A Framework for Collaborative Curation of Neuroscientific Literature

Area of Science:

• Computational Neuroscience
• Neuroscience Informatics

Background:

• Extracting parameters for large neuronal circuit models from literature is complex and prone to errors.
• A systematic approach is needed for curating and integrating neuroscientific data.

Purpose of the Study:

• To introduce a literature curation framework for systematic annotation and management of model parameters.
• To facilitate the creation of shareable, curated corpora of neuroscientific annotations.

Main Methods:

• Developed a framework including an annotation format, Python API (NeuroAnnotation Toolbox; NAT), and graphical interface (NeuroCurator).
• Implemented methods for annotating text, figures, equations, and tables, linking parameters to experimental values.
• Utilized ontological terms for context and bounding boxes for parameter provenance.

Main Results:

• Released a corpus of modeling parameters for thalamo-cortical circuitry.
• Demonstrated the framework's ability to systematically annotate and manage large parameter sets.

Conclusions:

• The framework enhances traceability and classification of literature information for neuronal models.
• Enables more efficient and systematic integration of neuroscientific data into models and analyses.