Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 19.
Published in final edited form as: Curr Biol. 2020 Dec 18;31(5):1029–1036.e2. doi: 10.1016/j.cub.2020.11.032

The Motor Representation of Sensory Experience

Celine Cont 1, Eckart Zimmermann 1,2,*
PMCID: PMC7611541  EMSID: EMS126693  PMID: 33290742

Summary

How do we estimate the position of an object in the world around us? Naturally, we would direct our gaze to that object. Accordingly, neural motor coordinates entail the distance of external objects and thus might be used to derive perceptual estimates. Several general frameworks in the history of perceptual science have offered such a view.1–4 However, a mechanism showing how motor and visual processes communicate remains elusive. Here, we report that every post-saccadic error biases visual localization in a serially dependent manner. In order to simulate a realignment of visual space through motor coordinates, we induced an artificial de-alignment between visual and motor space. We found that when performing saccades under this distortion, post-saccadic error information clearly realigned visual and motor space, again in a serially dependent manner. These results demonstrate that the consequences of every saccade directly influence where we see objects in the world. On a neural basis, this requires that motor signals, which generate close to the saccade production machinery, are reported to cortical areas and arrange visual space. This view is consistent with recent electrophysiological findings of post-saccadic error processing in posterior parietal cortex.5

Results

We investigated whether saccade coordinates contribute to perceptual estimates of external spatial distances. Once a saccade is finished, the difference between the intended landing position and the actual target position, i.e., the post-saccadic error, might be used to calibrate motor and visual space. Such a recalibration might occur in a serially dependent manner: the post-saccadic error of a saccade undershooting a target would signal that the desired target is further away than estimated. Consequently, visual or saccadic localization on the following occasion would then estimate a target at the same physical position to be more peripheral. Studying sensorimotor serial dependencies in isolation requires control of the attracting influence of purely perceptual serial dependencies. The latter biases participants to perceive objects closer to recently seen objects.6 To this end, we presented targets repeatedly at the very same physical position. We alternated trials that demanded participants to perform a saccade to a target with trials in which participants localized a target visually.

We asked observers to make a rightward saccade to a target that was presented as soon as the fixation point disappeared (see Figure 1B). During execution of the saccade, the target was—unbeknownst to the participant—displaced by a small degree to the left or right. Size and direction of this displacement were drawn from a normal distribution with a mean of 0° and a variance of 2°. With an average displacement size of 20%, relative to the required saccade amplitude, participants remained insensitive to the intra-saccadic target jump.7 After a saccade trial was finished, a visual localization trial was presented in which the participant was asked to localize a briefly flashed target with a mouse pointer while keeping gaze directed to the fixation point. The localization target was presented at the same eccentricity as the initial saccade target. We prohibited the use of visual landmarks by conducting the experiment in a dark room.

Figure 1. The Motor Representation of Visual Space.

Figure 1

Open in a new tab

(A) Saccades are performed to targets (indicated by the blue disk) in the external world. The distance of the eye position to the target after the saccade has landed defines the post-saccadic error. According to this model, the post-saccadic error is transferred to neural visual and motor maps and locally recalibrates the internal representation of external space. Gray disks in the visual and at the motor map indicate internal target representations before saccade execution, and the dashed circle indicates the recalibrated target representation after saccade performance. The results of this study demonstrate that the post-saccadic error calibrates visual localization spatially selective.

(B and C) Illustration of a trial sequence in the three conditions of experiment 1. Trial n-1 and n alternated until the end of an experimental session. Condition 1 (shown in blue) required participants in trial n-1 to fixate on a fixation point. After 1,000–1,500 ms, the fixation point disappeared, and a saccade target was presented. Participants were required to perform a rightward saccade to the target. In trial n, participants fixated throughout the entire trial. After 1,000–1,500 ms, participants had to localize a briefly shown (300 ms) target with a mouse pointer while fixating their gaze on the previously presented fixation point. In condition 2 (shown in red), participants had to perform a saccade to the target in the n-1 trial and n trial. In these trials, the target was displaced during execution of the saccade. In condition 3 (shown in yellow), participants in trial n-1 observed a target displacement the same size as those in the experimental trials described before. The localization trial n was identical to the visual localization trial in conditions 1 and 2.

(D) Average visual localization error in trial n as a function of final saccade target position in trial n-1. Positive numbers on the x axis represent rightward displacements of the saccade target. Post-saccadic errors in these trials thus indicated that saccade amplitudes were too small to reach the target. Negative numbers on the y axis indicate localization error in foveal direction. The positive slope of the regression reveals that positive post-saccadic errors in trial n-1 led participants to localize the target further in the periphery in trial n. Error bars represent SEM.

