Article | Open | Published:

# Face perception influences the programming of eye movements

## Introduction

Faces are very salient visual stimuli for humans and our visual system has developed efficient mechanisms to preferentially detect and process them. A particular status of face stimuli is supported by a large number of studies, which consistently showed that faces elicit fast and characteristic brain and behavioral responses, as compared to other categories of visual stimuli1,2,3,4,5,6. Eye-tracking studies for example showed that during free exploration of complex scenes containing faces, observers immediately direct their gaze toward them and spend a lot of the exploration time looking at them7,8,9,10,11.

Given previous reports suggesting that face stimuli elicit fast and involuntary saccades toward them, we expected that the presence of a face distractor would also impact amplitude of saccades directed toward another target stimulus. Furthermore, based on previous findings (e.g., McPeek et al.30, we expected error saccades to be shorter than correct saccades, suggesting their modulation by a concurrently programmed corrective saccade. Finally, we examined the possibility that concurrent programming of saccades was also influenced by the content of their targets.

## Experiment 1

### Materials and Methods

#### Participants

Twenty-four participants (three females; mean age ± SD = 25.5 ± 6.26 years) with normal or corrected-to-normal vision, recruited from University Grenoble Alpes, took part in the experiment. They all came twice to complete two experimental sessions, one with faces as target stimuli and one with vehicles as target stimuli. All participants gave their informed written consent before participating in the study, which was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and was approved by the ethic committee of University Grenoble Alpes.

#### Stimuli

Stimuli were 240 colored images taken among the stimuli used by Crouzet et al.3 and chosen from Corel Stock Photo Libraries (Corel Photo stock library 1996. Ottawa, Ontario, Canada) widely used in literature for visual recognition. There were 120 images of human faces and 120 images of vehicles of various size and taken from different viewpoints. Out of the 120 images used for each category, there were 42 images of faces and 24 images of vehicles for which the main object was symmetrical around the vertical midline. However, this difference did not impact our results (see Supplementary Analysis S1 and Table S1). Stimuli were carefully chosen so that faces and vehicles had on average the same spatial position and size. For each image, the main object, i.e. face or vehicle, was manually delineated using a rectangle box (size 8.35 × 7.96°). We then computed the center of the box and ensured that no significant difference was observed for the mean position or size of objects between faces and vehicles images (mean center Xface = 5.86°; mean center Xvehicle = 5.89°; t238 = −0.25; p = 0.8; mean center Yface = 5.86°; mean center Yvehicle = 5.83°; t238 = 0.21; p = 0.84; zero being the top left corner of the image). Stimuli (sized 300 × 300 pixels) subtended 11 × 11° of visual angle at a viewing distance of 60 cm. For each category, 10 images were used for training and the remaining 110 images were used in the main experiment. It should be noted that some of these images were previously used in Guyader et al.14.

#### Procedure

Stimuli were displayed using the Softeye software40 against a gray background (luminance of 128 on a 256 gray-level scale) on a 21-inch CRT monitor with a spatial resolution of 1024 × 768 pixels, a refresh rate of 85 Hz and a mean gray luminance of 68 cd/m2. Participants were seated 60 cm away from the display. Participants’ head was stabilized by a chin- and a forehead rest. Eye movements were recorded using an Eyelink 1000 eye-tracker (SR Research) with a sampling rate of 1000 Hz and a nominal spatial resolution of 0.01 degrees of visual angle. Only one eye was recorded using the “pupil-corneal reflexion” mode (right eye in 16 out of 24 participants). The Eyelink software automatically detected saccades with the following thresholds: speed >30 degrees/s, acceleration >8000 degrees/s2, and saccadic displacement >0.15 degrees. Fixations were detected when the pupil was visible, and no saccade was in progress. Blinks were detected during partial or total occlusion of the pupil. Each session was preceded by a calibration procedure during which participants had to orient their gaze toward nine separate targets appearing sequentially in a 3 × 3 grid that occupied the entire display. A drift correction was carried out every ten trials and a new calibration was done in the middle of the experiment and when the drift error was above 0.5°.

