Poster B63, Sunday, March 26, 8:00 – 10:00 am, Pacific Concourse
Using fNIRS to Investigate Speech-Language Tasks
Nicholas Wan1, Allison Hancock1, Ronald Gillam1; 1Utah State University
Multiple authors have claimed that functional near-infrared spectroscopy (fNIRS) is a useful imaging technique for tasks that involve speech production because data are minimally susceptible to motion artifacts. This study was designed to test how speech processes related to articulation and voicing influence patterns of cortical hemodynamics and their interpretation and if it is possible to further reduce motor-related activity from the speech-language signal. Participants completed three reading tasks (oral reading, silent mouthing, and silent reading) while undergoing fNIRS imaging. We compared three measures of the hemodynamic response function (amplitude, periodicity, and slope) for each task across five regions of interest (ROIs): primary motor cortex (M1), supplementary motor area (SMA), inferior frontal gyrus (IFG), superior temporal gyrus (STG), and inferior parietal lobule (IPL). There were significant main effects for task for M1, SMA, and IPL. Greater detail of how ROIs are functionally connected were computed via Granger Causality, revealing stronger networks between motor areas during oral reading and stronger networks between language areas during silent reading. These results suggest that motion artifacts related to speech processes have minimal effects on fNIRS measures of the hemodynamic response function across the parasylvian region. Regression was used to reduce shared variance between tasks involving jaw movement in order to reveal an underlying activity similar to silent reading. This technique was successful, strongly correlating with silent reading. This suggests it is possible to further reduce non-speech- and non-language-related activity from the speech-language network, improving the signal quality from fNIRS recordings.
Topic Area: LANGUAGE: Other