(E) Saccade localization in trial n-1 as function of saccade localization in trial n. Same conventions as in (D).

(F) Control condition: visual localization as function of target displacement presented during fixation. Same conventions as in (D).

(G) Average slopes retrieved from individual linear fits for all three conditions (same color code as in D—F). Error bars represent SEM.

(H) Average intercepts retrieved from individual linear fits for all three conditions (same color code as in D—F). Error bars represent SEM.

We estimated a linear regression between the position at which participants localized a target with a mouse pointer in trial n and the size and direction of the post-saccadic error in trial n-1 for each participant (see Figure 1D). Please note that positive numbers on the x axis represent rightward displacements of the saccade target. Post-saccadic errors in these trials thus indicated that saccade amplitudes were too small to reach the target. Moreover, we used the slopes to quantify the magnitude of serial dependencies. We found a clear serial dependency of visual localization on post-saccadic error (see Figures 1D and 1G, t[11] = 2.91, p = 0.014). In other words, visual localization judgments of participants were attractively biased by the post-saccadic error, i.e., the difference between saccade landing and final saccade target position. Please note that all data in the current study show the characteristic foveal bias that is comparable in strength for saccades and visual localization.8 This similarity is consistent with the view that the post-saccadic error calibrates both the visual and the motor map.

Next, we aimed to determine whether the transfer from postsaccadic error to vision is similar in strength to serial dependencies between saccade landings. To this end, in a separate session, participants had to perform a saccade in trial n-1 and in trial n. In agreement with previous studies,9–12 we found that post-saccadic errors in trial n-1 affected the saccade landing in trial n (see Figure 1E, t[5] = −3.49, p = 0.017). The magnitude of these dependencies was virtually identical to those for visual localization (compare Figure 1G).

To exclude the target jump itself and to see if it is sufficient to generate serial dependencies, we conducted a control condition in which participants had to keep ocular fixation throughout the entire experimental session. A target was presented at the same eccentricity and with the same duration as in the visual localization and the saccade condition and then displaced with the same sizes as before. In this control, no significant serial dependency was found (see Figure 1F, t[11] = 0.427, p = 0.677), thus ruling out an influence of visible target jumps on localization. One of us has previously demonstrated that the insensitivity to target displacements during saccades is equally strong in fixation when a mask covers the motion transient of the target jump.13 We therefore checked whether the current findings of serial dependencies are saccade related by testing whether similar results would be obtained in a fixation control condition that includes a masked target jump. In this condition, we found no serial dependency between the target jump direction in trial n-1 and visual localization in trial n (t[3] = −1.64, p = 0.20).

When comparing the intercepts of the regression lines (see Figure 1H), we found a significant decrease of localization eccentricity in the control condition compared to the main experiment (see Figure 1E) in which saccades alternated with visual localization (t[11] = 2.48, p = 0.03). According to our theory, this increase in foveal bias should disappear if a saccade is performed between localization trials. The post-saccadic error would indicate the physical target location and thereby reduce mislocalization.

We asked subjects to localize a target that was presented at the same position over 400 trials (see Figure 2A). In this experiment, subjects fixated in the screen center to avoid confounding influences of eye position adaptation14 and localized a briefly shown visual target by moving a rectangle with the arrow keys to the perceived position of the target. We found that prohibiting saccades led to an increasing drift toward the fovea (see Figure 2A, yellow curve). Comparing the average of the first five and the average of the last five trials across subjects revealed a significant difference in localization (see Figure 2C, t [3] = 3.55, p = 0.03). Next, we wondered whether saccade amplitudes were unaffected by this drift. To this end, we implemented an additional task in which after each localization trial a saccade trial was presented. However, in contrast to experiment 1, the saccade target was presented for 400 ms only and was not displaced during saccade execution. We found that saccade performance was mostly unaffected by the perceptual drift (see Figures 2B and 2C, red curve, t[3] = 0.58, p = 0.59). This allowed us to ask whether the more accurate saccadic coordinates might calibrate visual localization and thus realign visual with external space. Figure 2B (blue curve) shows that this was the case. In this condition, localization was statistically indistinguishable at the beginning and end of the session (see Figure 2C, t[3] = 2.27, p = 0.11).

Figure 2. Serial Sensorimotor Dependencies Recalibrate Visual Space.

Figure 2

Open in a new tab

(A) Visual localization of the target in the “localization only” condition and trial-by-trial structure. Localization data were binned into bins of 10 trials in the graph. The thick line shows the average over all subjects, and the shaded area represents the SEM.