All participants underwent two experimental sessions on two different days (separated by less than two weeks), one session for which the targets were human faces (and the distractors were vehicle images) and the other one for which the targets were images containing vehicles (the distractors were human face images). The order of sessions was counterbalanced between participants. The procedure and task were exactly the same as in Crouzet et al.3 and Guyader et al.14. For each session, a trial started with a white fixation cross subtending 0.73° of visual angle, displayed centrally for 800 to 1600 ms (duration sampled from a uniform distribution) and followed by a gap (mean gray-level screen) of 200 ms. Following the gap, two images (a target and a distractor) were simultaneously displayed on the left and the right of the central fixation cross for 400 ms. The center of each image was lateralized at 7.6° from the center of the screen. The inter-trial interval was fixed at 1000 ms (Fig. 1). Participants had to make a saccade as fast as possible toward the target image. There were 240 trials in each session, each image being seen twice, on the left and the right side, randomly. Participants completed a training session comprising 10 trials prior to the experiment in order to get familiarized with the stimuli and the task. The experiment lasted approximately 15 minutes in each session.

### Results

#### Accuracy

We performed a paired t-test on mean error rates with Target Category (Face, Vehicle) as within-subject factor. Results revealed that participants made significantly more error saccades when the target stimulus was a vehicle (i.e. face distractor mER ± SD: 23.37 ± 10.62%) than when it was a face (i.e. vehicle distractor; 12.46 ± 6.40%, t23 = 5.17, p < 0.001).

#### Latency and Amplitude of the first saccade

ANOVAs with Target Stimulus (Face, Vehicle) and Saccade Accuracy (Correct, Error) as within-subject factors were performed on mean SRT of the first saccade (in ms) and saccade amplitude (in degrees).

The ANOVA performed on mean SRT (Fig. 2a) revealed a main effect of Target Category (F1,23 = 18.21, p < 0.001, ηp2 = 0.442) and a main effect of saccade Accuracy (F1,23 = 34.37, p < 0.0001, ηp2 = 0.599). Participants initiated saccades faster when the target stimulus was a face (183 ± 21 ms) than when it was a vehicle (201 ± 22 ms) and they were slower to initiate correct (200 ± 19 ms) than incorrect saccades (184 ± 21 ms). There was a marginally significant interaction between Target Category and Saccade Accuracy (F1,23 = 3.33, p < 0.08, ηp2 = 0.126), suggesting that the difference in latencies according to the target stimulus was more pronounced for correct (25 ms difference with 187 ± 17 ms for Face target and 212 ± 26 ms for Vehicle target) than error saccades (14 ms difference with 175 ± 30 ms for Face target and 189 ± 21 ms for Vehicle target).

$$PS(a)={c}_{1}(1-{e}^{-\frac{a}{{c}_{2}}})$$

With PS the peak speed (in degrees/s), a the amplitude (in degrees) and c1 and c2 the parameters of the model (in degrees/s and in degrees, respectively). For each participant, a classical least-square fitting procedure was used to estimate the two parameters (mean R² = 0.69, SD = 0.21). For each trial, the 2-parameters function was used to estimate the predicted peak speed based on the saccade amplitude observed at this trial. This predicted peak speed was then used to normalize the observed peak speed of correct saccades and of error saccades followed by a corrective saccade (see Buonocore et al.43 for a similar procedure). Normalized peak speed higher than 1 indicates higher peak speed than what is predicted by the observed amplitude.

We then performed an ANOVA on normalized peak speed with Target Category (Face, Vehicle) and Saccade Accuracy (Correct, Error) as within-subject factors (see Fig. 3d). It should be noted that two subjects were excluded from this analysis due to aberrant fit parameters leading to biased estimation of predicted peak speed. Results revealed a main effect of Saccade Accuracy (F1,21 = 6.09, p < 0.05, ηp2 = 0.225), suggesting a violation of the main sequence for error saccades: normalized peak speed of error saccades (1.08 ± 0.13) were higher than that of correct saccades (1.04 ± 0.07). There was no main effect of Target Stimulus, nor interaction with Saccade Accuracy (both Fs =<1, ηp2 < 0.050).

