The authors have declared that no competing interests exist.
The knowledge about the usual size of objects—familiar size—is known to be a taken into account for distance perception. The influence of familiar size on action programming is less clear and has not yet been tested with regard to vergence eye movements. In two experiments, we stereoscopically presented everyday objects, such as a credit card or a package of paper tissues, and varied the distance as specified by binocular disparity and the distance as specified by familiar size. Participants had to fixate the shown object and subsequently reach towards it either with open or with closed eyes. When binocular disparity and familiar size were in conflict, reaching movements revealed a combination of the two depth cues with individually different weights. The influence of familiar size was larger when no visual feedback was available during the reaching movement. Vergence movements closely followed binocular disparity and were largely unaffected by familiar size. In sum, the results suggest that in this experimental setting familiar size is taken into account for programming and executing reaching movements while vergence movements are primarily based on binocular disparity.
The size of the retinal image created by an object is inversely proportional to its distance from the observer. If the object and therefore its size is familiar to the observer, the distance of the object can be calculated from the retinal image. For example, when we see a car and the car appears quite small, we conclude that the car is far away. This pictorial depth cue is called familiar size. Most objects we encounter are familiar to us, if not as an individual then as belonging to a category of objects. For example, we might not have seen the car before but we know how big cars typically are.
Familiar size is important for relative depth perception. For example, when pictures of a golf ball and a baseball with the same size are presented, the golf ball appears closer [
There is a number of studies on familiar size regarding grasping, yet only few concerning reaching movements. First, we will sum up the studies concerning grasping.
By directly investigating the use of familiar size for movement planning, Marotta and Goodale [
In a series of experiments, Sousa and colleagues [
In order to reduce the likelihood of being unfamiliar with the shown object, we stereoscopically presented an everyday object, namely a tennis ball supposing that its size was familiar to all the participants [
However, since we presented the same tennis ball in all trials, we cannot generalize the results to other objects. Furthermore, using only one object might have led to an over- or underestimation of the true effect of familiar size. An overestimation is possible since Sousa and colleagues [
The aim of the current study is to test whether the effect of familiar size on reaching movements with and without visual feedback can be replicated with other objects. Therefore, we stereoscopically presented six different objects and varied the distance as specified by binocular disparity and the size of the familiar object.
So far, we have discussed binocular disparity and (familiar) size as important depth cues for grasping and reaching. However, in personal space—the area surrounding the observer within arm’s reach and slightly beyond—vergence is another relevant depth cue [
Vergence movements are thought to be mainly driven by binocular disparity. Thus, most research on vergence movements has been done using either disparity step stimuli or disparity ramp stimuli, which means that an abstract object like a cross or a vertical line changes disparity either abruptly or smoothly. Consequently, most models on the control of vergence movements are based on those experiments and use disparity as the input signal [
In the studies mentioned above, binocular disparity was not available or it was congruent with monocular depth cues. There is ambiguous evidence for the influence on vergence movements when monocular depth cues or illusive perceptions are in conflict with binocular disparity. On the one hand, there are studies showing that vergence follows binocular disparity [
However, to our knowledge, so far it has not been examined whether the depth cue familiar size is used for the control of vergence movements. To investigate this question, we measured vergence movements in addition to reaching movements. The overall aim was to investigate whether familiar size is considered for controlling reaching and vergence movements.
To test the influence of familiar size on distance perception in personal space, we systematically varied the size of familiar everyday objects while measuring vergence eye movements and reaching movements. As results might depend on whether or not visual feedback is available during the reaching movement, we instructed participants to reach with eyes open (sighted reaches) or with eyes closed (blind reaches).
