographics and to [3] for details of the animal husbandry/data collection.
- The features include mouse positions within the cage, locations in the image and automatically generated behaviour labels: refer to [1, §6.2] for the automated labelling process.
The basic unit of processing is the Behavioural Time Interval (BTI) (1 second duration): within this, we are interested in the behaviour of each mouse individually.
The data is provided for reproducibility purposes and also to allow further research on the data: for this reason, it is provided in its complete form and with minimal processing.

2.1 Storage Organisation

The dataset is stored organised by age-group: Adult/Young.
Within each, we provide a dataframe of features for each cage, enumerated via an alphabetical code running from A through P.

2.2 Format of individual Dataframes

The data is stored as pandas dataframes in pickle format with bz2 compression. They can be read into python using pd.read_pickle('...', compression='bz2').
Each Dataframe is indexed by:
- Run: A continuous recording period around a light-to-dark or dark-to-light transition (see [1, §3.3.1] ). Run identifiers are unique across cages within the same age-group.
- Segment: Segment identifier. Segments are unique within each cage
- BTI: Sample index (since start of segment)
- Mouse: Red/Green/Blue identifier
The columns in each dataframe are organised in a conceptual two-level index:
- Sensors: 'Groundtruth' information from sensors.
  - RFID.Ant: Mode of the antenna-based location for the BTI --- integer from 1 to 18
  - RFID.Vel: Number of changes in the antenna pickup within the BTI.
  - Light: Boolean indicator of lights-on (True) or lights-off (False).
- TIM: Auto-detections and identifications of mice in the BTI using the Tracking & Identification Module ([1, §4] and [2]). This is the average Bounding-Box in the BTI, represented in CoCo format i.e. top-left corner co-ordinate (x,y) and size (w,h).
- ALM: Auto-classification of the Observability and Behaviour in the BTI using the Activity Labelling Module [1, §5].
  - Obs: Observability Label --- Boolean True/False
  - Beh: Behaviour Label --- 1 of 7 behaviour labels according to simplified schema in [1, §5.3.1]: N/Obs is used if not observable.
- ALM.Prob: Probability scores over the Behaviour labels provided by the ALM [1, §5]. This is the score per behaviour.

References

If you make use of this data, please cite our work, as below:

For the general dataset:

[1] M. P. J. Camilleri, “Automated Identification and Behaviour Classification for Modelling Social Dynamics in Group-Housed Mice,” PhD Thesis, University of Edinburgh, 2023.

For the Tracking and Identification Module:

[2] M. P. J. Camilleri, L. Zhang, R. S. Bains, A. Zisserman, and C. K. I. Williams, “Persistent Object Identification Leveraging Non-Visual Markers,” CoRR (arXiv), cs.CV (2112.06809), Dec. 2021. Available on arXiv

The raw data is courtesy of the Mary Lyon Centre at MRC Harwell, as described in:

[3] R. S. Bains, H. L. Cater, R. R. Sillito, A. Chartsias, D. Sneddon, D. Concas, P. Keskivali-Bond, T. C. Lukins, S. Wells, A. Acevedo Arozena, P. M. Nolan, and J. D. Armstrong. “Analysis of Individual Mouse Activity in Group Housed Animals of Different Inbred Strains using a Novel Automated Home Cage Analysis System”. In: Frontiers in Behavioral Neuroscience 10 (106) (June 2016). Available online

michael-camilleri / IMADGE