- Email: [email protected]
Inter-ocular and intra-ocular integration during prehension
Neuroscience Letters 487 (2011) 17–21
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Inter-ocular and intra-ocular integration during prehension Steve Hansen a , Spencer Hayes b , Simon J. Bennett b,∗ a b
Brock University, CANADA Liverpool John Moores University, UK
a r t i c l e
i n f o
Article history: Received 17 February 2010 Received in revised form 24 August 2010 Accepted 22 September 2010 Keywords: Prehension Visual control Binocular Monocular Alternating
a b s t r a c t This study examined the inter-ocular (alternating monocular samples) and intra-ocular (monocular or binocular samples) integration during a prehensile task with a range of occlusion intervals (0–75 ms). In the first experiment, participants were uncertain regarding the impending visual condition, as well as target size and location. In the second experiment, a pre-cue on target location was provided. Data from both experiments indicated that participants modified their movement kinematics when provided with alternating monocular samples, irrespective of whether or not there was an occlusion interval. Similar adaptations were found in conditions requiring intra-ocular integration but only following the introduction of an occlusion interval. These findings are consistent with participants having a general intolerance for alternating monocular samples and as a consequence using a more cautious reach and grasp strategy. Crown Copyright © 2010 Published by Elsevier Ireland Ltd. All rights reserved.
Briefly presented visual samples can be integrated over time when the intra-ocular occlusion period is shorter than the intrinsic persistence of the visual samples [5]. In motor tasks demanding high precision (i.e., manual aiming, one-handed catching), the persistence of binocular samples (20 ms duration) enables performance to be maintained with intra-ocular occlusion periods of no longer than approximately 40–80 ms [4,7]. On the other hand, monocular samples (20 ms duration) seem to have a shorter persistence of approximately 20 ms [2,14]. The implication is that the control of tasks involving high precision is not only dependent on the duration of the intra-ocular occlusion interval, but also whether the visual input is provided to one or both eyes. The finding of a difference in the persistence of binocular and monocular vision that facilitates performance of precision tasks has led to recent attempts to determine the duration over which the monocular input can be integrated between the eyes. In an experiment [14] based on the method developed by [8], participants made more one-handed ball catches when provided with 20 ms or 80 ms alternating monocular samples than in a condition of continuous monocular vision. It was suggested that individuals were capable of gaining useful information by integrating alternating monocular samples presented without an inter-ocular occlusion as long as the time between alternating samples to the same eye did not exceed 80 ms. Contradictory evidence was obtained with a variation of this methodology that included an inter-ocular occlusion between the alternating monocular samples [3,2]. In those experiments, the pro-
∗ Corresponding author at: Brain and Behaviour Research Group, Liverpool John Moores University, Byrom Street, Liverpool L3 3AF, UK. E-mail address: [email protected] (S.J. Bennett).
portion of balls caught in monocular and alternating monocular conditions did not differ regardless of the occlusion interval and was consistently worse compared to corresponding conditions of binocular vision. To better determine the temporal integration limits of binocular vision, a recent study investigated the integration of alternating monocular samples during prehension [18]. Providing participants with continuous binocular vision resulted in smaller average grip apertures than a continuous monocular vision condition, as well as conditions in which alternating monocular samples were separated by an inter-ocular occlusion as short as 14 ms. This result was replicated in a second experiment where visual feedback of the hand was occluded for the initial 80% of the movement. The authors suggested that the larger grip aperture in the alternating monocular vision conditions, which would have been “largely programmed before the reach begins” (p. 96), could not be accounted for by participants having difficulties integrating highly dissimilar retinal images as the hand approached the target. Moreover, it was concluded that the participants adopted a cautious reach and grasp strategy as a consequence of having no tolerance for the integration of alternating monocular samples to provide a binocular percept. In the above study, alternating monocular conditions had interocular occlusion that ranged from 14–58 ms. It remains unclear, therefore, if participants could have integrated alternate monocular samples presented consecutively (i.e., without an inter-ocular occlusion) [14]. Furthermore, to date there has been no detailed report on measures of movement kinematics as the hand is transported from the home position to the target [12,13]. For instance, it is unknown whether participants increase movement time and change the proportion of time spent in the accelerative or decelerative phases of the movement following the introduction of an
0304-3940/$ – see front matter. Crown Copyright © 2010 Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2010.09.065
18
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
