Neuroscience Cloud Analysis As a Service: An open-source platform for scalable, reproducible data analysis

Abe, Taiga; Kinsella, Ian; Saxena, Shreya; Buchanan, E. Kelly; Couto, João; Briggs, John D.; Kitt, Sian Lee; Glassman, Ryan; Zhou, John; Paninski, Liam; Cunningham, John P.

doi:10.1016/j.neuron.2022.06.018

Cited by 21 publications

(11 citation statements)

References 83 publications

(119 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Often, developers spend more time in hardware setup and installation than configuring, training, and evaluating pose tracking models (see e.g. [25] for further discussion). We therefore developed a cloud application that supports the full life cycle of animal pose estimation (Extended Data Fig.…”

Section: Resultsmentioning

confidence: 99%

“…5), the Lightning Pose semi-supervised context model (LP; middle column), and the EKS applied to the LP models (LP+EKS; right column). We define several quality metrics in this setting to quantify the accuracy of the different models 2 . One such quality metric can be computed by comparing the "vertical" pupil diameter (the difference of the vertical position of the top and bottom pupil keypoints) vs the "horizontal" diameter (the difference of the horizontal positions of the left and right keypoints).…”

Section: Significantly Improved Tracking Accuracy On Large-scale Publ...mentioning

confidence: 99%

“…We further improve our networks’ predictions using a general Bayesian post-processing approach, which we coin the Ensemble Kalman Smoother : we aggregate (“ensemble”) the predictions of multiple networks – which is known to improve their accuracy and robustness [24, 25] — and model those aggregated predictions with a spatially-constrained Kalman smoother that takes their collective uncertainty into account.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling, and cloud-native open-source tools

Biderman

Whiteway

Hurwitz

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Pose estimation algorithms are shedding new light on animal behavior and intelligence. Most existing models are only trained with labeled frames (supervised learning). Although effective in many cases, the fully supervised approach requires extensive image labeling, struggles to generalize to new videos, and produces noisy outputs that hinder downstream analyses. We address each of these limitations with a semi-supervised approach that leverages the spatiotemporal statistics of unlabeled videos in two different ways. First, we introduce unsupervised training objectives that penalize the network whenever its predictions violate smoothness of physical motion, multiple-view geometry, or depart from a low-dimensional subspace of plausible body configurations. Second, we design a new network architecture that predicts pose for a given frame using temporal context from surrounding unlabeled frames. These context frames help resolve brief occlusions or ambiguities between nearby and similar-looking body parts. The resulting pose estimation networks achieve better performance with fewer labels, generalize better to unseen videos, and provide smoother and more reliable pose trajectories for downstream analysis; for example, these improved pose trajectories exhibit stronger correlations with neural activity. We also propose a Bayesian post-processing approach based on deep ensembling and Kalman smoothing that further improves tracking accuracy and robustness. We release a deep learning package that adheres to industry best practices, supporting easy model development and accelerated training and prediction. Our package is accompanied by a cloud application that allows users to annotate data, train networks, and predict new videos at scale, directly from the browser.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Significantly Improved Tracking Accuracy On Large-scale Publ...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling, and cloud-native open-source tools

Biderman

Whiteway

Hurwitz

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Essential tasks, including manual curation of clustered spikes and artifact rejection, are often hidden or irretrievable from the written record. Efforts to reproduce findings are hampered by idiosyncratic data and code organization, poor documentation, and obscured vital details, including hardware requirements 7 . In collaborations among multiple scientists, these problems can be further exacerbated due to the variability in how each participant carries out analysis.…”

Section: Introductionmentioning

confidence: 99%

Spyglass: a framework for reproducible and shareable neuroscience research

Lee,

Denovellis,

et al. 2024

Preprint

View full text Add to dashboard Cite

Sharing data and reproducing scientific results are essential for progress in neuroscience, but the community lacks the tools to do this easily for large datasets and results obtained from intricate, multi-step analysis procedures. To address this issue, we created Spyglass, an open-source software framework designed to promote the shareability and reproducibility of data analysis in neuroscience. Spyglass integrates standardized formats with reliable open-source tools, offering a comprehensive solution for managing neurophysiological and behavioral data. It provides well-defined and reproducible pipelines for analyzing electrophysiology data, including core functions like spike sorting. In addition, Spyglass simplifies collaboration by enabling the sharing of final and intermediate results across custom, complex, multi-step pipelines as well as web-based visualizations. Here we demonstrate these features and showcase the potential of Spyglass to enable findable, accessible, interoperable, and reusable (FAIR) data management and analysis by applying advanced state space decoding algorithms to publicly available data from multiple laboratories.

show abstract

“…The electrophysiology community has therefore begun migrating data and workloads to flexible, scalable pipelines that run in the public cloud (e.g. DANDI (Halchenko et al 2022), NeuroCAAS (Abe et al 2022), DataJoint (Yatsenko et al 2015), and Code Ocean (Cheifet 2021)). Cloud platforms offer the ability to remotely launch standardized analysis environments that read from a single, immutable dataset, eliminating the need for individuals and labs to maintain local copies.…”

Section: Introductionmentioning

confidence: 99%

Compression strategies for large-scale electrophysiology data

Buccino,

Winter,

Bryant

et al. 2023

J. Neural Eng.

View full text Add to dashboard Cite

With the rapid adoption of high-density electrode arrays for recording neural activity, electrophysiology data volumes within labs and across the field are growing at unprecedented rates. For example, a one-hour recording with a 384-channel Neuropixels probe generates over 80 GB of raw data. These large data volumes carry a high cost, especially if researchers plan to store and analyze their data in the cloud. Thus, there is a pressing need for strategies that can reduce the data footprint of each experiment. Here, we establish set of benchmarks for comparing the performance of various compression algorithms on experimental and simulated recordings from Neuropixels 1.0 (NP1) and 2.0 (NP2) probes. For lossless compression, audio codecs (FLAC and WavPack) achieve compression ratios 6% higher for NP1 and 10% higher for NP2 than the best general-purpose codecs, at the expense of a slower decompression speed. For lossy compression, the WavPack algorithm in “hybrid mode” increases the compression ratio from 3.59 to 7.08 for NP1 and from 2.27 to 7.04 for NP2 (compressed file size of ∼14% for both types of probes), without adverse effects on spike sorting accuracy or spike waveforms. Along with the tools we have developed to make compression easier to deploy, these results should encourage all electrophysiologists to apply compression as part of their standard analysis workflows.

show abstract

Neuroscience Cloud Analysis As a Service: An open-source platform for scalable, reproducible data analysis

Cited by 21 publications

References 83 publications

Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling, and cloud-native open-source tools

Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling, and cloud-native open-source tools

Spyglass: a framework for reproducible and shareable neuroscience research

Compression strategies for large-scale electrophysiology data

Contact Info

Product

Resources

About