An Interactive Brain Atlas of Knowledge

Area of Science:

• Neuroscience
• Computational Biology
• Bioinformatics

Background:

• Biomedical knowledge is rapidly expanding, creating challenges for human comprehension.
• Existing literature mining tools struggle to integrate semantic methods with neuroimaging standards.
• This gap hinders the full utilization of biomedical data in brain research.

Purpose of the Study:

• To develop a semantic meta-analysis platform, TVBase, for projecting biomedical knowledge onto a standardized 3D brain.
• To bridge the gap between literature-based biomedical information and neuroimaging data.
• To support translational and multi-scale computational neuroscience.

Main Methods:

• Utilized the SCAIView platform to extract biomedical entities and their brain anatomy associations from PubMed.
• Employed entropy-based calculations to determine association strengths between concepts and anatomical terms.
• Developed a transformation matrix to map over 800 anatomical terms to voxel coordinates of a standard brain atlas.

Main Results:

• Successfully projected biomedical knowledge from over 36 million articles onto a standardized 3D brain.
• Enabled region-specific information extraction linking biomedical concepts to brain anatomy.
• Created a novel method for knowledge projection supporting computational neuroscience.

Conclusions:

• TVBase facilitates hypothesis-free neuroimaging pattern interpretation and hypothesis generation.
• The platform supports applications in personalized medicine by integrating literature data with brain structure.
• TVBase is an open-source Python library promoting accessibility in computational neuroscience research.