inter-ocular occlusion. This omission is important because binocular information from retinal disparities has been shown to be particularly important for the latter on-line control of prehensile movement [13]. Nine males (right hand, right eye dominant) from the host University provided informed consent prior to participating in this experiment. All procedures were conducted in accordance with the local ethical guidelines and the 1964 declaration of Helsinki. Participants were required to perform a prehensile task that involved reaching out to “grasp” 2D square targets located in their midsaggital plane. Specifically, they were instructed to place the index finger and thumb on the edges of the targets in the same way they would when grasping a real 3D square target. A visual stimulus generator (ViSaGe) with proprietary software (CRS Toolbox) operating in MATLAB (Mathworks Inc.) presented the targets on a 21-in. computer monitor (refresh rate of 160 Hz). The monitor was mounted horizontally in a wooden frame and had a clear Perspex overlay mounted 5 mm above the screen surface. White target squares were presented against a black background and contained a thin black cross (2 mm thick) that bisected their width. Either a small (24 mm × 24 mm) or a large (36 mm × 36 mm) square was presented at a near (285 mm) or far (385 mm) distance. The home position was located approximately 200 mm from the participant. Participants wore a pair of PLATO liquid crystal goggles throughout the protocol. The state of the liquid crystal lenses was manipulated to provide binocular, monocular or alternating monocular vision during the grasping movements. In the binocular condition, both lenses were transparent or opaque at the same time, whereas in the monocular conditions only the left lens was cycled between transparent and opaque states. In the alternating monocular conditions, the left and right lenses were alternately switched between transparent and opaque states such that vision was only available to one eye at a time. When vision was continuous there was no occlusion interval (0 ms) between 25 ms samples presented to both eyes (binocular condition), the left eye alone (monocular condition), or alternately between the right and left eyes (alternating monocular). For all other intermittent vision conditions, the state of the liquid crystal goggles was cycled such that participants were provided with a 25 ms sample (both eyes, left eye, alternate left or right eye) followed by a 12.5, 25, 50 or 75 ms occlusion interval: note that taking account of minimal delays inherent from switching states of the liquid crystal lenses (i.e., TranslucentTechnologies, Inc. technical report), it can be estimated that the opaque state was approximately 3 ms shorter than the driving signal. Participants completed 300 pseudo-randomly ordered trials (15 vision conditions, 2 target location, 2 target size, 5 repeats), which were separated into five blocks of 60. Each block comprised 15 trials to each of the two target locations and two target sizes. A trial began with the participant placing their finger and thumb together at the home position. The goggles were then switched opaque for 500 ms after which a target square appeared and simultaneously the goggles cycled between opaque and transparent states. Participants were asked to perform the prehensile task as quickly and as accurately as possible. Movement of markers attached to the distal end of the index finger, thumb, and radial-carpal joint was recorded at 200 Hz for the duration of the trial with a Qualysis ProReflex optoelectronic system; system accuracy has been measured at 0.5 mm [15]. Post experimentation, the resulting three-dimensional position data were filtered using a second order dual pass Butterworth filter with a low-pass cut-off frequency of 8 Hz [11]. The filtered position data were differentiated to acquire velocity data. From these data, we extracted reaction time, time after peak velocity, maximum grip aperture and movement time [11].
Fig. 1. Movement time (MT) and time after peak velocity (TAPV) in Experiment 1 as a function of vision condition (BINO: continuous binocular, Alt: alternating, 0, 12.5, 25, 50, 75, MONO: continuous monocular). Standard error of the mean indicated.
Inter-ocular integration was examined using separate one-way repeated measures ANOVA that compared the alternating monocular vision conditions (0, 12.5, 25, 50, 75 ms occlusion) to the continuous binocular and monocular vision conditions; see [14,18]. Intra-ocular integration was examined with separate 2-Vision Condition (binocular, monocular) by 5-Occlusion Interval (0, 12.5, 25, 50, 75 ms) ANOVA with repeated measures on both factors. Significant main effects and interactions were decomposed using Fisher LSD (P < 0.05). The different levels of target size (small, large) and location (near, far) were included in the experimental design in order to minimise advance planning. However, these factors did not interact with vision condition and hence were collapsed [see also 18]. Inter-ocular integration. There were no main effects for reaction time, F(6, 48) = 0.70, indicating that the amount of time spent planning the response was unaffected by vision condition and/or occlusion interval (grand mean = 256 ms). For maximum grip aperture, the effect of vision condition approached conventional levels of significance, F(6, 48) = 2.10, P < 0.07. Observation of the group means indicated that compared to continuous alternating vision there was a small increase in maximum grip aperture when the occlusion interval was longer than 12.5 ms (0 ms = 58.6 mm; 12.5 ms = 59.1 mm; 25 ms = 59.5 mm; 50 ms = 59.1 mm; 75 ms = 59.1 mm). Movement time, F(6, 48) = 5.38, P < 0.05, and the time after peak velocity, F(6, 48) = 3.66, P < 0.05, were significantly longer in all the alternating monocular conditions than in the continuous binocular or monocular vision condition (Fig. 1) Intra-ocular integration. There were no main effects for reaction time. However, compared to continuous binocular and monocular vision conditions, participants exhibited a significant increase in movement time, F(4, 32) = 5.27, P < 0.05, and time after peak velocity, F(4, 32) = 4.90, P < 0.05, when there was a 12.5 ms occlusion interval (Fig. 2). This difference was maintained across all occlusion intervals but there was not always a further significant increase. In addition, there was a significant interaction for maximum grip aperture, F(4, 32) = 3.57, P < 0.05. The introduction of an intra-ocular occlusion interval in the monocular vision condition resulted in a small but significant increase in maximum grip aperture (0 ms = 58.1 mm; 12.5 ms = 59.3 mm; 25 ms = 59.0 mm; 50 ms = 59.4 mm; 75 ms = 59.1 mm). For the comparison to continuous binocular vision, maximum grip aperture
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
19
Fig. 2. Movement time (MT) and time after peak velocity (TAPV) in Experiment 1 as a function of vision condition (MONO: monocular, BINO: binocular) and occlusion interval (0, 12.5, 25, 50, 75 ms). Standard error of the mean indicated.