### Discussion

Experiment 2 was thus designed to disambiguate the extent to which saccade amplitude was affected by the content of the target and distractor stimuli, as well as the extent to which saccades following an error compensated for the amplitude of the error, by constraining the ending location of saccadic responses. Design and procedure were the same as in Experiment 1 with the difference that this time, a white cross was added in the center of each image and participants were asked to perform their saccades toward the white cross in the center of the lateral image containing the target stimulus.

## Experiment 2

### Materials and Methods

#### Participants

Fourteen participants (eight females; mean age ± SD = 24.6 ± 4.9 years) with normal or corrected-to-normal vision, recruited from University Grenoble Alpes, took part in the experiment. They all came twice to complete two experimental sessions, one with faces as target stimuli and one with vehicles as target stimuli. All participants gave their informed written consent before participating in the study, which was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans and was approved by the ethic committee of University Grenoble Alpes.

#### Stimuli and Procedure

Stimuli and procedure were exactly the same as in Experiment 1. The right eye was recorded in 11 out of the 14 participants. The only difference was that a white fixation cross subtending 0.73° of visual angle was added in the center of each lateral image (Fig. 4a), displayed at 7.6° of eccentricity from the central fixation point, and participants were instructed to perform saccades as fast as possible toward the cross in the center of the image containing the target stimulus.

#### Data analyses

We extracted saccade parameters detected by the Eyelink software and did the same analyses as in Experiment 1. Trials where SRT was inferior to 50 ms or where a blink occurred during stimulus presentation were discarded from the analysis. This resulted in removing 1.21% of the trials.

### Results

#### Accuracy

We performed a paired t-test on mean error rates with Target Stimulus (Face, Vehicle) as a within-subject factor. Results revealed that participants made significantly more errors when the target stimulus was a vehicle than when it was a face (Vehicle as target: 26.59 ± 11.45%, Face as target: 11.29 ± 5.87%, t13 = 7.29, p < 0.001).

#### Latency and Amplitude of the first saccade

ANOVAs with Target Category (Face, Vehicle) and Saccade Accuracy (Correct, Error) as within-subject factors were performed on mean SRT of the first saccade (in ms) and saccade amplitude (in degrees).

The ANOVA performed on mean SRT (Fig. 4b) revealed a main effect of Target Category (F1,23 = 17.36, p < 0.005, ηp2 = 0.572) and a main effect of saccade Accuracy (F1,23 = 7.76, p < 0.05, ηp2 = 0.374). Participants initiated saccades faster when the target stimulus was a face (200 ± 37 ms) than when it was a vehicle (220 ± 30 ms) and they were slower to initiate correct (220 ± 24 ms) than error saccades (200 ± 43 ms). There was no interaction between the two factors (F1,23 = 1.68, p = 2.22, ηp2 = 0.115).

The ANOVA performed on mean Saccade Amplitude (Fig. 4c) revealed a significant main effect of Target Category (F1,23 = 6.44, p < 0.05, ηp2 = 0.331) and a main effect of Saccade Accuracy (F1,23 = 86.13, p < 0.0001, ηp2 = 0.868). Saccades were overall larger when the target stimulus was a vehicle (5.98 ± 0.87°) than when it was a face (5.68 ± 1.42°) and error saccades (5.23 ± 0.96°) were shorter than correct saccades (6.51 ± 0.74°). Furthermore, there was a significant interaction between these two factors (F1,23 = 59.74, p < 0.0001, ηp2 = 0.822). Pairwise comparisons showed that correct saccades were larger when the target stimulus was a face than a vehicle (face: 6.75 ± 0.65°, Vehicle: 6.26 ± 0.89°, p < 0.005) but this difference was reversed for error saccades which were larger when the target stimulus was a vehicle (i.e. face distractor: 5.70 ± 0.77°) than when it was a face (i.e. vehicle distractor: 4.61 ± 1.14°, p < 0.001), suggesting overall larger saccades directed toward faces (either as target or distractor) than toward vehicles. Finally, error saccades were shorter than correct saccades for both target categories (both ps < 0.005).

