Related Experiment Video
Updated: May 14, 2026
08:36
Dynamic Inter-subject Functional Connectivity Reveals Moment-to-Moment Brain Network Configurations Driven by Continuous or Communication Paradigms
Published on: March 21, 2019
Performance bounds for dynamic causal modeling of brain connectivity.
Shun Chi Wu1, A Lee Swindlehurst
1Department of Electrical Engineering and Computer Science, University of California, Irvine, CA 92697, USA. scwu@uci.edu
Summary
This study derives Cramér-Rao bounds for nonlinear dynamic causal models (DCM) used in brain imaging. These bounds assess the accuracy of estimating causal interactions between brain regions from EEG/MEG data under various conditions.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Biophysics
Background:
- Complex dynamical models are used to understand brain connectivity and causal interactions.
- Dynamic Causal Models (DCM) aim to mimic event-related potentials from EEG/MEG for brain functionality analysis.
- Accurate estimation of DCM parameters is crucial for analyzing effective connectivity.
Purpose of the Study:
- Derive Cramér-Rao performance bounds for nonlinear dynamic causal model (DCM) parameter estimates.
- Examine how operating conditions influence the accuracy of DCM parameter estimation.
- Provide a theoretical framework for assessing the reliability of brain connectivity analyses.
Main Methods:
- Focus on a class of nonlinear dynamic causal models (DCM) characterized by connectivity parameters.
- Inference of DCM parameters using simulated or empirical EEG/MEG data.
- Derivation and analysis of Cramér-Rao lower bounds (CRB) for parameter estimation accuracy.
Main Results:
- Established theoretical Cramér-Rao bounds for nonlinear DCM parameter estimation.
- Quantified the impact of factors like noise and sampling rate on estimation precision.
- Demonstrated how source localization accuracy affects the performance bounds.
Conclusions:
- The derived Cramér-Rao bounds provide a benchmark for the achievable accuracy in DCM parameter estimation.
- Understanding these bounds is essential for optimizing experimental designs and data acquisition parameters.
- This work contributes to more reliable inferences of causal brain network dynamics.
