Eye gaze is a strong cue constantly used in social interactions to point to objects or share our focus of attention. Although eye gaze and arrows similarly act as directional stimuli, recent research has found that eye gaze and arrows yield opposite behavioral congruency effects in a spatial interference paradigm (Cañadas & Lupiáñez, 2012; Marotta et al., 2018): arrows eliciting faster responses when their direction is congruent with their position (standard congruency effect), and gaze producing faster reaction times for incongruent conditions (reversed congruency effect). Although the nature of the reversion with eye-gaze targets is still unknown, the accumulated evidence (Marotra et al., 2019; Román-Caballero et al., 2021a,b; Aranda-Martín et al., 2022; Hemmerich et al., 2021) suggests that both social and non-social directional stimuli share a spatial interference component, whereas eye gaze triggers unique attentional effects of opposite direction (i.e., benefiting the response for incongruent trials). Therefore, any condition that reduces the spatial conflict with arrows will increase the magnitude of the reversion with eye gaze. In two recent studies (Román-Caballero et al., 2021a,b), we showed that both congruency effects can be affected in a similar way when the target is presented in a perceptually complex scenario and observers need to segregate the directional stimulus (i.e., eye gaze or arrow) from the background. Increasing the segregation demands, the irrelevant code of the target spatial position is weakened by the moment in which the response to the relevant dimension is available (i.e, target direction), thus reducing the interference. Another manipulation that has been proved to be effective in reducing spatial interference effect (spatial Stroop, specifically) is cueing the position of appearance of the target just before its display (with a peripheral onset; Funes et al., 2007). This reduction in the spatial interference is presumably the result of a similar mechanism, thus valid cued positions activate in advance the irrelevant location code and it will be weakened by the moment the target appears. The aim of the present study is to investigate whether a pre-cueing manipulation is also effective to reduce the standard cognruency effect wit arrows (in valid trials) and at the same time make more negative the reversed congruency effect with eye gaze. We therefore expect similar results to the ones observed in the two experiments in Román-Caballero et al. (2021b) with the manipulation of the target–background synchrony. <br> <br> <b>Hypotheses</b> <br> The use of a non-predictive cue (i.e., 50% of the times the cue and the target share the same position) immediately before the target onset will produce two different types of trials: valid and invalid location trials. We expect faster responses for valid than invalid location trials (main effect of validity), and the previously observed interaction between congruency and target type, with an standard congruency effect for arrows and a reversed congruency effect for gaze. More importantly, we expect peripheral cueing to reduce the congruency effect (validity x congruency interaction), but similarly for arrows and gaze (i.e., no three way interaction). More specifically, for invalid location trials, we expect the emergence of a standard congruency effect with arrow targets and a reversed congruency effect with eye gaze. However, we predict a reduction in the magnitude of the congruency effect with arrows in valid trials and a subsequent more negative reversion with eye-gaze stimuli. <br> <br> <h4><b>Method</b></h4> <b>Participants</b> <br> A sample of 45 adults will participate in the study. All of them will be students from the University of Granada who will sign an informed consent and receive partial course credit for participating. All will have to have a self-reported normal or corrected-to-normal vision and be naïve to the purpose of the experiment. <br> <br> We conducted several power analyses with different analytic methods to ensure enough statistical power regardless of the approach (see data base and R script in https://osf.io/a4nqh/). We chose the reversed congruency effect (d_z = 0.4) and the two-way interaction (partial eta squared = .16 / Cohen's f = 0.44) in Experiment 2 of Román-Caballero et al. (2021b) as references to estimate the required sample size. Thus, using the pwr (Champely et al., 2020) R package, we have found that at least 40 participants were necessary for an effect size of d_z = 0.4 in a one-sided paired t test (i.e., for the congruency effect for each stimulus type) with an alpha of .05 and a power of .80. Furthermore, using the Superpower R package (Caldwell et al., 2020), we found power of .76 with a sample size of 40 participants for the congruency-by-synchrony interaction (pre-cueing will produce an equivalent effect to the manipulation of synchrony in Román-Caballero et al., 2021b) using a 2 x 2 x 2 repeated measures ANOVA (same alpha). The power increased up to .85 with 45 participants in the directional t test analysis and .81 for the ANOVA interaction. In parallel, we assessed the power of a sample of 40 participants to detect the congruency-by-synchrony interaction in a single-trial data