Article Figures & Data
Figures
- Figure 1.
Task and behavior. a, Schematic of eye-tracking and display setup. Observers fixated on a central cross and all stimuli appeared in the lower right quadrant. b, Trial timeline for two-target tasks (Experiments 1 and 3). In Experiment 3, a probe grating appeared after the response cue, which the observer adjusted to estimate orientation (not shown). c, Trial timeline for three-target task (Experiment 2). d, Performance accuracy normalized to average neutral performance for each observer, mean and SEM. Experiments 1–3, n = 30. V, Valid; N, neutral; I, invalid. **p < 0.01; ***p < 0.001.
- Figure 2.
Microsaccade rate. a, Mean microsaccade rate across the trial. Data are combined across the common precue conditions of Experiments 1–3 (neutral, T1, T2, shown as separate colored lines). Dashed vertical lines show trial events. The response cue is not shown because its timing differs for two-target and three-target tasks. Light gray shading shows the pretarget inhibition period used for statistical analysis and arrow indicates the posttarget rebound. b, Enlargement of pretarget inhibition period labeled in a. Dark gray shading shows significant cluster-corrected time windows, T1 < neutral, p < 0.05.
- Figure 3.
Microsaccade timing. a, Raster plot showing microsaccade (MS) onset times (blue ticks) in 40 precue T1 trials for an example observer from Experiment 2. Dashed vertical lines show trial events. For each trial, the latencies of the last pre-T1 MS and first post-T1 MS were recorded to quantify inhibition and rebound timing, respectively. b, Distribution across trials of inhibition (last pre-T1 MS) latencies for another example observer from Experiment 2. Precue conditions are shown in different colors. c, Inhibition latency distributions (last pre-T1 MS) for the group of observers. Latencies were z-scored to combine across observers with different overall timings and MS rates. Colored lines and shaded regions show mean and SEM of the estimated probability density for each precue condition. Vertical lines show the medians of the group distributions. d, Summary of group latencies. Markers and error bars show mean and SEM of each observer's median z-scored latency for each precue condition. The absolute magnitude of the latency difference between precue T1 and neutral conditions was 37 ms. e and f correspond to c and d, but show the rebound (first post-T1 MS) latencies. Experiments 1–3, n = 30; Experiment 2, n = 9.
- Figure 4.
Directions of inhibition (last pre-T1) and rebound (first post-T1) microsaccades. Polar histograms show the proportion of trials with microsaccades in each direction of the total number of trials with pre-T1/post-T1 microsaccades. Target stimuli were positioned at 315°.
- Figure 5.
Relation between microsaccades and behavior. a, Test of behavioral microsaccadic suppression. Change in the accuracy of target report when a microsaccade occurred 0–100 ms before the target compared with when no microsaccade occurred in that interval. b, Effect of last pre-T1 and first post-T1 MS latency on behavior. Latencies are binned into 200 ms intervals (separated by gray vertical lines). Markers show change in accuracy when a MS occurred in a bin compared with mean accuracy across all trials for a given target. Mean and SEM are shown for each target (colored markers and lines). Dashed vertical lines show trial events. Experiments 1–3, n = 30; Experiment 2, n = 9. **p < 0.01. c, Same as the central portion of b, but with higher temporal resolution (100 ms latency bins, 10 ms steps) to assess MS-driven behavioral tradeoffs between T1 and T2. (T3 is not replotted at higher resolution because of lower reliability due to fewer observers.) Gray-shaded region shows significant cluster-corrected time window (bin centers ±50 ms) for the difference between T1 and T2, p < 0.05. n = 30.
