UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of visual sampling in obstacle compensation Vienneau, Opal Yvette 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1997-0491.pdf [ 2.99MB ]
JSON: 831-1.0077311.json
JSON-LD: 831-1.0077311-ld.json
RDF/XML (Pretty): 831-1.0077311-rdf.xml
RDF/JSON: 831-1.0077311-rdf.json
Turtle: 831-1.0077311-turtle.txt
N-Triples: 831-1.0077311-rdf-ntriples.txt
Original Record: 831-1.0077311-source.json
Full Text

Full Text

THE ROLE OF VISUAL SAMPLING IN OBSTACLE COMPENSATION by OPAL YVETTE VffiNNEAU B.P.E., The University of British Columbia, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES School of Human Kinetics We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 1997 © Opal Yvette Vienneau, 1997 In. presenting this thesis • in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I .agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the 'head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. * Department "of \r\ixr^\ar\ }<^t rl Z-H'c^^ The University of British Columbia Vancouver, Canada Date \j^2-r^M^J c P / ^ ^ -DE-6 (2/88) 11 ABSTRACT Most obstacle avoidance studies have identified the importance of visual information in choosing obstacle avoidance strategies and of kinesthetic information for successful clearance of obstacles. The intent of this investigation was to determine on what and how visual information is utilized when an obstacle is encountered in our travel path and on some of the subsequent gait adaptations made to accommodate it. Accommodation strategies are based on movement goals and, in the case of this experiment, the goal was to step up onto a platform of set height and width but placed at varying distances from the subject's start position. Twelve volunteers completed seven walking trials for each of four obstacle placement conditions. The kinematic properties of their gait and the corresponding eye movements were recorded and analyzed to determine whether there was an inherent pattern in accommodating obstacles. Subjects were found to slow down in mid path and make changes to their stride length. Smaller adjustments were applied within the last two strides before the obstacle. Stride length change was the primary method of adaptation. Subjects focussed primarily on the obstacle throughout the trials, however gaze time on the obstacle decreased as the subjects approached the point of step-up. The normal condition, set at a natural right-foot step-up position, presented the least change in stride and exhibited patterns similar to those observed in the control situation where there was no obstacle. The greatest effect of stride was observed for the close obstacle placement, while for the far obstacle placement participants displayed similar patterns that were less pronounced. The results of the present experiment indicated that multiple visual sampling was necessary in all strides to apply adaptations to gait. iii TABLE OF CONTENTS Eage Abstract ii Table of contents iii List of tables v List of figures vi Acknowledgement vii Dedication viii Introduction 1 Method 7 Subjects 7 Task and apparatus 7 Experimental conditions 12 Procedure and design 12 Data Analysis 14 Statistical Analysis 18 Results: 24 Right stride length 24 Right stride frequency 24 Right stride speed 28 Left stride length 28 Left stride frequency 31 Left stride speed 31 Results (continued) Page Gaze time - preliminary findings 31 Gaze time 36 Non-significant variables 38 Discussion 42 Conclusion 45 References 47 Appendix A: Mean and standard deviations for all dependent variables 50 Appendix B: Pilot experiment 55 V LIST OF TABLES Table Page 1. ANOVA results for all kinematic dependent variables for the last 4 25 strides 2. ANOVA results for all dependent gaze variables 39 3. ANOVA results for all kinematic dependent variables for first stride 39 vi LIST OF FIGURES Figure Page 1. Schematic diagram of Series 4000 Eye tracker system 9 2. Schematic diagram of experimental setup 11 3. Stride layout for kinematic analysis 15 4. Gaze locations identified by eye-line-of-gaze videotape (A, B, C & D) 17 5. Gaze locations identified by eye-line-of-gaze videotape (AD & BC) 19 6A. All gaze codes 20 6B. Gaze areas, without "E" 21 6C. Gaze areas - A and D pooled 22 6D. Gaze areas, AD pooled, BC pooled 23 7. Right - stride length by stride 26 8. Right - stride frequency by stride 27 9. Right - stride speed by stride 29 10. Left - stride length by stride 30 11. Left - stride frequency by stride 32 12. Left - stride speed by stride 33 13. Gaze time BC by stride 3 7 14. Gaze time AD by stride 40 15. Gaze time E by stride 41 ACKNOWLEDGEMENT Vll Throughout these past years many people have contributed in one way or another to the successful completion of this thesis. First and foremost, would like to thank my graduate advisor, Dave Sanderson, for sharing his thoughts and enthusiasm in the area of biomechanics. His continual support and high standards inspired me to accomplish what seemed impossible and present a project that I am pleased to have completed. I would also like to thank my other