Related Experiment Video
Updated: Sep 23, 2025
14:14
Targeting Neuronal Fiber Tracts for Deep Brain Stimulation Therapy Using Interactive, Patient-Specific Models
Published on: August 12, 2018
9.0K
Reliability and subject specificity of personalized whole-brain dynamical models
Justin W M Domhof1, Simon B Eickhoff1, Oleksandr V Popovych1
1Institute of Neuroscience and Medicine, Brain and Behaviour (INM-7), Research Centre Jülich, Jülich, Germany; Institute for Systems Neuroscience, Medical Faculty, Heinrich Heine University Düsseldorf, Düsseldorf, Germany.
Neuroimage
|May 17, 2022
Summary
Dynamical whole-brain models show variable reliability and subject specificity. Enhanced personalization improves parameter reliability, while simulated functional connectivity (FC) can outperform empirical FC.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Systems Neuroscience
Background:
- Dynamical whole-brain models integrate structural connectivity (SC) and functional connectivity (FC).
- These models explore brain dynamics and their relation to behavioral, clinical, and demographic traits.
- The reliability and subject specificity of these models, considering empirical FC variability, remain under-investigated.
Purpose of the Study:
- To comprehensively assess the reliability and subject specificity of dynamical whole-brain model parameters and simulated functional connectivity (FC).
- To investigate the impact of model personalization, complexity, and brain parcellation on modeling outcomes.
- To evaluate the structure-function relationship within these modeling frameworks.
Main Methods:
- Fitted parameters of dynamical whole-brain models under varying implementation paradigms.
- Evaluated model reliability and subject specificity across linear, phase oscillator, and neural mass network models.
- Compared simulated FC with empirical FC and assessed structure-function relationships using empirical SC.
Main Results:
- Model parameter reliability varied from poor to good, with enhanced personalization improving reliability.
- Model complexity did not consistently affect reliability; neural mass models showed poor reliability but enhanced subject specificity.
- Simulated FC outperformed empirical FC in reliability and subject specificity.
- Accurate identification of individual subject's simulated FC from SC reached 70% for non-linear models.
- Brain parcellation had a more pronounced effect on modeling results than on empirical data.
Conclusions:
- Dynamical whole-brain modeling offers valuable insights but requires careful consideration of reliability and subject specificity.
- Enhanced model personalization is crucial for improving parameter reliability.
- Simulated FC holds potential for superior reliability and subject specificity compared to empirical FC.
- Future applications should integrate reliability estimates for robust findings and consider the impact of parcellation choices.
Keywords:
Brain connectomeReliabilityResting-state brain dynamicsSubject specificityWhole-brain model
