Open Access
Article  |   January 2025
Impaired visual perceptual accuracy in the upper visual field induces asymmetric performance in position estimation for falling and rising objects
Author Affiliations
  • Takashi Hirata
    Department of Cognitive and Psychological Sciences, Graduate School of Informatics, Nagoya University, Aichi, Japan
    JSPS Research Fellowships for Young Scientists, Tokyo, Japan
    [email protected]
  • Nobuyuki Kawai
    Department of Cognitive and Psychological Sciences, Graduate School of Informatics, Nagoya University, Aichi, Japan
    Academy of Emerging Sciences, Chubu University, Kasugai, Aichi, Japan
    [email protected]
Journal of Vision January 2025, Vol.25, 1. doi:https://doi.org/10.1167/jov.25.1.1
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Takashi Hirata, Nobuyuki Kawai; Impaired visual perceptual accuracy in the upper visual field induces asymmetric performance in position estimation for falling and rising objects. Journal of Vision 2025;25(1):1. https://doi.org/10.1167/jov.25.1.1.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Humans can estimate the time and position of a moving object's arrival. However, numerous studies have demonstrated superior position estimation accuracy for descending objects compared with ascending objects. We tested whether the accuracy of position estimation for ascending and descending objects differs between the upper and lower visual fields. Using a head-mounted display, participants observed a target object ascending or descending toward a goal located at 8.7° or 17.1° above or below from the center of the monitor in the upper and lower visual fields, respectively. Participants pressed a key to match the time of the target's arrival at the goal, with the gaze kept centered. For goals (8.7°) close to the center, ascending and descending objects were equally accurate, whereas for goals (17.1°) far from the center, the ascending target's position estimation in the upper visual field was inferior to the others. Targets moved away from the center for goals further from the center and closer to the center for goals nearer to the center. As the positional accuracy of ascending and descending objects was not assessed for each of the four goals, it remains unclear which was more important for impaired accuracy: the proximity of the target position or direction of the upward or downward motion. However, taken together with previous studies, we suggest that estimating the position of objects moving further away from the central fovea of the upper visual field may have contributed to the asymmetry in position estimation for ascending and descending objects.

Introduction
Humans can accurately estimate the position of a moving object, as evidenced by their behavior to successfully catch or intercept a falling object at an appropriate moment (Brenner, Driesen, & Smeets, 2014). Interestingly, our performance in catching or intercepting moving objects exhibits an asymmetry in the vertical direction (upward and downward). Specifically, people estimate arrival time or time to contact more accurately for a descending object than for an ascending one (Baurés & Hecht, 2011; Hirata, Hirata, & Kawai, 2024; Le Séac'h, Senot, & McIntyre, 2010; Senot, Zago, Lacquaniti, & McIntyre, 2005; Senot et al., 2012; Zago, La Scaleia, Miller, & Lacquaniti, 2011). These studies suggest that humans have a superior ability to estimate the positions of descending objects. 
The asymmetry in the accuracy of position estimation may be caused by our long experience on Earth, where the objects we interact with fall more often than rise in everyday environments. Given that naturally falling objects always move in the direction of gravity, our enhanced ability to estimate the position of a descending object may become evident when the motion of the object aligns with the direction of gravity. However, Hirata et al. (2024) reported that this is not the case. In their study, participants saw a ball target moving either upward or downward via a head-mounted display (HMD) along their longitudinal body axis in both the upright and supine postures. The target moved either above or below the observer's visual field in both postures. The results showed that the estimation of arrival time for downward-moving objects was superior, regardless of postures, indicating that gravity information from vestibular inputs does not influence the observed vertical asymmetry in position estimation accuracy. 
In somewhat different literature, a vertical asymmetry in human visual perceptual accuracy has been documented. Visual perception is more accurate in the lower visual field than in the upper visual field. Previous studies reported that contrast sensitivity (Cameron, Tai, & Carrasco, 2002; Carrasco, Williams, & Yeshurun, 2002) and spatial resolution (Barbot, Xue, & Carrasco, 2021; Carrasco et al., 2002) are higher in the lower than in the upper visual field. These studies suggest that human performance in recognizing static objects is better in the lower than in the upper visual field. In fact, people discriminate static images more accurately in the lower than in the upper visual field (He, Cavanagh, & Intriligator, 1996; Levine & McAnany, 2005; Schütz, Braun, & Gegenfurtner, 2009; Zito, Cazzoli, Müri, & Mosimann, 2016). These performance asymmetries in the upper and lower visual fields may reflect retinotopic asymmetry (Corbett & Carrasco, 2011). Namely, it has been demonstrated that the density of ganglion cells is higher on the dorsal side of the retina (corresponding to the lower visual field) than on the ventral side (upper visual field) (Kupers, Benson, Carrasco, & Winawer, 2022). Asymmetry in the density of retinal ganglion cells is more pronounced on the farther side of the upper and lower visual fields (Kupers et al., 2022). Therefore, visual perceptual accuracy may show asymmetry in the upper and lower visual fields, particularly on the far side of these visual fields. 
Visual perceptual asymmetry in the upper and lower visual fields has also been reported in moving objects. Discriminating a specific moving object from many objects moving in various directions is more accurate in the lower than in the upper visual field (He et al., 1996; Zito et al., 2016). The interception performance of a horizontally moving visual target is also better when observers capture the target in the lower than in the upper visual field (Danckert & Goodale, 2001). These findings suggest that asymmetries in visual perceptual accuracy in the upper and lower visual fields may not only occur in the retina and primary visual cortex. In addition to the primary visual cortex, they may also be observed in visual motion processing, which occurs further downstream. 
Previous studies have evaluated the accuracy of position estimation for ascending and descending objects in various situations, such as different postures (Hirata et al., 2024; Senot et al., 2005; Zago et al., 2011), different accelerations (Hirata et al., 2024; Indovina et al., 2005; Le Séac'h et al., 2010; Miller et al., 2008; Senot et al., 2005), different visual scene context (Zago et al., 2011), or different object shapes (Gallagher, Torok, Klaas, & Ferré, 2020). However, no study has examined the accuracy of position estimation of vertically moving objects in the upper and lower visual fields separately. Given that with human visual perceptual accuracy, the lower visual field is superior to that in the upper visual field, we predicted that asymmetry in position estimation accuracy for ascending and descending objects might have occurred in previous studies because observers recognized ascending and descending objects in the upper and lower visual fields, respectively. In this study, we evaluated the accuracy of arrival time estimation for ascending and descending objects when observers recognized them in either the upper or the lower visual field. 
Methods
Participants
Forty-four people participated in this experiment (22 females and 22 males, mean age 29.5 ± 10.25 years). All the participants had normal or corrected vision. They provided written informed consent after the aims and procedures of the experiment were explained to them. The experimental procedures were designed and conducted in accordance with the Declaration of Helsinki and approved by the Committee for Human Research at Nagoya University (230308-C-01). 
Apparatus
Virtual reality (VR) visual stimuli were presented using an HMD (VIVE Pro Eye; HTCVIVE, New Taipei City, Taiwan) at a refresh rate of 90 Hz. The VR stimuli were created in Unity (Unity Technologies, San Francisco, CA, USA) using customized programs written in C#. Participants were seated on a chair in an upright position (Figure 1A). A chinrest was provided to ensure a stable posture and maintain head position. The participants placed the index finger of their dominant hand on a keypad to respond to the visual stimuli in the VR scenes. 
Figure 1.
 