(B) Localization in the “n-1: saccade and n: perception” condition that was measured in a separate session. Bars above represent the trial structure for this session. The graph shows saccade landing in the n-1 trial, shown in red, and visual localization data are shown in blue. Same conventions as in (A).

(C) Average difference in localization from the “localization only” condition and from the “n-1: saccade and n: perception” condition. Same color code as in (A and B). Error bars represent SEM.

(D and E) Average saccade error in trial n as a function of visual localization error in trial n-1. Positive numbers on the x axis indicate that the saccade undershot and landed to the left of the saccade target. Error bars represent SEM. Again, data shown in blue represent the realignment condition, and data shown in red the saccade landing in the realignment condition.

We then asked how positional coordinates were transferred from motor to visual processing. We found serial dependencies from post-saccadic errors to visual localization in this condition (t[3] = 8.52, p = 0.003). By contrast, no transfer of serial dependencies from visual localization to saccade landing was observed (see Figure 2D, t[3] = 1.11, p = 0.34). This result dem-onstrates a primacy of the saccade motor error for the calibration of motor and visual space.

In perception, serial dependencies are short lived, lasting only over the last three trials.15 We analyzed the dependencies in our data for up to three trials back. Figure 3 shows similar average slopes for the saccade and the visual localization conditions of experiment 1 (see Figures 3A and 3B). At least descriptively, serial dependencies could be observed for the last three trials, but only n-1 was statistically significant. For experiment 2 (see Figure 3C), serial dependencies were significant only for n-1. This slight difference of results from the serial dependency literature might occur because in our setup, only every second trial is an n-trial, in contrast to most other serial dependency experiments in which every trial is an n-trial.

Figure 3. Serial Dependencies in Different Trial Histories.

Figure 3

Open in a new tab

(A) Analysis of serial dependencies for n-1, n-2, and n-3 corresponding to the n-back (saccade) and n (visual condition) of experiment 1. Each bar represents the average slope and was tested with a one-sample t test. Error bars represent SEM.

(B) Analysis of serial dependencies for n-1, n-2, and n-3 corresponding to the n-back (saccade) and n (saccade condition) of experiment 1. Same conventions as in (A).

(C) Analysis of serial dependencies for n-1, n-2, and n-3 corresponding to the n-back (saccade) and n (saccade condition) of experiment 2. Same conventions as in (A).

A central question in spatial realignment is the specificity of the effect: will a post-saccadic error that is connected to a certain saccade amplitude, say 10°, shift positions in the entire visual map or be selective for the performed 10° amplitude? We aimed to answer that question by asking observers to perform 10° saccade amplitudes in trial n-1 and to localize a visual stimulus in trial n that could appear in one of five positions (6°, 8°, 10°, 12°, or 14°, see Figures 4A and 4B). We found that recalibration was spatially selective (see Figure 4C). Significant serial dependencies from post-saccadic error to visual localization were only found at 10°, i.e., the position that matched the required saccade amplitude (t[5] = 3.14, p = 0.026).

Figure 4. Spatial Specificity of Serial Sensori-motor Dependencies.

Figure 4

Open in a new tab

(A) In experiment 3, participants had to perform a saccade to the saccade target in trial n-1, which appeared after 1,000–1,500 ms. During saccade execution, the saccade target was displaced.

(B) In trial n, participants fixated throughout the entire trial. After 1,000–1,500 ms, participants had to localize a briefly shown (400 ms) target with a mouse pointer while fixating their gaze on the previously presented fixation point. The localization target appeared in one of five possible locations (6°, 8°, 10°, 12°, and 14°).

(C) Average slopes from the regression between post-saccadic error in trial n-1 and localization in trial n. Error bars represent SEM.

Discussion

In this study, we report that every execution of a saccade calibrates local visual space. Post-saccadic errors biased the localization of objects in space in a serially dependent way as a saccade served as a measure to estimate the physical eccentricity of objects. From a theoretical perspective, only motor actions produce the necessary error signal to recalibrate internal spatial estimates to external space. As saccades are usually directed to a visual signal, the post-saccadic error reflects the state of the external world with respect to the visual and the motor signal driving the saccade (see Figure 1A). The phenomenon of serial dependencies has been discovered quite recently15,16 and has generated its own research field. Serial dependencies represent a weighted learning from sensory or—as in our case—motor history. It is currently a matter of debate whether serial dependencies occur on a perceptual or decision level in neural processing.17,18 Our findings demonstrate a transfer of motor to visual information and therefore suggest that serial dependencies need not be restricted to one particular level in the neural hierarchy. Our study does not allow general inferences as to why the brain favors serial dependencies. However, conversely, looking at action and perception on a trial-by-trial basis unmasked the previously overlooked information transfer between the errors of saccades and visual space. This finding clearly suggests a neural mechanism for spatial calibration of vision that has not been considered in previous accounts of active vision. Studies demonstrating visual shifts following saccade adaptation4 have suggested a shared resource guiding localization and saccades. However, the magnitude of these visual shifts was only a fraction of the motor adaptation. A direct effect of motor adaptation on vision in terms of a shared resource would predict full transfer of adaptation, which has never been reported.

