The antisaccade task has proven highly useful in basic and clinical neuroscience, and the neural structures involved are well documented. However, the cognitive and neural mechanisms that mediate task performance are not yet understood. An event-related fMRI study was designed to dissociate the neural correlates of two putative key functions, volitional saccade generation and inhibition of reflexive saccades, and to investigate their interaction. Nineteen healthy volunteers performed a task that required (a) to initiate saccades volitionally, either with or without a simultaneous demand to inhibit a reflexive saccade; and (b) to inhibit a reflexive saccade, either with or without a simultaneous demand to initiate a saccade volitionally. Analysis of blood oxygen level-dependent signal changes confirmed a major role of the frontal eye fields and the supplementary eye fields in volitional saccade generation. Inhibition-related activation of a specific fronto-parietal network was highly consistent with previous evidence involved in inhibitory processes. Unexpectedly, there was little evidence of specific brain activation during combined generation and inhibition demands, suggesting that the neural processing of generation and inhibition in antisaccades is independent to a large extent.

Schizophrenia (SZ) patients showed increased volitional saccade latencies, suggesting deficient volitional initiation of action. Yet increased volitional saccade latencies may also result from deficits in attention shifts. To dissociate attention shifting and saccade initiation, we asked 25 SZ patients and 25 healthy subjects to make saccades toward newly appearing (onset) targets and toward the loci of disappearing (offset) targets. Similar onsets and offsets were also used as attention cues in a Posner-type manual task. As expected, onsets and offsets had similar effects on attention. In contrast, saccade latencies were considerably longer with offset compared to onset targets, reflecting additional time for volitional saccade initiation. Unexpectedly, SZ patients had normal saccade latencies. Presumably, the expected deficit was compensated by decreased fixation-related neural activity, which was induced by the disappearance of fixation stimuli.

Slowed initiation of volitional but not visually guided saccades indicates impaired volitional action control in schizophrenia patients (SZ). The present study aimed at identifying neural correlates of this specific deficit. Fourteen SZ and 13 healthy control participants (HC) underwent functional magnetic resonance imaging while performing volitional and visually guided saccades. SZ showed increased latencies in volitional but not in visually guided saccades. Brain activation during volitional saccades compared to visually guided saccades was increased in SZ compared to HC in several areas: the supplementary eye fields, suggesting inefficient production of volitional saccades; the prefrontal cortex, pointing to altered top down control on complex eye movements; and the left middle temporal area, suggesting changes in early sensory and attention processing during the volitional control of saccades in SZ.

Although externally as well as internally-guided eye movements allow us to flexibly explore the visual environment, their differential neural mechanisms remain elusive. A better understanding of these neural mechanisms will help us to understand the control of action and to elucidate the nature of cognitive deficits in certain psychiatric populations (e.g. schizophrenia) that show increased latencies in volitional but not visually-guided saccades. Both the superior precentral sulcus (sPCS) and the intraparietal sulcus (IPS) are implicated in the control of eye movements. However, it remains unknown what differential contributions the two areas make to the programming of visually-guided and internally-guided saccades. In this study we tested the hypotheses that sPCS and IPS distinctly encode internally-guided saccades and visually-guided saccades. We scanned subjects with fMRI while they generated visually-guided and internally-guided delayed saccades. We used multi-voxel pattern analysis to test whether patterns of cue related, preparatory and saccade related activation could be used to predict the direction of the planned eye movement. Results indicate that patterns in the human sPCS predicted internally-guided saccades but not visually-guided saccades in all trial periods and patterns in the IPS predicted internally-guided saccades and visually-guided saccades equally well. The results support the hypothesis that the human sPCS and IPS make distinct contributions to the control of volitional eye movements.

No abstract