In Psychological Science, 21(7), 1006-13 (2010) The Direction of Walking—but Not Throwing or Kicking—Is Adapted by Optic Flow Hugo Bruggeman and William H. Warren. Brown University Corresponding Author: Hugo Bruggeman, Department of Cognitive and Linguistic Sciences, Brown University, Providence, RI 02912 E-mail: hugo_bruggeman@brown.edu Abstract Optic flow is known to adapt the direction of walking, but the locus of adaptation remains unknown. The effect could be due to realignment of anatomical eye, head, trunk, and leg coordinate frames or to recalibration of a functional mapping from the visual direction of the target to the direction of locomotor thrust. We tested whether adaptation of walking to a target, with optic flow displaced by 10°, transfers to facing, throwing, and kicking a ball to the target. A negative aftereffect for initial walking direction failed to transfer to head orientation or throwing or kicking direction. Thus, participants effectively threw or kicked the ball to the target, and then walked in another direction to retrieve it. These findings are consistent with recalibration of a task-specific visuo-locomotor mapping, revealing a functional level of organization in perception and action. Keywords: perception and action, locomotion, human Received 8/12/09; Revision accepted 11/5/09

A process of perceptual-motor calibration is essential to keep everyday actions coordi-nated with environmental targets. Without such calibration—in situations in which per-ceptual feedback is removed—goal-directed movements drift over time and lose preci-sion and accuracy (Bingham & Pagano, 1998; Vindras & Viviani, 1998). Presuma-bly, the perceptual-motor mappings that un-derlie adaptive behavior are maintained by continuous calibration throughout life (Bingham, Coats, & Mon-Williams, 2007). We recently demonstrated that optic flow serves as a “teaching signal” to adapt the mapping between the visual direction of a goal and the direction of walking (Brugge-man, Zosh, & Warren, 2007). During adap-tation, participants repeatedly walked to a target in a textured virtual environment while the heading direction specified by op-tic flow (virtual heading) was displaced by 10° to the right of their actual walking di-rection (locomotor axis). On the first few trials, the virtual heading at the onset of walking missed the target by a comparable amount (~10°), leading to a curved path (Fig. 1). But after several trials, this initial virtual heading error was reduced by half, even though participants were unaware of the error. In the test phase, when the heading displacement was removed and the target was a single line in a dark field, we observed a negative aftereffect in the initial walking direction (about 4°–5° to the left), which led to a significantly curved path in the opposite direction that revealed the adaptation (Fig. 1).
--Insert Fig. 1. here-- Two of our findings are critical. First, optic flow appears to be essential for adaptation of the initial walking direction, as adaptation occurred only in a visually structured envi-ronment with rich optic flow, not in a sparse environment with the target alone. Optic flow ordinarily specifies the direction of travel through the environment (Gibson, 1950; Warren, 2007), and thus provides a reliable teaching signal for locomotor adap-tation. Second, contrary to previous sugges-tions (Rushton, 2004; Rushton & Salvucci, 2001), the adaptation cannot be ascribed to a shift in the visual straight-ahead. Partici-pants were instructed to face the target at the beginning of each trial, and their head ori-entation did not show any sign of adaptation. What, then, is the locus of adaptation? Two hypotheses are suggested by the literature (Redding, Rossetti, & Wallace, 2005; Rie-ser, Pick, Ashmead, & Garing, 1995). First, the anatomical hypothesis proposes that ad-aptation is due to proprioceptive realignment of the eye-head, head-trunk, or trunk-legs coordinate frames (or some combination of these frames), and that this realignment al-ters coordinate transformations between body segments. Second, the functional hy-pothesis proposes that optic flow recali-brates a task-specific mapping from the visual direction of the target to the direction of locomotor thrust. The anatomical hypothesis stems from prism adaptation studies of pointing to a visual target (C.S. Harris, 1963; Redding & Wallace, 1997). A participant wearing later-

