The influence of the bullseye versus standard fixation cross on eye movements and classifying natural images from EEG

Abstract

Eye movements are a ubiquitous and natural behavior, but in many tightly controlled experimental visual paradigms, eye movements are undesirable. Their occurrence can pose challenges to the interpretation of behavioral and neuroscientific data, in particular for magneto- and electroencephalography (M/EEG), which is sensitive to signals created by eye muscle movement. Here we compared the effect of two different fixation symbols - the standard fixation cross and the bullseye fixation cross - in the context of a visual paradigm with centrally presented naturalistic object images. We investigated eye movements and EEG data recorded simultaneously using behavioral and multivariate analysis techniques. Our findings comparing the bullseye to the standard fixation cross are threefold. First, the bullseye fixation cross reduces the number of saccades and amplitude size of microsaccades. Second, the bullseye subtly reduces classification accuracy in both eye tracking and EEG data for the classification of single object images, but not for the superlevel category animacy. Third, using representational similarity analysis, we found a systematic relationship between eye tracking and EEG data at the level of single images for the standard, but not for the bullseye fixation cross. In conclusion, we recommend the bullseye fixation cross in experimental paradigms with fixation when particularly tight control of fixation is beneficial.