Data from 26 participants were obtained. All participants were right-handed as indicated by the Edinburgh Handedness Inventory [
Participants were seated individually in a dimly lit room. The experiment was presented on a Mitsubishi WD-60737 screen with a refresh rate of 60 Hz. This display is a rear-projection DLP™ TV. Since the whole screen exceeded the visual field, we used a screen resolution of 1064 by 598 pixels on a visible screen size of approximately 77 by 43 cm. Participants were seated 90 cm in front of the display with the eyes in the height of the center of the screen; the resulting viewing angle was approximately 46° horizontally and 22° vertically. However the stimuli were presented in the center of the screen. For image separation, participants wore shutter glasses (3D Vision glasses by Nvidia Corporation, Santa Clara, USA) which were synchronized with the display.
To control the experiment, participants used a Cedrus Response Pad RB-840 (cedrus, San Pedro, USA). The response pad was placed on the table in front of the participant with the button for the right index finger 65 cm in front of the display and 48 cm below the middle of the display where the objects were presented.
Eye movements were recorded binocularly with a sample rate of 500 Hz using the SR Research Ltd. EyeLink 1000 Plus eye-tracking system (SR Research, Ontario, Canada), with a spatial resolution below 0.01° and a spatial accuracy of better than 0.5°. Head movements were restrained by a chin and forehead rest.
Reaching estimates were recorded using the marker-based optical tracking system OptiTrack V120: TRIO (NaturalPoint, Inc.), which was mounted above the participants’ seat. It consists of three 640 * 480 VGA sensors, has a capture speed of 120 FPS and—according to the manufacturer—sub-millimeter accuracy. A reflective sphere with a diameter of 9.5 mm was attached to the right index finger. The distance from the tip of the finger to the marker was measured individually. The position of the marker was streamed via the NatNet SDK to Vizard 5.8 (WorldViz, Santa Barbara, USA), a Python based VR development platform. Vizard was also used for the creation and rendering of the objects, for handling the experimental flow and recording experimental data. Rendering was done using participants’ individual interpupillary distances. These were measured with a pupilometer (Essilor PRC), which in an earlier study had an interrater reliability of
Schematic representation of the experimental setup (not drawn to scale, forehead rest not shown); adapted from [
The stimuli in the experiment consisted of six different everyday objects. The six objects were selected after a pilot study with 20 participants (11 female) between 18 and 41 years old (
Once the informed consent was given, the vision tests were run, and a demographic questionnaire was completed using LimeSurvey with a local server [
The experiment was divided into two blocks, one for sighted and one for blind reaches. The order of the blocks was balanced across participants. Before each block the eye-tracker was calibrated using a standard nine-point binocular calibration, followed by a nine-point validation procedure. Each block consisted of 6 trials for exercise, which were not analyzed, followed by a new calibration and validation and 42 experimental trials. During the exercise trials, the objects had the correct size and were shown with a distance of either 25, 35, or 45 cm from the participant’s eyes. Each object was presented once, whereby the distance-object combination varied randomly for each participant. In the experimental trials the objects were presented in the distance as specified by disparity of either 30 or 40 cm. Each object was presented once at 30 and once at 40 cm in the correct size. To test the influence of familiar size, each object was also presented in the distance of 30 cm with a smaller size suggesting a distance of 40 cm and in the distance of 40 cm with a larger size suggesting a distance of 30 cm. Additionally, to prevent participants from remembering the distances and to reduce the rate of trials in which there was a conflict between disparity-specified distance and familiar size-specified distance, each object was presented at the distance of 25, 35, and 45 cm once in the correct size. Thus, in 28.6% of the experimental trials, there was a conflict of the distance specified by binocular disparity or specified by familiar size. Objects always appeared in the center of the screen at eye height with the side facing the participant at the intended distance.