ally displacing prisms exhibits a rapid ad-aptation of pointing direction in the absence of changes in gaze direction (Bingham & Romack, 1999). The adaptation in pointing direction transfers to auditory targets and is attributed to a realignment of the felt posi-tion of the arm relative to the head. It is ac-companied by a more gradual adaptation of the visual straight-ahead, which is attributed to realignment of the felt eye position rela-tive to the head (Hay & Pick, 1966b): ok.. Analogously, optic flow might serve to rea-lign the proprioceptive coordinate frames of body segments during locomotion. The ini-tiation of walking to a visual target involves a transformation of the visual direction of the target from oculocentric to headcentric, bodycentric, and effector coordinate frames. Thus, optic flow might adapt eye-head, head-trunk, or trunk-legs coordinate trans-formations. We have already shown that the eye-head frames are not realigned, because head orientation to face the goal is unaf-fected by locomotor adaptation (Bruggeman et al., 2007). However, a realignment of the head-trunk or trunk-legs frames could ac-count for adaptation of the initial walking direction. Alternatively, optic flow might serve to re-calibrate a functional mapping from the tar-get’s visual direction to the direction of locomotor thrust. We call such a perceptual-motor relation a control law (Warren, 1988, 2006), a task-specific mapping from an in-formational variable to a control parameter that governs forces for a class of actions. A control law for steering is functionally spe-cific to the task of locomotion, not anatomi-
cally specific to particular body segments, and may apply to any goal-directed dis-placement of the body regardless of the ef-fectors involved. Although the mapping depends on proprioceptive relations among body segments, according to this view they are not the sites of adaptation. Of course, it is possible that both anatomical realignment and functional recalibration occur during locomotor adaptation. In the study reported in this article, we tested the anatomical and functional hy-potheses by determining whether the adap-tation of initial walking direction transfers to other goal-directed actions performed with the head, trunk, and leg segments: (a) facing a distant target, (b) throwing a ball to the target, and (c) kicking a ball to the target. If the locus of adaptation is realignment of the eye-head coordinate frames, the negative aftereffect for walking should transfer to all three tasks, but our previous data are incon-sistent with this hypothesis (Bruggeman et al., 2007). If there is a head-trunk realign-ment, the aftereffect should transfer to throwing and kicking, but not facing the tar-get. And if there is a trunk-leg realignment, it should transfer only to kicking. In con-trast, if the locus of adaptation is a task-spe-cific control law, the aftereffect for walking should not transfer to any of the nonloco-motor tasks. Surprisingly, this would create a situation in which participants effectively throw or kick a ball to the target, and then walk in a different direction to retrieve it. Method Participants

Sixteen participants (5 females and 11 males) with normal or corrected-to-normal vision participated in the experiment. They ranged in age from 18 to 28 years (M = 21.1 years, SD = 2.6). None of them had any pre-vious experience in this kind of study. Sub-jects gave informed consent in accordance with the human-subjects policies of Brown University. Materials Participants walked freely in the 12 m 12 m Virtual Environment Navigation Lab at Brown University while wearing a head-mounted display, which presented stereo images generated at 60 Hz with a 63° (hori-zontal) 53° (vertical) field of view (a shield blocked all peripheral vision, includ-ing downward vision). Head position and orientation were recorded at 30 Hz and used to update the display with a latency of 70 ms. In the line environment, used on pre- and posttest trials, the target was a vertical red line in a dark field; the line, which had a diameter of 0.008 m, ran off the top and bottom of the display. In the texture envi-ronment, used on adaptation trials, the target was a doorway that was 0.5 m wide and po-sitioned in a frontal wall of a virtual room with a ground plane and ceiling; these sur-faces were mapped with a gray-scale ran-dom-noise texture. In addition, an array of vertical textured posts (0.05 m in diameter) was randomly positioned on the ground, with approximately 19 posts visible in the first frame. This visually structured envi-ronment provided rich optic flow, including motion parallax and target drift, to a walking participant.
The virtual heading was dissociated from the physical heading by displacing the simulated direction of travel through the virtual environment by 10° to the right (or left) of the participant’s actual direction of walking in the physical environment, as in our previous study (Bruggeman et al., 2007). Thus, an observer walking in the egocentric direction of the target would see an optic-flow pattern that specified a heading 10° to the right of the target, equivalent to a virtual heading error of +10° (or a pattern that specified a heading 10° to the left of the tar-get, an error of –10°). Steering so as to place the visually specified heading (e.g., the fo-cus of expansion in the optic-flow pattern) on the target would bring the virtual heading error to 0°. Design and procedure A pretest-adaptation-posttest design was used. During the 38 adaptation trials, the virtual heading was displaced by 10° from the actual walking direction; for half of the participants, it was displaced to the right (+10°), and for the other half, it was dis-placed to the left (–10°). At the beginning of each trial, participants positioned themselves at a starting post, turned to face a distant marker, and began walking when the marker disappeared and the texture environment ap-peared; the target doorway was at an initial distance of 9 m. Participants were instructed to walk at a normal pace on the most direct path to the target and to walk through any posts that obstructed their way; the envi-ronment remained visible throughout each trial, providing continuous (error) feedback.

