DeepEthogram, a machine learning pipeline for supervised behavior classification from raw pixels
- Harvard Medical School, United States
- Boston Children's Hospital, United States
- Massachusetts General Hospital, United States
- Champalimaud Center for the Unknown, Portugal
Abstract
Videos of animal behavior are used to quantify researcher-defined behaviors-of-interest to study neural function, gene mutations, and pharmacological therapies. Behaviors-of-interest are often scored manually, which is time-consuming, limited to few behaviors, and variable across researchers. We created DeepEthogram: software that uses supervised machine learning to convert raw video pixels into an ethogram, the behaviors-of-interest present in each video frame. DeepEthogram is designed to be general-purpose and applicable across species, behaviors, and video-recording hardware. It uses convolutional neural networks to compute motion, extract features from motion and images, and classify features into behaviors. Behaviors are classified with above 90% accuracy on single frames in videos of mice and flies, matching expert-level human performance. DeepEthogram accurately predicts rare behaviors, requires little training data, and generalizes across subjects. A graphical interface allows beginning-to-end analysis without end-user programming. DeepEthogram's rapid, automatic, and reproducible labeling of researcher-defined behaviors-of-interest may accelerate and enhance supervised behavior analysis.
Data availability
Code is posted publicly on Github and linked in the paper. Video datasets and human annotations are publicly available and linked in the paper.
Article and author information
Author details
Christopher D Harvey
Funding
National Institutes of Health (R01MH107620)
- Christopher D Harvey
National Science Foundation (GRFP)
- Nivanthika K Wimalasena
Fundacao para a Ciencia ea Tecnologia (PD/BD/105947/2014)
- Tomás Cruz
Harvard Medical School Dean's Innovation Award
- Christopher D Harvey
Harvard Medical School Goldenson Research Award
- Christopher D Harvey
National Institutes of Health (DP1 MH125776)
- Christopher D Harvey
National Institutes of Health (R01NS089521)
- Christopher D Harvey
National Institutes of Health (R01NS108410)
- Christopher D Harvey
National Institutes of Health (F31NS108450)
- James P Bohnslav
National Institutes of Health (R35NS105076)
- Clifford J Woolf
National Institutes of Health (R01AT011447)
- Clifford J Woolf
National Institutes of Health (R00NS101057)
- Lauren L Orefice
National Institutes of Health (K99DE028360)
- David A Yarmolinsky
European Research Council (ERC-Stg-759782)
- M Eugenia Chiappe
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
Ethics
Animal experimentation: All experimental procedures were approved by the Institutional Animal Care and Use Committees at Boston Children's Hospital (protocol numbers 17-06-3494R and 19-01-3809R) or Massachusetts General Hospital (protocol number 2018N000219) and were performed in compliance with the Guide for the Care and Use of Laboratory Animals.
Copyright
© 2021, Bohnslav et al.
This article is distributed under the terms of the Creative Commons Attribution License permitting unrestricted use and redistribution provided that the original author and source are credited.
Metrics
-
- 13,219
- views
-
- 1,257
- downloads
-
- 105
- citations
Views, downloads and citations are aggregated across all versions of this paper published by eLife.
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
DeepEthogram, a machine learning pipeline for supervised behavior classification from raw pixels
eLife 10:e63377.
https://doi.org/10.7554/eLife.63377
Further reading
-
- Neuroscience
Control of voluntary limb movement is predominantly attributed to the contralateral motor cortex. However, increasing evidence suggests the involvement of ipsilateral cortical networks in this process, especially in motor tasks requiring bilateral coordination, such as locomotion. In this study, we combined a unilateral thoracic spinal cord injury (SCI) with a cortical neuroprosthetic approach to investigate the functional role of the ipsilateral motor cortex in rat movement through spared contralesional pathways. Our findings reveal that in all SCI rats, stimulation of the ipsilesional motor cortex promoted a bilateral synergy. This synergy involved the elevation of the contralateral foot along with ipsilateral hindlimb extension. Additionally, in two out of seven animals, stimulation of a sub-region of the hindlimb motor cortex modulated ipsilateral hindlimb flexion. Importantly, ipsilateral cortical stimulation delivered after SCI immediately alleviated multiple locomotor and postural deficits, and this effect persisted after ablation of the homologous motor cortex. These results provide strong evidence of a causal link between cortical activation and precise ipsilateral control of hindlimb movement. This study has significant implications for the development of future neuroprosthetic technology and our understanding of motor control in the context of SCI.
-
- Neuroscience
Sound location coding has been extensively studied at the central nucleus of the mammalian inferior colliculus (CNIC), supporting a population code. However, this population code has not been extensively characterized on the single-trial level with simultaneous recordings or at other anatomical regions like the dorsal cortex of inferior colliculus (DCIC), which is relevant for learning-induced experience dependent plasticity. To address these knowledge gaps, here we made in two complementary ways large-scale recordings of DCIC populations from awake mice in response to sounds delivered from 13 different frontal horizontal locations (azimuths): volumetric two-photon calcium imaging with ~700 cells simultaneously recorded at a relatively low temporal resolution, and high-density single-unit extracellular recordings with ~20 cells simultaneously recorded at a high temporal resolution. Independent of the method, the recorded DCIC population responses revealed substantial trial-to-trial variation (neuronal noise) which was significantly correlated across pairs of neurons (noise correlations) in the passively listening condition. Nevertheless, decoding analysis supported that these noisy response patterns encode sound location on the single-trial basis, reaching errors that match the discrimination ability of mice. The detected noise correlations contributed to minimize the error of the DCIC population code of sound azimuth. Altogether these findings point out that DCIC can encode sound location in a similar format to what has been proposed for CNIC, opening exciting questions about how noise correlations could shape this code in the context of cortico-collicular input and experience-dependent plasticity.