Experimental design. (A) Experimental setup. All participants sat in the chair and put their chins on the chinrest. An HMD was used to show the visual stimulus. Participants placed the index finger of their dominant hand on the keypad. (B) Virtual reality scenes depicted the following movements from the left side of the view: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend. The magenta lines indicate the direction of the ball's target motion in each VR scene. The red dot was the fixation point. (C) Experimental schedule and presentation order.
Figure 1.
 
Experimental design. (A) Experimental setup. All participants sat in the chair and put their chins on the chinrest. An HMD was used to show the visual stimulus. Participants placed the index finger of their dominant hand on the keypad. (B) Virtual reality scenes depicted the following movements from the left side of the view: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend. The magenta lines indicate the direction of the ball's target motion in each VR scene. The red dot was the fixation point. (C) Experimental schedule and presentation order.
Visual stimuli
All participants observed four types of VR scenes: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend (Figure 1B). A white ball target with a diameter of 0.2 m in the virtual world (horizontal and vertical visual angles: 0.88°) and a white ring-shaped goal (horizontal and vertical visual angles: 1.56° and 0.22°, respectively) were presented at 13 m away from the participant's viewpoint in all VR scenes. The target moved from 8.7° to 17.1° or from 17.1° to 8.7° above (in the upper visual field) or below (in the lower visual field) the center of the scene (0°). The distance of the target motion was the same at 4 m (8.4°) for all VR scenes. In each VR scene, the goal was located at 17.1° and 8.7° in the upper and lower visual fields. The target moved upward toward the goal in the Upper-Ascend and Lower-Ascend scenes or downward toward the goal in the Upper-Descend and Lower-Descend scenes. When the target arrived at the goal in all VR scenes, it overlapped with the ring-shaped goal, resembling the appearance of Saturn. The target did not make physical contact with the goal in either scene. 
In all VR scenes, the target remained stationary at its initial position for 3,000 ms before moving toward the goal. The target was accelerated by 9.81 m/s2 in all VR scenes. The initial velocity of the target was 0 m/s. Hence, the target arrived at the goal of 670 ms for all VR scenes. The participants were not informed of the motion duration or target kinematics. After arrival at the goal, the target disappeared for 330 ms (intertrial interval [ITI]). Each trial lasted 4 s (3,000 ms pause + 670 ms motion + 330 ms ITI), and each type of VR scene consisted of 20 trials. 
Procedure
Before the experiment, the participants adjusted their chair and chinrest heights to comfortable positions and wore the HMD. A 5-point calibration was performed to determine the participants’ eye positions before each VR scene was presented. Each VR scene was presented separately in four blocks: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend. Four blocks were presented following the Greco-Latin square (Figure 1C) across participants. For each VR scene, the participants were instructed to press the “0” key of the numerical pad with the index finger of their dominant hand as accurately as possible to coincide with the timing of the target's arrival at the goal. Participants did not receive feedback from their key responses. Their key responses did not affect the target motion in the VR scenes. Furthermore, the target did not stop when the participant pressed the key before it arrived at the goal. They were also instructed to keep looking at the red fixation dot at the center of the VR scene (Figure 1B) during the trial and not move their heads while viewing each VR scene. Blinking was not allowed during target motion for 670 ms. 
Data analysis and exclusion criteria
All data were analyzed using MATLAB (MathWorks, Natick, MA, USA). Twenty key response data were recorded for each of the four VR scenes. To ensure that the participants maintained their gaze at the fixation point, eye positions were recorded during the experiment using Tobii eye cameras with an HMD (VIVE Pro Eye with Tobii eye tracking) at a sampling rate of 120 Hz. Figures 2A and 2B illustrate the horizontal and vertical eye positions during target motion (670 ms) for two typical participants in all 20 trials of the Lower-Ascend scene. As shown in Figure 2A, Participant 1 succeeded in looking close to (within the ± 3° acceptance window) the fixation point in every trial (blue dots). However, Participant 2 in Figure 2B looked at the goal in some trials (gray dots). Based on the eye position results, we excluded the key response data in which the participant's eye position once moved more than 3° horizontally or vertically from the fixation point (area) during the target motion in a trial. For some participants, all 20 trials of key response data from a VR scene were excluded when they violated the fixation area in more than half the trials in one VR scene (n = 1 from the Upper- and Lower-Descend scenes, and n = 2 from the Upper- and Lower-Ascend scenes). The mean excluded the trials of key response data for each participant, except the participant whose data were excluded from more than half of all 20 trials in one VR scene, as follows: 1.27 trials (SD = ± 2.17) in the Upper-Ascend scene, 0.82 trials (SD = ± 1.84) in the Upper-Descend scene, 1.0 trial (SD = ± 1.66) in the Lower-Ascend scene, and 1.0 trial (SD = ± 1.76) in the Lower-Descend scene. The excluded key response data were not used for further analysis. 
Figure 2.
 