committee members, Ian Franks, Digby Elliott and Tony Hodgson for their sincere interest and support, and for their helpful advice throughout the completion of my research. I appreciate the help and support I received from staff and from my peers at the University of British Columbia who have made my life as a graduate student an enjoyable, memorable experience. I am pleased to acknowledge the graduate secretary, Carole Loughren, who always seemed to have the answers. My appreciation extends further to Paul Nagelkerke for his technical help in the experimentation and to Liu Yuanlong for the quick statistics recap. The old lab rats, Michelle Johanssen, Alec Black, and Peggy McBride, all shared their friendship, support and intellect throughout the completion of research papers and oral presentations. And to the newest lab rats, Kelly Moore and Dale Tiessen, who have made my last year all that much more enjoyable, I extend a warm welcome and best wishes. I would like to extend my deepest gratitude to everyone at Curve Lake First Nation, without whose support I could not have been here. In particular I would like to extend my special thanks to Gayle Taylor and Lori Jacobs for their patience and understanding when fielding all the questions of a stressed student. It is not an easy job and the effort has been greatly appreciated. Of course, I would be remiss were I to fail to mention the great support extended to me by my close friends and family, who, I am sure, at times thought I had dropped off the face of the earth. And finally to my husband, Eric, for his graceful patience and understanding, and for reminding me each day that there was a light, however dim, at the end of this long road. Thanks for being there. To my parents, who accepted nothing less than what I was capable of. INTRODUCTION 1 The versatility of legged travel in animals, and therefore humans, is clearly expressed in their ability to negotiate obstacles within their path and to adapt to non-uniform terrain. Apart from reading, independent mobility is one of our most significant skills for occupational and social well being and often this ability goes unappreciated until a loss of or damage to our visual or motor functions compromises it. Although there is abundant information that describes the motor requirements for mobility, visual requirements for safe locomotion are poorly understood. Most early studies of locomotion concentrated on the basic features of gait, and were conducted on flat, uncluttered surfaces. However, even where level ground provides secure, adequate footing, precise control is still necessary to avoid falling. Walking and running require a shifting of the body's center of gravity towards each foot in turn. Balance becomes disrupted if the center of gravity is shifted too much or too little in any direction or if unexpected forces act on the body, as occur when the foot strikes an uneven, soft or loose surface. And what happens when we encounter an obstacle in our path? The skill of negotiating obstacles during everyday locomotion inherently involves "controlling the dynamic flux of the body parts relative to each other so as to control the dynamic flux of the body as a whole relative to the environment" (Lee and Lishman, 1977). Requirements for successful locomotion within any environment involve production of a basic locomotor rhythm that is able to support the body against gravity and propel it in the intended direction. Basic locomotor patterns are adequate for straight paths with level, even surfaces, but for travel over variable terrain these locomotor patterns must be changed. Research has provided a variety of explanations for and about these movement adaptations. Investigators have addressed the importance of the vestibular system for balance and postural control (Nashner, 1980; Patla, 1986; Georgopoulos and Grillner, 1989; Patla, Prentice et al., 1994) and the visual system for environmental feedback (Gibson, 1958; Baumberger et al., 1992; Patla and Rietdyk, 1993; Patla, Rietdyk et al., 1994; Patla, Prentice et al., 1994; Berg, Wade and Greer, 1994). Adaptational strategies are goal-dependent and the requirements for achieving these goals will often involve stepping over, onto or around an obstacle. Recent studies have described changes to locomotor patterns at the kinetic and kinematic levels during the presentation of obstacles. The results of this research has shown that strategies, or movement options, are used to go safely over obstacles of different sizes and shapes and that it is possible to implement obstacle avoidance strategies within the same step cycle in which the obstacle is presented. In contrast, research has also found that when subjects were asked to go around an obstacle as an avoidance strategy they were unable to change direction in the on-going step (Patla, Prentice, et al., 1991; Spaulding and Patla, 1991; Patla and Rietdyk, 1993). Experimenters concluded that step length regulations required alteration of vertical and anteroposterior impulse components (Patla, Robinson, et al, 1989) and that changing