linear mixed-effects model with simr R package (Green et al., 2022). Using the trial data of Experiment 2 of Román-Caballero et al. (2021b), the achieved power for the congruency-by-synchrony interaction was .73. Increasing the sample to 45 participants the reached power for a similar design would be .84. <br> <br> The study will be conducted following the ethical guidelines laid down by the University of Granada (2442/CEIH/2021), in accordance with the ethical standards of the 1964 Declaration of Helsinki. Outlier detection will be based on performance (i.e., mean RTs, and accuracy) identified as poor in terms of meeting all the following indices: standard deviation from the mean (> 2.5), studentized deleted residuals (> 3), and Cook’s Di (> 4/n). <br> <br> <b>Stimuli and procedure</b> <br> The study will be carried out online during the Spring of 2022 (April–June). The task will be programmed using OpenSesame and administered online with JATOS. Social stimuli are neutral Caucasian obtained from photographing eight volunteers (four women) aged between 22 and 32 years (see Experiment 2 of Román-Caballero et al., 2021b). <br> <br> There were three pictures for each face, one with closed eyes, other looking leftward, and another rightward, in which the eyes were controlled to be at the same height. For arrow targets, we made the backgrounds by randomly selecting a 70 x 80 matrix of pixels from the pictures of faces with closed eyes. We added a gray line at the height of the arrow display to form the corresponding closed-eyes condition for arrows. For the directional conditions with nonsocial stimuli, two arrows were embedded in the pixel background at the same height as the eyes in the social counterparts. Additionally, a gray rectangle was added behind the arrows to control that all the targets had the same contrast with the background. <br> <br> The participants will be instructed to perform the task in a moment of the day between 10 a.m. to 8 p.m., in a quiet room with dim light, using glasses or contact lenses if needed, and to run the experiment with the Chrome browser. Each trial will begin with a white fixation cross presented in the center of a black screen for 1,000 ms and two closed-eye faces / pixelated backgrounds on each side of the fixation point (regarding the type of block, eye-gaze or arrow block). Participants will be instructed to fixate the cross. After a period of 900 ms since the onset of the fixation point and the two backgrounds, the edges of one of the two lateral stimuli will be highlighted in red for 50 ms, cueing one of the two sides in the form of a peripheral sudden onset. After 100 ms of the cue onset, the target (i.e., arrows or eye gaze) will be presented for 2,000 ms in one of the two positions, looking/pointing to the right or the left. Although the background/placeholder display will be bilateral, note that the target onset will be always lateralized to ensure the emergence of spatial conflict. The cue and the target will appear 50% of the times in the same position (valid trials) and 50% on opposite sides (invalid trials). Participants will be instructed to press the z key in response to targets indicating the left and the m key in response to targets indicating the right, independent of the location of the target. Importantly, this design produced congruent trials (e.g., a right-indicating target presented on the right) and incongruent (e.g., a left-indicating target presented on the right). Since the target onset, participants will be required to discriminate stimulus direction as fast as possible to avoid committing errors. Auditory feedback for incorrect key presses will be provided to participants, in the form of a 220-Hz tone presented for 1,500 ms. Participants will be instructed to use headphones during the task and to adjust the volume to a comfortable value. <br> <br> The experiment will be composed of two halves of two blocks each (one for arrow targets and the other for gaze). Thus, there will be four blocks in total, with 96 trials each, preceded by a 15-trial practice block. The order of target type will be counterbalanced across participants. Moreover, the cue and target locations, and the target direction will be randomly selected within each block. <br> <br> <b>Design and statistical analysis</b> <br> The experiment have a (2 x 2 x 2) three-factor repeated measures design: target type (gaze vs. arrows), congruency (congruent vs. incongruent trials), and validity (valid vs. invalid trials). Accordingly, a linear and a logistic mixed-effects model on single-trial data will be respectively conducted for RT and accuracy with the R packages lme4 (Bates et al., 2015) and lmerTest (Kuznetsova et al., 2017). Target type, congruency, and validity with all interactions will be included as fixed effects, and participant as random effect. In addition, paired t tests (with mean RT and accuracy) will be used for comparison between two conditions. To estimate the effect size and its precision, we will use Cohen’s dz (Mdiff/SDdiff representing the mean and the standard of the difference score) and its 95% confidence interval for paired t tests. For individual effects from the linear mixed-effects model, partial eta squared and its 90% confidence will be estimated.