It should be noted that the mean amplitude of correct saccades (6.75° for saccades toward faces, 6.26° for saccades toward vehicles) were below the amplitude of the target crosses added in the center of images (7.6°, see Fig. 4c,d). In order to test whether correct saccades significantly undershot the targets, we computed the mean gain of correct saccades (i.e. ratio of observed amplitude on the amplitude required to reach the target cross, based on the starting point of the saccade) for each participants and compared it against 1 (i.e. equivalence between the amplitude of the saccade and the eccentricity of the target) using one-sample t-tests. Results revealed that the gain of correct saccades toward faces and vehicles were significantly below 1, suggesting that they undershot the targets by 10 and 16% respectively (Face as target: mean gain ± SD: 0.90 ± 0.09, t13 = −4.13, p < 0.05; Vehicle as target: 0.84 ± 0.12, t13 = −5.24, p < 0.05, Fig. 5b,e).

As for Experiment 1, we examined whether error saccades (representing 18.2% of the trials) were followed by a corrective one. A second saccade was considered as corrective if its ending point was on the side of the display where the target was present (i.e. if they crossed the central fixation point). This represented 88.81% of the saccades following an error. Again, these corrective saccades had very short latencies with a median of 96 ms (min = 21 ms; max = 1010 ms; mean = 147 ms, SD = 140 ms), suggesting that more than half of corrective saccades had latencies below 100 ms (see Fig. 5a). In order to test whether corrective saccades differed according to their target, we performed paired t-tests comparing the proportion, amplitude and latency of corrective saccades for Face and Vehicle. However, as in Experiment 1, there was no significant difference between the proportion (t13 = −0.19, p = 0.85), latency (t13 = 1.42, p = 0.18) or amplitude (t13 = 0.89, p = 0.39) of corrective saccades according to their target stimulus.

As for Experiment 1, we tested whether hypometric error saccades also deviated from correct saccades in terms of their main sequence, by examining normalized peak speed of saccades (see Fig. 5c and Supplementary Fig. S3). Here again, results revealed a main effect of Saccade Accuracy (F1,13 = 5.07, p < 0.05, ηp2 = 0.280), suggesting a violation of the main sequence for error saccades: normalized peak speed of error saccades (1.11 ± 0.16) were higher than that of correct saccades (1.03 ± 0.04). There was no main effect of Target Category (F1,13 = 2.38, p = 0.15 ηp2 = 0.155), and no interaction between this factor and Saccade Accuracy (F1,13 = 3.63, p = 0.08, ηp2 < 0.218).

## General Discussion

### Saccade amplitude is modulated by the content of the stimuli

Throughout two experiments, we consistently observed that saccades toward faces were larger than saccades toward vehicles. This effect persisted even if participants were explicitly required to perform their saccades toward a specific location within the images (Experiment 2) indicating it could not be voluntarily controlled. Results of Experiment 2 revealed that all saccades actually tended to undershoot the target point, but to a greater extent when they were directed toward vehicle than face stimuli, suggesting that saccades toward vehicles were rather hypometric relative to saccades toward faces. These results therefore indicate that the content of visual stimuli influences the programming of saccade amplitude in a saccadic choice task.