By contrast, serial dependencies offer a more dynamic view of spatial localization that does not rely on reading out the state of a static motor map. If every post-saccadic error instructs a weighted reformation of local position codes in a perceptual map, the brain is in a constant state of learning. This learning might also be described as minimization of prediction errors. It has been suggested that the sensorimotor system can accurately predict saccade landing positions and that prediction errors, i.e., differences to the actual post-saccadic error, are reduced by adaptation.19–21 The dynamic view of a spatial representation, presented here, offers an action-oriented account of localization in which the goal of vision is the distal control of movements.22 In this view, a strong separation between spatial perception and action, which introspection wrongly suggests, does not exist. Our phenomenological experience suggests that the function of perception is to create a rich mental image of the outside world. In stark contrast, change detection experiments have continuously demonstrated how little information is represented in a given visual snapshot.23 Perception requires the constant performance of saccades getting updated. We found that saccade errors calibrate visual perception, while the reverse, i.e., serial dependencies from visual localization to saccade targeting, was not observed. It is true that under this view, spatial perception will inherit execution noise that contributes to the post-saccadic error. However, execution noise will be small and outbalanced by the recalibration through the comparison of predicted and post-saccadic error, as demonstrated by findings from saccade adaptation.19–21

A number of illusions24–30 or adaptation procedures31–36 distort the visual representation of space. It has been shown that saccades with very low latencies (<140 ms) are not affected by the visual distortion, whereas saccades with longer latencies are influenced by it. The frequent distortions of visual space imply a need for calibration. Dissociations between perceptual and action coding—as reported in Figure 2C—are in line with dual systems theories that proclaim a division between vision for recognition and vision for action.37–39 Our findings suggest that motor coordinates from the vision-for-action processing stream are dominant with regard to the calibration of spatial coding in perceptual maps. This dominance might consist simply because only actions can receive feedback about the actual position of objects in the external world. There is thus no reason for recalibration to be restricted to saccade-induced errors. Comparable recalibration of visual localization has been reported in the domain of reaching movements40 and through the acquisition of motor skills.41

Our results also allow us to quantify the claim of formulations of the predictive coding account, in which saccades are described as hypotheses to estimate the state of the external world.3 Given the high frequency in which we perform saccades in real life (3 Hz), small attractive serial effects are sufficient to guarantee alignment as demonstrated by the results shown in Figure 2C. An important question concerns the specificity of sensorimotor serial dependencies to visual distances. Does the post-saccadic error shift the position code in the entire map or specific locations within the map separately? Our results clearly indicate that recalibration is spatially selective. Spatial specificity is well known for saccade adaptation in which gain learning from systematic post-saccadic error manifests in a so-called adaptation field, i.e., a spatial window surrounding the adapted saccade.42 Selectivity of spatial recalibration is necessary to compensate for local distortions in the perceptual map that might occur, for instance, through adaptation processes. In a uniform shift of all positions, local distortions would survive. Spatial selectivity requires that, in addition to the retinal post saccadic error, the size of the intended saccade amplitude is processed. There is now evidence that both of these signals are represented in the activity of certain neurons in posterior parietal cortex (PPC): recent findings in electrophysiology have discovered cortical areas involved in saccade error processing.5 Neurons with persistent pre- and post-saccadic responses in the PPC reflect the intended saccade landing based on efference copy information. Neurons with late post-saccadic responses represent the actual saccade ending position. Postsaccadic error detection is a primary function of the oculomotor vermis in the cerebellum.43 The functional role of saccade error processing in the PPC is not yet clear. We suggest that this activity might calibrate motor and visual space likewise. The PPC is an ideal candidate for this role, given its well-known function in coordinating visual space44 and the more recently discovered function in representing recent sensory history.45

Star⋆Methods

Key Resources Table

REAGENT or RESOURCE SOURCE IDENTIFIER
Software and Algorithms
Mathematica Wolfram Mathematica RRDI :SCR_014448
R R Foundation for Statistical Computing N/A
OSF HOME this manuscript https://osf.io/hkuv7/
Open in a new tab

Resource Availability

Lead Contact

Further information and requests should be directed to and will be fulfilled by the Lead Contact, Eckart Zimmermann (eckart.zimmermann@hhu.de).

