Related Experiment Video
Updated: Jul 24, 2025
08:45
Mapping Cortical Dynamics Using Simultaneous MEG/EEG and Anatomically-constrained Minimum-norm Estimates: an Auditory Attention Example
Published on: October 24, 2012
14.7K
Temporal Mapper: Transition networks in simulated and real neural dynamics
Mengsen Zhang1,2, Samir Chowdhury1, Manish Saggar1
1Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA, USA.
Network Neuroscience (Cambridge, Mass.)
|July 3, 2023
Summary
Researchers developed Temporal Mapper, a novel method using topological data analysis to map brain state transitions from time series data. This approach bridges data-driven and mechanistic modeling, showing significant links to behavioral performance in fMRI studies.
Area of Science:
- Neuroscience
- Complex Systems
- Computational Biology
Background:
- Characterizing large-scale brain dynamics requires integrating data-driven and mechanistic modeling approaches.
- A significant challenge lies in the conceptual translation between these modeling paradigms due to varying prior knowledge assumptions.
- Brain dynamics are conceptualized as a complex landscape where modulations can induce transitions between stable brain states (attractors).
Purpose of the Study:
- To bridge the gap between data-driven and mechanistic modeling of brain dynamics.
- To introduce a novel method, Temporal Mapper, for reconstructing attractor transition networks from time series data.
- To validate the method's efficacy and explore its empirical relevance in neuroscience.
Main Methods:
- Utilized topological data analysis tools to develop the Temporal Mapper method.
- Employed a biophysical network model to generate simulated time series data with a ground-truth attractor transition network for theoretical validation.
- Applied the Temporal Mapper to functional magnetic resonance imaging (fMRI) data from a continuous multitask experiment for empirical validation.
Main Results:
- Temporal Mapper successfully reconstructed ground-truth attractor transition networks from simulated time series data, outperforming existing time-varying methods.
- Analysis of fMRI data revealed significant associations between the occupancy of high-degree nodes and cycles in the transition network and subjects' behavioral performance.
- The study demonstrated the method's capability to identify dynamic organizational principles in brain activity.
Conclusions:
- Temporal Mapper provides a novel, data-driven approach to inferring brain state transition networks from time series.
- This method offers a significant step towards integrating data-driven and mechanistic modeling frameworks for understanding brain dynamics.
- The findings highlight the link between network topology of brain state transitions and cognitive performance, suggesting potential biomarkers for brain function.