The procedure of one trial was as follows: After a drift correction with the stimulus (a grey ring) on the level of the screen (with zero parallax), the object was presented. Participants were allowed to look at the object for unlimited time. When they felt they had a good image of the object and its position, they looked in the center of the object and pressed a button with the right index finger. 1000 ms after the button press, there was an acoustic signal telling the participants that they could start the reaching movement. In blind reaching trials, the object was blanked out with the acoustic signal and participants closed their eyes to prevent visual feedback. In sighted reaching trials the object and the fingers were visible throughout the whole reaching movement. Participants performed the reach with their right index finger outstretched to the position where they would touch the surface of the object with their finger tip. Participants were instructed to make a “fast and natural” movement. The participants terminated their reach by pressing the button with their left index finger. Subsequently, they laid their right hand back on the response pad and for blind reaches opened their eyes again. Next, a text message was displayed on the level of the screen, asking the participants whether they saw a 3D image or a double image. Participants answered by pressing with their left or right middle finger, respectively, and confirmed their response via the same button. Then participants started the next trial with a button press of the right index finger.
Data preparation and analyses were done using SPSS 20 [
Raw data from the eye-tracking system were transformed from pixel to mm and to a coordinate system with the origin at the center of the screen. All samples corresponding to blinks were eliminated and replaced via interpolation by 3rd order splines using the R package zoo [
Raw data from the motion tracker was corrected by the offset between the tip of the finger and the marker. The onset of the reaching movement was defined as the first time when the marker was lifted more than 2 cm above the button. Reaching duration was calculated as the time from onset of the reaching movement to the button press indicating the subjectively defined end position. To calculate the reaching distance, we used the sample containing the button press of the participant, 11 samples before and 12 samples after it and calculated the median. Since we measured reaching movements with 120 Hz, we thus averaged over 200 ms. Data were then transformed into an egocentric coordinate system with the unit cm.
In the process of data filtering, all trials, in which a participant did not follow the instructions (e.g., pressed the buttons in the wrong order), reported to have seen a double image, or in which there were technical problems, were excluded from the analyses (8%). For vergence movement analyses, trials containing a blink in the first 50 ms and trials with a vergence latency shorter than 50 ms were excluded as this distorts the baseline correction described above. Also, vergence distance had to be between 0 and 60 cm, vergence latency between 50 and 650 ms. Further, for vergence latency and maximal vergence velocity, trials containing a blink in the first 650 ms were excluded. For reaching movement analyses, we excluded trials in which the finger was still on the table when reaching distance was collected as apparent in the height of the marker on the fingertip. After data filtering, 86% of the trials remained for analyses in the relevant distances of 30 and 40 cm for vergence distance; 74% of the trials remained for vergence latency and maximal vergence velocity; 89% for reaching distance and reaching duration.
All analyses were done using linear mixed models (LMMs). The disparity and size manipulations were coded as factors named disparity-specified distance (30, 40) and familiar size-specified distance (30, 40), which could thus be congruent with no conflict between the two depth cues or incongruent with the two depth cues specifying different distances (conflict trials). Please note that this coding is equivalent to other studies [
LMMs were used for the analyses since they allow to simultaneously account for multiple sources of (random) variability and thus generalize the experimental effects and interactions across both subjects and items [
Choosing the appropriate random effects structure is crucial when using LMMs [
We modeled the main effect of reaching type (sighted or blind reaches), the main effect of disparity-specified distance (30 or 40), the main effect of familiar size-specified distance (30 or 40) as well as all interactions between them as fixed effects. We then added random intercepts, random slopes, as well as correlations between them for participants, objects, and participant-object-combinations. For the three-way interaction there was no random slope for participant-object combinations, as each object was only presented once for each participant in each combination of conditions. For participants, random intercepts account for the fact that participants may differ in their general tendency to perceive distances—further away or closer; random slopes account for the fact that the manipulations (reaching type, disparity-specified distance, familiar size-specified distance) might affect some participants more than others (e.g., in previous studies participants differed in the weight they assigned to familiar size in their reaching estimates). For objects, random intercepts account for the fact that objects may differ in their general tendency to be perceived as closer or further away; random slopes account for the fact that the manipulations might have an effect for some objects but not for others. Random intercepts and random slopes for participant-object-combinations account for the fact that the mean reaching distance might vary between different participant-object-combinations and that the experimental effects might not be of the same size for all combinations of participants