Integrated multi-echo denoising strategy improves identification of inherent language laterality

Area of Science:

• Neuroimaging
• Brain Mapping
• Functional Magnetic Resonance Imaging

Background:

• Resting-state fMRI (rs-fMRI) is widely used in neuroscience and clinical research.
• A key challenge in rs-fMRI is distinguishing neuronal BOLD signal from noise.
• Effective noise reduction is crucial for accurate brain activity mapping.

Purpose of the Study:

• To evaluate a novel denoising technique, integrated multi-echo rs-fMRI analysis, for language mapping.
• To assess the impact of this approach on the specificity and laterality of language representation.
• To compare the proposed method against conventional denoising strategies.

Main Methods:

• Acquired multiband multi-echo rs-fMRI data from 12 subjects.
• Utilized language task fMRI to define language areas.
• Compared seed-based correlation analysis results from three rs-fMRI preprocessing approaches: integrated multi-echo ICA, TE-dependent denoising, and conventional denoising.
• Computed laterality index automatically to ensure objectivity.

Main Results:

• The optimal combination of 3-echo data enhanced correlations within the language system.
• The multi-echo denoising approach revealed synchronization areas more specific to the language system compared to conventional methods.
• This improved the identification of language system laterality.

Conclusions:

• The integrated multi-echo approach effectively reduces widespread non-specific correlations.
• This method enhances language mapping accuracy and the identification of language system laterality using rs-fMRI.
• The findings support the utility of advanced denoising techniques for precise brain function analysis.