Implicit and explicit timing in oculomotor control

Abstract

The passage of time can be estimated either explicitly, e.g. before leaving home in the morning, or implicitly, e.g. when catching a flying ball. In the present study, the latency of saccadic eye movements was used to evaluate differences between implicit and explicit timing. Humans were required to make a saccade between a central and a peripheral position on a computer screen. The delay between the extinction of a central target and the appearance of an eccentric target was the independent variable that could take one out of four different values (400, 900, 1400 or 1900 ms). In target trials, the delay period lasted for one of the four durations randomly. At the end of the delay, a saccade was initiated by the appearance of an eccentric target. Cue&target trials were similar to target trials but the duration of the delay was visually cued. In probe trials, the duration of the upcoming delay was cued, but there was no eccentric target and subjects had to internally generate a saccade at the estimated end of the delay. In target and cue&target trials, the mean and variance of latency distributions decreased as delay duration increased. In cue&target trials latencies were shorter. In probe trials, the variance increased with increasing delay duration and scalar variability was observed. The major differences in saccadic latency distributions were observed between visually-guided (target and cue&target trials) and internally-generated saccades (probe trials). In target and cue&target trials the timing of the response was implicit. In probe trials, the timing of the response was internally-generated and explicitly based on the duration of the visual cue. Scalar timing was observed only during probe trials. This study supports the hypothesis that there is no ubiquitous timing system in the brain but independent timing processes active depending on task demands.

Figure 1. Schematic drawing of the sequence of events in the different trial types presented to subjects.

Target trial: The trial started with the appearance of a fixation cross for a randomized duration followed by the appearance of two empty ‘boxes’, one at the center of the screen and at a 9-deg eccentric position. After the appearance of the two boxes, a target was flashed in the central one for 50 milliseconds. Extinction of the central target marked the beginning of the delay period that could last either 400, 900, 1400 or 1900 ms. At the end of the delay period, a target appeared for 50 ms in the eccentric box and the subject had to make a saccade to the eccentric box within a 400 ms grace period. Cue&target trial: The trial started with a fixation cross (same duration as target condition) that was followed by the cue period when a red disk was presented on the screen for one of the four durations tested randomly. A short fixation period followed disk appearance and the two empty boxes appeared on the screen. The end of the trial was similar as in a target trial but delay duration was always equal to cue duration. Probe trial: same as sequence as cue&target trial but the eccentric target appeared with a fixed probability at the end of the delay period.

Figure 2. Histograms of saccadic absolute latencies in target trials.

Time zero on the abscissa represents the beginning of the delay period. The time elapsed until the appearance of the eccentric target is represented with vertical dashed lines for the four different durations tested. The ordinate represents the number of saccades in the 10-ms bins.

Figure 3. Group data target trials.

A: Mean relative latency (±2 SE) as a function of delay duration in target trials. B: Latency variance as a function of delay duration.

Figure 4. Comparison of variable and fixed foreperiods.

Mean latency (±2 SE) as a function of delay duration in target trials. Variable foreperiod (Variable) and fixed foreperiod blocks of trials (Fixed). In the variable foreperiod condition, mean latencies were longer for 400 ms delay duration. An opposite trend was found in the fixed foreperiod condition. Group data from 6 subjects (6/9) who participated in this control experiment.

Figure 5. Influence of previous delay duration.

A: Mean relative latency during the current trial (‘n’) as a function of delay duration during the previous trial (‘n-1’, colored curves). B: Data circled with the dashed ellipse in figure A (400 ms delay). Group data.

Figure 6. Histograms of absolute saccadic latencies in cue&target trials.

The ordinate represents the percentage of saccades in the 100-ms bins for each of the 4 delay durations independently. The abscissa represents the time elapsed until the appearance of the eccentric target (vertical dashed lines).

Figure 7. Comparison of target and cue&target trials.

A: Mean relative latency (±2 SE) as a function of delay duration in cue&target (dashed line) and, for comparison, target trials (continuous line). B: Latency variance in cue&target and target trials.

Figure 8. Histograms of saccadic absolute latencies in the probe trials.

X-axis: saccadic absolute latencies; Y-axis: number of saccades in the 100-ms bins. Note the increasing spread of the latencies with increasing delay duration. Vertical dashed lines: time of target appearance in cue&target trials.

Figure 9. Group data in probe trials.

A: mean absolute latency (±2 SE) as a function of delay duration. B: Latency variance for the same data. C: Mean relative latency (±2 SE). The horizontal dashed line represents the transition between saccades occurring after the end of the cued duration (positive values) or before (negative values).

Figure 10. Influence of previous trial type on sequence effect.

Y-axis: mean relative latency in milliseconds. X-axis: duration of the delay (cue) during trial ‘n’. Colors: delay (cue) duration during trial ‘n-1’. A: Probe trials preceded by a cue&target trial. B: Probe trials preceded by a probe trial.

Figure 11. Saccadic latencies in relative time.

A: Normalized saccadic latencies in relative time (mean saccadic latency divided by delay duration). B: Median relative time as a function of delay duration. The horizontal dotted line represents the boundary between overestimation and underestimation.

Figure 12. Slope analysis.

A. Variance as a function of latency squared in target trials. B. Variance as a function of latency squared in probe trials.