and objects. Since this is a study on familiar size, it is crucial that the participants recognize the object and activate a concept of its usual size but this is most likely not the case for all objects for all participants. For example, some participants might play tennis and know exactly the size of a tennis ball and this should influence their behavior; whereas those who have not seen or touched a tennis ball for years are not influenced by its size, but might be influenced by the size of other objects. Additionally, all correlations between random slopes and random intercepts within the same grouping factor can be modeled. This maximal model would require estimating a total of 23 variance components plus 77 correlations. To estimate this random effects structure, there are not enough data points in our study; therefore, the estimation algorithm does not converge when estimating a maximal model. Even if it would converge, the estimated LMMs would be still likely to be over-parameterized relative to the data and their estimated parameters thus not reliable. Therefore, we followed the recommendation by Bates and colleagues [
LMM analyses were estimated with orthogonal contrasts in R using the package afex [
We always started with the maximal model. Since this does usually not converge (given the high number of model parameters that need to be estimated), in the next step, we specified a zero-correlation parameter model including all variance components, but no covariance parameters. We then did a principal component analysis (PCA) of the variance-covariance matrices of the random effects structure using the function rePCA() from the package RePsychLing [
After model reduction, the final model was used to exclude outliers. As the presence of outliers may cause stress in the model and distort the estimated parameters, we removed data points with absolute standardized residuals greater than 3 (between 0.68 and 1.92% of the data).
After data reduction, significance values for the fixed effects of the final model were computed via the Kenward-Roger approximation for degrees of freedom [
First of all, we looked at all congruent trials. We aggregated vergence distance and reaching distance over the six objects using the median. As expected, the repeated measures correlation indicated that there was a significant association between aggregated vergence distance and object distance,
Mean vergence distance for object distance; each data point represents the median for one participant (aggregated over six objects); the dashed black line represents the object distance as specified by binocular disparity and familiar size.
For reaching distance, there was a significant repeated measures correlation between aggregated reaching distance and object distance for sighted reaches,
Mean sighted and blind reaches for object distance; each data point represents the median for one participant (aggregated over six objects); the dashed black line represents the object distance as specified by binocular disparity and familiar size.
Next, we compared the vergence movements in congruent and conflict trials for the relevant distances of 30 and 40 cm.
Mean vergence distance for a binocular disparity-specified distance of 30 cm (left) and 40 cm (right) and a familiar size-specified distance of 30 cm (solid orange) and 40 cm (dashed blue) as a function of time from object onset.
For vergence movements, the parameters vergence latency, maximal vergence velocity, and vergence distance were tested. Mean vergence latency, the time from object onset until the vergence movement started, was 216 ms. Using linear mixed models, we tested whether vergence latency was influenced by disparity-specified distance or familiar size-specified distance. Neither the main effect of disparity-specified distance,
Next, we tested the effect of disparity-specified distance and familiar size-specified distance on maximal vergence velocity. The main effect of disparity-specified distance was significant,
The effect of disparity-specified distance and familiar size-specified distance on vergence distance is displayed in
Estimated marginal means of vergence distance for a binocular disparity-specified distance of 30 cm and 40 cm and a familiar size-specified distance of 30 cm (solid orange) and 40 cm (dashed blue). Error bars are model-based 95% confidence intervals indicating which values of the estimated means are likely and do not permit comparisons across repeated-measures factors [
Next, we compared reaching distance and reaching duration in congruent and conflict trials for the relevant distances of 30 and 40 cm. The effect of disparity-specified distance and familiar size-specified distance on reaching distance is displayed in
Estimated marginal means of reaching distance for sighted reaches (left) and blind reaches (right), for a binocular disparity-specified distance of 30 cm and 40 cm and a familiar size-specified distance of 30 cm (solid orange) and 40 cm (dashed blue). Error bars are model-based 95% confidence intervals indicating which values of the estimated means are likely and do not permit comparisons across repeated-measures factors [
Mean reaching duration for sighted reaches (left) and blind reaches (right) for congruent trials (disparity-specified distance corresponded to familiar size-specified distance; dark gray) and conflict trials (disparity-specified distance was opposed to familiar size-specified distance; light gray) of the by-subject aggregated data. Error bars are within-subjects 95% confidence intervals according to Morey [
Reaching duration, the time from the start of the reaching movement until the button press indicating the end of the movement, was on average 2738 ms for sighted reaches and 2232 ms for blind reaches (
The research question was whether familiar size is considered for controlling reaching and vergence movements. To investigate this, we systematically varied the size of stereoscopically presented familiar everyday objects while measuring vergence eye movements and reaching movements.
