Probing the neural basis of visual working memory: A validation study using fMRI and fNIRS

摘要

Visual working memory (VWM) plays a key role in visual cognition, comparing percepts and identifying changes in the world as they occur. Previously, fMRI has identified activation in frontal, parietal and temporal areas involved in VWM. Here, we conducted a cross-modal neuroimaging study to determine whether functional near-infrared spectroscopy (fNIRS) was an effective tool to measure changes in activation during VWM processing. We used fNIRS in conjunction with fMRI during an event-related color change detection task with set sizes (SS) 2, 4 and 6. Half of the trials were change trials. Thirteen subjects participated. Positions of sources and detectors were digitized and transformed to a common adult atlas. MonteCarlo simulations generated probability distributions of photon migration for all channels that were then, transformed to MNI space and combined to create subject-specific masks. The thirteen subject-specific masks were combined to create a union mask. fMRI and fNIRS signals were corrected for motion, de-convolved to create maps of beta coefficients, and weighted by the union mask. Voxel-based correlations were computed between fMRI and fNIRS beta coefficients and analyzed. Stronger hemodynamic activation was reported across frontal, parietal, and temporal regions for SS6 than for SS2 for hits, misses, false alarms and correct rejections. Specifically, strong correlations between fMRI and fNIRS were observed in the intra-parietal sulcus (IPS) for Hits at SS 6 than for SS2. Correct Rejections at SS6 elicited greater activation in the middle frontal gyrus than SS2. Further, IPS also showed greater activation on False Alarms than for Misses, for SS2 and SS6. Robust voxel-based correlations between fNIRS and fMRI signals demonstrated that fNIRS is an effective tool to measure functional activation in the VWM network. This is significant because fNIRS is cheap, portable, and can be used with infants and aging and clinical populations.