increased significantly but only when there was a 75 ms occlusion interval (0 ms = 57.8 mm; 12.5 ms = 58.0 mm; 25 ms = 58.4 mm; 50 ms = 58.5 mm; 75 ms = 59.4 mm). Participants were unable to maintain their normal movement kinematics (i.e., continuous binocular vision) when visual samples were presented alternately between the eyes with a short occlusion interval (i.e., 12.5 ms). More importantly, similar changes to movement kinematics were evident when there was no inter-ocular occlusion. Specifically, participants lengthened the deceleration phase of the movement as the hand approached the target, and as consequence took longer to complete the prehensile action. This adaptation to the timing of movement control was concomitant with a small increase in grip aperture and is consistent with individuals being unable to integrate the disparate retinal inputs from alternate monocular samples [13]. A similar adaptation to the movement kinematics was evident when an intra-ocular occlusion interval of 12.5 ms was introduced in the binocular and monocular vision conditions. Compared to research on one-handed catching [2,3] these data suggest that the intra-ocular persistence of binocular and/or monocular information for a prehensile action upon a static target is of a shorter duration. It is noteworthy, however, that the kinematic measures reported here are more sensitive to perturbations of the available information than the outcome measures used in one-handed catching. Further work is required to better determine task differences in informational persistence. Contrary to previous work in which the visual condition was known after completing the first trial within each block [3,18], participants in our first experiment were uncertain regarding the impending visual condition as well as the target location and size. In combination with the fact that participants were instructed to respond as quickly and accurately as possible to the appearance of the target, this uncertainty could have influenced the manner in which movements were controlled [12,17]. For instance, participants could have been biased towards responding rapidly after
target appearance (i.e., shorter duration planning) and then use vision to modify the reach and grasp if necessary. To provide reliable information regarding target location and potentially reduce the reliance on depth information available during the trial, we repeated the first protocol but with the addition of a pre-cue. Eight males from the host university provided informed consent prior to completing the protocol. The group had not participated in the first experiment and were all right hand and right eye dominant. The apparatus and protocol were the same as in the initial procedure with the exception that a pre-cue (1 mm diameter white circle) was presented for 500 ms at the target location while both lenses were clear. After this foreperiod, the target square was presented and the goggles began cycling between the transparent and opaque states. Inter-ocular integration. There was a significant main effect for RT, F(6, 42) = 5.59, which was a result of a longer RT in the continuous binocular vision condition compared to all other conditions. Movement time, F(6, 42) = 2.90, P < 0.05, and time after peak velocity, F(6, 42) = 3.43, P < 0.05, were significantly longer in the alternating monocular vision conditions than the continuous binocular vision condition. There were no differences in the duration of movement time and time after peak velocity between the continuous monocular and binocular conditions (Fig. 3). Maximum grip aperture in all of the alternating monocular vision conditions (0 ms = 61.3 mm; 12.5 ms = 61.7 mm; 25 ms = 61.4 mm; 50 ms = 61.1 mm; 75 ms = 61.3 mm) was significantly larger, F(6, 42) = 5.10, P < 0.05, than that exhibited in the continuous binocular vision condition (59.9 mm); the larger grip aperture in Experiment 2 compared to Experiment 1 was most likely due to anatomical differences between the groups Intra-ocular integration. One again, RT in the continuous binocular vision condition was significantly longer than RT in all other conditions, F(4,28) = 6.04, P < 0.05. Moreover, compared to when both binocular and monocular vision was continuous, movement time, F(4, 28) = 5.77, P < 0.05, and time after peak velocity, F(4,
20
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
Fig. 3. Movement time (MT) and time after peak velocity (TAPV) in Experiment 2 as a function of vision condition (BINO: continuous binocular, Alt: alternating 0, 12.5, 25, 50, 75 ms, MONO: continuous monocular). Standard error of the mean indicated.
28) = 7.31, P < 0.05,were longer, and maximum grip aperture, F(4, 28) = 3.52, P < 0.05, was larger, when visual samples were separated by a 25 ms occlusion interval (578 ms, 359 ms, 61.0 mm) compared to when vision was continuous (558 ms, 337 ms, 60.1 mm). There was no further change in these dependent variables as the occlusion interval lengthened and no interaction between vision conditions Consistent with previous research [18], we found in both experiments reported here that participants modified the kinematics of their normal prehensile movement when a 12.5 ms inter-ocular occlusion was introduced between alternating monocular samples. In addition, we found that the same adaptation to the movement kinematics was made when alternating monocular samples were presented without an inter-ocular occlusion. Specifically, there was an increase in maximum grip aperture and a lengthening of movement time and deceleration time in the alternating monocular vision conditions compared to the continuous binocular and monocular vision conditions. Providing participants with advance information on target location did not ameliorate these effects. Importantly, we also showed that these adapted movement kinematics were not simply a consequence of participants planning for the worse-case scenario prior to trial onset (e.g., alternating monocular samples separated by a 75 ms occlusion interval). For instance, in the second experiment participants did not know the impending visual condition prior to trial onset but still they exhibited different movement kinematics when provided with continuous binocular or monocular vision. The implication is that participants took advantage of the available information on a trialby-trial basis in order to more effectively plan and control their prehensile movement. Notably, with the exception of the continuous binocular vision condition in the second experiment, there was no evidence of a change in reaction time. Thus, it would appear that the lengthening of TAPV and MT were not an adaptive response to shorter duration planning phase. In combination, the results of the current study confirm the suggestion that participants have a general intolerance to alternating monocular samples in prehension. In both experiments of the current study, there was a more gradual change in movement kinematics as the intra-ocular occlusion interval in conditions of binocular and monocular vision was lengthened, hence indicating some tolerance for intra-ocular integration [2,7]. The finding of differential effects in these visual conditions concurs with the suggestion from perceptual studies [8,16] that inter-ocular integration of disparate monocular inputs between the left and right eyes does not entail the same underlying