Materials Availability

The stimuli that support the findings of this study are available at https://osf.io/hkuv7/.

Experimental Model and Subject Details

Participants

All participants were recruited through the Heinrich-Heine University Düsseldorf and received either course credit or payment of 10€/h besides of the authors and participants working in the lab. Experimental procedures were approved the local ethics committee of the psychological department of the Heinrich-Heine University Düsseldorf. Written informed consent was obtained prior to each experiment in accordance with the declaration of Helsinki.

Apparatus

The subject was placed at a 57 cm distance from a FlexScan T57S, MA-1790, EIZO (Corporation, Japan) monitor. The visible screen diagonal was 40.5 cm, resulting in a visual field of 40° x30°. To avoid visual references the room was completely dark. A transparent foil reduced the luminance of the monitor by 2 log units and prevented the visibility of the monitor borders. All stimuli were presented with a refresh rate of 120 Hz and a resolution of 600 × 800 pixels. All stimuli were red or green squares (0.75° × 0.75°). Eye movements were recorded with the EyeLink 1000 system (SR Research Ltd., Mississauga, Ontario, Canada), which sampled eye positions at a rate of 1000 Hz. The head was sustained with a chin- and forehead-rest. For all subjects the left eye was recorded. Viewing was binocular. At the beginning of the session, the Eyelink was calibrated with the standard nine-point Eyelink procedure. Before each trial, fixation was checked.

Method Details

Experiment 1

Experiment 1 contained 4 conditions. In condition 1 and 3 the same 12 healthy subjects (five male, seven female, mean age 31.3 year, incl. one author) were measured. In condition 2, 6 different healthy subjects (three male, three female, mean age 27.5 year) were tested. In condition 4, four healthy subjects (two female, two male, mean age 26.6 year) All had normal or corrected-to-normal vision without any visual impairments. Prior to the experiment, all gave their informed consent which included all ethical standards. All of them received course credit or payment of 10€/h except the participating author.

Each session started with the presentation of a fixation square (0.75×0.75°), 10° to the left of screen center, to which participants had to direct their gaze. The color of the fixation square alternated between trials. A red fixation square signaled that a rightward saccade had to be performed as soon as a saccade target appeared and a green square indicated to keep fixation throughout the entire trial. The fixation square disappeared after 1000-1500 ms. In a saccade trial, a saccade target (0.75x0.75°, red color) then appeared in screen center. When the eye tracker detected that eye position deviated more than 2.5° from the fixation square, the saccade target was displaced to the left or right. Size and direction of this displacement was drawn each trial from a normal distribution with a μ = 0° and σ2 = 2°. The saccade target disappeared after 1260 ms and a new trial started automatically. Each subject performed 800 trials in an approximately 1 h session.

All conditions in Experiment 1 had a repeatedly alternating trial structure where a specific trial kind (n-1) was always followed by another trial kind (n). In the first session, participants had to perform a saccade to a target that was displaced intra-saccadically in trial n-1 and had to localize a target during fixation with a mouse pointer in trial n. In the second session, participants performed a saccade to a target that was displaced intra-saccadically in both, trial n-1 and trial n. The third session served as a control session. In trial n-1, participants were required to keep gaze directed to the fixation square while a target appeared in screen center after 1000-1500 ms. After 300 ms the target jumped to a new location that was determined on each trial by a normal distribution (μ = 0°, σ2 = 2°). Condition 4 was identical as condition 3, except that a whole-field random texture mask was presented for 60 ms at the time of the target displacement.

Experiment 2

In Experiment 2 two conditions were measured. In both conditions, the same 4 healthy subjects (three male, one female, mean age 32 year, incl. one author) participated. All had normal or corrected-to-normal vision without any visual impairments. Prior to the experiment, all gave their informed consent which included all ethical standards. All of them received course credit or payment of 10€/h except the participating author.

