Systematic Survey of Public Datasets for Behavioral Research in Invertebrate Models: Toward FAIR and Standardized Data Sharing
Behavioral datasets for invertebrate model organisms are rapidly expanding alongside automated imaging, tracking, and artificial intelligence (AI) based phenotyping, yet their technical structure and compliance with the Findable, Accessible, Interoperable and Reusable (FAIR) principles remain heterogeneous. We present a two-stage survey of openly available behavioural datasets for major invertebrate models Caenorhabditis elegans (C. elegans), Drosophila melanogaster (D. melanogaster), Galleria mellonella (G. mellonella), and planarians Schmidtea mediterranea (S. mediterranea) with larval zebrafish (Danio rerio) included as a vertebrate comparator. Stage 1 comprised a PRISMA-guided literature review (from 2015 to 2025) across indexed databases and complementary non-indexed sources, yielding 12 eligible publications describing 12 open behavioural datasets. Stage 2 independently screened and technically evaluated repository deposits (from June 2022 to July 2025), producing a final corpus of 20 datasets scored on a four-dimension ordinal rubric capturing usability, annotation richness, technical quality and AI-readiness. All extracted descriptors, repository search logs, and scoring sheets are released as public data records enabling full regeneration of figures and summary statistics. Across Stage 2 deposits, multimodality and open file formats were common, whereas interoperability and AI-readiness were most constrained by limited machine-readable metadata, weak raw-to-derived provenance, and sparse adoption of formal standards or ontologies. This Data Descriptor provides a reproducible, dataset-centred overview of behavioural resources for invertebrate models and practical guidance for FAIR-aligned publication, secondary biological analyses, and AI benchmarking.