Cortical Dynamics during the Preparation of Antisaccadic and Prosaccadic Eye Movements in Humans in a Gap Paradigm

Abstract

To compare the cortical dynamics of different oculomotor tasks, EEG and eye movements were recorded in 21 volunteers. Using a comprehensive approach, subjects were asked to perform saccadic tasks, which included a saccadic eye movement to a peripheral target (prosaccadic), a movement to the opposite side (antisaccadic), or maintain the gaze fixed (no-go). In mixed trials, prosaccadic, antisaccadic and no-go tasks were indicated by a color square (S1) present for 1900–2500 ms (instructive period). S1 disappeared for 370 ms (gap) and a black dot at 8 deg at right or left indicated the beginning of the task. Reaction times, amplitude of eye movements and number of errors were greatest in antisaccadic tasks, suggesting a greater difficulty. The EEG showed a contingent negativity variation (CNV) that increased progressively along the instructive period and suddenly during the gap: higher in antisaccadic, followed by prosaccadic and no-go tasks. Principal component analysis (PCA) disentangled fronto-central and occipital CNV-related and fronto-central gap-related components. The instructive period was characterized by fronto-central and occipital beta desynchronization (ERD) higher in antisaccadic than in no-go and parieto-occipital alpha synchronization higher in no-go than in antisaccadic tasks. During the gap, parieto-occipital beta and alpha ERD were higher in antisaccadic compared to no-go. The gap was further characterized by a fronto-central increase of inter-trial coherence in theta: highest during antisaccadic, followed by prosaccadic and no-go tasks. This phase locking in theta was also accompanied by theta ERS, which was significantly higher in antisaccadic than in the other two tasks. In PCA of spectral power two main components had dynamics similar to those extracted from voltage data, suggesting cross-frequency coupling. These results suggest that the more difficult saccadic tasks are associated with top-down control mediated by frontal cortex, while simpler tasks rely more on bottom-up control mediated by posterior cortices.