Recent studies suggested that this bias for face stimuli could be mediated by rapid processing of visual information via a subcortical retino-tectal pathway, through which part of visual information exiting ganglion retinal cells directly projects to the superficial layers of the superior colliculus (sSC) before reaching the pulvinar and the amygdala53,54,55,56. Nakano et al.53 for example argued that the speed of processing along in the main retino-geniculo-cortical pathway (the earliest visual response latencies observed in human visual cortical areas are of 56 ms in V1, and 70–80 ms in V3 and V457 cited in53, together with the time required to generate a saccade (20–30 ms) is too slow to account for the fast saccadic responses toward faces in just 100–110 ms, as previously observed by Crouzet et al.3 However, electrophysiological recordings in monkeys54,56 showed that neurons in the sSC and in the pulvinar exhibit distinct response to face-like stimuli within the first 25 ms and 50 ms following stimulus onset, respectively. As the pulvinar directly projects to the lateral intraparietal cortex, involved in saccade generation with latencies of 30 ms, such timings would be more compatible with the speed of face processing previously observed. Interestingly, the retino-tectal pathway is thought to be particularly involved in the detection of behaviorally-relevant stimuli such as faces See for a review55. However, exactly how fast detection of faces via the sSC in turn modulates the weight of a saccade program toward this stimulus in the iSC remains to be addressed.

### Error saccades are corrected on-line

Overall, the data obtained in the present study point to the ability for the oculomotor system to efficiently perform an online correction of saccadic response, through the development of a saccade program while a first erroneous saccade is being executed, resulting in its interruption, and then the execution of an accurate saccade toward the target. It should be noted that the notion of ‘correction’ does not necessarily imply that participants were aware of an error and voluntarily corrected for it. On the contrary, a couple of studies indicated that hypometric error saccades followed by short-latency corrective saccades were more likely to be observed when participants were unaware of the initial error47,68.

## Conclusion

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## References

1. 1.

Liu, J., Harris, A. & Kanwisher, N. Stages of processing in face perception: an MEG study. Nat. Neurosci. 5, 910–916 (2002).

2. 2.

Liu, J., Higuchi, M., Marantz, A. & Kanwisher, N. The selectivity of the occipitotemporal M170 for faces. Neuroreport 11, 337–341 (2000).

3. 3.

Crouzet, S. M., Kirchner, H. & Thorpe, S. J. Fast saccades toward faces: face detection in just 100 ms. J. Vis. 10(16), 1–17 (2010).

4. 4.

Morand, S. M., Grosbras, M. H., Caldara, R. & Harvey, M. Looking away from faces: influence of high-level visual processes on saccade programming. J Vis 10, 1–10 (2010).

5. 5.

Haxby, J. V., Hoffman, E. A. & Gobbini, M. I. The distributed human neural system for face perception (Record Supplied By Publisher). Trends Cogn Sci 4, 223–233 (2000).

6. 6.

Farah, M. J., Wilson, K. D., Drain, M. & Tanaka, J. N. What is ‘Special’ about Face Perception? Psychol. Rev. 105, 482–498 (1998).

7. 7.

Foulsham, T., Cheng, J. T., Tracy, J. L., Henrich, J. & Kingstone, A. Gaze allocation in a dynamic situation: Effects of social status and speaking. Cognition 117, 319–331 (2010).

8. 8.

Hirvenkari, L. et al. Influence of Turn-Taking in a Two-Person Conversation on the Gaze of a Viewer. PLoS One 8, (2013).

9. 9.

Coutrot, A. & Guyader, N. How saliency, faces, and sound influence gaze in dynamic social scenes. J. Vis. 14, 5–5 (2014).

10. 10.

Tilke, J., Ehinger, K., Durand, F. & Torralba, A. Learning to predict where humans look. Proc. IEEE Int. Conf. Comput. Vis. 2106–2113, https://doi.org/10.1109/ICCV.2009.5459462 (2009).

11. 11.

Marat, S., Rahman, A., Pellerin, D., Guyader, N. & Houzet, D. Improving Visual Saliency by Adding ‘Face Feature Map’ and ‘Center Bias’. Cognit. Comput. 5, 63–75 (2013).

12. 12.

Crouzet, S. M. & Thorpe, S. J. Low-level cues and ultra-fast face detection. Front. Psychol. 2 (2011).

13. 13.

Boucart, M. et al. Finding faces, animals, and vehicles in far peripheral vision. J. Vis. 16, 10 (2016).