Experiment 2 contained a “localization only” and a “n-1: saccade / n: localization” condition. In the “localization only” condition, participants were instructed to fixate the fixation square throughout the entire session. The fixation square was located in the middle of the computer screen. After 1000-1500 ms the fixation square disappeared and a target was presented for 400 ms. After disappearance of the target the screen remained blank and another red localization square appeared 600 ms later. The square appeared at a random location within a spatial window of 5° around the target position. Participants could move the localization square by clicking the arrow keys. Every click on the corresponding arrow key moved the localization square 0.25° to the left or to the right. Participants were instructed to move the square to the perceived position of the target. In Experiment 2, the saccade target was not displaced, i.e., pre-saccadic and post-saccadic target location remained identical. Each subject performed 400 trials in one sessions. In the “n-1: saccade / n: localization” condition, there was a repeatedly alternating trial structure where participants had to perform a saccade to a saccade target in trial n-1 and localize the target while keeping gaze at the fixation point in the localization trial. The localization trials in the “n-1: saccade / n: localization” condition were identical to the localization trials in the “localization only” condition. The saccade target appeared after 1000-1500 ms and disappeared after 400 ms. All subjects finished their saccades before the target was extinguished (S1: latency: 235.66 ± 8.01, duration: 43.11 ± 0.21; S2: latency: 253.09 ± 3.14, duration: 47.78 ± 0.31; S3: latency: 218.76 ± 4.77, duration: 47.22 ± 0.57; S4: latency: 288.18 ± 40.02, duration: 44.35 ± 1.63). In trial n, participants had to localize a target that appeared after 1000-1500 ms and was shown for 400 ms with the arrow key and pressed space to continue to the next trial. The re-alignment condition contained 400 trials in one session and lasted in total 20 min.

Experiment 3

In Experiment 3 one condition was measured. 6 subjects (three female, three male, mean age 31.6 year, incl. one author) participated. All had normal or corrected-to-normal vision without any visual impairments. Participants gave their informed consent which included all ethical standards and received course credit or payment of 10€/h except the participating author.

Experiment 3 had a repeatedly alternating trial structure. Participants had to perform a saccade to a target that was displaced intra-saccadically in trial n-1 and had to localize a target during fixation with a mouse pointer in trial n. The target was presented on the screen center after 1000-1500 ms. The saccade target in trial n-1 always appeared at the same location, i.e., in screen center, such that the required saccade amplitude was 10°. The localization target in trial n appeared in 1 of 5 different possible positions (–6°,-8°,10°,12°,14°), randomly varied, across trials.

Quantification and Statistical Analysis

In the analysis of all experimental conditions we considered either saccade landing positions or visual localization data, i.e., the position of the mouse click at which participants reported the appearance of the target. In all conditions the targets (saccade targets and localization targets) were presented in screen center, i.e., at 0°. Leftward deviations of saccade landing or visual landing from the target were expressed as negative numbers and rightward deviations as positive numbers. All trials went into analysis in which participants performed the saccade to the target in the saccade trials and in which participants kept fixation in the visual localization trials. Saccade were classified as correct if landing of the horizontal component fell within a range of ± 5° from the horizontal saccade target coordinate. Fixations in visual localization trials were classified as correct if no saccade with a horizontal component larger than 2° was observed. Of those trials with correct saccade or localization performance we analyzed all trials with consecutive trial numbers, given the n-back logic of the experimental conditions.

Acknowledgments

This research was supported by the European Research Council (project moreSense grant agreement 757184) and by the Deutsche Forschungsgemeinschaft (DFG) (ZI 1456/5-1).

Footnotes

Author Contributions

Both authors conceptualized the study and contributed to the design. C.C. tested the participants and analyzed the data under the supervision of E.Z. Both authors contributed to the writing of the manuscript.

Declaration of Interests

The authors declare no competing interests.

Data and Code Availability

For data analysis, we used custom built scripts in Mathematica and R. including the ARTtoolbox. Our individual subject dataset is available in Excel format as Data S1. The data that support the findings of this study are available at https://osf.io/hkuv7/.