processing as that required in conditions of intermittent binocular or monocular vision, where the visual samples only need be integrated over time (i.e., intra-ocular integration). It is relevant to consider, therefore, how the intolerance for alternating monocular samples during prehension might be explained. Given that there was no evidence of a binocular advantage in the continuous binocular vision condition compared to the continuous monocular vision condition, it could be suggested that participants did not use binocular information and instead relied on monocular information alone. It would follow, therefore, that the adaptation to movement kinematics in the alternating monocular conditions was not a direct consequence of being unable to integrate disparate monocular inputs to gain a binocular percept. It is far more likely, however, that participants did use binocular information when available [18] and this resulted in similar movement kinematics to those exhibited when using monocular information alone. Indeed, in separate tests we have found that participants discriminate target squares embedded in a randomdot stereogram from continuous alternating monocular samples, and hence perceive depth from disparate monocular inputs. Accepting that binocular information would have been used if perceived, one possibility is that intolerance to alternating monocular vision was a result of the 25 ms sample duration limiting the temporal integration of luminance, which precluded the cross-correlation of disparate retinal images and the perception of stereomotion [9,10]. In this respect, it is notable that similar effects (on grasp aperture) as those reported in the current study were found when alternating monocular samples were presented for 50 ms [18]; this should have been sufficient to permit temporal integration of luminance. Moreover, the duration of the visual samples was long enough to perceive a static depth cue from alternating monocular samples [16], which should have proved useful for movement control. For instance, the perception of form in depth can be maintained from 10 ms alternating monocular samples if they are separated by an inter-ocular delay of no more than 50–70 ms. If the duration of the visual samples was not the limiting factor, it remains a possibility that the dissimilarity between alternating monocular samples disrupted the perception of coarse binocular disparity that provides information on the closing gap as the hand approaches the target; for more a detailed explanation of disparity processing see [13]. It is known from work using dichoptic displays that independent image motion rather than fused image motion tends to be perceived as images become increasingly dissimilar [1]. This is consistent with the suggestion that stimuli (e.g., alternate monocular samples) that are related to each other are more likely to be combined than disjoint stimuli [6]. However, while a disruption to processing of coarse binocular disparity could explain the differences found for the kinematic measures associated with on-line control in the current study, it does not follow that alternating monocular samples should have disrupted the perception of fine binocular disparity that provides size information from the static target, which can be used to scale grip aperture prior to reach onset [18]. We suggest, therefore, that a more parsimonious explanation for the effects on movement kinematics observed in conditions of alternating monocular vision is that participants recognised prior to initiating the reach that they were receiving alternating monocular samples, which they understood (from previous trials) disrupted the perception of motion in depth. Consequently, this resulted in a strategic increase in grip aperture as well as changes in temporal aspects of the movement. This interpretation is consistent with the suggestion that a wider grip aperture is used to provide a greater margin for error when vision is perturbed [19] and not that it represents an adaptation to a misperception of target size per se. Indeed, if participants had imperfectly superimposed the two laterally displaced alternate
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
monocular images, it would be expected that the resulting target image was perceived to be slightly larger than its true size. In this case, grip aperture would be increased but there should be no modification to variables associated with on-line control. Our experiments confirmed and extended upon the finding that participants are generally intolerant to alternating monocular samples for the control of prehension. We suggest that the changes in movement kinematics are consistent with the breakdown of a motion in depth signal that is particularly important for controlling the closure of the gap between the hand and target during the latter stage of the movement. References [1] T.J. Andrews, C. Blakemore, Integration of motion information during binocular rivalry, Vis. Res. 42 (2002) 301–309. [2] S.J. Bennett, D. Ashford, D. Elliott, Intermittent vision and one-handed catching: the temporal limits of binocular and monocular integration, Motor Control 7 (2003) 378–387. [3] S.J. Bennett, D. Ashford, N. Rioja, J. Coull, D. Elliott, Integration of intermittent visual samples over time and between the eyes, J. Mot. Behav. 38 (2006) 439–450. [4] S.J. Bennett, D. Elliott, D.J. Weeks, D. Keil, The effects of intermittent on prehension under binocular and monocular viewing, Motor Control 7 (2003) 46–56. [5] M. Coltheart, Iconic memory and visible persistence, Percept. Psychophys. 27 (1980) 1959–1975. [6] P. Dixon, V. Di Lollo, Beyond visible persistence: an alternative account of temporal integration and segregation in visual processing, Cognit. Psychol. 26 (1994) 33–63.
21
[7] D. Elliott, R. Chua, B.J. Pollock, The influence of intermittent vision on manual aiming, Acta Psychol. 85 (1994) 1–13. [8] G.R. Engel, An investigation of visual responses to brief stereoscopic stimuli, Q. J. Exp. Psychol. 22 (1970) 148–166. [9] E. Gheorghiu, C.J. Erkelens, Differences in perceived depth for temporally correlated and uncorrelated dynamic random-dot stereograms, Vis. Res. 45 (2005) 1603–1614. [10] E. Gheorghiu, C.J. Erkelens, Temporal properties of disparity processing revealed by random-dot stereograms, Perception 34 (2005) 1205–1219. [11] S. Hansen, D. Elliott, M.A. Khan, Comparing derived and acquired acceleration profiles: three dimensional optical electronic data analyses, Behav. Res. Methods 39 (2007) 748–754. [12] S. Hansen, C.M. Glazebrook, J.G. Anson, D.J. Weeks, D. Elliott, The influence of advance information about target location and visual feedback on movement planning and execution, Can. J. Exp. Psychol. 60 (2006) 200–208. [13] D.R. Melmoth, M. Storoni, G. Todd, A.L. Finlay, S. Grant, Dissociation between vergence and binocular disparity cues in the control of prehension, Exp. Brain Res. 183 (2007) 283–298. [14] I. Olivier, D.J. Weeks, J.L. Lyons, K.L. Ricker, D. Elliott, Monocular and binocular vision in one-handed ball catching: interocular integration, J. Mot. Behav. 30 (1998) 343–351. [15] J.G. Richards, The measurement of human motion: a comparison of commercially available systems, Hum. Mov. Sci. 18 (1999) 509–602. [16] J. Ross, J.H. Hogben, Short term memory in stereopsis, Vis. Res. 14 (1974) 1195–1201. [17] E.J. Schlicht, P.R. Schrater, Effects of visual uncertainty on grasping movements, Exp. Brain Res. 182 (2007) 47–57. [18] K.R. Wilson, P.M. Pearson, H.E. Matheson, J.J. Marotta, Temporal integration limits of stereovision in reaching and grasping, Exp. Brain Res. 189 (2008) 91–98. [19] A.M. Wing, C. Fraser, The contribution of the thumb to reaching movements, Q. J. Exp. Psychol. 35A (1983) 297–309.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Inter-ocular and intra-ocular integration during prehension Steve Hansen a , Spencer Hayes b , Simon J. Bennett b,∗ a b
Brock University, CANADA Liverpool John Moores University, UK
a r t i c l e
i n f o
Article history: Received 17 February 2010 Received in revised form 24 August 2010 Accepted 22 September 2010 Keywords: Prehension Visual control Binocular Monocular Alternating
a b s t r a c t This study examined the inter-ocular (alternating monocular samples) and intra-ocular (monocular or binocular samples) integration during a prehensile task with a range of occlusion intervals (0–75 ms). In the first experiment, participants were uncertain regarding the impending visual condition, as well as target size and location. In the second experiment, a pre-cue on target location was provided. Data from both experiments indicated that participants modified their movement kinematics when provided with alternating monocular samples, irrespective of whether or not there was an occlusion interval. Similar adaptations were found in conditions requiring intra-ocular integration but only following the introduction of an occlusion interval. These findings are consistent with participants having a general intolerance for alternating monocular samples and as a consequence using a more cautious reach and grasp strategy. Crown Copyright © 2010 Published by Elsevier Ireland Ltd. All rights reserved.
