Trial-to-trial variability differentiates motor imagery during observation between low versus high responders: a...

Trial-to-trial variability is a well-known issue in brain signals measured using functional near-infrared spectroscopy (fNIRS). We aimed to investigate whether trial-to-trial variability does provide information about individual performance. Seventeen subjects observed a virtual reality grasping task in first-person view while either imagining (motor imagery during observation, MIO) or imitating (motor execution, ME) the movements. Each condition was performed with the display in one of two positions, a conventional vertical position and a mirrored horizontal position which placed the virtual arm in the correct position relative to the viewpoint. Averaged oxy-hemoglobin concentration Δ[O(2)Hb] showed that the responses could be differentiated into two distinct groups: low responders (LR) and high responders (HR). Within groups, two main sources of trial-to-trial variability were identified: (a) the Δ[O(2)Hb] amplitude, with largest amplitudes in ME conditions (group HR) and smallest amplitudes in MIO conditions (group LR), and (b) the sign of Δ[O(2)Hb], with positive responses occurring most frequently during ME (group HR) and negative responses most frequently during MIO (group LR). Furthermore, the trial-to-trial dynamics differed between groups and could be described in group LR as inverted polynomial U-shaped curve in the mirror conditions (ME-mirror, MIO-mirror). Last, trial-to-trial variability was significantly dependent on task modality, i.e. ME (group HR) versus MIO (group LR), and/or the mirrored display positions (group LR). Our results show a relationship of trial-to-trial variability to individual MI performance, which may be of significance for neurorehabilitation applications. Although the sources of trial-to-trial variability remain unknown, we suggest that they may contribute to future neurofeedback applications.