Intra-subject test-retest reliability for auditory-evoked functional near-infrared spectroscopy responses: effects of systemic physiology correction

• Published in: Neurophotonics (Print) Vol. 12; no. 1; p. 015015
• Main Authors: Sinfield, Victoria C., Aaker, Dalton, Metzger, Abigail, Tong, Yunjie, Shader, Maureen J.
• Format: Journal Article
• Language: English
• Published: United States Society of Photo-Optical Instrumentation Engineers 01.01.2025; SPIE
• Subjects: Hemoglobin; Infrared spectroscopy; Physiological aspects; Research Papers; hemodynamic signal denoising; test-retest reliability; systemic physiology augmented functional near-infrared spectroscopy; systemic physiology; intra-subject variability
• ISSN: 2329-423X; 2329-4248
• DOI: 10.1117/1.NPh.12.1.015015

Functional near-infrared spectroscopy (fNIRS) is a valuable neuroimaging tool for non-invasively measuring hemodynamic changes in response to neural activity, particularly in auditory research. Although fNIRS shows strong test-retest reliability at the group level, individual-subject level reliability is often compromised by systemic noise. We investigate how correcting for systemic-physiological signals affects reliability in single-subject fNIRS data. fNIRS data were collected from one participant over 10 sessions during a passive auditory task. Using general linear modeling, six correction approaches were compared: no correction, physiology correction, short-channel correction, short-channel + physiology correction, short-channel + physiology + lag correction, and short-channel + tCCA correction. Intraclass correlation coefficient analysis revealed that physiology correction yielded the highest test-retest reliability score, whereas short-channel correction had the lowest. These results align with previous findings suggesting that global systemic artifacts bolster reliability, and regressing such artifacts enhances the clarity of the observed neuronal response, as supported by visual comparisons of raw and denoised signals. We highlight the impact of correcting for extra-cerebral signals in single-subject auditory research and demonstrate that, while incorporating short channels in fNIRS data collection may reduce reliability, it offers a more accurate representation of the neuronal response.