Functional overestimation due to spatial smoothing of fMRI data.

Area of Science:

• Neuroimaging
• Functional Magnetic Resonance Imaging (fMRI)
• Statistical Analysis

Background:

• Pearson correlation is a fundamental technique in neuroimage analysis.
• Spatial smoothing's impact on functional overestimation in fMRI lacks complete theoretical understanding.
• Correlation scale invariance offers a novel perspective to explain these effects.

Purpose of the Study:

• To provide a theoretical explanation for functional overestimation caused by spatial smoothing in fMRI.
• To investigate the relationship between image noise, spatial smoothing, and temporal correlation (tcorr) maps.
• To offer guidance on optimal spatial smoothing parameters for correlation-based fMRI analysis.

Main Methods:

• Utilized theoretical analysis based on correlation scale invariance.
• Employed numerical simulations and subject experiments with fMRI data.
• Calculated temporal correlations (tcorr) of voxel timecourses against task timecourses.

Main Results:

• Spatial smoothing reduces noise, which smooths the tcorr map and expands estimated neuroactivity.
• fMRI data smoothing can lead to overestimation of activation spots in correlation maps.
• Minimal smoothing (FWHM ≤ 2 voxels) is recommended for correlation-based functional mapping.

Conclusions:

• Spatial smoothing's effect on correlation-based fMRI analysis is understood through noise reduction and smoothing effects.
• Image noise shapes functional activity blobs in tcorr maps.
• Careful application of spatial smoothing is crucial for accurate interpretation of fMRI results.