14. 14.

Guyader, N., Chauvin, A., Boucart, M. & Peyrin, C. Do low spatial frequencies explain the extremely fast saccades towards human faces? Vision Res., https://doi.org/10.1016/j.visres.2016.12.019 (2017).

15. 15.

Kirchner, H. & Thorpe, S. J. Ultra-rapid object detection with saccadic eye movements: Visual processing speed revisited. Vision Res. 46, 1762–1776 (2006).

16. 16.

Fischer, B. & Weber, H. Express saccades and visual attention. Behav. Brain Sci. 16, 553 (1993).

17. 17.

Kalesnykas, R. P. & Hallett, P. E. The differentiation of visually guided and anticipatory saccades in gap and overlap paradigms. Exp. Brain Res. 68, 115–121 (1987).

18. 18.

Gilchrist, I. D. & Proske, H. Anti-saccades away from faces: Evidence for an influence of high-level visual processes on saccade programming. Exp. Brain Res. 173, 708–712 (2006).

19. 19.

Bindemann, M., Burton, A. M., Langton, S. R. H., Schweinberger, S. R. & Doherty, M. J. The control of attention to faces. J. Vis. 7, 15 (2007).

20. 20.

Quaia, C., Lefèvre, P. & Optican, L. M. Model of the control of saccades by superior colliculus and cerebellum. J. Neurophysiol. 82, 999–1018 (1999).

21. 21.

Kapoula, Z. & Robinson, D. A. Saccadic undershoot is not inevitable: Saccades can be accurate. Vision Res. 26, 735–743 (1986).

22. 22.

Kowler, E., Anderson, E., Dosher, B. & Blaser, E. The role of attention in the programming of saccades. Vision Res. 35, 1897–1916 (1995).

23. 23.

Collewijn, H., Erkelens, C. & Steinman, R. Binocular co-ordination fo human horisontal saccadic eye movements. J. Physiol. 404, 157–182 (1988).

24. 24.

Coren, S. & Hoenig, P. Effect of non-target stimuli upon length of voluntary saccades. Percept. Mot. Ski. 34, 499–508 (1972).

25. 25.

Van der Stigchel, S. & Nijboer, T. C. The global effect: what determines where the eyes land? J. Eye Mov. Res. 4, 1–13 (2011).

26. 26.

Vitu, F. About the global effect and the critical role of retinal eccentricity: Implications for eye movements in reading. J. Eye Mov. Res. 2, 1–18 (2008).

27. 27.

Findlay, J. M. Global visual processing for saccadic eye movements. Vision Res. 22, 1033–1045 (1982).

28. 28.

Chou, I. han, Sommer, M. A. & Schiller, P. H. Express averaging saccades in monkeys. Vision Res. 39, 4200–4216 (1999).

29. 29.

Findlay, J. M. & Blythe, H. I. Saccade target selection: Do distractors affect saccade accuracy? Vision Res. 49, 1267–1274 (2009).

30. 30.

McPeek, R. M., Skavenski, A. A. & Nakayama, K. Concurrent processing of saccades in visual search. Vision Res. 40, 2499–2516 (2000).

31. 31.

Weber, H., Dürr, N. & Fisher, B. Effects of pre-cues on voluntary and reflexive saccade generation. I. Anti-cues for pro-saccades. Exp. Brain Res. 120, 403–416 (1998).

32. 32.

Godijn, R. & Theeuwes, J. Oculomotor capture and Inhibition of Return: Evidence for an oculomotor suppression account of IOR. Psychol. Res. 66, 234–246 (2002).

33. 33.

Findlay, J. M., Brown, V. & Gilchrist, I. D. Saccade target selection in visual search: The effect of information from the previous fixation. Vision Res. 41, 87–95 (2001).

34. 34.

Godijn, R. & Theeuwes, J. Programming of Endogenous and Exogenous Saccades: Evidence for a Competitive Integration Model. J. Exp. Psychol. Hum. Percept. Perform. 28, 1039–1054 (2002).

