Test-retest reliability of functional near infrared spectroscopy in infants.

There has been a rapid rise in the number of publications using functional near infrared spectroscopy (fNIRS) for human developmental research over the past decade. However test-retest reliability of this measure of brain activation in infants remains unknown. To assess this, we utilized data from a longitudinal cohort who participated in an fNIRS study on social perception at two age points. Thirteen infants had valid data from two sessions held 8.5 months apart (4 to 8 months and 12 to 16 months). Inter- and intrasession fNIRS test-retest reliability was assessed at the individual and group levels using the oxyhemoglobin ([Formula: see text]) signal. Infant compliance with the study was similar in both sessions (assessed by the proportion of time infants looked to the stimuli), and there was minimal discrepancy in sensor placement over the targeted area between sessions. At the group level, good spatial overlap of significant responses and signal reliability was seen (spatial overlap was 0.941 and average signal change within an region of interest was [Formula: see text]). At participant level, spatial overlap was acceptable ([Formula: see text] on average across infants) although signal reliability varied between participants. This first study of test-retest reliability of fNIRS in infants shows encouraging results, particularly for group-based analysis.