Partitioning variability in animal behavioral videos using semi-supervised variational autoencoders DOI Creative Commons
Matthew R Whiteway, Dan Biderman, Yoni Friedman

et al.

PLoS Computational Biology, Journal Year: 2021, Volume and Issue: 17(9), P. e1009439 - e1009439

Published: Sept. 22, 2021

Recent neuroscience studies demonstrate that a deeper understanding of brain function requires behavior. Detailed behavioral measurements are now often collected using video cameras, resulting in an increased need for computer vision algorithms extract useful information from data. Here we introduce new analysis tool combines the output supervised pose estimation (e.g. DeepLabCut) with unsupervised dimensionality reduction methods to produce interpretable, low-dimensional representations videos more than estimates alone. We this by extracting interpretable features three different head-fixed mouse preparations, as well freely moving open field arena, and show how these can facilitate downstream neural analyses. also produced our model improve precision interpretation analyses compared outputs either fully or

Language: Английский

Brain-wide representations of prior information in mouse decision-making DOI Creative Commons

Charles Findling,

Félix Hubert,

Luigi Acerbi

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: July 4, 2023

The neural representations of prior information about the state world are poorly understood. To investigate them, we examined brain-wide Neuropixels recordings and widefield calcium imaging collected by International Brain Laboratory. Mice were trained to indicate location a visual grating stimulus, which appeared on left or right with probability alternating between 0.2 0.8 in blocks variable length. We found that mice estimate this thereby improve their decision accuracy. Furthermore, report subjective is encoded at least 20% 30% brain regions which, remarkably, span all levels processing, from early sensory areas (LGd, VISp) motor (MOs, MOp, GRN) high level cortical (ACAd, ORBvl). This widespread representation consistent model Bayesian inference involving loops areas, as opposed incorporated only decision-making areas. study offers first perspective encoding cellular resolution, underscoring importance using large scale single standardized task.

Language: Английский

Citations

39
Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling and cloud-native open-source tools DOI
Dan Biderman, Matthew R Whiteway, Cole Hurwitz

et al.

Nature Methods, Journal Year: 2024, Volume and Issue: 21(7), P. 1316 - 1328

Published: June 25, 2024

Language: Английский

Citations

15
Task specificity in mouse parietal cortex DOI Creative Commons
Julie J. Lee, Michael Krumin, Kenneth D. Harris

et al.

Neuron, Journal Year: 2022, Volume and Issue: 110(18), P. 2961 - 2969.e5

Published: Aug. 12, 2022

Parietal cortex is implicated in a variety of behavioral processes, but it unknown whether and how its individual neurons participate multiple tasks. We trained head-fixed mice to perform two visual decision tasks involving steering wheel or virtual T-maze recorded from the same parietal during these Neurons that were active task typically inactive steering-wheel vice versa. Recording apparatus without stimuli yielded specificity as task, suggesting depends on physical context. To confirm this, we some third combining context with environment T-maze. This hybrid engaged those task. Thus, participation by mouse specific, this determined

Language: Английский

Citations

34
An open-source platform for head-fixed operant and consummatory behavior DOI Creative Commons
Adam Gordon-Fennell,

Joumana M Barbakh,

MacKenzie T Utley

et al.

eLife, Journal Year: 2023, Volume and Issue: 12

Published: Aug. 9, 2023

Head-fixed behavioral experiments in rodents permit unparalleled experimental control, precise measurement of behavior, and concurrent modulation neural activity. Here, we present OHRBETS (Open-Source Rodent Behavioral Experimental Training System; pronounced 'Orbitz'), a low-cost, open-source platform hardware software to flexibly pursue the basis variety motivated behaviors. mice tested with displayed operant conditioning for caloric reward that replicates core phenotypes observed during freely moving conditions. also permits optogenetic intracranial self-stimulation under positive or negative procedures real-time place preference like assays. In multi-spout brief-access consumption task, licking as function concentration sucrose, quinine, sodium chloride, modulated by homeostatic circadian influences. Finally, highlight functionality OHRBETS, measured mesolimbic dopamine signals task display strong correlations relative solution value magnitude consumption. All designs, programs, instructions are provided online. This customizable enables replicable consummatory behaviors can be incorporated methods perturb record dynamics vivo.

Language: Английский

Citations

18
Neural representational geometries reflect behavioral differences in monkeys and recurrent neural networks DOI Creative Commons
V. Fascianelli, Aldo Battista, Fabio Stefanini

et al.

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Aug. 1, 2024

Abstract Animals likely use a variety of strategies to solve laboratory tasks. Traditionally, combined analysis behavioral and neural recording data across subjects employing different may obscure important signals give confusing results. Hence, it is essential develop techniques that can infer strategy at the single-subject level. We analyzed an experiment in which two male monkeys performed visually cued rule-based task. The their performance shows no indication they used strategy. However, when we examined geometry stimulus representations state space activities recorded dorsolateral prefrontal cortex, found striking differences between monkeys. Our purely results induced us reanalyze behavior. new showed representational are associated with reaction times, revealing were unaware of. All these analyses suggest using strategies. Finally, recurrent network models trained perform same task, show correlate amount training, suggesting possible explanation for observed differences.

