Open science and data sharing in cognitive neuroscience with MouseBytes and MouseBytes+

Neuroscience data sharing has been challenging due to non-standardized formats and metadata issues. Open Science and FAIR principles (Findable, Accessible, Interoperable, Reusable) aim to address this, with success seen in neuroimaging and genomics. However, behavioral data from animal models, especially rodents, lack similar standardization. Traditional behavioral experiments lack automation and standardized outputs, limiting data sharing. Computerized, automated behavioral testing using touchscreen technology offers a solution by generating standardized digital data. This technology, used in over 400 labs worldwide, allows for easy synchronization with neuro-technologies such as fiber photometry, miniscopes, and optogenetics, enabling investigation of neural activity alongside behavior. MouseBytes is a web application designed to facilitate data sharing and analysis of these standardized cognitive datasets, supporting transparency and reproducibility in rodent cognitive neuroscience research.

Existing platforms like PRIME Data Exchange (for neuroimaging) and ENCODE (for genomics) showcase the benefits of open science data sharing in neuroscience. The authors cite several papers demonstrating the use of open data repositories and their impact on research reproducibility and collaboration. The lack of similar integration in behavioral data analysis in animal models is highlighted as a significant gap that hinders the progress of neuroscience research. The authors review the existing literature and the various attempts at developing such open-access repositories within neuroscience.

MouseBytes is a web application built using ASP.NET WEB API framework, Angular 5.0 (client-side), and .NET Core 2.0 (server-side). It connects to a Microsoft SQL Server 2017 database. Data is collected using touchscreen technology and exported as XML files (containing all necessary features) or CSV files (allowing for customized extraction). A quality control procedure checks for errors before data entry. The platform allows users to create experiments, sub-experiments, and animal profiles with associated metadata. Data visualization is implemented using TIBCO Spotfire, accessible to both public and private datasets (for logged-in users). MouseBytes+ extends the platform to incorporate complementary data from various neuro-technologies (fiber photometry, MRI, miniscopes, etc.) linked to the cognitive behavioral data in MouseBytes. DataCite is used to generate DOIs for public datasets, enhancing data discoverability and citability. The platform also integrates with the Canadian Open Neuroscience Platform (CONP) via DataLad, ensuring public datasets remain updated and are easily accessible. The code repository for MouseBytes is available on Github under the GPL 3.0 license.

MouseBytes currently houses data from over 3000 mice, spanning a range of animal disease models, cognitive tasks, ages, and sexes. The website has been viewed over 30,000 times since 2019, with 13,000 homepage visits from 4600 unique users. Datasets have been downloaded 868 times, and 1200 unique download links have been generated. The MouseBytes dashboard provides aggregated key metrics and allows interactive filtering of data based on task, age, sex, and genotype. The platform currently includes 88 datasets covering 10 cognitive tasks, with 31 public and 57 private datasets. Four cognitive tasks (5-CSRTT, PVD, PAL, CPT) account for over 90% of data points, primarily from amyloid and TDP-43 mouse models. Data from younger mice (3-6 months) are most prevalent. MouseBytes+ allows users to upload various data types (processed datasets, software, video/audio, other files) linked to MouseBytes data and external repositories via DOIs or unique links. The platform's search functionality allows users to find relevant data through multiple parameters such as Author, PIs, Repository Title, File Type, and dataset features, promoting data sharing and utilization.

MouseBytes and MouseBytes+ meet the FAIR data principles and the NIH scientific data management and sharing policy. The platforms offer unique persistent identifiers, open access to public data, quality control, and secure data storage. The touchscreen technology and the associated platforms are widely used. The integration of MouseBytes with Touchscreen Cognition (an online community) and CONP improves data discoverability and community engagement. The platform enables meta-analyses across different laboratories, facilitating assessment of robustness and reproducibility. Furthermore, MouseBytes enables the formation of scientific consortia without formal membership requirements. It assists researchers in meeting journal and funding agency data availability mandates and promotes data equity by providing free access to curated cognitive behavioral data. The multimodal integration of data in MouseBytes+ is a significant advancement for understanding brain-behavior relationships.

MouseBytes and MouseBytes+ represent significant advancements in open science for cognitive neuroscience. They provide a robust, FAIR-compliant platform for sharing and analyzing rodent cognitive data, integrating behavioral and neuro-technology data. The platforms' wide adoption highlights their value in promoting reproducibility, collaboration, and equitable access to valuable research data. Future work could focus on expanding the range of supported neuro-technologies and cognitive tasks, and on further developing data analysis tools within the platform.

While the platform offers comprehensive functionalities, there might be limitations related to the ongoing evolution of the touchscreen technology. Future updates and improvements will be necessary to ensure compatibility with novel technologies and research methods. Furthermore, the reliance on user input for metadata could lead to inconsistencies. Strategies for improving metadata quality and consistency could include automation or standardized templates.