Eye position results and number of excluded data. The horizontal and vertical axes represent the horizontal and vertical eye positions (°), respectively. Red squares indicate the criteria for an acceptable fixation area. The blue and gray dots represent the eye position data points during the target motion in each trial. A total of 1,608 data points (670 ms × 120 Hz × 20 trials) are plotted in each figure. The blue dots represent acceptable eye positions during the target motion in each trial. The gray dots represent unacceptable eye positions, where the eye positions moved more than 3° vertically during the target motion. The circle in A indicates the goal position under the Lower-Ascend condition. (A) Horizontal and vertical eye positions of Participant 1. (B) Horizontal and vertical eye positions of Participant 2.
Figure 2.
 
Eye position results and number of excluded data. The horizontal and vertical axes represent the horizontal and vertical eye positions (°), respectively. Red squares indicate the criteria for an acceptable fixation area. The blue and gray dots represent the eye position data points during the target motion in each trial. A total of 1,608 data points (670 ms × 120 Hz × 20 trials) are plotted in each figure. The blue dots represent acceptable eye positions during the target motion in each trial. The gray dots represent unacceptable eye positions, where the eye positions moved more than 3° vertically during the target motion. The circle in A indicates the goal position under the Lower-Ascend condition. (A) Horizontal and vertical eye positions of Participant 1. (B) Horizontal and vertical eye positions of Participant 2.
Timing differences (TDs) were calculated for each key response to assess the accuracy of arrival time estimation for each VR scene. TD was defined as the difference between the actual timing of target arrival at the goal (670 ms) and the keypress time (Hirata et al., 2024). A positive TD value indicates that the key is pressed before the target arrives at the goal, whereas a negative TD value indicates that the key is pressed after the target reaches the goal. After calculating the TDs for all the trials, the mean TDs for each participant were calculated for each VR scene. Finally, we averaged the mean TDs of all the participants in each VR scene. A two-way analysis of variance (ANOVA) was performed to evaluate the differences in mean TDs with the target direction (ascending and descending) and goal location (upper and lower). Tukey's honest significant difference (HSD) test was performed as a post hoc test, and the level of statistical significance was set at p < 0.05 for all analyses. 
Results
Figure 3 shows the mean TDs for the Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend scenes. All mean TDs were positive, indicating that, on average, the key was pressed before the target reached the goal. A two-way ANOVA with target direction and goal location conditions showed a significant main effect of the target direction condition (F(1, 163) = 7.7, p < 0.01, η2 = 0.04) and a significant interaction between target direction and goal location conditions (F(1, 163) = 4.86, p < 0.01, η2 = 0.02). However, the main effect of goal location condition was not significant (F(1, 163) = 2.39, p = 0.1, η2 < 0.01). A post hoc analysis with Tukey's HSD correction indicated that the mean TD for the Upper-Ascend scene was significantly larger than that for the other scenes (Upper-Ascend vs. Upper-Descend, p < 0.01; Upper-Ascend vs. Lower-Ascend, p = 0.03; Upper-Ascend vs. Lower-Descend, p = 0.01), revealing that the arrival time estimation for the Upper-Ascend scene was inaccurate compared with the others. 
Figure 3.
 
The mean TDs for ascending and descending targets in the upper and lower visual fields. The vertical axis on the left side indicates the difference between the arrival times at the goal (670 ms) and keypress times. The right side of the vertical axis indicates the keypress time. The orange and blue bars represent the mean TDs for the ascending and descending targets, respectively. The orange and blue bars on the left side represent the upper visual conditions (Upper-Ascend and Upper-Descend scenes), and those on the right side represent the lower visual conditions (Lower-Ascend and Lower-Descend scenes). Error bars indicate standard error. Asterisks indicate the results of post hoc analysis using Tukey's HSD (*p < 0.05, **p < 0.01).
Figure 3.
 
The mean TDs for ascending and descending targets in the upper and lower visual fields. The vertical axis on the left side indicates the difference between the arrival times at the goal (670 ms) and keypress times. The right side of the vertical axis indicates the keypress time. The orange and blue bars represent the mean TDs for the ascending and descending targets, respectively. The orange and blue bars on the left side represent the upper visual conditions (Upper-Ascend and Upper-Descend scenes), and those on the right side represent the lower visual conditions (Lower-Ascend and Lower-Descend scenes). Error bars indicate standard error. Asterisks indicate the results of post hoc analysis using Tukey's HSD (*p < 0.05, **p < 0.01).
Figure 4 shows the mean TD results for each participant for ascending targets in the upper and lower visual fields (Upper-Ascend and Lower-Ascend scenes; Figure 4A) and descending targets in both visual fields (Upper-Descend and Lower-Descend scenes; Figure 4B). The orange (A) and blue (B) dots represent the mean TD for ascending and descending motions, respectively. The dashed diagonal line indicates equal performance of the arrival time estimation between the upper and lower visual fields. In the ascending motion (A), 70% of each mean TD for the upper visual field was larger than that for the lower visual field. This indicates that most participants estimated the arrival time for ascending targets more inaccurately in the upper visual field than in the lower visual field. In the descending motion (B), 54.7% of each mean TD for the upper visual field was larger than that for the lower visual field. Furthermore, each mean TD and the average of the mean TDs for each participant (green dots) are distributed around the diagonal line in Figure 4B, suggesting that the accuracy of arrival time estimation for descending targets is similar between the upper and lower visual fields. 
Figure 4.
 