References

  • 1.Berkeley G. History of Psychology (Appleton-Century-Crofts) 1948. An essay toward a new theory of vision, 1709; pp. 69–80. [Google Scholar]
  • 2.Mach E. The analysis of sensation. Forgotten Books; 1886. [Google Scholar]
  • 3.Parr T, Friston KJ. The active construction of the visual world. Neuropsychologia. 2017;104:92–101. doi: 10.1016/j.neuro-psychologia.2017.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Zimmermann E, Lappe M. Visual Space Constructed by Saccade Motor Maps. Front Hum Neurosci. 2016;10:225. doi: 10.3389/fnhum.2016.00225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Zhou Y, Liu Y, Lu H, Wu S, Zhang M. Neuronal representation of saccadic error in macaque posterior parietal cortex (PPC) eLife. 2016;5:e10912. doi: 10.7554/eLife.10912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Manassi M, Liberman A, Kosovicheva A, Zhang K, Whitney D. Serial dependence in position occurs at the time of perception. Psychon Bull Rev. 2018;25:2245–2253. doi: 10.3758/s13423-018-1454-5. [DOI] [PubMed] [Google Scholar]
  • 7.Zimmermann E, Morrone MC, Burr DC. Buildup of spatial information over time and across eye-movements. Behav Brain Res. 2014;275:281–287. doi: 10.1016/j.bbr.2014.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Musseler J, van der Heijden AHC, Mahmud SH, Deubel H, Ertsey S. Relative mislocalization of briefly presented stimuli in the retinal periphery. Percept Psychophys. 1999;61:1646–1661. doi: 10.3758/bf03213124. [DOI] [PubMed] [Google Scholar]
  • 9.Desmurget M, Pélisson D, Grethe JS, Alexander GE, Urquizar C, Prablanc C, Grafton ST. Functional adaptation of reactive saccades in humans: a PET study. Exp Brain Res. 2000;132:243–259. doi: 10.1007/s002210000342. [DOI] [PubMed] [Google Scholar]
  • 10.Srimal R, Diedrichsen J, Ryklin EB, Curtis CE. Obligatory adaptation of saccade gains. J Neurophysiol. 2008;99:1554–1558. doi: 10.1152/jn.01024.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Havermann K, Lappe M. The influence of the consistency of postsaccadic visual errors on saccadic adaptation. J Neurophysiol. 2010;103:3302–3310. doi: 10.1152/jn.00970.2009. [DOI] [PubMed] [Google Scholar]
  • 12.Collins T. Trade-off between spatiotopy and saccadic plasticity. J Vis. 2014;14:28. doi: 10.1167/14.12.28. [DOI] [PubMed] [Google Scholar]
  • 13.Zimmermann E, Born S, Fink GR, Cavanagh P. Masking produces compression of space and time in the absence of eye movements. J Neurophysiol. 2014;112:3066–3076. doi: 10.1152/jn.00156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Harris LR, Smith AT. The coding of perceived eye position. Exp Brain Res. 2008;187:429–437. doi: 10.1007/s00221-008-1313-0. [DOI] [PubMed] [Google Scholar]
  • 15.Fischer J, Whitney D. Serial dependence in visual perception. Nat Neurosci. 2014;17:738–743. doi: 10.1038/nn.3689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Cicchini GM, Anobile G, Burr DC. Compressive mapping of number to space reflects dynamic encoding mechanisms, not static logarithmic transform. Proc Natl Acad Sci USA. 2014;111:7867–7872. doi: 10.1073/pnas.1402785111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Cicchini GM, Mikellidou K, Burr DC. The functional role of serial dependence. Proc Biol Sci. 2018;285:20181722. doi: 10.1098/rspb.2018.1722. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Fritsche M, Mostert P, de Lange FP. Opposite effects of recent history on perception and decision. Curr Biol. 2017;27:590–595. doi: 10.1016/j.cub.2017.01.006. [DOI] [PubMed] [Google Scholar]
  • 19.Chen-Harris H, Joiner WM, Ethier V, Zee DS, Shadmehr R. Adaptive control of saccades via internal feedback. J Neurosci. 2008;28:2804–2813. doi: 10.1523/JNEUROSCI.5300-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Collins T, Wallman J. The relative importance of retinal error and prediction in saccadic adaptation. J Neurophysiol. 2012;107:3342–3348. doi: 10.1152/jn.00746.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Wong AL, Shelhamer M. Sensorimotor adaptation error signals are derived from realistic predictions of movement outcomes. J Neurophysiol. 2011;105:1130–1140. doi: 10.1152/jn.00394.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Goodale MA. Transforming vision into action. Vision Res. 2011;51:1567–1587. doi: 10.1016/j.visres.2010.07.027. [DOI] [PubMed] [Google Scholar]
  • 23.Simons DJ, Rensink RA. Change blindness: past, present, and future. Trends Cogn Sci. 2005;9:16–20. doi: 10.1016/j.tics.2004.11.006. [DOI] [PubMed] [Google Scholar]
