Traditional methods study non-human sexual behaviour by manual annotations of selected sexual behaviour parameters, which can create errors. These limitations can be addressed using the multi-animal pose-estimation toolbox, DeepLabCut. It automatically identifies body parts that can be used to infer behaviour. Some sexual behaviour recordings are very low-resolution. This is problematic for DeepLabCut because the annotator cannot accurately identify the body parts. To circumvent this, we labelled frames from high-resolution videos, followed by customised data augmentation during neural network training. Simple Behavioral Analysis was used to generate random forest classifiers for male sexual behaviours. There was a wide range of errors between the human-labelled and machine-identified body parts, and the behavioural classifiers did not match manual annotations. In addition to the lab, neuroscientists need to study sexual behaviour in the wild, to facilitate the understanding of sexual diversity across species, ecosystems and evolution. Camera traps are commonly used to capture behaviour in the wild, but it is extremely time-consuming to manually review camera trap datasets that are usually in hundreds of thousands to millions of images. To address this, we used MegaDetector to identify animals in a camera trap dataset from Wellington, New Zealand. Following that, we used DeepLabCut Model Zoo to identify body parts. This pose estimation enabled us to screen images where animals were physically interacting. However, the potential of DeepLabCut had not been fully realised in this use case, due to the difficulty for the model to identify body parts in these images.
Cornil Charlotte A., Court Lucas. Neuroestrogens in the control of sexual behavior: Past, present, and future. Current Opinion in Endocrine and Metabolic Research. Vol. 24:2022. Elsevier BV. [Cross Ref]
Kleanthous Natasa, Hussain Abir Jaafar, Khan Wasiq, Sneddon Jennifer, Al-Shamma'a Ahmed, Liatsis Panos. A survey of machine learning approaches in animal behaviour. Neurocomputing. Vol. 491:442–463. 2022. Elsevier BV. [Cross Ref]
Jiang Le, Lee Caleb, Teotia Divyang, Ostadabbas Sarah. Animal pose estimation: A closer look at the state-of-the-art, existing gaps and opportunities. Computer Vision and Image Understanding. Vol. 222:2022. Elsevier BV. [Cross Ref]
Hardin Abigail, Schlupp Ingo. Using machine learning and DeepLabCut in animal behavior. acta ethologica. Vol. 25(3):125–133. 2022. Springer Science and Business Media LLC. [Cross Ref]
Mazzucato Luca. Neural mechanisms underlying the temporal organization of naturalistic animal behavior. eLife. Vol. 11:2022. eLife Sciences Publications, Ltd. [Cross Ref]
David Caroline D., Wyrosdic Brianna N., Park Jin Ho. Strain differences in post-castration sexual and aggressive behavior in male mice. Behavioural Brain Research. Vol. 422:2022. Elsevier BV. [Cross Ref]
Lauer Jessy, Zhou Mu, Ye Shaokai, Menegas William, Schneider Steffen, Nath Tanmay, Rahman Mohammed Mostafizur, Di Santo Valentina, Soberanes Daniel, Feng Guoping, Murthy Venkatesh N., Lauder George, Dulac Catherine, Mathis Mackenzie Weygandt, Mathis Alexander. Multi-animal pose estimation, identification and tracking with DeepLabCut. Nature Methods. Vol. 19(4):496–504. 2022. Springer Science and Business Media LLC. [Cross Ref]
Mathuru Ajay S., Libersat Frédéric, Vyas Ajai, Teseo Serafino. Why behavioral neuroscience still needs diversity?: A curious case of a persistent need. Neuroscience & Biobehavioral Reviews. Vol. 116:130–141. 2020. Elsevier BV. [Cross Ref]
Cisek Paul, Hayden Benjamin Y.. Neuroscience needs evolution. Philosophical Transactions of the Royal Society B: Biological Sciences. Vol. 377(1844)2022. The Royal Society. [Cross Ref]
Clutton-Brock Tim. Social evolution in mammals. Science. Vol. 373(6561)2021. American Association for the Advancement of Science (AAAS). [Cross Ref]
Sliwa Julia. Toward collective animal neuroscience. Science. Vol. 374(6566):397–398. 2021. American Association for the Advancement of Science (AAAS). [Cross Ref]
Delisle Zackary J., Flaherty Elizabeth A., Nobbe Mackenzie R., Wzientek Cole M., Swihart Robert K.. Next-Generation Camera Trapping: Systematic Review of Historic Trends Suggests Keys to Expanded Research Applications in Ecology and Conservation. Frontiers in Ecology and Evolution. Vol. 9:2021. Frontiers Media SA. [Cross Ref]
Hawkes L. A., Fahlman A., Sato K.. Introduction to the theme issue: Measuring physiology in free-living animals. Philosophical Transactions of the Royal Society B: Biological Sciences. Vol. 376(1830)2021. The Royal Society. [Cross Ref]
Hahn Nathan R., Bombaci Sara P., Wittemyer George. Identifying conservation technology needs, barriers, and opportunities. Scientific Reports. Vol. 12(1)2022. Springer Science and Business Media LLC. [Cross Ref]
Ye Shaokai, Mathis Alexander, Mathis Mackenzie Weygandt. Panoptic animal pose estimators are zero-shot performers. 2022. arXiv. [Cross Ref]