Prior to saccade initiation several cortical areas are activated and this pattern of activation depends on the type of eye movement. Accordingly, the organisation of a visually guided saccade requires the activation of parietal areas and this phenomenon is expressed on the cortical potential obtained at about 200 ms prior to saccade onset over the parietal areas. In the case of remembered saccades the pattern of cortical activation should be more complex than that of visually guided saccades.

We recorded the cortical potential in an experimental paradigm with three variants. First, the pure remembered saccade task (MEM task) in which the target was flashed followed by a variable delay of 1 - 6 s. At the end of delay period the go signal permits the execution of a saccade towards the position of the flash. In the second task (DEL task) the target remained on for a period of 1 - 6 s and the saccade was made towards the still visible target. The third task (DEL₁) was a variant of the DEL task but the target disappears with the appearance of the go signal.

The EEG was recorded with 25 electrodes (10/20 IO system) and the movements of the right eye with the IRIS device. The off-line analysis consisted of averaging procedures with the trigger on the sensory (peripheral target) as well as the motor (saccade onset) events.

The pattern of cortical activation linked with the peripheral target was qualitatively similar between tasks. At about 500 - 600 ms after the peripheral target the EEG activity reaches a maximum over the central -frontal areas and then starts to decline in a such a way that at the end of long delay trials the EEG activity is found to oscillate near the base line. On the other hand the EEG activity with the MEM task is significantly larger over middle-left anterior areas than with the two delay tasks and this difference shows a weak but significantly correlation (R = 0.21) with the accuracy of the remembered saccade. As for the potentials obtained with triggering the saccade onset, the two delay tasks are followed by larger activation than the MEM task over the posterior areas. It seems that the presence of the peripheral target enhances the cortical activation for, additionally, the DEL task produces larger activity than the DEL₁ task.

In summary I conclude that the anterior activation obtained with the MEM task could be considered as a memory related effect. However, it remains to be seen whether this phenomenon can been assigned to working memory mechanisms supposed to participate in the programming of remembered saccades.