Each participant's result of mean TD between ascending and descending targets in the upper and lower visual fields. The horizontal axes in A and B indicate the TD in the Lower-Ascend and Lower-Descend scenes, respectively. The vertical axes in A and B indicate the TD in the Upper-Ascend and the Upper-Descend scenes, respectively. The dashed diagonal line represents the equal performance of arrival time estimation between the upper and lower visual fields. (A) Individual mean TDs for ascending motion in the upper and lower visual fields. The orange dots represent the individual results of the mean TDs for ascending targets. (B) Individual mean TDs for descending motion in the upper and lower visual fields. The blue dots represent the individual results of the mean TDs for the descending targets. The green dots in A and B represent the average of the individual mean TDs.
Figure 4.
 
Each participant's result of mean TD between ascending and descending targets in the upper and lower visual fields. The horizontal axes in A and B indicate the TD in the Lower-Ascend and Lower-Descend scenes, respectively. The vertical axes in A and B indicate the TD in the Upper-Ascend and the Upper-Descend scenes, respectively. The dashed diagonal line represents the equal performance of arrival time estimation between the upper and lower visual fields. (A) Individual mean TDs for ascending motion in the upper and lower visual fields. The orange dots represent the individual results of the mean TDs for ascending targets. (B) Individual mean TDs for descending motion in the upper and lower visual fields. The blue dots represent the individual results of the mean TDs for the descending targets. The green dots in A and B represent the average of the individual mean TDs.
Discussion
The arrival time estimates were more accurate for the Lower-Descend scene than for the Upper-Ascend scenes. This asymmetry in the arrival time estimation accuracy for ascending and descending targets is consistent with the results of previous studies on arrival time estimation (Hirata et al., 2024) and interception performance (Senot et al., 2005; Zago et al., 2011) under various conditions. Importantly, the cause of asymmetry in the arrival time estimation for the Upper-Ascend and Lower-Descend scenes differs from those identified in previous studies and might be attributed to differences in visual perceptual accuracy between the upper and lower visual fields. A novel finding was that the accuracy of the arrival time estimation for the ascending and descending targets did not show a significant difference between the Lower-Ascend and Upper-Descend scenes. This result implies that visual perception in the upper and lower visual fields, as well as the direction of the target motion (ascending and descending), produces an asymmetry in the accuracy of position estimation for a moving object. The accuracy of arrival time estimation was impaired only when the target moved upward in the upper visual field. This result suggests that inaccurate perception of an ascending object in the upper visual field contributes to the well-documented asymmetry in position estimation accuracy between ascending and descending objects. 
Hirata et al. (2024) suggested that the different tracking performances of smooth pursuit eye movements induce asymmetry in the accuracy of position estimation for ascending and descending objects. Smooth pursuit eye movements are known for the accurate pursuit of a downward object rather than an upward object (Akao, Kumakura, Kurkin, Fukushima, & Fukushima, 2007; Hirata et al., 2024; Ke, Lam, Pai, & Spering, 2013). In previous studies, smooth pursuit eye movements occurred during position estimation for a visual target moving toward the lower or upper visual fields because participants were asked to track the target (Hirata et al., 2024) or were not inhibited from tracking it (Zago et al., 2011). Importantly, however, asymmetry in the position estimation accuracy for the Upper-Ascend and Lower-Descend scenes was observed in a situation without smooth pursuit eye movements in this study. Visual perceptual asymmetry in the upper and lower visual fields, as well as different performances of smooth pursuit eye movements, might independently produce asymmetric accuracy in position estimation. Thus, when observers do not track ascending and descending objects with smooth pursuit eye movements, the asymmetric performance in position estimation arises mainly from differences in visual perceptual accuracy between the upper and lower visual fields. 
In this study, the target always moved at the same acceleration (1 G) across all VR scenes. Consequently, the performance of arrival time estimation was potentially improved by learning the duration of the target motion (670 ms) across 20 trials within each VR scene. To clarify this possibility, the mean TD for each trial was calculated for each VR scene (Figure 5, gray dots). Moreover, we compared the mean TDs for the beginning (1st–7th), middle (8th–13th), and end (14th–20th trials) phases (Figure 5, blue bars). A one-way ANOVA with three phases (beginning, middle, and end) shows no significant differences in all VR scenes (Upper-Ascend: F(2, 17) = 2.52, p = 0.1, η2 = .22; Upper-Descend: F(2, 17) = 0.26, p = 0.77, η2 = 0.03; Lower-Ascend: F(2, 17) = 0.97, p = 0.39, η2 = 0.01; Lower-Descend: F(2, 17) = 1.95, p = 0.17, η2 = 0.08). These results suggested that the participants did not learn the duration of the target motion during the 20 trials for each VR scene. 
Figure 5.
 