Briefly presented visual samples can be integrated over time when the intra-ocular occlusion period is shorter than the intrinsic persistence of the visual samples [5]. In motor tasks demanding high precision (i.e., manual aiming, one-handed catching), the persistence of binocular samples (20 ms duration) enables performance to be maintained with intra-ocular occlusion periods of no longer than approximately 40–80 ms [4,7]. On the other hand, monocular samples (20 ms duration) seem to have a shorter persistence of approximately 20 ms [2,14]. The implication is that the control of tasks involving high precision is not only dependent on the duration of the intra-ocular occlusion interval, but also whether the visual input is provided to one or both eyes. The finding of a difference in the persistence of binocular and monocular vision that facilitates performance of precision tasks has led to recent attempts to determine the duration over which the monocular input can be integrated between the eyes. In an experiment [14] based on the method developed by [8], participants made more one-handed ball catches when provided with 20 ms or 80 ms alternating monocular samples than in a condition of continuous monocular vision. It was suggested that individuals were capable of gaining useful information by integrating alternating monocular samples presented without an inter-ocular occlusion as long as the time between alternating samples to the same eye did not exceed 80 ms. Contradictory evidence was obtained with a variation of this methodology that included an inter-ocular occlusion between the alternating monocular samples [3,2]. In those experiments, the pro-
∗ Corresponding author at: Brain and Behaviour Research Group, Liverpool John Moores University, Byrom Street, Liverpool L3 3AF, UK. E-mail address: [email protected] (S.J. Bennett).
portion of balls caught in monocular and alternating monocular conditions did not differ regardless of the occlusion interval and was consistently worse compared to corresponding conditions of binocular vision. To better determine the temporal integration limits of binocular vision, a recent study investigated the integration of alternating monocular samples during prehension [18]. Providing participants with continuous binocular vision resulted in smaller average grip apertures than a continuous monocular vision condition, as well as conditions in which alternating monocular samples were separated by an inter-ocular occlusion as short as 14 ms. This result was replicated in a second experiment where visual feedback of the hand was occluded for the initial 80% of the movement. The authors suggested that the larger grip aperture in the alternating monocular vision conditions, which would have been “largely programmed before the reach begins” (p. 96), could not be accounted for by participants having difficulties integrating highly dissimilar retinal images as the hand approached the target. Moreover, it was concluded that the participants adopted a cautious reach and grasp strategy as a consequence of having no tolerance for the integration of alternating monocular samples to provide a binocular percept. In the above study, alternating monocular conditions had interocular occlusion that ranged from 14–58 ms. It remains unclear, therefore, if participants could have integrated alternate monocular samples presented consecutively (i.e., without an inter-ocular occlusion) [14]. Furthermore, to date there has been no detailed report on measures of movement kinematics as the hand is transported from the home position to the target [12,13]. For instance, it is unknown whether participants increase movement time and change the proportion of time spent in the accelerative or decelerative phases of the movement following the introduction of an
0304-3940/$ – see front matter. Crown Copyright © 2010 Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2010.09.065
18
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
inter-ocular occlusion. This omission is important because binocular information from retinal disparities has been shown to be particularly important for the latter on-line control of prehensile movement [13]. Nine males (right hand, right eye dominant) from the host University provided informed consent prior to participating in this experiment. All procedures were conducted in accordance with the local ethical guidelines and the 1964 declaration of Helsinki. Participants were required to perform a prehensile task that involved reaching out to “grasp” 2D square targets located in their midsaggital plane. Specifically, they were instructed to place the index finger and thumb on the edges of the targets in the same way they would when grasping a real 3D square target. A visual stimulus generator (ViSaGe) with proprietary software (CRS Toolbox) operating in MATLAB (Mathworks Inc.) presented the targets on a 21-in. computer monitor (refresh rate of 160 Hz). The monitor was mounted horizontally in a wooden frame and had a clear Perspex overlay mounted 5 mm above the screen surface. White target squares were presented against a black background and contained a thin black cross (2 mm thick) that bisected their width. Either a small (24 mm × 24 mm) or a large (36 mm × 36 mm) square was presented at a near (285 mm) or far (385 mm) distance. The home position was located approximately 200 mm from the participant. Participants wore a pair of PLATO liquid crystal goggles throughout the protocol. The state of the liquid crystal lenses was manipulated to provide binocular, monocular or alternating monocular vision during the grasping movements. In the binocular condition, both lenses were transparent or opaque at the same time, whereas in the monocular conditions only the left lens was cycled between transparent and opaque states. In the alternating monocular conditions, the left and right lenses were alternately switched between transparent and opaque states such that vision was only available to one eye at a time. When vision was continuous there was no occlusion interval (0 ms) between 25 ms samples presented to both eyes (binocular condition), the left eye alone (monocular condition), or alternately between the right and left eyes (alternating monocular). For all other intermittent vision conditions, the state of the liquid crystal goggles was cycled such that participants were provided with a 25 ms sample (both eyes, left eye, alternate left or right eye) followed by a 12.5, 25, 50 or 75 ms occlusion interval: note that taking account of minimal delays inherent from switching states of the liquid crystal lenses (i.e., TranslucentTechnologies, Inc. technical report), it can be estimated that the opaque state was approximately 3 ms shorter than the driving signal. Participants completed 300 pseudo-randomly ordered trials (15 vision conditions, 2 target location, 2 target size, 5 repeats), which were separated into five blocks of 60. Each block comprised 15 trials to each of the two target locations and two target sizes. A trial began with the participant placing their finger and thumb together at the home position. The goggles were then switched opaque for 500 ms after which a target square appeared and simultaneously the goggles cycled between opaque and transparent states. Participants were asked to perform the prehensile task as quickly and as accurately as possible. Movement of markers attached to the distal end of the index finger, thumb, and radial-carpal joint was recorded at 200 Hz for the duration of the trial with a Qualysis ProReflex optoelectronic system; system accuracy has been measured at 0.5 mm [15]. Post experimentation, the resulting three-dimensional position data were filtered using a second order dual pass Butterworth filter with a low-pass cut-off frequency of 8 Hz [11]. The filtered position data were differentiated to acquire velocity data. From these data, we extracted reaction time, time after peak velocity, maximum grip aperture and movement time [11].