35. 35.

Becker, W. & Jürgens, R. An analysis of the saccadic system by means of double step stimuli. Vision Res. 19, 967–983 (1979).

36. 36.

McPeek, R. M. & Keller, E. L. Superior Colliculus Activity Related to Concurrent Processing of Saccade Goals in a Visual Search Task. J. Neurophysiol. 87, 1805–1815 (2002).

37. 37.

Findlay, J. M. & Walker, R. A model of saccade generation based on parallel processing and competitive inhibition. Behav. Brain Sci. 22, 661–74; discussion 674–721 (1999).

38. 38.

Walker, R. & McSorley, E. The parallel programming of voluntary and reflexive saccades. Vision Res. 46, 2082–2093 (2006).

39. 39.

Ramakrishnan, A., Chokhandre, S. & Murthy, A. Voluntary Control of Multisaccade Gaze Shifts During Movement Preparation and Execution. J. Neurophysiol. 103, 2400–2416 (2010).

40. 40.

Ionescu, G., Guyader, N. & Guérin-dugué, A. SoftEye software. IDDN. FR 1 (2009).

41. 41.

Viviani, P. & Swensson, R. G. Saccadic eye movements to peripherally discriminated visual targets. J. Exp. Psychol. Hum. Percept. Perform. 8, 113–126 (1982).

42. 42.

Bahill, A. T., Clark, M. R. & Stark, L. The Main Sequence, A Tool for Studying Human Eye Movements. Math. Biosci. 24, 191–204 (1975).

43. 43.

Buonocore, A., McIntosh, R. D. & Melcher, D. Beyond the point of no return: effects of visual distractors on saccade amplitude and velocity. J. Neurophysiol. 115, 752–762 (2016).

44. 44.

Evdokimidis, I., Tsekou, H. & Smyrnis, N. The mirror antisaccade task: direction-amplitude interaction and spatial accuracy characteristics. Exp. Brain Res. 174, 304–311 (2006).

45. 45.

Allik, J., Toom, M. & Luuk, A. Planning of saccadic eye movements. Psychol Res 67, 10–21 (2003).

46. 46.

Massen, C. Parallel programming of exogenous and endogenous components in the antisaccade task. Q. J. Exp. Psychol. Sect. A Hum. Exp. Psychol. 57, 475–498 (2004).

47. 47.

Mokler, A. & Fischer, B. The recognition of errors and corrections in an antisaccade task. Exp. Brain Res. 125, 511–516 (1999).

48. 48.

Trappenberg, T. P., Dorris, M. C., Munoz, D. P. & Klein, R. M. A Model of Saccade Initiation Based on the Competitive Integration of Exogenous and Endogenous Signals in the Superior Colliculus. J. Cogn. Neurosci. 13, 256–271 (2001).

49. 49.

Dorris, M. C., Olivier, E. & Munoz, D. P. Competitive Integration of Visual and Preparatory Signals in the Superior Colliculus during Saccadic Programming. J. Neurosci. 27, 5053–5062 (2007).

50. 50.

Meeter, M., Van Der Stigchel, S. & Theeuwes, J. A competitive integration model of exogenous and endogenous eye movements. Biol. Cybern. 102, 271–291 (2010).

51. 51.

Lee, C., Rohrer, W. H. & Sparks, D. L. Population coding of saccadic eye movements by neurons in the superior colliculus. Nature 332, 357–360 (1988).

52. 52.

Munoz, D. P. & Schall, J. D. Concurrent, Distributed Control of saccade initiation in the frontal eye field and superior colliculus. Super. Colliculus New approaches stydying sensorimotor Integr. 55–82, https://doi.org/10.1201/9780203501504 (2004).

53. 53.

Nakano, T., Higashida, N. & Kitazawa, S. Facilitation of face recognition through the retino-tectal pathway. Neuropsychologia 51, 2043–2049 (2013).

54. 54.

