Sprecher
Beschreibung
Recent advances in imaging and computational methods have led to a rapid expansion of video-based tracking in the biological sciences. The increased adoption of video tracking assays (VTAs) has created new opportunities and challenges for quantitative analysis across multiple scales of biological systems. However, there is a lack of coherent community standards and a metadata reporting schema which are necessary for reproducibility, interoperability and consistency in experimental findings.
The project Behavioral Standard Metadata (BeStMeta) aims to build a generalized reporting schema for VTAs. In our vision, the schema needs to be adaptable to different experimental setups, organisms (e.g., zebrafish, rodents) and analysis pipelines ensuring flexibility for different VTA setups. All metadata needs to be reported in machine-readable formats in order to support interoperability and further processing.
In this work, we present initial results of a comprehensive survey of existing metadata standards on published studies in ecotoxicology and neuroscience using VTAs. Relevant publications were identified through a structured literature search using the open-source PubMed and PMC APIs. We then combined manual extraction with an automated pipeline based on regular expressions and large language models (LLM). These approaches were employed to systematically survey the papers and
to create a comprehensive overview of all the metadata that is reported consistently or inconsistently across behavioral studies. Preliminary results on 84 publications indicate major gaps in metadata and data reporting: only 18 studies shared video files in public repositories and less than half reported technical metadata such as camera model, frame rate, or resolution highlighting systemic shortcomings in current practices. In summary, this survey highlights the necessity of establishing a unified metadata reporting schema for VTAs.
| Alternative Track | 1. Advancing FAIR Metadata with AI: Methods, Challenges, and Synergies |
|---|