**Participants.** 15 participants will be recruited through word of mouth and campus mailing lists. Each participant will complete two sessions of approximately one hour each. Participants will receive 7 Euros per session and 4 Euros as a bonus for the completion of both sessions. We will obtain written informed consent from all subjects prior to inclusion in the study. **Apparatus.** Stimuli will be projected onto a standard 16:9 (200 x 113 cm) video-projection screen (Celexon HomeCinema, Tharston, Norwich, UK), mounted on a wall, 270 cm in front of the participant. The projector is a ProPixx (Vpixx Technologies, Saint-Bruno, QC, Canada) running at 1440 Hz vertical refresh rate and a resolution of 960 x 540 pixels. The experimental code is implemented in MATLAB (Mathworks, Natick, MA, USA), using the Psychophysics and Eyelink toolboxes (Kleiner et al., 2007; Cornelissen et al., 2002) and runs on a Dell Precision T7810 Workstation with a Debian 8 operating system. Eye movements of both eyes are recorded via an EyeLink 2 head-mounted system (SR Research, Osgoode, ON, Canada) at a sampling rate of 500 Hz, while participants rest their head on a chin rest. Responses are collected with a standard keyboard. **Procedure.** The experiment comprises two conditions, i.e. the saccade/active condition and the replay/passive condition, each run in a separate session. The trial structure is described in *Figure 1*. The following factors will be varied throughout the two sessions (and interleaved across trials): 1. Stimulus movement duration. Based on the number of frames displayed to present a stimulus movement of 4 dva. Durations: 4.2, 5.6, 7.0, 8.4, 9.8, 11.2, 12.6, 14.0, 15.4, 16.8 milliseconds (6, 8, 10, 12, 14, 16, 18, 20, 22, or 24 frames with a duration of 0.7 ms each). 2. Spatial frequency range of the noise patches. Random noise patches can either involve low spatial frequencies of 0.25 - 1 cycles/dva or high spatial frequencies of 1 - 4 cycles/dva. 3. Saccade direction. Saccades of a 16 dva amplitude will either be made from a fixation dot 8 dva right of the screen center to a target stimulus 8 dva left of the screen center, or vice versa. 4. Direction of stimulus movement. Noise patches moved either vertically upwards or downwards relative to the initial target location. In the saccade condition, participants will complete 10 trials in each experimental cell, thus resulting in 800 trials in total in the first session. In case of temporal contraints or delays resulting in test session significantly longer than one hour, the experiment can be terminated earlier. In the replay condition, all valid trials (i.e., fixation control passed, saccade made and detected in time, valid response key pressed when in target area) collected during the saccade session will be used for replay. Each session will be subdivided in ten blocks. In breaks between blocks participants will have the opportunity to rest. ![Figure 1. Trial Procedure.][1] ***Figure 1.*** Trial Procedures in Saccade/Active (left column) and in Replay/Passive (right column) Condition. ***A.*** In the saccade condition, participants fixate a dot in the left or right half of the screen at 8 dva horizontal eccentricity from the screen center. In the replay condition, the fixation dot is at screen center. The extinction of the fixation dot after successful fixation is the cue to make a saccade to the target. Fixation control is passed after 500 milliseconds of fixation within a 1.5 dva radius around the fixation dot (dashed line). After 5 seconds without fixation or 25 re-fixations, the trial will be aborted and a new calibration requested. ***B.*** In the saccade condition, participants make saccades to the target location, which is also located at a 8 dva horizontal eccentricity from the screen center. Therefore, they are required to make saccades of approximately 16 dva to reach the location of the target. When the saccade is detected, the target moves vertically upwards or downwards, traveling a total distance of 4 dva with varying speeds from 240 to 950 dva/s. The speed of the movement depends on the number of frames displayed between start and end location of the target. In the replay condition, participants are required to remain in the inital fixation area, while the target stimulus moves from the periphery of the screen towards the central fixation point. To successfully imitate the movement of the stimulus across the visual field during a saccade, eye movement data recorded in each trial of the saccade condition will be saved, smoothed, and resampled to 1440 Hz to be finally replayed at screen refresh rate during the replay condition. Thus, the saccadic velocity profile, the stimulus characteristics, and the stimulus movement in a specific trial remain the same between saccade and replay condition. If possible, to reduce the temporal uncertainty in the replay condition, the saccade onset will be set to a fixed duration after cue onset which is determined by the median saccade latency minus 100 ms, as recorded in the saccade condition. ***C.*** When the target stimulus reaches its final location, the screen will remain blank for 250 ms. ***D.*** After the blanking period (and after the eye movement has ended), a distractor stimulus of the same properties as the target stimulus will be displayed on the opposite side of the initial target area (dotted line). Both identical stimuli are now located at a 4 dva vertical eccentricity from the initial target location, which is defined by a circular boundary of a 2 dva radius around the initial target area (dotted line). Trials in which stimulus movement is still ongoing although the eyes have already crossed this boundary will be excluded. By pressing either Arrow-up or Arrow-down keys on a standard keyboard, participants indicate whether the target stimulus has moved upwards or downwards. **Saccade Detection.