direction required an additional change of the mediolateral impulse component (Patla, Prentice, et al, 1991). From these findings it was felt that to change direction a more complex order of neural and motor interaction was needed than was used to coordinate step length changes Thus adaptive movements necessary to change direction required more time to plan and implement. Although obstacle avoidance strategies are important, many of our daily activities require our ability to step onto objects to attain our goals, be they to step onto a stool to change a light bulb, to climb stairs or to step off the street onto a curb. Research has shown that subjects had the ability to accurately determine the dimensions of obstacles and to change their limb trajectory in order to provide an adequate margin of safety. In one case, an average toe clearance of approximately 10cm was observed while subjects stepped over the obstacles provided (Patla and Rietdyk, 1993; Patla, Rietdyk et al., 1995) and obstacles up to 8cm in height were successfully cleared within the time constraint of one step cycle (Patla, Prentice et al., 1991). 3 Humans guide their activities based on environmental attributes and knowledge of their own action capabilities. This implies that we are capable of perceiving an object's functional properties based on an intrinsic measurement system used for analyzing visual information and perceptual motor control. In deciding which accommodation strategy to use, subjects appear to make choices based on body-scaled environmental features. This was demonstrated in Warren's (1984) affordance study of subjects climbing a staircase of variable riser heights where stairs were perceived climbable in relation to individual subject's leg length. In an study of subjects going through doorways of different widths subjects were found to twist their upper body while going through an opening when the ratio of opening width to shoulder width reached a critical threshold (Warren and Whang, 1987). Another experiment by Patla (1995) found that subjects, when given the choice to go over or around obstacles, chose to step over obstacles that were equal to or shorter than the subject's lower leg segment length. Although these authors agree on the importance of visual information to locomotor control, there is little research that directly addresses the nature of visual information. Gaining information about the visual requirements of common activities is important to our understanding of skilled locomotor behavior and is the focus of the present study. Most findings in the area of adaptive locomotion are quite recent and there is still a great deal we do not know about how humans adapt to changes in terrain. How and when do people alter their gait as they approach an obstacle, for example, a curb, and what changes or alterations do they make? Pertinent cues about our environment are sampled so that goals may be reached by modification of gait (Corlett, 1992). Our visual processes guide these modifications during locomotion. It has been shown that visual input provides information about the obstacle such as size and surface characteristics (Patla and Rietdyk, 1993). It also provides information concerning self-motion and an individual's dynamic relation to their surroundings. This 4 information is used to determine time-to-contact with the object (Lee and Young, 1986). This aspect of visual input has been exhibited in studies of the approach in long jump. Strides taken early in the approach phase of the jump were uniform in length and possessed little variance, when compared from jump to jump. The last 5 strides showed much greater variability from trial to trial, demonstrating that these strides were being were used to compensate for errors in position that were produced in the earlier strides. The final stride produced the greatest variability and was evidently under visual control. It was concluded that jumpers use visual information to modulate step length near the final approach phase in order to enhance take-off accuracy (Hay, 1988; Lee, Lishman and Thompson, 1982). Recently, Vienneau and Sanderson (1996) investigated the existence of similar locomotor patterns in response to an obstacle. The experiment was set up to simulate the actions of someone walking across the street and stepping up onto a curb. Subjects were asked to walk down a prepared path and step up onto a platform of standard curb dimensions. From the data collected, no changes were observed between the first stride and the last stride before the subject stepped up onto the platform and it was inferred that gait adaptations were incorporated during the first and last stride before meeting the platform. This supported the findings by Lee, Lishman and Thomson (1982), Jeannerod and Prablanc (1983) and Lee and Young (1986) that movement is first made in the general direction and then is fine-tuned later, closer to the obstacle. When visual input is reduced, locomotion is compromised. The characteristics of visually guided locomotor pattern adjustments have been a focus of investigation (Warren, Young and Lee, 1986; Patla, Robinson et al, 1989; Patla, 1991; Patla, Prentice et al, 1991). Results showed that, even for an irregular path, intermittent sampling (in this case, less than 50% of the travel time) of the environment was sufficient for safe locomotion. Thomson (1980, 1983, 1986) demonstrated the intermittent nature of visual sampling in guiding locomotion to a specific target location by testing subject memory over various target distances with no vision delays. 