Purpose: To locate cortical activity correlated with pro- and anti-saccade task performance using event-related fMRI.

Method: subjects viewed, through a mirror, stimuli projected onto a screen attached to the back end of a MR-scanner (Siemens Vision 1.5 T). Each 17 s trial started with a central fixation point, which disappeared after 15.3 s. After a temporal gap (200 ms), a saccade target (1500 ms) appeared either 10 deg to the left or right of central fixation. The subject's task was to either make pro-saccades to the target or anti- saccades to the target's mirror location. During task performance, the subject's eye-movements were recorded using an infra-red limbus tracking technique (Kimmig et al., 1999, Exp. Brain. Res., (in press). During performance of a single trial 12 scans (6x128x128 voxels, 4 mm slices, TE = 66 ms, TR = 1.5 s) were made. During each run, 10 trials were presented. Five subjects participated. Eye-movement recordings were used to verify correct performance of the pro- and anti-saccade tasks. Incorrect trials were evaluated separately in the final analysis.

Results: We could establish task-related activity in primary visual cortex, the frontal eye fields and the supplementary motor area.

Conclusion: Using event-related fMRI it is possible to establish saccade-related patterns of cortical activity, while retaining better control over subjects' actual task performance then in conventional block designs. We will discuss the present results in relation to saccade-related activity as found in block designs.