During the 8 pretest and 8 posttest trials, the optical displacement was removed, and the (initial) direction of throwing, kicking, walking, and head orientation was assessed. On each trial, the line environment appeared with the target line at an initial distance of 9 m. Participants first threw a 0.150-kg la-crosse ball underhand toward the target (es-sentially rolling the ball on the ground) and then kicked a 2-kg rubber ball toward the target (using the sole of their foot to roll the ball on the ground, while holding a walking stick to help maintain balance). Finally, the target line disappeared, and they walked to-ward its location for about 5 m. Just prior to the onset of walking on every trial, partici-pants were instructed to face the target, and head orientation was recorded. Participants could not see the ball (the experimenter placed it in their hand or under their foot) and did not receive any visual feedback during pre- and posttest trials (between tri-als, the experimenter led participants to the start position of the next trail). Prior to the experiment, participants received 5 practice trials in the line environment (without the optical displacement) to familiarize them-selves with the throwing, kicking, and walking procedure. The research protocol was approved by Brown University’s insti-tutional review board. Data processing and analysis To collapse data across the two groups of participants, we converted the head-position data (walking task) and head-orientation data (facing task) of the group with a –10° offset (to the left) to the corresponding val-ues for a +10° offset (to the right). The fac-
ing direction prior to walking was measured as the instantaneous head orientation 0.05 s after trial onset (triggered by the disappear-ance of the distant marker, on adaptation trials, or of the target, on pre- and posttest trials). The head-position data during walk-ing were filtered with a second-order But-terworth filter (0.6-Hz cutoff) to reduce the effects of gait oscillations. The first 0.25 m and last 1 m of each path were removed to eliminate transients, and the intervening temporal samples were normalized in space by binning them into 35 segments, each about 0.23 m in length; these were ensem-ble-averaged to yield a mean path about 7.75 m long. Path statistics were computed on the segment means and used for analysis of the virtual heading error, which was defined as the angle between the direction of head dis-placement and the direction of the target. The initial walking direction was computed as the heading error during the 5th and 6th bins (0.38 to 0.45 m from the start). Throw-ing and kicking accuracy were determined by measuring the point at which the ball crossed a line on the floor immediately be-hind the virtual target; the lateral distance from the target was converted to an angular error. Like the head-position and orientation data, the throwing and kicking data were collapsed across the two groups of partici-pants; the data of the group with a –10° off-set were converted to the corresponding values for a +10° offset. Data for the initial walking direction during adaptation trials were fitted with a first-order exponential decay function. Post hoc con-trasts, based on within-subjects one-way

