Spatially and Temporally Predictive Saccades and Their Neural Correlates

Abstract

Prediction is needed to compensate for neural delays between a sensory input and an appropriate motor output. We designed an eye movement task to clarify the behavioural control and neural correlates that are involved in both temporal and spatial prediction. A task involving temporally and spatially predictive and non-predictive saccades was employed in an MRI machine in which four conditions were tested: spatially/temporally predictive (ST-PRED), temporally predictive/spatially non-predictive (T-PRED), spatially predictive/ temporally non-predictive (S-PRED), and spatially/temporally non-predictive (NON). Data from 24 normal human participants (mean age = 22.4 yrs) showed distinct behavioural differences between conditions. All participants elicited primarily predictive saccades (saccadic reaction time: SRT < 100ms) in the ST-PRED condition. The NON condition elicited primarily reactive saccades (SRT > 100ms). The average SRT of the S-PRED condition fell between the average of the ST-PRED and NON conditions, and no significant differences in SRT were observed between the T-PRED and the NON conditions. Analysis of the functional imaging data identified regions with activations that correlated to the predictive conditions. Contrasts of predictive conditions isolating both spatially and temporally predictive areas showed activation of the PEF, insular cortex, and DLPFC which may play a role in the control of predictive saccades. Contrasts that isolated spatially predictive areas also showed activation of the PEF, insular cortex, and DLPFC while contrasts that isolated temporally predictive areas showed select activation of the cerebellum which may serve as the internal clock that drives the regular rhythmic behaviour observed for the temporal aspect of predictive saccades. Surprisingly, activation or frontal areas responsible for saccadic control such as the Frontal and Supplementary Eye fields were equal among all conditions. The behavioural differences validated the activity of the contrasts to isolate brain areas that are correlated with both spatial and temporal prediction. The results from these contrasts indicated that brain activation in the ST-PRED and S-PRED conditions reflects predictive responses to visual stimuli while brain activation in the T-PRED condition reflects motor-timed responses. This suggests that utilizing a predictive saccade task is a valuable tool for simultaneously testing both spatial and temporal prediction that involves fast internally-guided responses.