Exploring effective multiplicity in multichannel functional near-infrared spectroscopy using eigenvalues of...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Statistical Analysis

Background:

• Multichannel functional near-infrared spectroscopy (fNIRS) enables broad cortical coverage but necessitates controlling for family-wise errors (FWEs) due to increased statistical multiplicity.
• Traditional Bonferroni methods strictly control Type I errors (false positives) but can increase Type II errors (false negatives), especially with numerous channels.
• Balancing Type I and Type II errors is crucial for accurate interpretation of fNIRS data.

Purpose of the Study:

• To explore the feasibility of using effective multiplicity (M) in multichannel fNIRS studies.
• To evaluate M as an alternative to Bonferroni-based methods for error control in fNIRS.
• To maintain a balance between Type I and Type II errors in fNIRS data analysis.

Main Methods:

• Applied the effective multiplicity (M) method, derived from correlation matrix eigenvalues, to multichannel fNIRS data.
• Utilized resampling simulations on three experimental datasets with varying activation patterns.
• Assessed M's performance in a 44-channel fNIRS setting.

Main Results:

• Effective multiplicity (M) was controlled between 10 and 15 in the 44-channel fNIRS setup.
• The number of significantly activated channels remained consistent irrespective of the total number of measured channels.
• The M method demonstrated effective control over statistical errors.

Conclusions:

• The effective multiplicity (M) approach is a feasible and effective alternative to Bonferroni-based methods for multichannel fNIRS.
• This method offers a better balance between controlling false positives and false negatives in fNIRS studies.
• Implementing M can improve the statistical rigor and interpretability of multichannel fNIRS findings.