analysis of variance with repeated measures on trial, were applied to test whether the ini-tial walking direction on subsequent adapta-tion trials differed from that on the first adaptation trial . In addition, pre- and post-test data were used to compute a difference in error between pretest and posttest, and these error differences were analyzed using a within-subjects two-way ANOVA (Task Trial) with repeated measures on task and trial. A significant difference in error (i.e., a shift in the direction of movement from pretest to posttest) indicates an aftereffect of adaptation. Results Adaptation Phase During adaptation trials, optic flow was used for both on-line control of locomotion and recalibration of the visuo-locomotor map-ping. When the optic flow was displaced to the right of the locomotor axis, walking in the egocentric direction of the target caused the target to drift leftward, yielding a path that curved to the left as the participant “chased” the target. Alternatively, if the participant used the optic-flow pattern to control steering (e.g., by shifting the focus of expansion onto the target), this would bring the virtual heading error down to zero and yield a straighter path. Such on-line steering control was evident on the first adaptation trial. The path was much less curved than predicted by the egocentric-direction strategy, and straightened out after the initiation of walking (Fig. 2a). A detailed analysis of the virtual heading error (the an-gle between the virtual heading and the vis-
ual target) over the first trial revealed an early correction (Fig. 2b): Within the first 3 m of walking, following the onset of the op-tic flow, the virtual heading error dropped from +10° to +3° (reducing the error by about 70%). This early steering adjustment recurred in subsequent trials, even after ad-aptation of the initial walking direction. This result replicates previous findings (Brugge-man et al., 2007; M.G. Harris & Carre, 2001; Turano, Yu, Hao, & Hicks, 2005; Warren, Kay, Zosh, Duchon, & Sahuc, 2001; Wood, Harvey, Young, Beedie, & Wilson, 2000) showing that people primarily rely on optic flow for the on-line control of locomotion in visually structured environ-ments. --Insert Fig. 2. here-- Adaptation was evident in both the reduction of path curvature (Fig. 2a) and the decline in initial heading error (Fig. 2b) over trials. The latter specifically reflects a change in the mapping from target direction to initial walking direction. On the first adaptation trial, participants started walking in the tar-get’s egocentric direction, and initial head-ing error was not statistically different from the predicted value of +10°, t(15) = 1.6, p = .14. But the initial walking direction quickly changed, such that by the fourth trial, the initial heading error differed significantly from +10°, t(15) = 8.8, p < .0001. Over the first 20 adaptation trials, there was a main effect of trial, F(19, 285) = 2.6, p < .0003, p2 = .15. Post hoc contrasts (Bonferroni-adjusted p cutoff = .0025) showed that ini-

tial heading error on the first trial was mar-ginally different from initial heading error on the third trial (p = .0026), and was sig-nificantly different from initial heading error on the fourth trial (p = .001) and nearly all subsequent trials (median p = .0021). Fitting the decreasing initial heading error over tri-als with a first-order exponential decay function (Fig. 3) revealed an asymptotic value of 4.9°, a 50% reduction from the original 10° displacement (the proportion of variance accounted for, r?, was .59). --Inset Fig. 3. here-- In contrast, facing direction prior to walking did not change over adaptation trials, but was comparatively stable. The mean head orientation relative to the distant marker was 0.4° (SD = 1.9°) for the first five adaptation trials and 0.24° (SD = 1.7°) for the last five adaptation trials (Fig. 3). This replicates our previous finding that the visuo-locomotor adaptation cannot be attributed to adaptation of the visual straight-ahead or to realignment of the eye-head coordinate frames (Brugge-man et al., 2007). Pre- and posttest phases The data from the pre- and posttests (see Fig. 3) allowed us to evaluate whether ad-aptation of the initial walking direction transferred to the facing, throwing, or kick-ing direction. On pretest trials, all four tasks were comparably accurate relative to the target, F(3, 45) = 0.09, p = .96. The mean directional error was 0.18° (SD = 2.8°) for walking, 0.23° (SD = 2.1°) for facing, –
0.14° (SD = 3.4°) for throwing, and 0.28° (SD = 2.7°) for kicking. To evaluate the posttest trials for transfer effects of adapta-tion of the initial walking direction, we computed an error difference score (posttest – pretest) for each task (see Fig. 4). Initial walking direction shifted by –4.4° (SD = 2.5°) on the first posttest trial relative to the pretest (Fig. 4). This shift was signifi-cantly different from 0°, t(15) = 3.1, p < .0001, and demonstrated a large negative aftereffect, corresponding to 86% of the original 5.1° adaptation. In contrast, none of the other three tasks revealed a significant shift on the first posttest trial; the mean shift was 0.33° (SD = 1.6°) for facing, 0.45° (SD = 4.9°) for throwing, and 0.94° (SD = 6.0°) for kicking. A within-subjects repeated measures ANOVA confirmed that the shifts were significantly different across the tasks, F(3, 45) = 6.6, p = .0008, p2 = .31. Post hoc contrasts showed that the shift in walking direction was significantly greater than the shifts for facing, throwing, and kicking, p < .0001, p = .0002, and p = .0032, respec-tively. --Insert Fig. 4. here-- This pattern of results held up across all eight posttest trials, F(3, 45) = 4.1, p = .01, p2 = .22. The mean shift in initial walking direction was –3.5° (SD = 2.5°), which was significantly different from 0°, t(15) = 5.7, p < .0001, and corresponded to about 70% of the original adaptation of 5.1°. The other