Language: Английский

Citations

9
Rastermap: a discovery method for neural population recordings DOI Creative Commons
Carsen Stringer, Lin Zhong, Atika Syeda

et al.

Nature Neuroscience, Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 16, 2024

Abstract Neurophysiology has long progressed through exploratory experiments and chance discoveries. Anecdotes abound of researchers listening to spikes in real time noticing patterns activity related ongoing stimuli or behaviors. With the advent large-scale recordings, such close observation data become difficult. To find neural data, we developed ‘Rastermap’, a visualization method that displays neurons as raster plot after sorting them along one-dimensional axis based on their patterns. We benchmarked Rastermap realistic simulations then used it explore recordings tens thousands from mouse cortex during spontaneous, stimulus-evoked task-evoked epochs. also applied whole-brain zebrafish recordings; wide-field imaging data; electrophysiological rat hippocampus, monkey frontal various cortical subcortical regions mice; artificial networks. Finally, illustrate high-dimensional scenarios where similar algorithms cannot be effectively.

Language: Английский

Citations

9
Dimensionality reduction beyond neural subspaces with slice tensor component analysis DOI Creative Commons
Arthur Pellegrino, Heike Stein, N. Alex Cayco-Gajic

et al.

Nature Neuroscience, Journal Year: 2024, Volume and Issue: 27(6), P. 1199 - 1210

Published: May 6, 2024

Abstract Recent work has argued that large-scale neural recordings are often well described by patterns of coactivation across neurons. Yet the view variability is constrained to a fixed, low-dimensional subspace may overlook higher-dimensional structure, including stereotyped sequences or slowly evolving latent spaces. Here we argue task-relevant in data can also cofluctuate over trials time, defining distinct ‘covariability classes’ co-occur within same dataset. To demix these covariability classes, develop sliceTCA (slice tensor component analysis), new unsupervised dimensionality reduction method for tensors. In three example datasets, motor cortical activity during classic reaching task primates and recent multiregion mice, show capture more structure using fewer components than traditional methods. Overall, our theoretical framework extends population incorporating additional classes variables capturing structure.

Language: Английский

Citations

8
Beyond nature, nurture, and chance: Individual agency shapes divergent learning biographies and brain connectome DOI Creative Commons
Warsha Barde,

J. Renner,

Brett Addison Emery

et al.

Science Advances, Journal Year: 2025, Volume and Issue: 11(2)

Published: Jan. 10, 2025

Individual choices shape life course trajectories of brain structure and function beyond genes environment. We hypothesized that individual task engagement in response to a learning program results individualized biographies connectomics. Genetically identical female mice living one large shared enclosure freely engaged self-paced, automatically administered monitored tasks. discovered growing increasingly stable interindividual differences trajectories. Adult hippocampal neurogenesis connectivity as assessed by high-density multielectrode array positively correlated with the variation exploration efficiency. During some tasks, divergence transiently collapsed, highlighting sustained significance context for individualization. Thus, equal environments do not result because confronts individuals lead divergent paths.

Language: Английский

Citations

1
An adaptable, reusable, and light implant for chronic Neuropixels probes DOI Open Access
Célian Bimbard, Flóra Takács, Joana Catarino

et al.

Published: Jan. 15, 2025

Electrophysiology has proven invaluable to record neural activity, and the development of Neuropixels probes dramatically increased number recorded neurons. These are often implanted acutely, but acute recordings cannot be performed in freely moving animals neurons tracked across days. To study key behaviors such as navigation, learning, memory formation, must chronically. An ideal chronic implant should (1) allow stable for weeks; (2) reuse after explantation; (3) light enough use mice. Here, we present “Apollo Implant”, an open-source editable device that meets these criteria accommodates up two 1.0 or 2.0 probes. The comprises a “payload” module which is attached probe recoverable, “docking” cemented skull. design adjustable, making it easy change distance between probes, angle insertion, depth insertion. We tested eight labs head-fixed mice, rats. days was stable, even repeated implantations same probe. Apollo provides inexpensive, lightweight, flexible solution reusable recordings.

Language: Английский

Citations

1
Reproducibility ofin vivoelectrophysiological measurements in mice DOI Open Access

Kush Banga,

Julius Benson, Jai Bhagat

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2022, Volume and Issue: unknown

Published: May 9, 2022

Abstract Understanding brain function relies on the collective work of many labs generating reproducible results. However, reproducibility has not been systematically assessed within context electrophysiological recordings during cognitive behaviors. To address this, we formed a multi-lab collaboration using shared, open-source behavioral task and experimental apparatus. Experimenters in ten laboratories repeatedly targeted Neuropixels probes to same location (spanning secondary visual areas, hippocampus, thalamus) mice making decisions; this generated total 121 replicates, unique dataset for evaluating electrophysiology experiments. Despite standardizing both procedures, some outcomes were highly variable. A closer analysis uncovered that variability electrode targeting hindered reproducibility, as did limited statistical power routinely used analyses, such single-neuron tests modulation by parameters. Reproducibility was enhanced histological quality-control criteria. Our observations suggest data from systems neuroscience is vulnerable lack but across-lab standardization, including metrics propose, can serve mitigate this.

Language: Английский

Citations

27