Concerning reaching movements, we found that participants reached in the expected distance when binocular disparity and familiar size specified the same distance (congruent trials). When the two depth cues specified different distances (conflict trials), both binocular disparity and familiar size significantly influenced reaching distance. When visual feedback was available (sighted reaches), binocular disparity was given more weight while without visual feedback (blind reaches) reaching distances were more affected by familiar size. Thus, the results suggest that both depth cues were combined with the weight depending on the reaching type (with or without visual feedback) and the participant’s strategy. These results are in line with the findings of an earlier experiment [
Concerning vergence movements, the mean of the vergence distances in congruent trials was smaller than expected for all distances. On average, computed vergence distance was about 3 cm shorter than object distance as specified by binocular disparity and familiar size. On the one hand, this is not surprising as other researchers have also found non-perfect vergence distances and considerable individual differences [
When familiar size-specified distance was in conflict with disparity-specified distance, vergence distance followed largely the distance as specified by binocular disparity. This is in line with models on the control of vergence movements [
Experiment 2 was basically a replication of experiment 1. As in experiment 1, we systematically varied the size of familiar everyday objects while measuring vergence eye movements and reaching movements.
We conducted a power analysis based on simulations from an LMM (using the R package simr [
The experimental setup was exactly the same as in experiment 1 except for the lighting conditions. While experiment 1 was conducted in a dimly lit room, all lights were turned on in experiment 2 resulting in a brightly lit room. The stimuli were the same as in experiment 1 with the exception that the six objects were presented on a bright background. The procedure of the experiment and the instructions for the participants were exactly the same as in experiment 1. Data preparation was also exactly the same as in experiment 1. After data filtering, 88% of the trials remained for analyses in the relevant distances of 30 and 40 cm for vergence distance; 75% of the trials remained for vergence latency and maximal vergence velocity; 89% for reaching distance and reaching duration. Data were analyzed using LMMs with the procedure described above. For comparability to experiment 1, we used log-transformed values for the dependent variables maximal vergence velocity and reaching duration. The final models can be found in
Results for experiment 2 are reported following the same structure as for experiment 1.
For congruent trials, repeated measures correlation indicated that there was a significant association between aggregated vergence distance and object distance,
Mean vergence distance for object distance; each data point represents the median for one participant (aggregated over six objects); the dashed black line represents the object distance as specified by binocular disparity and familiar size.
For reaching distance, there was a significant repeated measures correlation between aggregated reaching distance and object distance for sighted reaches,
Mean sighted and blind reaches for object distance; each data point represents the median for one participant (aggregated over six objects); the dashed black line represents the object distance as specified by binocular disparity and familiar size.