three tasks did not reveal significant shifts; the mean shifts were 0.45° (SD = 1.4°) for facing, –0.63° (SD = 4.7°) for throwing, and 0.43° (SD = 4.9°) for kicking. Thus, the ad-aptation of initial walking direction deci-sively failed to transfer to facing, throwing, or kicking direction. Discussion The finding that the adaptation of walking direction does not transfer to facing, kicking, or throwing contradicts the anatomical hy-pothesis. The dissociation of walking from the other three tasks demonstrates that -visuo-locomotor adaptation cannot be ex-plained by realignment of eye-head, head-trunk, or trunk-leg coordinate frames. In contrast, the present results are entirely con-sistent with the functional hypothesis. Spe-cifically, optic flow appears to recalibrate a task-specific mapping from the visual direc-tion of the target to the direction of loco-motor thrust. This recalibration is functionally specific to locomotion, does not transfer to other goal-directed actions, and is not mediated by local realignment of ana-tomical coordinate frames. In effect, after throwing or kicking a ball to the target, par-ticipants indeed walked in a different direc-tion to fetch it. This study replicated our previous findings that optic flow is used for on-line steering control (Warren et al., 2001) and drives visuo-locomotor adaptation, without affect-ing the visual straight-ahead (Bruggeman et al., 2007). The absence of a shift in facing, throwing, or kicking direction clearly indi-cates that adaptation of walking direction
does not alter the direction of the visual straight-ahead (i.e., perceived egocentric di-rection). This result counters arguments by Rushton and his colleagues (Rushton, 2004; Rushton & Salvucci, 2001) that steering control relies exclusively on the egocentric direction of the target, and that optic flow plays only an indirect role by influencing perceived egocentric direction. In fact, our study demonstrates the opposite effect: Steering control relies primarily on optic flow in visually structured environments, and the egocentric-direction strategy de-pends on optic flow to keep the visuo-loco-motor mapping calibrated. Most other studies of the adaptation of walking direction have employed laterally displacing prisms (Brandwood, Rushton, & Charron, 2009; Held & Bossom, 1961; Morton & Bastian, 2004; Redding & Wallace, 1985a, 1985b; Rogers & Spencer, 2005). In contrast to our experiment, such studies have found a partial shift in the vis-ual straight-ahead. Morton and Bastian (2004) found a 5% shift in the visual straight-ahead after 25 trials of walking a 2-m-long path that was marked by two bound-ary lines, and also reported transfer from walking adaptation to the direction of reaching (30% of prism displacement). Redding and Wallace (1985b) observed that the visual straight-ahead shifted by 9.3% of the prism displacement after 10 min of walking with prisms in a hallway, and Held and Bossom (1961) found a similar 10% shift in the straight-ahead after 2 hr of walking with prisms on a tree-lined path. Note that this magnitude of shift in the

