1. Computational and Systems Biology
  2. Neuroscience
Download icon

Low-dimensional learned feature spaces quantify individual and group differences in vocal repertoires

  1. Jack Goffinet
  2. Samuel Brudner
  3. Richard Mooney
  4. John Pearson  Is a corresponding author
  1. Duke University, United States
Research Article
  • Cited 4
  • Views 1,190
  • Annotations
Cite this article as: eLife 2021;10:e67855 doi: 10.7554/eLife.67855

Abstract

Increases in the scale and complexity of behavioral data pose an increasing challenge for data analysis. A common strategy involves replacing entire behaviors with small numbers of handpicked, domain-specific features, but this approach suffers from several crucial limitations. For example, handpicked features may miss important dimensions of variability, and correlations among them complicate statistical testing. Here, by contrast, we apply the variational autoencoder (VAE), an unsupervised learning method, to learn features directly from data and quantify the vocal behavior of two model species: the laboratory mouse and the zebra finch. The VAE converges on a parsimonious representation that outperforms handpicked features on a variety of common analysis tasks, enables the measurement of moment-by-moment vocal variability on the timescale of tens of milliseconds in the zebra finch, provides strong evidence that mouse ultrasonic vocalizations do not cluster as is commonly believed, and captures the similarity of tutor and pupil birdsong with qualitatively higher fidelity than previous approaches. In all, we demonstrate the utility of modern unsupervised learning approaches to the quantification of complex and high-dimensional vocal behavior.

Data availability

Dataset 1---------Online, publicly available MUPET dataset: ~5GB Available at: https://github.com/mvansegbroeck/mupet/wiki/MUPET-wiki Figs: 2, 3, 4d-eDataset 2----------Single zebra finch data: ~200-400 MB of audio generated as part of work in progress in Mooney Lab. Figs: 2e-f, 4a-c, 5a, 5d, 6b-eDataset 3---------Mouse USV dataset: ~30-40 GB of audio generated as part of work in progress in Mooney Lab. Figs: 4fDataset 5---------This is a subset of dataset 3, taken from a single mouse: ~1GB of audio. Figs: 5b-e, 6aDataset 6---------10 zebra finch pupil/tutor pairs: ~60 GB of audio generated as part of work in progress in Mooney Lab. Figs: 7Upon acceptance, all Datasets 2-6 will be archived in the Duke Digital Repository (https://research.repository.duke.edu). DOI in process.

The following previously published data sets were used

Article and author information

Author details

  1. Jack Goffinet

    Computer Science, Duke University, Durham, United States
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6729-0848
  2. Samuel Brudner

    Neurobiology, Duke University, Durham, United States
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6043-9328
  3. Richard Mooney

    Department of Neurobiology, Duke University, Durham, United States
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-3308-1367
  4. John Pearson

    Biostatistics & Bioinformatics, Neurobiology, Center for Cognitive Neuroscience, Psychology and Neuroscience, Electrical and Computer Engineering, Duke University, Durham, United States
    For correspondence
    john.pearson@duke.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9876-7837

Funding

National Institute of Mental Health (R01-MH117778)

  • Richard Mooney

National Institute of Neurological Disorders and Stroke (R01-NS118424)

  • Richard Mooney
  • John Pearson

National Institute on Deafness and Other Communication Disorders (R01-DC013826)

  • Richard Mooney
  • John Pearson

National Institute of Neurological Disorders and Stroke (R01-NS099288)

  • Richard Mooney

Eunice Kennedy Shriver National Institute of Child Health and Human Development (F31-HD098772)

  • Samuel Brudner

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 data generated in conjunction for this study were generated by experiments performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All of the animals were handled according to approved institutional animal care and use committee (IACUC) protocols of Duke University, protocol numbers A171-20-08 and A172-20-08.

Reviewing Editor

  1. Jesse H Goldberg, Cornell University, United States

Publication history

  1. Received: February 24, 2021
  2. Accepted: May 12, 2021
  3. Accepted Manuscript published: May 14, 2021 (version 1)
  4. Version of Record published: June 18, 2021 (version 2)

Copyright

© 2021, Goffinet 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

  • 1,190
    Page views
  • 148
    Downloads
  • 4
    Citations

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

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)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Computational and Systems Biology
    2. Medicine
    Ryan P O'Hara et al.
    Research Article

    Hypertrophic cardiomyopathy (HCM) is associated with risk of sudden cardiac death (SCD) due to ventricular arrhythmias (VAs) arising from the proliferation of fibrosis in the heart. Current clinical risk stratification criteria inadequately identify at-risk patients in need of primary prevention of VA. Here, we use mechanistic computational modeling of the heart to analyze how HCM-specific remodeling promotes arrhythmogenesis and to develop a personalized strategy to forecast risk of VAs in these patients. We combine contrast-enhanced cardiac magnetic resonance imaging and T1 mapping data to construct digital replicas of HCM patient hearts that represent the patient-specific distribution of focal and diffuse fibrosis and evaluate the substrate propensity to VA. Our analysis indicates that the presence of diffuse fibrosis, which is rarely assessed in these patients, increases arrhythmogenic propensity. In forecasting future VA events in HCM patients, the imaging-based computational heart approach achieved 84.6%, 76.9%, and 80.1% sensitivity, specificity, and accuracy, respectively, and significantly outperformed current clinical risk predictors. This novel VA risk assessment may have the potential to prevent SCD and help deploy primary prevention appropriately in HCM patients.

    1. Computational and Systems Biology
    2. Neuroscience
    Elliot H Smith et al.
    Research Article

    Interictal epileptiform discharges (IEDs), also known as interictal spikes, are large intermittent electrophysiological events observed between seizures in patients with epilepsy. Though they occur far more often than seizures, IEDs are less studied, and their relationship to seizures remains unclear. To better understand this relationship, we examined multi-day recordings of microelectrode arrays implanted in human epilepsy patients, allowing us to precisely observe the spatiotemporal propagation of IEDs, spontaneous seizures, and how they relate. These recordings showed that the majority of IEDs are traveling waves, traversing the same path as ictal discharges during seizures, and with a fixed direction relative to seizure propagation. Moreover, the majority of IEDs, like ictal discharges, were bidirectional, with one predominant and a second, less frequent antipodal direction. These results reveal a fundamental spatiotemporal similarity between IEDs and ictal discharges. These results also imply that most IEDs arise in brain tissue outside the site of seizure onset and propagate toward it, indicating that the propagation of IEDs provides useful information for localizing the seizure focus.