Mean TD in each trial under each VR scene. The horizontal axes in all four figures represent the trial number (1st–20th). The left vertical axes in all four figures represent the mean TD for each trial (gray dots). The right vertical axes in all four figures indicate the mean TD for the beginning (1st–7th), middle (8th–13th), and end (14th–20th) phases (blue dots). (A) Upper-Ascend scene. (B) Upper-Descend scene. (C) Lower-Ascend scene. (D) Lower-Descend scene. Error bars indicate the standard error.
Figure 5.
 
Mean TD in each trial under each VR scene. The horizontal axes in all four figures represent the trial number (1st–20th). The left vertical axes in all four figures represent the mean TD for each trial (gray dots). The right vertical axes in all four figures indicate the mean TD for the beginning (1st–7th), middle (8th–13th), and end (14th–20th) phases (blue dots). (A) Upper-Ascend scene. (B) Upper-Descend scene. (C) Lower-Ascend scene. (D) Lower-Descend scene. Error bars indicate the standard error.
Previous studies have reported that foveal-petal motion was more accurately detected than foveal-fugal motion (Edwards & Badcock, 1993; Giaschi, Zwicker, Young, & Bjornson, 2007). In our study, the target motion could be classified into foveal-petal and foveal-fugal motion, suggesting that these motions might contribute to the accuracy of arrival time estimation. Indeed, the accuracy of arrival time estimation revealed asymmetry between the Upper-Descend (foveal-petal) and the Upper-Ascend (foveal-fugal) scenes. However, the accuracy of the arrival time estimation was similar between the foveal-petal (Lower-Ascend) and foveal-fugal (Lower-Descend) motions in the lower visual field. While previous studies (Edwards & Badcock, 1993; Giaschi et al., 2007) have evaluated the accuracy of motion detection, our study evaluated the arrival time estimation performance. Owing to the differences in visual tasks, the asymmetry in foveal-petal and foveal-fugal motion observed in previous studies might not be replicated in the context of arrival time estimation. 
This study presents the goal at 8.7° from the fixation point in the Upper-Descend and Lower-Ascend scenes and at 17.1° in the Upper-Ascend and Lower-Descend scenes. These two different goal locations could contribute to the accuracy of arrival time estimation. Previous studies have found that human visual spatial resolution is highest at the fovea and decreases drastically from the fovea to the peripheral visual field (Anton-Erxleben & Carrasco, 2013; Strasburger, Rentschler, & Jüttner, 2011). Therefore, the similar accuracy of arrival time estimation observed in the Upper-Descend and Lower-Ascend scenes can be attributed to the proximity of the goal location (8.7°) to the fovea. On the other hand, the asymmetric performance of the arrival time estimation for the Upper-Ascend and Lower-Descend scenes may be interpreted as a visual perceptual difference between the upper and lower visual fields. A previous study reported that visual perception was more accurate at 15° from the central visual field in the lower visual field than at the same eccentricity in the upper visual field (Barbot et al., 2021). Therefore, the goal position at 17.1° in the upper visual field (Upper-Ascend) might be perceived inaccurately compared to that in the lower visual field (Lower-Descend), inducing the asymmetric performance of arrival time estimation for the Upper-Ascend and Lower-Descend scenes. 
The biased distribution of ganglion cells between the ventral (upper visual field) and dorsal (lower visual field) sides of the retina is one of the factors inducing the perceptual asymmetry in the upper and lower visual fields. In humans, the density of ganglion cells on the ventral side of the retina is approximately 60% lower than that on the dorsal side (Curcio & Allen, 1990). However, this asymmetry emerges significantly beyond the peripheral visual field at approximately 30° or further (Curcio & Allen, 1990). Thus, a biased distribution of ganglion cells could not account for the asymmetry in arrival time estimation accuracy for the Upper-Ascend and Lower-Descend scenes because the final position of the target motion was either 8.7° or 17.1° in this study. 
Apart from the nonhomogeneous distribution of ganglion cells in the retina, processing in the middle temporal area (MT) and area V6, which are considered the primary brain regions for visual motion processing, may reflect the asymmetric accuracy of arrival time estimation. This is because motion-sensitive neurons in the macaque MT (Brewer, Press, Logothetis, & Wandell, 2002; Gattass & Gross, 1981; Van Essen, Newsome, & Maunsell, 1984) and area V6 (Galletti, Fattori, Gamberini, & Kutz, 1999; Pitzalis, Fattori, & Galletti, 2013) have more receptive fields in the lower visual field than in the upper one. The inhomogeneous distribution of these receptive fields suggests that a lower visual field can facilitate finer visual image analysis, resulting in a lower visual field with an advantage in the visual perception of moving objects. In other words, recognition of moving objects in the upper visual field is disadvantageous for visual motion perception. Importantly, in this study, even when descending targets were presented in the upper visual field (Upper-Descend), the accuracy of the arrival time estimation was not impaired. Consistent with this result, Ezzo, Winawer, Carrasco, and Rokers (2023) observed a similar accuracy for vertical motion detection in 7° of upper and lower visual fields. These results indicate that the superiority of visual motion processing is not limited to the lower visual field but extends at least to approximately 8.7° in the upper visual field. Nevertheless, as shown by inaccurate arrival time estimation in the Upper-Ascend scene, visual motion processing appears to be impaired somewhere between approximately 8.7° and 17.1° in the upper visual field, leading to inaccurate visual perception and position estimation of moving objects. Danckert and Goodale (2001) reported that pointing performance for horizontally moving objects was inaccurate at approximately 14° above the center of the visual field than at approximately 14° below the center. Taken together with Danckert and Goodale's study, the boundary would lie between 8.7° and 14° in the upper visual field. To clarify the specific boundaries of changes in visual motion processing performance in the upper visual field, future studies need to investigate the accuracy of position estimation for moving objects at different vertical locations in the upper visual field. 