straight-ahead (equivalent to 1°) would ac-count for less than a quarter of the shift in walking direction we observed in the first posttest trial (4.4°). The discrepancy in the literature is likely related to the different methodologies used to create a visual displacement. First, whereas we computed a new direction of travel through the virtual environment, which displaced only the optic-flow pattern, prisms displace the entire visual scene, in-cluding both optic flow and the images of objects. Consequently, participants' inci-dental views of their own limbs or body during prism adaptation could have induced a shift in the visual straight-ahead or reach-ing direction. Second, whereas our head-mounted display created a symmetrical pin-cushion distortion that was independent of the optical displacement, lateral prisms cre-ate asymmetrical optical distortions that in-crease with the degree of displacement. Specifically, wedge prism glasses introduce image stretching (during horizontal eye movements) and shearing (during vertical eye movements) that depend on the direction of gaze (Hay & Pick, 1966a; Pick & Hay, 1966), and hence offer a possible basis for eye-head realignment. Despite the reports from prism studies, our finding of a disso-ciation between walking direction on the one hand and facing, throwing, and kicking on the other establishes the existence of a func-tional visuo-locomotor mapping that can be recalibrated without adaptation of the straight-ahead or realignment of coordinate frames.
The present results raise the question of whether this recalibration is specific to the act of walking, or might generalize to other forms of overground locomotion. One may hypothesize several levels of organization in perception and action, including a neuro-musculoskeletal level defined by particular segments and muscle groups, a biomechani-cal level defined by joint moments and movement patterns, and a functional level determined by goal-directed forces and out-comes (Bernstein, 1996). A functional class of actions may be characterized by the fam-ily of movements that direct forces toward a common perceptually specified outcome, regardless of the particular segments or movements involved (Reed, 1982; Withagen & Michaels, 2005). Walking, hopping, crawling, swimming, and even rollerboarding with the arms may be-long to the same family of locomotor actions because they involve applying directional thrust to displace the body toward a goal, and produce corresponding perceptual in-formation. If so, adaptation of walking di-rection should transfer to forms of locomotion that involve steering by the ap-plication of thrust, but not to locomotion by means of a tricycle with handlebars, a vehi-cle with a steering wheel, or possibly a skateboard that is steered by leaning. More narrowly, if the action class is defined by the directional displacement of the center of mass relative to the feet, then adaptation of walking should transfer to hopping and side-stepping, but not to crawling. More broadly, if optic flow recalibrates a general mapping between the visual direction of the target

and the direction of the locomotor axis, re-gardless of the manner of force application, then adaptation should transfer to all of these forms of locomotion. Related evidence for a functional level of visuo-locomotor organization includes ex-periments in which the visual gain of walk-ing was manipulated (Mohler et al., 2007; Rieser et al., 1995; Withagen & Michaels, 2002). Participants adapted to conditions in which the optic-flow rate was visually slower (or visually faster) than the physical walking speed. When transferred to normal conditions and instructed to walk to a target with eyes closed, they overshot (or under-shot) its distance. This effect transferred from walking to sidestepping and crawling, but not to throwing a ball to the target (Rie-ser et al., 1995; Withagen & Michaels, 2002). Thus, recalibration of the mapping from visual target distance to walking dis-tance appears to generalize to a family of related locomotor actions, but not to differ-ent goal-directed tasks. This pattern of func-tional dissociation suggests that instead of a single multipurpose representation of space in the brain, there may be multiple percep-tual-motor mappings that are specialized for different families of actions (Ingle, 1973). We conclude that optic flow continually calibrates a task-specific control law for lo-comotion, specifically, a functional mapping from the visual direction of a goal to the di-rection of locomotor thrust. These results provide evidence for a functional level of organization in perception and action that is not tied to particular anatomical loci. They imply that optic flow not only plays a direct
role in the on-line control of steering, but also plays an indirect role in calibrating the egocentric-direction strategy for steering within a family of locomotor actions. Acknowledgments: The authors thank Chaz Firestone and two anonymous reviewers for their comments on an earlier version of this manuscript. Declaration of Conflicting Interests: The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article. ? Funding: This research was supported by the U.S. National Institutes of Health (NIH EY10923) and by the Center for Vision Re-search at Brown University. References Bernstein, N.A. (1996). On dexterity and its development. In M.L. Latash & M.T. Turvey (Eds.), Dexterity and its develop-ment (pp. 3–244). Mahwah, NJ: Erlbaum. Bingham, G., Coats, R., & Mon-Williams, M. (2007). Natural prehension in trials without haptic feedback but only when calibration is allowed. Neuropsychologia, 45, 288–294. Bingham, G.P., & Pagano, C.C. (1998). The necessity of a perception-action approach to definite distance perception: Monocular distance perception to guide reaching. Journal of Experimental Psychol-ogy: Human Perception and Performance, 24, 145–168. Bingham, G.P., & Romack, J.L. (1999). The rate of adaptation to displace-