Next, we compared the vergence movements in congruent and conflict trials for the relevant distances of 30 and 40 cm. Mean vergence latency, the time from object onset until the vergence movement started, was 227 ms. We tested the effect of disparity-specified distance and familiar size-specified distance on vergence latency. Vergence latency was not significantly affected by disparity-specified distance,
Next, we tested the effect of disparity-specified distance and familiar size-specified distance on maximal vergence velocity. The main effect of disparity-specified distance was significant,
The effect of disparity-specified distance and familiar size-specified distance on vergence distance is displayed in
Estimated marginal means of vergence distance for a binocular disparity-specified distance of 30 cm and 40 cm and a familiar size-specified distance of 30 cm (solid orange) and 40 cm (dashed blue). Error bars are model-based 95% confidence intervals indicating which values of the estimated means are likely and do not permit comparisons across repeated-measures factors [
Next, we compared reaching distance and reaching duration in congruent and conflict trials for the relevant distances of 30 and 40 cm. The effect of disparity-specified distance and familiar size-specified distance on reaching distance is displayed in
Estimated marginal means of reaching distance for sighted reaches (left) and blind reaches (right), for a binocular disparity-specified distance of 30 cm and 40 cm and a familiar size-specified distance of 30 cm (solid orange) and 40 cm (dashed blue). Error bars are model-based 95% confidence intervals indicating which values of the estimated means are likely and do not permit comparisons across repeated-measures factors [
Mean reaching duration for sighted reaches (left) and blind reaches (right) for congruent trials (disparity-specified distance corresponded to familiar size-specified distance; dark gray) and conflict trials (disparity-specified distance was opposed to familiar size-specified distance; light gray) of the by-subject aggregated data. Error bars are within-subjects 95% confidence intervals according to Morey [
Reaching duration, the time from the start of the reaching movement until the button press indicating the end of the movement, was on average 2049 ms for sighted reaches and 1971 ms for blind reaches (
The aim of experiment 2 was to replicate the results of experiment 1 concerning reaching movements and to test the possibility that the significant effect of familiar size on vergence distance found in experiment 1 was not due to familiar size itself but a measurement error caused by the pupil size artefact [
Concerning reaching movements, we found that the mean reaching distances corresponded to the expected distance when binocular disparity and familiar size specified the same distance (congruent trials). Both binocular disparity and familiar size significantly influenced reaching distance when the two depth cues specified different distances (conflict trials). The results replicate the findings of experiment 1 and of a previously published experiment [
Concerning vergence movements, the mean of the vergence distances in congruent trials corresponded quite precisely to the expected distances. When familiar size-specified distance was in conflict with disparity-specified distance, vergence distance was significantly affected by binocular disparity. The effect of familiar size-specified distance was far from reaching significance and only explained less than 1% of the variance. Thus, the change in luminance conditions compared to experiment 1 did change the results concerning vergence movements. This suggests that the effect of familiar size-specified distance on vergence distance found in experiment 1 indeed has to be attributed to the pupil size artefact.
The size of familiar objects can be used as a cue to their distance. The influence of familiar size on relative depth perception and grasping movements has been shown. Here, two experiments were conducted to examine whether familiar size is taken into account for the control of reaching and vergence movements. Six different objects with familiar size were stereoscopically presented. The distance to the objects as specified by binocular disparity and the distance as specified by familiar size were manipulated. Participants made vergence movements and reaching movements with eyes open (sighted reaches) or with eyes closed (blind reaches) towards the objects. We begin with a discussion of the influence of familiar size on reaching movements before elaborating on the influence on vergence movements which is followed by a discussion of the limitations of the current experiment as well as open research questions.
In both experiments, reaching distances for sighted reaches were highly correlated with object distance as specified by binocular disparity and familiar size (congruent trials) and distributed around the expected values. This supports the basic assumption that most participants in the given setting are indeed able to reach correctly to stereoscopically presented objects. When no visual feedback was available during the reaching movement (blind reaches), reaching distances were more dispersed between subjects, but still highly correlated with object distance. This is in line with previous work demonstrating larger variation and more errors when visual feedback is not available [
When familiar size-specified distance was in conflict with disparity-specified distance, results from both experiments suggest that information from both depth cues was combined for reaching movements. Since we did not find evidence for bimodal distributions with means close to 30 and 40, respectively, of reaching distances, participants did likely not decide which depth cue to follow but made a compound distance estimate. Participants varied in the weight they gave to each cue for this compound estimate as indicated by the variance of the random slope of familiar size-specified distance for participants. While some participants oriented themselves more on binocular disparity, others gave more weight to familiar size. These results replicate the findings of an earlier experiment [
Also in line with our earlier experiment [
To sum up, familiar size influences reaching movements towards stereoscopically presented familiar objects, especially when no visual feedback is available. But what about the influence on vergence movements?