  • 24.Aglioti S, DeSouza JF, Goodale MA. Size-contrast illusions deceive the eye but not the hand. Curr Biol. 1995;5:679–685. doi: 10.1016/s0960-9822(95)00133-3. [DOI] [PubMed] [Google Scholar]
  • 25.Burr DC, Morrone MC, Ross J. Separate visual representations for perception and action revealed by saccadic eye movements. Curr Biol. 2001;11:798–802. doi: 10.1016/s0960-9822(01)00183-x. [DOI] [PubMed] [Google Scholar]
  • 26.Pavani F, Boscagli I, Benvenuti F, Rabuffetti M, Farne A. Are perception and action affected differently by the Titchener circles illusion? Exp Brain Res. 1999;127:95–101. doi: 10.1007/s002210050777. [DOI] [PubMed] [Google Scholar]
  • 27.Wong E, Mack A. Saccadic programming and perceived location. Acta Psychol (Amst) 1981;48:123–131. doi: 10.1016/0001-6918(81)90054-8. [DOI] [PubMed] [Google Scholar]
  • 28.Bridgemen B, Kirch M, Sperling A. Segregation of cognitive and motor aspects of visual function using induced motion. Percept Psychophys. 1981;29:336–342. doi: 10.3758/BF03207342. [DOI] [PubMed] [Google Scholar]
  • 29.de’Sperati C, Baud-Bovy G. Blind saccades: an asynchrony between seeing and looking. J Neurosci. 2008;28:4317–4321. doi: 10.1523/JNEUROSCI.0352-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Zimmermann E, Morrone MC, Burr D. Visual motion distorts visual and motor space. J Vis. 2012;12:10. doi: 10.1167/12.2.10. [DOI] [PubMed] [Google Scholar]
  • 31.Whitaker D, Mcgraw PV, Levi DM. The influence of adaptation on perceived visual location. Vision Res. 1997;37:2207–2216. doi: 10.1016/S0042-6989(97)00030-8. [DOI] [PubMed] [Google Scholar]
  • 32.Snowden RJ. Shifts in perceived position following adaptation to visual motion. Curr Biol. 1998;8:1343–1345. doi: 10.1016/S0960-9822(07)00567-2. [DOI] [PubMed] [Google Scholar]
  • 33.Nishida S, Johnston A. Influence of motion signals on the perceived position of spatial pattern. Nature. 1999;397:610–612. doi: 10.1038/17600. [DOI] [PubMed] [Google Scholar]
  • 34.McGraw PV, Whitaker D, Skillen J, Chung ST. Motion adaptation distorts perceived visual position. Curr Biol. 2002;12:2042–2047. doi: 10.1016/s0960-9822(02)01354-4. [DOI] [PubMed] [Google Scholar]
  • 35.Whitney D. The influence of visual motion on perceived position. Trends Cogn Sci. 2002;6:211–216. doi: 10.1016/S1364-6613(02)01887-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Zimmermann E, Morrone MC, Burr D. Adaptation to size affects saccades with long but not short latencies. J Vis. 2016;16:2. doi: 10.1167/16.7.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Goodale MA, Milner AD, Jakobson LS, Carey DP. A neurological dissociation between perceiving objects and grasping them. Nature. 1991;349:154–156. doi: 10.1038/349154a0. [DOI] [PubMed] [Google Scholar]
  • 38.Milner AD, Goodale MA. Two visual systems reviewed. Neuropsychologia. 2008;46:774–785. doi: 10.1016/j.neuropsychologia.2007.10.005. [DOI] [PubMed] [Google Scholar]
  • 39.Westwood DA, Goodale MA. Converging evidence for diverging pathways: neuropsychology and psychophysics tell the same story. Vision Res. 2011;51:804–811. doi: 10.1016/j.visres.2010.10.014. [DOI] [PubMed] [Google Scholar]
  • 40.Volcic R, Fantoni C, Caudek C, Assad JA, Domini F. Visumotor adaptaion changes stereoscopic depth perception and tactile discrimination. J Neuosci. 2013;33:17081–17088. doi: 10.1523/JNEUROSCI.2936-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Ostry DJ, Gribble PL. Sensory Plasticity in Human Motor Learning. Trends Neurosci. 2016;39:114–123. doi: 10.1016/j.tins.2015.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Frens MA, van Opstal AJ. Transferofshort-term adaptation in human saccadic eye movements. Exp Brain Res. 1994;100:293–306. doi: 10.1007/BF00227199. [DOI] [PubMed] [Google Scholar]
  • 43.Herzfeld DJ, Kojima Y, Soetedjo R, Shadmehr R. Encoding of error and learning to correct that error by the Purkinje cells of the cerebellum. Nat Neurosci. 2018;21:736–743. doi: 10.1038/s41593-018-0136-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Husain M, Nachev P. Space and the parietal cortex. Trends Cogn Sci. 2007;11:30–36. doi: 10.1016/j.tics.2006.10.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Akrami A, Kopec CD, Diamond ME, Brody CD. Posterior parietal cortex represents sensory history and mediates its effects on behaviour. Nature. 2018;554:368–372. doi: 10.1038/nature25510. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

For data analysis, we used custom built scripts in Mathematica and R. including the ARTtoolbox. Our individual subject dataset is available in Excel format as Data S1. The data that support the findings of this study are available at https://osf.io/hkuv7/.

RESOURCES