DeepEthogram, a machine learning pipeline for supervised behavior classification from raw pixels

  1. James P Bohnslav
  2. Nivanthika K Wimalasena
  3. Kelsey J Clausing
  4. Yu Y Dai
  5. David A Yarmolinsky
  6. Tomás Cruz
  7. Adam D Kashlan
  8. M Eugenia Chiappe
  9. Lauren L Orefice
  10. Clifford J Woolf
  11. Christopher D Harvey  Is a corresponding author
  1. Harvard Medical School, United States
  2. Boston Children's Hospital, United States
  3. Massachusetts General Hospital, United States
  4. 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.

The following previously published data sets were used

Article and author information

Author details

  1. James P Bohnslav

    Neurobiology, Harvard Medical School, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  2. Nivanthika K Wimalasena

    F.M. Kirby Neurobiology Center, Boston Children's Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  3. Kelsey J Clausing

    Molecular Biology, Massachusetts General Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  4. Yu Y Dai

    Molecular Biology, Massachusetts General Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  5. David A Yarmolinsky

    F.M. Kirby Neurobiology Center, Boston Children's Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  6. Tomás Cruz

    Champalimaud Neuroscience Programme, Champalimaud Center for the Unknown, Lisbon, Portugal
    Competing interests
    The authors declare that no competing interests exist.
  7. Adam D Kashlan

    F.M. Kirby Neurobiology Center, Boston Children's Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  8. M Eugenia Chiappe

    Champalimaud Neuroscience Porgramme, Champalimaud Center for the Unknown, Lisbon, Portugal
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1761-0457
  9. Lauren L Orefice

    Molecular Biology, Massachusetts General Hospital, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  10. Clifford J Woolf

    Department of Neurobiology, Harvard Medical School, Boston, United States
    Competing interests
    The authors declare that no competing interests exist.
  11. Christopher D Harvey

    Neurobiology, Harvard Medical School, Boston, United States
    For correspondence
    harvey@hms.harvard.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9850-2268

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.

Reviewing Editor

  1. Mackenzie W Mathis, EPFL, Switzerland

Version history

  1. Received: September 23, 2020
  2. Preprint posted: September 25, 2020 (view preprint)
  3. Accepted: September 1, 2021
  4. Accepted Manuscript published: September 2, 2021 (version 1)
  5. Version of Record published: September 21, 2021 (version 2)

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

  • 9,575
    Page views
  • 966
    Downloads
  • 35
    Citations

Article citation count generated by polling the highest count across the following sources: Scopus, Crossref, PubMed Central.

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)

  1. James P Bohnslav
  2. Nivanthika K Wimalasena
  3. Kelsey J Clausing
  4. Yu Y Dai
  5. David A Yarmolinsky
  6. Tomás Cruz
  7. Adam D Kashlan
  8. M Eugenia Chiappe
  9. Lauren L Orefice
  10. Clifford J Woolf
  11. Christopher D Harvey
(2021)
DeepEthogram, a machine learning pipeline for supervised behavior classification from raw pixels
eLife 10:e63377.
https://doi.org/10.7554/eLife.63377

Further reading

    1. Neuroscience
    Jing Wang, Hamid Azimi ... Gregor Rainer
    Research Article

    The lateral geniculate nucleus (LGN), a retinotopic relay center where visual inputs from the retina are processed and relayed to the visual cortex, has been proposed as a potential target for artificial vision. At present, it is unknown whether optogenetic LGN stimulation is sufficient to elicit behaviorally relevant percepts, and the properties of LGN neural responses relevant for artificial vision have not been thoroughly characterized. Here, we demonstrate that tree shrews pretrained on a visual detection task can detect optogenetic LGN activation using an AAV2-CamKIIα-ChR2 construct and readily generalize from visual to optogenetic detection. Simultaneous recordings of LGN spiking activity and primary visual cortex (V1) local field potentials (LFP) during optogenetic LGN stimulation show that LGN neurons reliably follow optogenetic stimulation at frequencies up to 60 Hz, and uncovered a striking phase locking between the V1 local field potential (LFP) and the evoked spiking activity in LGN. These phase relationships were maintained over a broad range of LGN stimulation frequencies, up to 80 Hz, with spike field coherence values favoring higher frequencies, indicating the ability to relay temporally precise information to V1 using light activation of the LGN. Finally, V1 LFP responses showed sensitivity values to LGN optogenetic activation that were similar to the animal's behavioral performance. Taken together, our findings confirm the LGN as a potential target for visual prosthetics in a highly visual mammal closely related to primates.

    1. Neuroscience
    Yuk-Hoi Yiu, Christian Leibold
    Research Article

    Hippocampal place cell sequences have been hypothesized to serve as diverse purposes as the induction of synaptic plasticity, formation and consolidation of long-term memories, or navigation and planning. During spatial behaviors of rodents, sequential firing of place cells at the theta timescale (known as theta sequences) encodes running trajectories, which can be considered as one-dimensional behavioral sequences of traversed locations. In a two-dimensional space, however, each single location can be visited along arbitrary one-dimensional running trajectories. Thus, a place cell will generally take part in multiple different theta sequences, raising questions about how this two-dimensional topology can be reconciled with the idea of hippocampal sequences underlying memory of (one-dimensional) episodes. Here, we propose a computational model of cornu ammonis 3 (CA3) and dentate gyrus (DG), where sensorimotor input drives the direction-dependent (extrinsic) theta sequences within CA3 reflecting the two-dimensional spatial topology, whereas the intrahippocampal CA3-DG projections concurrently produce intrinsic sequences that are independent of the specific running trajectory. Consistent with experimental data, intrinsic theta sequences are less prominent, but can nevertheless be detected during theta activity, thereby serving as running-direction independent landmark cues. We hypothesize that the intrinsic sequences largely reflect replay and preplay activity during non-theta states.