Related Experiment Video
Updated: May 10, 2026
Conducting Hyperscanning Experiments with Functional Near-Infrared Spectroscopy
Published on: January 19, 2019
Statistical analysis of fNIRS data: a comprehensive review.
1Bio Imaging & Signal Processing Lab., Dept. of Bio & Brain Engineering, Korea Advanced Institute of Science and Technology (KAIST), 373-1 Guseong-dong, Yuseong-gu, Daejon 305-701, Republic of Korea; Rotman Research Institute at Baycrest Centre, University of Toronto, Toronto, Ontario M6A 2E1, Canada.
Functional near-infrared spectroscopy (fNIRS) offers direct brain activity measurement but requires robust statistical analysis. This review details fNIRS signal processing, artifact correction, and advanced statistical methods for accurate neuronal activity extraction.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Functional near-infrared spectroscopy (fNIRS) is a non-invasive brain imaging technique measuring hemodynamic responses.
- fNIRS offers high temporal resolution and portability compared to fMRI, PET, and MEG.
- fNIRS signals are susceptible to noise and physiological interference, necessitating advanced statistical analysis.
Purpose of the Study:
- To provide a comprehensive review of statistical analysis methods for fNIRS data.
- To discuss historical developments and current techniques in fNIRS signal processing and inference.
- To present a unified framework for understanding various statistical approaches in fNIRS.
Main Methods:
- Review of statistical techniques including motion artifact correction, PCA/ICA, and FDR.
- Explanation of inference methods like t-test, ANOVA, and SPM.
- Introduction of a linear mixed-effects model with ReML for a unified statistical view.
Main Results:
- Detailed overview of historical and contemporary statistical approaches for fNIRS.
- Demonstration that common fNIRS inference methods are special cases of the linear mixed-effects model.
- Identification of open issues and future directions in fNIRS statistical analysis.
Conclusions:
- Effective statistical analysis is crucial for accurate neuronal signal extraction from noisy fNIRS data.
- A unified statistical framework, such as linear mixed-effects models, can encompass various existing fNIRS analysis techniques.
- Further research is needed to address remaining challenges in fNIRS statistical methodology.