** Saccades will be detected online using a custom-made velocity-based detection algorithm, inspired by Engbert & Mergenthaler (2006). Eye position is sampled online in all trials. With the onset of the cue, which is presented after a 500 ms period of successful fixation within the target area; all valid samples collected since the beginning of that fixation period serve as input for the algorithm. From cue onset until the successful detection of the saccade, the detection algorithm is executed after every retrieval of a sample. As a first step, eye position data is transformed into a 2D velocity space for x and y coordinates separately. In order to compute a velocity detection threshold, the median velocity and median-based standard deviation are computed for each dimension. The threshold will be set at median velocity plus the standard deviation multiplied by a factor of ten. To detect a saccade, at least two most recent samples above threshold have to be registered. As an additional criterion, the direction of the samples above threshold is computed. A horizontal rightward saccade will only be detected when the direction of the samples is in the range of 360+/-25 deg. The direction of a leftward saccade has to be in the range of 180+/-25 deg. In pre-tests, we have found this method to detect saccades earlier and more reliably than boundary-based techniques. **Stimuli.** As shown in *Figure 2*, target stimuli will be random noise patches (pixel noise) of low or high spatial frequencies, i.e., bandpass-filtered either from 0.25 to 1 cycles/dva spatial frequency or from 1 to 4 cycles/dva spatial frequency. All noise patches are scaled to an amplitude of 0.5, thus reducing their contrast to 50%. Finally, bandpass-filtered noise patches, initially of 3 dva diameter, where enveloped in a Gaussian window with a standard deviation of 0.5 dva. Each noise patch generated in the saccade condition will be saved and reused in the replay condition. The fixation dot used in both sessions is a white circle of 0.3 dva radius. When fixated, the area within the circle will be filled by another white circle of 0.1 dva radius. ![Figure 2. Two examples of noise patches used in the experiment.][2] ***Figure 2.*** Two examples of noise patches used in the experiment. *Left:* High spatial frequency (1-4 cycles/dva) noise patch. *Right:* Low spatial frequency (0.25-1 cycles/dva) noise patch. **Data Analysis.** Task performance will be analyzed offline. Trials will be included for analysis, if: - fixation control is passed. - the saccade is successfully detected and the eye crossed the boundary around the initial target location (i.e., target area). - the saccade is not falsely detected. - there are no blinks during saccade flight. - a response was given after saccade landing in the target area and by pressing the valid response keys. - the movement of the stimulus (as indicated by trial timestamps) is concluded before the eyes reach the target area. - in the replay condition the eyes do not leave the area of central fixation (radius of 2 dva). In addition, participants may be excluded from the analysis in case their understanding of the task was wrong or if too few trials can be collected, e.g., due to insufficient calibration. At least half of the planned trials will have to be completed and valid to justify the inclusion of the participant's data. For the main data analysis, results for each combination of task type, stimulus movement duration, and spatial frequency range, task performance will be assessed by computing each participant's ability to correctly identify the movement of the target stimulus (upwards or downwards). Both the percentage of correct reports and a measure of visual sensitivity (d') will be computed as criterion variables, while the factors saccade direction and direction of stimulus movement will be pooled. Reaction times (time of response - time of saccade offset) will be evaluated to rule out potential speed-accuracy tradeoffs. In addition to the analysis of spatial frequency effects, it is envisaged to extract specific stimulus features (such as components of spatial frequency and orientation) from the noise patch matrices to perform reverse-correlation analyses (e.g., Li et al., 2016). It is planned to use repeated-measures ANOVA and general linear mixed-effects models (Bates et al., 2014) to statistically evaluate the effects of fixed factors like task type, stimulus movement duration, and spatial frequency on percentage of correct reports and visual sensitivity. Low-level stimulus features may serve as additional random factors. **References** Bates, D., Maechler, M., Bolker, B., & Walker, S. (2014). lme4: Linear mixed-effects models using Eigen and S4. R package version, 1(7). Engbert, R., & Mergenthaler, K. (2006). Microsaccades are triggered by low retinal image slip. Proceedings of the National Academy of Sciences, 103(18), 7192-7197. Li, H. H., Barbot, A., & Carrasco, M. (2016). Saccade Preparation Reshapes Sensory Tuning. Current Biology. [1]: https://mfr.osf.io/export?url=https://osf.io/nrxpb/?action=download%26direct%26mode=render&initialWidth=848&childId=mfrIframe&format=1200x1200.jpeg [2]: https://mfr.osf.io/export?url=https://osf.io/9mmr4/?action=download&direct&mode=render&initialWidth=848&childId=mfrIframe&format=1200x1200.jpeg