Real-time, low-latency closed-loop feedback using markerless posture tracking

  1. Gary A Kane
  2. Gonçalo Lopes
  3. Jonny L Saunders
  4. Alexander Mathis
  5. Mackenzie W Mathis  Is a corresponding author
  1. Harvard University, United States
  2. NeuroGears, United Kingdom
  3. University of Oregon, United States
  4. EPFL, Switzerland

Abstract

The ability to control a behavioral task or stimulate neural activity based on animal behavior in real-time is an important tool for experimental neuroscientists. Ideally, such tools are noninvasive, low-latency, and provide interfaces to trigger external hardware based on posture. Recent advances in pose estimation with deep learning allows researchers to train deep neural networks to accurately quantify a wide variety of animal behaviors. Here we provide a new DeepLabCut-Live! package that achieves low-latency real-time pose estimation (within 15 ms, >100 FPS), with an additional forward-prediction module that achieves zero-latency feedback, and a dynamic-cropping mode that allows for higher inference speeds. We also provide three options for using this tool with ease: (1) a stand-alone GUI (called DLC-Live! GUI, and integration into (2) Bonsai and (3) AutoPilot. Lastly, we benchmarked performance on a wide range of systems so that experimentalists can easily decide what hardware is required for their needs.

Data availability

All models, data, test scripts and software is already released and made freely available on GitHub: https://github.com/DeepLabCut/DeepLabCut-live

Article and author information

Author details

  1. Gary A Kane

    The Rowland Institute at Harvard, Harvard University, Cambridge, United States
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7703-5055
  2. Gonçalo Lopes

    NeuroGears, London, United Kingdom
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0731-4945
  3. Jonny L Saunders

    Institute of Neuroscience, Department of Psychology, University of Oregon, Eugene, United States
    Competing interests
    The authors declare that no competing interests exist.
  4. Alexander Mathis

    Life Sciences, EPFL, Geneva, Switzerland
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3777-2202
  5. Mackenzie W Mathis

    Brain Mind Institute, EPFL, Genève, Switzerland
    For correspondence
    mackenzie.mathis@epfl.ch
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7368-4456

Funding

Chan Zuckerberg Initiative (EOSS)

  • Alexander Mathis
  • Mackenzie W Mathis

National Science Foundation (1309047)

  • Jonny L Sanders

The Rowland Institute at Harvard, Harvard University

  • Gary A Kane
  • Alexander Mathis
  • Mackenzie W Mathis

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 mouse work was carried out under the permission of the IACUC at Harvard University (#17-07-309). Dog videos and feedback was exempt from IACUC approval (with conformation with IACUC).

Reviewing Editor

  1. Gordon J Berman, Emory University, United States

Publication history

  1. Received: August 8, 2020
  2. Accepted: December 6, 2020
  3. Accepted Manuscript published: December 8, 2020 (version 1)
  4. Accepted Manuscript updated: December 9, 2020 (version 2)
  5. Version of Record published: January 4, 2021 (version 3)

Copyright

© 2020, Kane 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

  • 10,090
    Page views
  • 1,138
    Downloads
  • 44
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, Scopus, 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. Gary A Kane
  2. Gonçalo Lopes
  3. Jonny L Saunders
  4. Alexander Mathis
  5. Mackenzie W Mathis
(2020)
Real-time, low-latency closed-loop feedback using markerless posture tracking
eLife 9:e61909.
https://doi.org/10.7554/eLife.61909

Further reading

    1. Cancer Biology
    2. Computational and Systems Biology
    Matthew Roberts, Julia Ogden ... Carlos Lopez-Garcia
    Tools and Resources Updated

    Lung squamous cell carcinoma (LUSC) is a type of lung cancer with a dismal prognosis that lacks adequate therapies and actionable targets. This disease is characterized by a sequence of low- and high-grade preinvasive stages with increasing probability of malignant progression. Increasing our knowledge about the biology of these premalignant lesions (PMLs) is necessary to design new methods of early detection and prevention, and to identify the molecular processes that are key for malignant progression. To facilitate this research, we have designed XTABLE (Exploring Transcriptomes of Bronchial Lesions), an open-source application that integrates the most extensive transcriptomic databases of PMLs published so far. With this tool, users can stratify samples using multiple parameters and interrogate PML biology in multiple manners, such as two- and multiple-group comparisons, interrogation of genes of interests, and transcriptional signatures. Using XTABLE, we have carried out a comparative study of the potential role of chromosomal instability scores as biomarkers of PML progression and mapped the onset of the most relevant LUSC pathways to the sequence of LUSC developmental stages. XTABLE will critically facilitate new research for the identification of early detection biomarkers and acquire a better understanding of the LUSC precancerous stages.

    1. Computational and Systems Biology
    2. Neuroscience
    Rebecca Elizabeth Carlisle, Arthur D Kuo
    Research Article Updated

    Humans make a number of choices when they walk, such as how fast and for how long. The preferred steady walking speed seems chosen to minimize energy expenditure per distance traveled. But the speed of actual walking bouts is not only steady, but rather a time-varying trajectory, which can also be modulated by task urgency or an individual’s movement vigor. Here we show that speed trajectories and durations of human walking bouts are explained better by an objective to minimize Energy and Time, meaning the total work or energy to reach destination, plus a cost proportional to bout duration. Applied to a computational model of walking dynamics, this objective predicts dynamic speed vs. time trajectories with inverted U shapes. Model and human experiment (N=10) show that shorter bouts are unsteady and dominated by the time and effort of accelerating, and longer ones are steadier and faster and dominated by steady-state time and effort. Individual-dependent vigor may be characterized by the energy one is willing to spend to save a unit of time, which explains why some may walk faster than others, but everyone may have similar-shaped trajectories due to similar walking dynamics. Tradeoffs between energy and time costs can predict transient, steady, and vigor-related aspects of walking.