[Strategy on dealing with noisy NIRS data: implications on functional neuroimaging on swallowing]

Area of Science:

• Neuroimaging techniques
• Physiological monitoring
• Signal processing in neuroscience

Context:

• Functional near-infrared spectroscopy (fNIRS) is a promising tool for monitoring physiological functions like swallowing due to its motion artifact resilience.
• However, motion artifacts remain a significant technical challenge in fNIRS, particularly during dynamic tasks such as swallowing.
• Existing research offers technical insights into managing motion artifacts in fNIRS data.

Purpose:

• To present two distinct analytical strategies for processing fNIRS data contaminated with motion artifacts.
• To demonstrate how temporal data structures influence the choice of optimal analytical methods for fNIRS.
• To highlight effective approaches for extracting meaningful cortical activation signals despite movement-related noise.

Summary:

• Two case studies illustrate methods for analyzing fNIRS data with motion artifacts: a general linear model with hemodynamic response function regression for continuous data (language tasks) and signal averaging for disrupted data (ADHD go/no-go tasks).
• The first approach leverages temporal continuity for detailed cortical activation analysis during overt naming.
• The second approach simplifies temporal structures via averaging to isolate robust signals in data affected by unexpected motion artifacts, as seen in ADHD children's task performance.

Impact:

• Establishes that the optimal fNIRS data analysis strategy is contingent upon the specific temporal characteristics of the recorded data.
• Provides practical methodologies for researchers to mitigate motion artifact challenges in fNIRS studies.
• Enhances the reliability and interpretability of fNIRS findings across diverse experimental paradigms and populations.