Inter-session reliability of prefrontal cortex functional near-infrared spectroscopy measures during dual-task walking protocols of different duration in stroke survivors and healthy controls.

To advance understanding and support recovery of walking abilities after stroke, functional near-infrared spectroscopy (fNIRS) is a promising tool to quantify brain activity during walking. However, the reliability of fNIRS-derived parameters recorded during walking has not been sufficiently studied. This study aimed to determine the relative and absolute inter-session reliability of oxy- and deoxyhemoglobin concentrations (HbO/HbR) in the prefrontal cortex (PFC) during two different motor-cognitive dual-task walking protocols. These protocols consisted of one 2 min and four 30 s walking intervals during which spatio-temporal gait parameters were additionally recorded. Twelve stroke survivors (64.1 ± 7.4 years; 6 females) and nineteen healthy controls (71.3 ± 6.6 years; 14 females) participated in the study. Inter-session reliability was determined using intra-class correlation coefficients and Bland-Altman analysis. The reliability of HbO/HbR was poor during the 2 min protocol and fair to excellent during the 30 s protocol for half of the PFC subareas in both groups. This may be due to a relatively increased task-related prefrontal cortex activation during gait initiation, the averaging of multiple blocks and potentially reduced mind-wandering during the shorter baseline measurement of the 30 s protocol. In contrast, the reliability of most spatio-temporal gait parameters was slightly better for the 2 min protocol, but overall good to excellent for both protocols and cohorts. Based on these findings, it is recommended to use shorter protocols consisting of multiple blocks. Future studies might also adopt strategies to minimize potential mind-wandering during fNIRS baseline measurements. Additionally, implementing a familiarization session prior to the measurements might reduce practice effects.