A limitation of our study is that we evaluated the arrival time estimation accuracy for only one type of target motion (ascending or descending) at each goal position (8.7° and 17.1°) in the upper and lower visual fields. Due to this experimental design, our results cannot clearly determine whether the differences in the arrival time estimation accuracy are attributable to the goal position (eccentricity from the center) or the target motion (ascending and descending). It remains unclear whether the impaired arrival time estimation was specific to the ascending motion or affected both ascending and descending motions in the far upper visual field (17.1°). Future studies should evaluate the arrival time estimation accuracy for ascending and descending targets at the same goal position in the upper and lower visual fields. 
Our findings provide additional insight into the vertically asymmetric performance of acceleration perception. Phan, Jörges, Harris, and Kingdom (2024) reported that a downward motion requires more acceleration than an upward motion to be judged as moving at constant velocity. In that study, participants were required to fixate on a central point on a screen. Thus, the upward and downward visual stimuli moved toward the observer's upper and lower visual fields, respectively (Phan et al., 2024). Although Smith and Hammond (1986) found no significant difference in velocity perception between the upper and lower visual fields at 5°, the velocity perception may differ at a farther visual field (17.1°). In our study, the keypress timing was earlier in the Upper-Ascend scene than in other VR scenes, indicating that ascending targets in the upper visual field might be perceived faster than those in the lower visual field. However, it remains unclear whether this observation is consistent with previous findings on the acceleration bias for upward and downward motions. To better understand how visual perceptual asymmetry influences both acceleration and velocity perception, further research needs to estimate these parameters in the far peripheral (17.1°) upper and lower visual fields. 
Conclusions
This study evaluated the arrival time estimation for ascending and descending targets in the upper and lower visual fields separately. Based on previous findings, arrival time estimation for the descending target in the lower visual field (Lower-Descend) was more accurate than that for the ascending target in the upper visual field (Upper-Ascend). However, the accuracy of arrival time estimation was similar when the participants recognized ascending targets in the lower visual field (Lower-Ascend) and descending targets in the upper visual field (Upper-Descend). These results indicate that not only the direction of object motion (ascending and descending) but also the location of object recognition in the observer's visual field (upper and lower) contribute to position estimation performance. The arrival time estimation accuracy was evaluated for one type of target motion (ascending or descending) at each goal position (8.7° and 17.1°) in the upper and lower visual fields, respectively. Due to the limitations of the experimental design, it remains unclear whether an object ascending at a higher position in the upper visual field led to inaccurate position estimation or whether the direction of motion at the position was irrelevant to the accuracy. However, this study at least shows that the arrival time estimation accuracy of an object ascending at a higher position (17.1°) in the upper visual field is more impaired than that of an object descending at the same position (17.1°) in the lower visual field. This result suggests that the inaccurate position estimation for ascending objects in the upper visual field is one of the causes of the asymmetric performance of position estimation for ascending and descending objects. 
Acknowledgments
Supported by JSPS Research Fellow (JP22J1402) and JST CREST (Grant No. JPMJCR225P5). 
Commercial relationships: none. 
Corresponding authors: Takashi Hirata and Nobuyuki Kawai. 
Address: Nagoya University, Chikusaku, Furoucho, Nagoya, Aichi 464-8601, Japan. 
References
Akao, T., Kumakura, Y., Kurkin, S., Fukushima, J., & Fukushima, K. (2007). Directional asymmetry in vertical smooth-pursuit and cancellation of the vertical vestibulo-ocular reflex in juvenile monkeys. Experimental Brain Research, 182, 469–478. [CrossRef] [PubMed]
Anton-Erxleben, K., & Carrasco, M. (2013). Attentional enhancement of spatial resolution: Linking behavioural and neurophysiological evidence. Nature Reviews Neuroscience, 14(3), 188–200. [CrossRef] [PubMed]
Barbot, A., Xue, S., & Carrasco, M. (2021). Asymmetries in visual acuity around the visual field. Journal of Vision, 21(1), 2, https://doi.org/10.1167/jov.21.1.2. [PubMed]
Baurès, R., & Hecht, H. (2011). The effect of body posture on long-range time-to-contact estimation. Perception, 40(6), 674–681. [CrossRef] [PubMed]
