Efficient fully Bayesian approach to brain activity mapping with complex-valued fMRI data

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Biophysics

Background:

• Functional magnetic resonance imaging (fMRI) detects brain activity using the blood-oxygen-level-dependent (BOLD) signal.
• Traditional fMRI analysis uses only the real-valued magnitude of the BOLD signal.
• Complex-valued fMRI (cv-fMRI) signal analysis, incorporating both real and imaginary components, may enhance detection of neuronal activation.

Purpose of the Study:

• To propose a fully Bayesian model for brain activity mapping using cv-fMRI data.
• To develop a computationally efficient algorithm for processing cv-fMRI data.
• To demonstrate the model's effectiveness and efficiency in simulated and real cv-fMRI experiments.

Main Methods:

• Development of a fully Bayesian model for brain activity mapping with cv-fMRI data.
• Incorporation of temporal and spatial dynamics into the model.
• Implementation of a computationally efficient sampling algorithm using image partitioning and parallel computation.

Main Results:

• The proposed Bayesian model effectively maps brain activity from cv-fMRI data.
• The sampling algorithm significantly enhances processing speed through image partitioning.
• The approach demonstrates competitive performance against state-of-the-art methods in simulated and real data.

Conclusions:

• Analyzing the complex-valued fMRI signal provides a more holistic and potentially powerful approach to brain activity detection.
• The developed Bayesian model and efficient algorithm offer a viable and competitive method for cv-fMRI data analysis.
• This work supports the utility of cv-fMRI for enhanced neuroimaging research.