Generating dynamical neuroimaging spatiotemporal representations (DyNeuSR) using topological data analysis.

Area of Science:

• Neuroscience
• Data Science
• Computational Biology

Background:

• Standard neuroimaging analyses often average data, potentially obscuring individual brain dynamics.
• Topological Data Analysis (TDA) offers novel methods to represent complex, high-dimensional neuroimaging data.
• Techniques like Mapper distill brain data into simplified, interpretable graphical representations.

Purpose of the Study:

• To present an open-source neuroinformatics platform for TDA-based neuroimaging analysis.
• To facilitate the exploration, analysis, and validation of TDA-derived graphical representations.
• To enable researchers and clinicians to gain biological insights into complex mental disorders.

Main Methods:

• Development of a Python-based platform using Jupyter notebooks.
• Implementation of tools for visualizing and interacting with TDA-generated graphical representations.
• Utilizing the Mapper algorithm to distill high-dimensional neuroimaging data.

Main Results:

• The platform allows interactive exploration of Mapper's intermediate stages.
• It provides methods to assess the influence of Mapper parameters on data representations.
• The tools facilitate grounding TDA representations in neurophysiology and behavior.

Conclusions:

• The open-source platform democratizes the use of TDA in neuroimaging.
• It empowers researchers to explore topological representations of brain data.
• This approach holds promise for generating novel biological insights into neurological and psychiatric conditions.