Fig. 1. Movement time (MT) and time after peak velocity (TAPV) in Experiment 1 as a function of vision condition (BINO: continuous binocular, Alt: alternating, 0, 12.5, 25, 50, 75, MONO: continuous monocular). Standard error of the mean indicated.
Inter-ocular integration was examined using separate one-way repeated measures ANOVA that compared the alternating monocular vision conditions (0, 12.5, 25, 50, 75 ms occlusion) to the continuous binocular and monocular vision conditions; see [14,18]. Intra-ocular integration was examined with separate 2-Vision Condition (binocular, monocular) by 5-Occlusion Interval (0, 12.5, 25, 50, 75 ms) ANOVA with repeated measures on both factors. Significant main effects and interactions were decomposed using Fisher LSD (P < 0.05). The different levels of target size (small, large) and location (near, far) were included in the experimental design in order to minimise advance planning. However, these factors did not interact with vision condition and hence were collapsed [see also 18]. Inter-ocular integration. There were no main effects for reaction time, F(6, 48) = 0.70, indicating that the amount of time spent planning the response was unaffected by vision condition and/or occlusion interval (grand mean = 256 ms). For maximum grip aperture, the effect of vision condition approached conventional levels of significance, F(6, 48) = 2.10, P < 0.07. Observation of the group means indicated that compared to continuous alternating vision there was a small increase in maximum grip aperture when the occlusion interval was longer than 12.5 ms (0 ms = 58.6 mm; 12.5 ms = 59.1 mm; 25 ms = 59.5 mm; 50 ms = 59.1 mm; 75 ms = 59.1 mm). Movement time, F(6, 48) = 5.38, P < 0.05, and the time after peak velocity, F(6, 48) = 3.66, P < 0.05, were significantly longer in all the alternating monocular conditions than in the continuous binocular or monocular vision condition (Fig. 1) Intra-ocular integration. There were no main effects for reaction time. However, compared to continuous binocular and monocular vision conditions, participants exhibited a significant increase in movement time, F(4, 32) = 5.27, P < 0.05, and time after peak velocity, F(4, 32) = 4.90, P < 0.05, when there was a 12.5 ms occlusion interval (Fig. 2). This difference was maintained across all occlusion intervals but there was not always a further significant increase. In addition, there was a significant interaction for maximum grip aperture, F(4, 32) = 3.57, P < 0.05. The introduction of an intra-ocular occlusion interval in the monocular vision condition resulted in a small but significant increase in maximum grip aperture (0 ms = 58.1 mm; 12.5 ms = 59.3 mm; 25 ms = 59.0 mm; 50 ms = 59.4 mm; 75 ms = 59.1 mm). For the comparison to continuous binocular vision, maximum grip aperture
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
19
Fig. 2. Movement time (MT) and time after peak velocity (TAPV) in Experiment 1 as a function of vision condition (MONO: monocular, BINO: binocular) and occlusion interval (0, 12.5, 25, 50, 75 ms). Standard error of the mean indicated.