Nguyen, M. N. et al. Neuronal responses to face-like and facial stimuli in the monkey superior colliculus. Front. Behav. Neurosci. 8, 1–18 (2014).

55. 55.

Soares, S. C., Maior, R. S., Isbell, L. A., Tomaz, C. & Nishijo, H. Fast detector/first responder: Interactions between the superior colliculus-pulvinar pathway and stimuli relevant to primates. Front. Neurosci. 11, 1–19 (2017).

56. 56.

Nguyen, M. N. et al. Neuronal responses to face-like stimuli in the monkey pulvinar. Eur. J. Neurosci. 37, 35–51 (2013).

57. 57.

Yoshor, D., Bosking, W. H., Ghose, G. M. & Maunsell, J. H. R. Receptive fields in human visual cortex mapped with surface electrodes. Cereb. Cortex 17, 2293–302 (2007).

58. 58.

McSorley, E., McCloy, R. & Williams, L. The concurrent programming of saccades. PLoS One 11, 1–17 (2016).

59. 59.

Sharika, K. M., Ramakrishnan, A. & Murthy, A. Control of Predictive Error Correction During a Saccadic Double-Step Task. J. Neurophysiol. 100, 2757–2770 (2008).

60. 60.

Murthy, A. et al. Frontal Eye Field Contributions to Rapid Corrective Saccades. J. Neurophysiol. 97, 1457–1469 (2006).

61. 61.

Edelman, J. A. & Xu, K. Z. Inhibition of Voluntary Saccadic Eye Movement Commands by Abrupt Visual Onsets. J. Neurophysiol. 101, 1222–1234 (2009).

62. 62.

Castet, E. & Masson, G. S. Motion perception during saccadic eye movements. Nat. Neurosci. https://doi.org/10.1038/72124(2000).

63. 63.

Krekelberg, B. Quick guide Saccadic suppression. Curr. Biol. 20, 228–229 (2010).

64. 64.

Castet, E., Jeanjean, S. & Masson, G. S. Motion perception of saccade-induced retinal translation. Proc. Natl. Acad. Sci. 99, 15159–15163 (2002).

65. 65.

Fabius, J. H., Fracasso, A. & Van Der Stigchel, S. Spatiotopic updating facilitates perception immediately after saccades. Sci. Rep. 6, 1–11 (2016).

66. 66.

Van der Stigchel, S. & Hollingworth, A. Visuospatial Working Memory as a Fundamental Component of the Eye Movement System. Curr. Dir. Psychol. Sci. 27, 136–143 (2018).

67. 67.

Duhamel, J., Colby, C. L. & Goldberg, M. E. The Updating of the Representation of Visual representation. Science (80-.). 255, 90–92 (1992).

68. 68.

Nieuwenhuis, S., Richard Ridderinkhof, K., Blom, J., Band, G. P. H. & Kok, A. Error-related brain potentials are differentially related to awareness of response errors: Evidence from an antisaccade task. Psychophysiology 38, 752–760 (2001).

## Acknowledgements

This work was supported by NeuroCoG IDEX UGA in the framework of the “Investissements d’avenir” program (ANR-15-IDEX-02) and by a grant from the LabEx PERSYVAL-Lab (ANR-11-LABX-0025-01).

## Author information

### Affiliations

1. #### Univ. Grenoble Alpes, CNRS, Grenoble INP, GIPSA-lab, 38000, Grenoble, France

• Louise Kauffmann
• , Léa Entzmann
• , Camille Breuil
2. #### Univ. Grenoble Alpes, Univ. Savoie Mont Blanc, CNRS, LPNC, 38000, Grenoble, France

• Louise Kauffmann
• , Carole Peyrin
•  & Alan Chauvin

### Contributions

N.G. and C.P. and A.C. designed research and received funding; N.G., L.E., C.B. performed research; L.K., N.G. and A.C. analyzed data; L.K. wrote the paper. All authors reviewed the manuscript.

### Competing Interests

The authors declare no competing interests.

### Corresponding author

Correspondence to Louise Kauffmann.