Does brain functional connectivity alter across similar trials during imaging experiments?

In a typical functional brain imaging experiment, brain activities in response to several trials are measured. Recorded signals are then averaged across similar trials for each channel/voxel to obtain time-courses associated with the task of interest. Majority of functional connectivity studies employ these averaged signals to study brain's functional connections across channels/voxels. The assumption here, however, is that functional connections among brain networks do not change across trials during the course of the experiment. In this paper, we argue that this assumption may not always hold true. Using Functional Near-Infrared Spectroscopy (fNIRS), brain activities of five healthy adults in response to modified visual oddball task are recorded. Wavelet transform coherence (WTC) is then used to assess functional connectivity by considering recordings corresponding to three scenarios: i) the first half of total number of similar trials, ii) the second half of total number similar trials, and iii) the entire number of similar trials. Nonparametric permutation testing is utilized to examine the statistical difference in functional connectivity when assessed in these three scenarios. Observed differences suggest that brain's functional connectivity across similar trials changes during the course of the experiment, potentially due to changes in functional connections among brain networks as a result of task repetition.