[Brain functional network reconstruction based on compressed sensing and fast iterative shrinkage-thresholding...

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Signal Processing

Background:

• Resting-state functional magnetic resonance imaging (fMRI) is crucial for mapping brain functional networks.
• Conventional methods often suffer from significant noise, compromising the accuracy of analysis and interpretation.
• Overfitting is a common issue in network reconstruction, leading to unreliable results.

Purpose of the Study:

• To introduce a novel approach for reconstructing brain functional networks using compressed sensing.
• To enhance the accuracy of noise reduction and network reconstruction in fMRI data.
• To improve the reliability of brain function analysis in noisy environments.

Main Methods:

• Utilized the least absolute shrinkage and selection operator (LASSO) model, a compressed sensing technique.
• Employed the L1-norm penalty term to induce sparsity and prevent overfitting.
• Implemented the fast iterative shrinkage-thresholding algorithm (FISTA) for efficient model optimization and convergence to a global optimum.

Main Results:

• Achieved over 98% accuracy in noise reduction and brain functional network reconstruction.
• Demonstrated effective suppression of noise in fMRI data.
• Showcased superior performance compared to existing methods in experimental evaluations.

Conclusions:

• The LASSO-based compressed sensing method offers a robust solution for reconstructing accurate brain functional networks.
• This approach effectively mitigates noise interference, enhancing the study of human brain function.
• The findings support the utility of advanced signal processing techniques in neuroimaging research.