increased significantly but only when there was a 75 ms occlusion interval (0 ms = 57.8 mm; 12.5 ms = 58.0 mm; 25 ms = 58.4 mm; 50 ms = 58.5 mm; 75 ms = 59.4 mm). Participants were unable to maintain their normal movement kinematics (i.e., continuous binocular vision) when visual samples were presented alternately between the eyes with a short occlusion interval (i.e., 12.5 ms). More importantly, similar changes to movement kinematics were evident when there was no inter-ocular occlusion. Specifically, participants lengthened the deceleration phase of the movement as the hand approached the target, and as consequence took longer to complete the prehensile action. This adaptation to the timing of movement control was concomitant with a small increase in grip aperture and is consistent with individuals being unable to integrate the disparate retinal inputs from alternate monocular samples [13]. A similar adaptation to the movement kinematics was evident when an intra-ocular occlusion interval of 12.5 ms was introduced in the binocular and monocular vision conditions. Compared to research on one-handed catching [2,3] these data suggest that the intra-ocular persistence of binocular and/or monocular information for a prehensile action upon a static target is of a shorter duration. It is noteworthy, however, that the kinematic measures reported here are more sensitive to perturbations of the available information than the outcome measures used in one-handed catching. Further work is required to better determine task differences in informational persistence. Contrary to previous work in which the visual condition was known after completing the first trial within each block [3,18], participants in our first experiment were uncertain regarding the impending visual condition as well as the target location and size. In combination with the fact that participants were instructed to respond as quickly and accurately as possible to the appearance of the target, this uncertainty could have influenced the manner in which movements were controlled [12,17]. For instance, participants could have been biased towards responding rapidly after
target appearance (i.e., shorter duration planning) and then use vision to modify the reach and grasp if necessary. To provide reliable information regarding target location and potentially reduce the reliance on depth information available during the trial, we repeated the first protocol but with the addition of a pre-cue. Eight males from the host university provided informed consent prior to completing the protocol. The group had not participated in the first experiment and were all right hand and right eye dominant. The apparatus and protocol were the same as in the initial procedure with the exception that a pre-cue (1 mm diameter white circle) was presented for 500 ms at the target location while both lenses were clear. After this foreperiod, the target square was presented and the goggles began cycling between the transparent and opaque states. Inter-ocular integration. There was a significant main effect for RT, F(6, 42) = 5.59, which was a result of a longer RT in the continuous binocular vision condition compared to all other conditions. Movement time, F(6, 42) = 2.90, P < 0.05, and time after peak velocity, F(6, 42) = 3.43, P < 0.05, were significantly longer in the alternating monocular vision conditions than the continuous binocular vision condition. There were no differences in the duration of movement time and time after peak velocity between the continuous monocular and binocular conditions (Fig. 3). Maximum grip aperture in all of the alternating monocular vision conditions (0 ms = 61.3 mm; 12.5 ms = 61.7 mm; 25 ms = 61.4 mm; 50 ms = 61.1 mm; 75 ms = 61.3 mm) was significantly larger, F(6, 42) = 5.10, P < 0.05, than that exhibited in the continuous binocular vision condition (59.9 mm); the larger grip aperture in Experiment 2 compared to Experiment 1 was most likely due to anatomical differences between the groups Intra-ocular integration. One again, RT in the continuous binocular vision condition was significantly longer than RT in all other conditions, F(4,28) = 6.04, P < 0.05. Moreover, compared to when both binocular and monocular vision was continuous, movement time, F(4, 28) = 5.77, P < 0.05, and time after peak velocity, F(4,
20
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
Fig. 3. Movement time (MT) and time after peak velocity (TAPV) in Experiment 2 as a function of vision condition (BINO: continuous binocular, Alt: alternating 0, 12.5, 25, 50, 75 ms, MONO: continuous monocular). Standard error of the mean indicated.
28) = 7.31, P < 0.05,were longer, and maximum grip aperture, F(4, 28) = 3.52, P < 0.05, was larger, when visual samples were separated by a 25 ms occlusion interval (578 ms, 359 ms, 61.0 mm) compared to when vision was continuous (558 ms, 337 ms, 60.1 mm). There was no further change in these dependent variables as the occlusion interval lengthened and no interaction between vision conditions Consistent with previous research [18], we found in both experiments reported here that participants modified the kinematics of their normal prehensile movement when a 12.5 ms inter-ocular occlusion was introduced between alternating monocular samples. In addition, we found that the same adaptation to the movement kinematics was made when alternating monocular samples were presented without an inter-ocular occlusion. Specifically, there was an increase in maximum grip aperture and a lengthening of movement time and deceleration time in the alternating monocular vision conditions compared to the continuous binocular and monocular vision conditions. Providing participants with advance information on target location did not ameliorate these effects. Importantly, we also showed that these adapted movement kinematics were not simply a consequence of participants planning for the worse-case scenario prior to trial onset (e.g., alternating monocular samples separated by a 75 ms occlusion interval). For instance, in the second experiment participants did not know the impending visual condition prior to trial onset but still they exhibited different movement kinematics when provided with continuous binocular or monocular vision. The implication is that participants took advantage of the available information on a trialby-trial basis in order to more effectively plan and control their prehensile movement. Notably, with the exception of the continuous binocular vision condition in the second experiment, there was no evidence of a change in reaction time. Thus, it would appear that the lengthening of TAPV and MT were not an adaptive response to shorter duration planning phase. In combination, the results of the current study confirm the suggestion that participants have a general intolerance to alternating monocular samples in prehension. In both experiments of the current study, there was a more gradual change in movement kinematics as the intra-ocular occlusion interval in conditions of binocular and monocular vision was lengthened, hence indicating some tolerance for intra-ocular integration [2,7]. The finding of differential effects in these visual conditions concurs with the suggestion from perceptual studies [8,16] that inter-ocular integration of disparate monocular inputs between the left and right eyes does not entail the same underlying
