Combining optogenetic stimulation and fMRI to validate a multivariate dynamical systems model for estimating causal...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Causal estimation models like state-space multivariate dynamical systems (MDS) are used to infer directed functional interactions in the brain.
• The accuracy and validity of these computational methods using fMRI data remain poorly understood.
• Previous evaluations often rely on simplified simulations with limited biological relevance.

Purpose of the Study:

• To evaluate the accuracy of MDS models for estimating causal functional interactions in human brain imaging data.
• To leverage optogenetic fMRI (ofMRI) for in vivo validation of causal inference techniques.
• To assess the reliability of MDS in detecting directed connectivity from stimulated to downstream brain regions.

Main Methods:

• Optogenetic fMRI (ofMRI) was employed to selectively stimulate the right primary motor cortex (M1) in two independent cohorts.
• High-resolution whole-brain fMRI data were recorded during M1 stimulation.
• State-space multivariate dynamical systems (MDS) were applied to analyze the ofMRI data and estimate causal interactions.

Main Results:

• MDS accurately identified causal interactions from M1 to the thalamus in Cohort 1 and to the caudate-putamen (CPu) in Cohort 2.
• No significant causal influences were detected in the reverse direction (thalamus/CPu to M1).
• Control analyses confirmed the specificity of the estimated causal links between stimulated and activated regions.

Conclusions:

• State-space multivariate dynamical systems (MDS) can accurately and reliably estimate causal functional interactions in fMRI data.
• Optogenetic fMRI (ofMRI) provides a powerful in vivo approach for validating computational causal inference methods.
• This study validates MDS for causal interaction analysis in fMRI and highlights the utility of ofMRI for evaluating neuroimaging analysis techniques.