In both experiments, vergence distances were highly correlated with object distances as specified by both binocular disparity and familiar size (congruent trials). While in experiment 1, computed vergence distance was on average about 3 cm shorter than expected, in experiment 2, vergence distances in congruent trials corresponded quite precisely to the expected distances. Thus, the deviations in experiment 1 were probably due to the difference in luminance between the calibration screen (small grey ring on black background) and the stimulus screen (larger colored object on black background).
When familiar size-specified distance was in conflict with disparity-specified distance, vergence distance followed the distance as specified by binocular disparity. The effect of familiar size-specified distance on vergence distance was significant in experiment 1, but could not be replicated in experiment 2. This was despite the sufficiently large power of experiment 2. The two experiments differed only in luminance conditions. It is therefore likely, that the effect of familiar size-specified distance on vergence distance found in experiment 1 was due to the pupil size artefact described by Hooge and colleagues [
Vergence latency, the time from object onset until the vergence movement started, was neither influenced by binocular disparity nor by familiar size in experiment 1. In experiment 2, there was a significant effect of familiar size on vergence latency. However, the effect explained less than 1% of the variance and the mean difference between a familiar-size specified distance of 30 cm compared to 40 cm was only 2 ms. Thus, taken together the results suggest that vergence latency is not substantially influenced by familiar size. This is opposed to the results of Grosjean and colleagues [
Maximal vergence velocity depended on binocular disparity with larger velocities for smaller object distances in both experiments. This result was expected as vergence movements with larger amplitudes are known to have higher maximal velocities [
There are several limitations of the current experiment, which restrict the generalization of the results and pose questions for further research:
We used virtual instead of real objects in order to be able to present objects in several sizes and several distances with high precision, and to have a high level of experimental control. However, this raises the question whether the results can be generalized to reaching and vergence movements towards real objects. There are only few studies on reaching or grasping presenting both virtual and real objects. In general, the results for virtual and real objects are similar: manipulations had the same effect, although the movements were in part less precise towards virtual objects and the magnitude of the effects could differ [
All studies, in which familiar objects are repeatedly presented, have the limitation that the method does not allow us to differentiate between the effect of long-term familiarity with stored sizes in long-term memory as opposed to new associations learned from the congruent trials. This means that participants might have learned how big the tennis ball was in most of the trials instead of referring to their knowledge about tennis balls. Indeed, experiments with a priori unknown objects have shown that such associations can be acquired and used for programming grasping movements when feedback is available, objects are fully textured, and only few distances are used [
In our experiments, participants differed in the weight they assigned to each depth cue for the control of their reaching movement. While some gave more weight to binocular disparity, others oriented themselves mostly on familiar size. The reasons for this different weighting are still an open research question.
We did not find an effect of familiar size on vergence movements. This does, of course, not rule out the possibility that there is an effect under differing experimental conditions. At least since the work of Yarbus [
Using stereoscopic images of everyday familiar objects we replicated the result that familiar size is taken into account for programming and executing reaching movements towards familiar objects, especially when no visual feedback is available. On the other hand, vergence movements were largely unaffected by familiar size in the same setting and closely followed binocular disparity. This novel finding emphasizes differences in mechanisms of eye movement control and those of manual actions such as reaching and grasping. Since this is the first study that explores the effect of familiar size on vergence movements, many questions concerning neurophysiological mechanisms behind these differences are open for future investigations.
(DOCX)
The authors thank Susanne Jaster, Theresia Fippel, and Adam Kneitz for their help with the data collection of the experiment, Adam Kneitz for his help with literature research, Stefan Vogt for the data collection of the preliminary study for stimuli selection, and Rica Bönsel and Anneke Hamann for data collection, preliminary analysis, and discussion of a pilot study as part of their student research project.