processing as that required in conditions of intermittent binocular or monocular vision, where the visual samples only need be integrated over time (i.e., intra-ocular integration). It is relevant to consider, therefore, how the intolerance for alternating monocular samples during prehension might be explained. Given that there was no evidence of a binocular advantage in the continuous binocular vision condition compared to the continuous monocular vision condition, it could be suggested that participants did not use binocular information and instead relied on monocular information alone. It would follow, therefore, that the adaptation to movement kinematics in the alternating monocular conditions was not a direct consequence of being unable to integrate disparate monocular inputs to gain a binocular percept. It is far more likely, however, that participants did use binocular information when available [18] and this resulted in similar movement kinematics to those exhibited when using monocular information alone. Indeed, in separate tests we have found that participants discriminate target squares embedded in a randomdot stereogram from continuous alternating monocular samples, and hence perceive depth from disparate monocular inputs. Accepting that binocular information would have been used if perceived, one possibility is that intolerance to alternating monocular vision was a result of the 25 ms sample duration limiting the temporal integration of luminance, which precluded the cross-correlation of disparate retinal images and the perception of stereomotion [9,10]. In this respect, it is notable that similar effects (on grasp aperture) as those reported in the current study were found when alternating monocular samples were presented for 50 ms [18]; this should have been sufficient to permit temporal integration of luminance. Moreover, the duration of the visual samples was long enough to perceive a static depth cue from alternating monocular samples [16], which should have proved useful for movement control. For instance, the perception of form in depth can be maintained from 10 ms alternating monocular samples if they are separated by an inter-ocular delay of no more than 50–70 ms. If the duration of the visual samples was not the limiting factor, it remains a possibility that the dissimilarity between alternating monocular samples disrupted the perception of coarse binocular disparity that provides information on the closing gap as the hand approaches the target; for more a detailed explanation of disparity processing see [13]. It is known from work using dichoptic displays that independent image motion rather than fused image motion tends to be perceived as images become increasingly dissimilar [1]. This is consistent with the suggestion that stimuli (e.g., alternate monocular samples) that are related to each other are more likely to be combined than disjoint stimuli [6]. However, while a disruption to processing of coarse binocular disparity could explain the differences found for the kinematic measures associated with on-line control in the current study, it does not follow that alternating monocular samples should have disrupted the perception of fine binocular disparity that provides size information from the static target, which can be used to scale grip aperture prior to reach onset [18]. We suggest, therefore, that a more parsimonious explanation for the effects on movement kinematics observed in conditions of alternating monocular vision is that participants recognised prior to initiating the reach that they were receiving alternating monocular samples, which they understood (from previous trials) disrupted the perception of motion in depth. Consequently, this resulted in a strategic increase in grip aperture as well as changes in temporal aspects of the movement. This interpretation is consistent with the suggestion that a wider grip aperture is used to provide a greater margin for error when vision is perturbed [19] and not that it represents an adaptation to a misperception of target size per se. Indeed, if participants had imperfectly superimposed the two laterally displaced alternate
S. Hansen et al. / Neuroscience Letters 487 (2011) 17–21
monocular images, it would be expected that the resulting target image was perceived to be slightly larger than its true size. In this case, grip aperture would be increased but there should be no modification to variables associated with on-line control. Our experiments confirmed and extended upon the finding that participants are generally intolerant to alternating monocular samples for the control of prehension. We suggest that the changes in movement kinematics are consistent with the breakdown of a motion in depth signal that is particularly important for controlling the closure of the gap between the hand and target during the latter stage of the movement. References [1] T.J. Andrews, C. Blakemore, Integration of motion information during binocular rivalry, Vis. Res. 42 (2002) 301–309. [2] S.J. Bennett, D. Ashford, D. Elliott, Intermittent vision and one-handed catching: the temporal limits of binocular and monocular integration, Motor Control 7 (2003) 378–387. [3] S.J. Bennett, D. Ashford, N. Rioja, J. Coull, D. Elliott, Integration of intermittent visual samples over time and between the eyes, J. Mot. Behav. 38 (2006) 439–450. [4] S.J. Bennett, D. Elliott, D.J. Weeks, D. Keil, The effects of intermittent on prehension under binocular and monocular viewing, Motor Control 7 (2003) 46–56. [5] M. Coltheart, Iconic memory and visible persistence, Percept. Psychophys. 27 (1980) 1959–1975. [6] P. Dixon, V. Di Lollo, Beyond visible persistence: an alternative account of temporal integration and segregation in visual processing, Cognit. Psychol. 26 (1994) 33–63.
21
[7] D. Elliott, R. Chua, B.J. Pollock, The influence of intermittent vision on manual aiming, Acta Psychol. 85 (1994) 1–13. [8] G.R. Engel, An investigation of visual responses to brief stereoscopic stimuli, Q. J. Exp. Psychol. 22 (1970) 148–166. [9] E. Gheorghiu, C.J. Erkelens, Differences in perceived depth for temporally correlated and uncorrelated dynamic random-dot stereograms, Vis. Res. 45 (2005) 1603–1614. [10] E. Gheorghiu, C.J. Erkelens, Temporal properties of disparity processing revealed by random-dot stereograms, Perception 34 (2005) 1205–1219. [11] S. Hansen, D. Elliott, M.A. Khan, Comparing derived and acquired acceleration profiles: three dimensional optical electronic data analyses, Behav. Res. Methods 39 (2007) 748–754. [12] S. Hansen, C.M. Glazebrook, J.G. Anson, D.J. Weeks, D. Elliott, The influence of advance information about target location and visual feedback on movement planning and execution, Can. J. Exp. Psychol. 60 (2006) 200–208. [13] D.R. Melmoth, M. Storoni, G. Todd, A.L. Finlay, S. Grant, Dissociation between vergence and binocular disparity cues in the control of prehension, Exp. Brain Res. 183 (2007) 283–298. [14] I. Olivier, D.J. Weeks, J.L. Lyons, K.L. Ricker, D. Elliott, Monocular and binocular vision in one-handed ball catching: interocular integration, J. Mot. Behav. 30 (1998) 343–351. [15] J.G. Richards, The measurement of human motion: a comparison of commercially available systems, Hum. Mov. Sci. 18 (1999) 509–602. [16] J. Ross, J.H. Hogben, Short term memory in stereopsis, Vis. Res. 14 (1974) 1195–1201. [17] E.J. Schlicht, P.R. Schrater, Effects of visual uncertainty on grasping movements, Exp. Brain Res. 182 (2007) 47–57. [18] K.R. Wilson, P.M. Pearson, H.E. Matheson, J.J. Marotta, Temporal integration limits of stereovision in reaching and grasping, Exp. Brain Res. 189 (2008) 91–98. [19] A.M. Wing, C. Fraser, The contribution of the thumb to reaching movements, Q. J. Exp. Psychol. 35A (1983) 297–309.