Brenner, E., Driesen, B., & Smeets, J. B. (2014). Precise timing when hitting falling balls. Frontiers in Human Neuroscience, 8, 342. [CrossRef] [PubMed]
Brewer, A. A., Press, W. A., Logothetis, N. K., & Wandell, B. A. (2002). Visual areas in macaque cortex measured using functional magnetic resonance imaging. Journal of Neuroscience, 22(23), 10416–10426. [CrossRef]
Cameron, E. L., Tai, J. C., & Carrasco, M. (2002). Covert attention affects the psychometric function of contrast sensitivity. Vision Research, 42(8), 949–967. [CrossRef] [PubMed]
Carrasco, M., Williams, P. E., & Yeshurun, Y. (2002). Covert attention increases spatial resolution with or without masks: Support for signal enhancement. Journal of Vision, 2, 467, https://doi.org/10.1167/2.6.4. [PubMed]
Corbett, J. E., & Carrasco, M. (2011). Visual performance fields: Fames of reference. PLoS One, 6(9), e24470. [CrossRef] [PubMed]
Curcio, C. A., & Allen, K. A. (1990). Topography of ganglion cells in human retina. Journal of Comparative Neurology, 300(1), 5–25. [CrossRef]
Danckert, J., & Goodale, M. A. (2001). Superior performance for visually guided pointing in the lower visual field. Experimental Brain Research, 137, 303–308. [PubMed]
Edwards, M., & Badcock, D. R. (1993). Asymmetries in the sensitivity to motion in depth: A centripetal bias. Perception, 22(9), 1013–1023. [CrossRef] [PubMed]
Ezzo, R., Winawer, J., Carrasco, M., & Rokers, B. (2023). Asymmetries in the discrimination of motion direction around the visual field. Journal of Vision, 23(3), 1–19, https://doi.org/10.1167/jov.23.3.19.
Gallagher, M., Torok, A., Klaas, J., & Ferrè, E. R. (2020). Gravity prior in human behaviour: A perceptual or semantic phenomenon? Experimental Brain Research, 238(9), 1957–1962. [CrossRef] [PubMed]
Galletti, C., Fattori, P., Gamberini, M., & Kutz, D. F. (1999). The cortical visual area V6: Brain location and visual topography. European Journal of Neuroscience, 11(11), 3922–3936. [CrossRef]
Gattass, R., & Gross, C. G. (1981). Visual topography of striate projection zone (MT) in posterior superior temporal sulcus of the macaque. Journal of Neurophysiology, 46(3), 621–638. [CrossRef] [PubMed]
Giaschi, D., Zwicker, A., Young, S. A., & Bjornson, B. (2007). The role of cortical area V5/MT+ in speed-tuned directional anisotropies in global motion perception. Vision Research, 47(7), 887–898. [CrossRef] [PubMed]
He, S., Cavanagh, P., & Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. Nature, 383(6598), 334–337. [CrossRef] [PubMed]
Hirata, T., Hirata, Y., & Kawai, N. (2024). Human estimates of descending object's motion are more accurate than those of ascending objects regardless of gravity information. Journal of Vision, 24(3), 1–15, https://doi.org/10.1167/jov.24.3.2.
Indovina, I., Maffei, V., Bosco, G., Zago, M., Macaluso, E., & Lacquaniti, F. (2005). Representation of visual gravitational motion in the human vestibular cortex. Science, 308(5720), 416–419. [CrossRef] [PubMed]
Ke, S. R., Lam, J., Pai, D. K., & Spering, M. (2013). Directional asymmetries in human smooth pursuit eye movements. Investigative Ophthalmology & Visual Science, 54(6), 4409–4421. [PubMed]
Kupers, E. R., Benson, N. C., Carrasco, M., & Winawer, J. (2022). Asymmetries around the visual field: From retina to cortex to behavior. PLoS Computational Biology, 18(1), e1009771. [PubMed]
Le Séac'h, A. B. H., Senot, P., & McIntyre, J. (2010). Egocentric and allocentric reference frames for catching a falling object. Experimental Brain Research, 201, 653–662. [PubMed]
Levine, M. W., & McAnany, J. J. (2005). The relative capabilities of the upper and lower visual hemifields. Vision Research, 45(21), 2820–2830. [PubMed]
Miller, W. L., Maffei, V., Bosco, G., Iosa, M., Zago, M., Macaluso, E., … Lacquaniti, F. (2008). Vestibular nuclei and cerebellum put visual gravitational motion in context. Journal of Neurophysiology, 99(4), 1969–1982. [PubMed]
Phan, M. H., Jörges, B., Harris, L. R., & Kingdom, F. A. (2024). A visual bias for falling objects. Perception, 53(3), 197–207. [PubMed]
Pitzalis, S., Fattori, P., & Galletti, C. (2013). The functional role of the medial motion area V6. Frontiers in Behavioral Neuroscience, 6, 91. [PubMed]
Senot, P., Zago, M., Lacquaniti, F., & McIntyre, J. (2005). Anticipating the effect of gravity when intercepting moving objects: Differentiating up and down based on nonvisual cues. Journal of Neurophysiology, 94(6), 4471–4480. [PubMed]
Senot, P., Zago, M., Le Séac'h, A., Zaoui, M., Berthoz, A., Lacquaniti, F., … McIntyre, J. (2012). When up is down in 0g: how gravity sensing affects the timing of interceptive action. Journal of Neuroscience, 32(6), 1969–1973.
Schütz, A. C., Braun, D. I., & Gegenfurtner, K. R. (2009). Object recognition during foveating eye movements. Vision Research, 49(18), 2241–2253. [PubMed]
Smith, A. T., & Hammond, P. (1986). Hemifield differences in perceived velocity. Perception, 15(2), 111–117. [PubMed]
Strasburger, H., Rentschler, I., & Jüttner, M. (2011). Peripheral vision and pattern recognition: A review. Journal of Vision, 11(5), 1–13, https://doi.org/10.1167/11.5.13.
Van Essen, D. C., Newsome, W. T., & Maunsell, J. H. (1984). The visual field representation in striate cortex of the macaque monkey: asymmetries, anisotropies, and individual variability. Vision Research, 24(5), 429–448. [PubMed]
Zago, M., La Scaleia, B., Miller, W. L., & Lacquaniti, F. (2011). Coherence of structural visual cues and pictorial gravity paves the way for interceptive actions. Journal of Vison, 11(10), 1–13, https://doi.org/10.1167/11.10.13.
Zito, G. A., Cazzoli, D., Müri, R. M., Mosimann, U. P., & Nef, T. (2016). Behavioral differences in the upper and lower visual hemifields in shape and motion perception. Frontiers in Behavioral Neuroscience, 10, 128. [PubMed]
Figure 1.
 
