Dynamic time warping-based averaging framework for functional near-infrared spectroscopy brain imaging studies

Abstract

We investigate the problem related to the averaging procedure in functional near-infrared spectroscopy (fNIRS) brain imaging studies. Typically, to reduce noise and to empower the signal strength associated with task-induced activities, recorded signals (e.g., in response to repeated stimuli or from a group of individuals) are averaged through a point-by-point conventional averaging technique. However, due to the existence of variable latencies in recorded activities, the use of the conventional averaging technique can lead to inaccuracies and loss of information in the averaged signal, which may result in inaccurate conclusions about the functionality of the brain. To improve the averaging accuracy in the presence of variable latencies, we present an averaging framework that employs dynamic time warping (DTW) to account for the temporal variation in the alignment of fNIRS signals to be averaged. As a proof of concept, we focus on the problem of localizing task-induced active brain regions. The framework is extensively tested on experimental data (obtained from both block design and event-related design experiments) as well as on simulated data. In all cases, it is shown that the DTW-based averaging technique outperforms the conventional-based averaging technique in estimating the location of task-induced active regions in the brain, suggesting that such advanced averaging methods should be employed in fNIRS brain imaging studies.

Dyadic Sex Composition and Task Classification Using fNIRS Hyperscanning Data

“…The similarity score computed using DTW might thus more accurately describe how well two time series match each other when compared to a more conventional Euclidean distance measurement, as similar features will still be detected even if they do not line up exactly. Zhu et al [37] successfully employed DTW to average fNIRS signals and localize cortical brain activity. Fig.…”