ment prisms remains constant despite acqui-sition of rapid calibration. Journal of Experimental Psychology: Human Percep-tion and Performance, 25, 1331–1346. Brandwood, T., Rushton, S.K., & Charron, C. (2009). Exposure to displaced optic flow results in adaptation of visual straight ahead. Journal of Vision, 9(8), Ab-stract 825. Retrieved April 20, 2010, from http://journalofvision.org/9/8/825/ Bruggeman, H., Zosh, W.D., & War-ren, W.H. (2007). Optic flow drives human visuo-locomotor adaptation. Current Biol-ogy, 17, 2035–2040. Gibson, J.J. (1950). The perception of the visual world. Boston: Houghton Mif-flin. Harris, C.S. (1963). Adaptation to displaced vision: Visual, motor, or proprio-ceptive change? Science, 140, 812–813. Harris, M.G., & Carre, G. (2001). Is optic flow used to guide walking while wearing a displacing prism? Perception, 30, 811–818. Hay, J.C., & Pick, H.L., Jr. (1966a). Gaze-contingent prism adaptation: Optical and motor factors. Journal of Experimental Psychology, 72, 640–648. Hay, J.C., & Pick, H.L., Jr. (1966b). Visual and proprioceptive adaptation to op-tical displacement of the visual stimulus. Journal of Experimental Psychology, 71, 150–158. Held, R., & Bossom, J. (1961). Neo-natal deprivation and adult rearrangement: Complementary techniques for analyzing plastic sensory-motor coordinations. Journal of Comparative and Physiological Psychol-ogy, 54, 33–37.
Ingle, D. (1973). Two visual systems in the frog. Science, 181, 1053–1055. Mohler, B.J., Thompson, W.B., Creem-Regehr, S.H., Willemsen, P., Pick, H.L., Jr., & Reiser, J.J. (2007). Calibration of locomotion resulting from visual motion in a treadmill-based virtual environment. ACM Transactions on Applied Perception, 4, 20–32. Morton, S.M., & Bastian, A.J. (2004). Prism adaptation during walking generalizes to reaching and requires the cerebellum. Journal of Neurophysiology, 92, 2497–2509. Pick, H.L., Jr., & Hay, J.C. (1966). Gaze-contingent adaptation to prismatic spectacles. American Journal of Psychology, 79, 443–450. Redding, G.M., Rossetti, Y., & Wallace, B. (2005). Applications of prism adaptation: A tutorial in theory and method. Neuroscience & Biobehavioral Reviews, 29, 431–444. Redding, G.M., & Wallace, B. (1985a). Cognitive interference in prism ad-aptation. Perception & Psychophysics, 37, 225–230. Redding, G.M., & Wallace, B. (1985b). Perceptual-motor coordination and adaptation during locomotion: Determinants of prism adaptation in hall exposure. Per-ception & Psychophysics, 38, 320–330. Redding, G.M., & Wallace, B. (1997). Adaptive spatial alignment. Hillsdale, NJ: Erlbaum. Reed, E.S. (1982). An outline of a theory of action systems. Journal of Motor Behavior, 14, 98–134.