Experimental design. (A) Experimental setup. All participants sat in the chair and put their chins on the chinrest. An HMD was used to show the visual stimulus. Participants placed the index finger of their dominant hand on the keypad. (B) Virtual reality scenes depicted the following movements from the left side of the view: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend. The magenta lines indicate the direction of the ball's target motion in each VR scene. The red dot was the fixation point. (C) Experimental schedule and presentation order.
Figure 1.
 
Experimental design. (A) Experimental setup. All participants sat in the chair and put their chins on the chinrest. An HMD was used to show the visual stimulus. Participants placed the index finger of their dominant hand on the keypad. (B) Virtual reality scenes depicted the following movements from the left side of the view: Upper-Ascend, Upper-Descend, Lower-Ascend, and Lower-Descend. The magenta lines indicate the direction of the ball's target motion in each VR scene. The red dot was the fixation point. (C) Experimental schedule and presentation order.
Figure 2.
 
Eye position results and number of excluded data. The horizontal and vertical axes represent the horizontal and vertical eye positions (°), respectively. Red squares indicate the criteria for an acceptable fixation area. The blue and gray dots represent the eye position data points during the target motion in each trial. A total of 1,608 data points (670 ms × 120 Hz × 20 trials) are plotted in each figure. The blue dots represent acceptable eye positions during the target motion in each trial. The gray dots represent unacceptable eye positions, where the eye positions moved more than 3° vertically during the target motion. The circle in A indicates the goal position under the Lower-Ascend condition. (A) Horizontal and vertical eye positions of Participant 1. (B) Horizontal and vertical eye positions of Participant 2.
Figure 2.
 
Eye position results and number of excluded data. The horizontal and vertical axes represent the horizontal and vertical eye positions (°), respectively. Red squares indicate the criteria for an acceptable fixation area. The blue and gray dots represent the eye position data points during the target motion in each trial. A total of 1,608 data points (670 ms × 120 Hz × 20 trials) are plotted in each figure. The blue dots represent acceptable eye positions during the target motion in each trial. The gray dots represent unacceptable eye positions, where the eye positions moved more than 3° vertically during the target motion. The circle in A indicates the goal position under the Lower-Ascend condition. (A) Horizontal and vertical eye positions of Participant 1. (B) Horizontal and vertical eye positions of Participant 2.
Figure 3.
 
The mean TDs for ascending and descending targets in the upper and lower visual fields. The vertical axis on the left side indicates the difference between the arrival times at the goal (670 ms) and keypress times. The right side of the vertical axis indicates the keypress time. The orange and blue bars represent the mean TDs for the ascending and descending targets, respectively. The orange and blue bars on the left side represent the upper visual conditions (Upper-Ascend and Upper-Descend scenes), and those on the right side represent the lower visual conditions (Lower-Ascend and Lower-Descend scenes). Error bars indicate standard error. Asterisks indicate the results of post hoc analysis using Tukey's HSD (*p < 0.05, **p < 0.01).
Figure 3.
 
The mean TDs for ascending and descending targets in the upper and lower visual fields. The vertical axis on the left side indicates the difference between the arrival times at the goal (670 ms) and keypress times. The right side of the vertical axis indicates the keypress time. The orange and blue bars represent the mean TDs for the ascending and descending targets, respectively. The orange and blue bars on the left side represent the upper visual conditions (Upper-Ascend and Upper-Descend scenes), and those on the right side represent the lower visual conditions (Lower-Ascend and Lower-Descend scenes). Error bars indicate standard error. Asterisks indicate the results of post hoc analysis using Tukey's HSD (*p < 0.05, **p < 0.01).
Figure 4.
 
Each participant's result of mean TD between ascending and descending targets in the upper and lower visual fields. The horizontal axes in A and B indicate the TD in the Lower-Ascend and Lower-Descend scenes, respectively. The vertical axes in A and B indicate the TD in the Upper-Ascend and the Upper-Descend scenes, respectively. The dashed diagonal line represents the equal performance of arrival time estimation between the upper and lower visual fields. (A) Individual mean TDs for ascending motion in the upper and lower visual fields. The orange dots represent the individual results of the mean TDs for ascending targets. (B) Individual mean TDs for descending motion in the upper and lower visual fields. The blue dots represent the individual results of the mean TDs for the descending targets. The green dots in A and B represent the average of the individual mean TDs.
Figure 4.
 
Each participant's result of mean TD between ascending and descending targets in the upper and lower visual fields. The horizontal axes in A and B indicate the TD in the Lower-Ascend and Lower-Descend scenes, respectively. The vertical axes in A and B indicate the TD in the Upper-Ascend and the Upper-Descend scenes, respectively. The dashed diagonal line represents the equal performance of arrival time estimation between the upper and lower visual fields. (A) Individual mean TDs for ascending motion in the upper and lower visual fields. The orange dots represent the individual results of the mean TDs for ascending targets. (B) Individual mean TDs for descending motion in the upper and lower visual fields. The blue dots represent the individual results of the mean TDs for the descending targets. The green dots in A and B represent the average of the individual mean TDs.
Figure 5.
 
Mean TD in each trial under each VR scene. The horizontal axes in all four figures represent the trial number (1st–20th). The left vertical axes in all four figures represent the mean TD for each trial (gray dots). The right vertical axes in all four figures indicate the mean TD for the beginning (1st–7th), middle (8th–13th), and end (14th–20th) phases (blue dots). (A) Upper-Ascend scene. (B) Upper-Descend scene. (C) Lower-Ascend scene. (D) Lower-Descend scene. Error bars indicate the standard error.
Figure 5.
 
Mean TD in each trial under each VR scene. The horizontal axes in all four figures represent the trial number (1st–20th). The left vertical axes in all four figures represent the mean TD for each trial (gray dots). The right vertical axes in all four figures indicate the mean TD for the beginning (1st–7th), middle (8th–13th), and end (14th–20th) phases (blue dots). (A) Upper-Ascend scene. (B) Upper-Descend scene. (C) Lower-Ascend scene. (D) Lower-Descend scene. Error bars indicate the standard error.
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×