Rieser, J.J., Pick, H.L., Jr., Ashmead, D.H., & Garing, A.E. (1995). Calibration of human locomotion and models of percep-tual-motor organization. Journal of Experi-mental Psychology: Human Perception and Performance, 21, 480–497. Rogers, B.J., & Spencer, O.E. (2005). Heading toward distant targets: Op-tic flow and the recalibration of visual di-rection. Journal of Vision, 5(8), Abstract 385. Retrieved April 20, 2010, from http://journalofvision.org/5/8/385/ Rushton, S.K. (2004). Egocentric di-rection and locomotion. In V. Lucia, S.A. Beardsley, & S.K. Rushton (Eds.), Optic flow and beyond (pp. 339–362). Dordrecht, The Netherlands: Kluwer Academic. Rushton, S.K., & Salvucci, D.D. (2001). An egocentric account of the visual guidance of locomotion. Trends in Cognitive Sciences, 5, 6–7. Turano, K.A., Yu, D., Hao, L., & Hicks, J.C. (2005). Optic-flow and egocen-tric-direction strategies in walking: Central vs peripheral visual field. Vision Research, 45, 3117–3132. Vindras, P., & Viviani, P. (1998). Frames of reference and control parameters in visuomanual pointing. Journal of Experi-mental Psychology: Human Perception and Performance, 24, 569–591. Warren, W.H. (2006). The dynamics of perception and action. Psychological Re-view, 113, 358–389. Warren, W.H. (2007). Optic flow. In R.R. Hoy, G.M. Shepherd, A.I. Basbaum, A. Kaneko, & G. Westheimer (Eds.), The senses: A comprehensive reference (pp. 219–230). Oxford, England: Elsevier.
Warren, W.H., Jr. (1988). Action modes and laws of control for the visual guidance of action. In O. Meijer & K. Roth (Eds.), Movement behavior: The motor-ac-tion controversy (pp. 339–380). Amsterdam: North-Holland. Warren, W.H., Jr., Kay, B.A., Zosh, W.D., Duchon, A.P., & Sahuc, S. (2001). Optic flow is used to control human walk-ing. Nature Neuroscience, 4, 213–216. Withagen, R., & Michaels, C.F. (2002). The calibration of walking transfers to crawling: Are action systems calibrated? Ecological Psychology, 14, 223–234. Withagen, R., & Michaels, C.F. (2005). On ecological conceptualizations of perceptual systems and action systems. The-ory & Psychology, 15, 603–620. Wood, R.M., Harvey, M.A., Young, C.E., Beedie, A., & Wilson, T. (2000). Weighting to go with the flow? Current Bi-ology, 10, R545–R546.it

Legend figure 1. Adaptation to displaced optic flow and the aftereffect in subsequent walking direction among participants asked to walk toward a target in a textured virtual environment (ad-aptation) and in an otherwise dark field en-vironment (posttest). The illustration shows a plan view of the mean walking paths on the first (solid blue curve) and last (dashed red curve) of 38 adaptation trials and the first posttest trial (solid black curve). The corresponding vectors starting at the origin of the graph represent the initial walking di-rection. The path and vector data are taken from Bruggeman, Zosh, and Warren (2007). Legend figure 2. Results for the adaptation phase: (a) plan view of mean walking paths and (b) mean virtual heading error as a function of walked distance. Results for the first adaptation trial are shown in blue, and average results for the last three adaptation trials are shown in red. In (a), the dotted black curve corre-sponds to the prediction of the egocentric-direction strategy, and the y-axis corre-sponds to the prediction of the optic-flow strategy. In (b), the dotted black line corre-sponds to the egocentric-direction prediction of a 10° heading error, and the x-axis corre-sponds to the optic-flow prediction of a 0° error.
Legend figure 3. Direction error on each trial. Results are shown separately for initial walking direc-tion, kicking, throwing, and facing direction. Participants experienced normal optic flow in the pretest and posttest trials, and dis-placed (all data converted to +10°) optic flow during adaptation. The solid curve rep-resents the exponential fit of the decay in heading error during adaptation, Y(t), with t = the number of the adaptation trial. Legend figure 4. Shift in direction error between pretest and posttest as a function of task. The mean shift in direction error from pretest to the first posttest trial is a measure of the negative aftereffect. Error bars indicate 1 SE, based on between-subjects variance.

Bruggeman & Warren
Figure 1.
ADAPTATIONAFTEREFFECT

Bruggeman & Warren
Figure 2.
a b
VIRTUAL HEADING ERRORPATH

Bruggeman & Warren
Figure 3.

Bruggeman & Warren
Figure 4.