5 Subjects were successfully able to estimate short target distances, suggesting that visual sampling need only be intermittent based on subject memory. The present experiment extended this investigation of intermittency by analyzing the effects of obstacle distance on visual sampling and locomotor kinematics. An interesting debate ensued when Elliott (1986,1987) and Steenhuis and Goodale (1988) successfully challenged the constraints placed on Thomson's intermittency experiments. The debate forced a re-evaluation of the mechanics of memory in association with task performance. It was concluded that continuous visual sampling was not necessary in order to achieve relatively accurate targeting results. Based on these conclusions, the present experiment attempts to identify characteristics of visual sampling during goal-directed locomotion. By looking at precise lines-of-gaze and correlating them with specific modes of response, more insight would be gained into the specific role of the visual system during everyday activities such as walking. Particular factors that may be addressed include: duration of gaze, number and spatio-temporal location of visual samples taken during the performance; scanning region necessary (spatial domain of terrain perceived); and how far ahead we look (while walking) in order to safely implement changes to gait patterns. Such information would not only enhance our understanding of control of locomotion but also provide insight into addressing treatment strategies for increasing the mobility of visually impaired individuals. For the present experiment, interest was not simply on physical adaptations as it was in the pilot experiment, but also on how vision was involved in the adaptation process. The purpose was to determine whether characteristics of visual sampling and locomotor kinematics were affected by the distance between the subject and a platform placed within their path of locomotion. The experimenters analyzed all strides from when the platform was first seen to when the subject stepped up. All eye movements were monitored while subjects performed the walking task. For the task, the platform, representing a double-height curb, was placed in the path and the subjects were given specific instructions on how to respond to it (i.e. step up on it). Locomotor changes including step-length, stride duration, stride frequency and eye-line-of-gaze were monitored according to the placement of the curb in relation to the subject's starting position. This provided information on (1) visual sampling, or, where, when and how often the subject targeted the obstacle or their surroundings before modifying their gait to avoid tripping on it, and (2) the nature and time at which locomotor changes were applied. Based on the pilot experiments and on the findings in literature it was hypothesized that obstacle presence would affect gait adaptations and corresponding eye-line-of-gaze. Specifically: (1) visual sampling would occur in the first step after the curtain is opened, and in the last step when the subject stepped up onto the platform, (2) multiple visual samples would not be necessary for gait adaptations, and (3) any kinematic alterations would be observed only in the first step and in the last step during the step-up onto the obstacle. In addition, kinematic and visual results were investigated for relationships. Based on the pilot experiment and the literature presented here it was expected that subjects would produce all stride changes near the platform, with no alterations in the first steps. It will also be expected that, based on the simplicity of the task, multiple visual samples would not be necessary for safe clearance. 7 METHOD Subjects Twelve volunteers between the ages of 20 to 35 years participated as subjects in this experiment. Although there were no limitations on height or weight, subjects had to be in good physical health and possess no physical or sensory disabilities that could possibly affect normal walking. Excluding all subjects with self-reported medications and health problems minimized any variability that might have affected results. Eyeglasses and hard contact lenses interfered with the eye-tracking equipment and therefore were not acceptable for this experiment. All subjects were naive to the hypothesis being tested and were inexperienced at the walking task. The experiment was carried out according to the ethical guidelines laid down by the University of British Columbia Behavioral Sciences Screening Committee for research and other studies involving human subjects. Task and Apparatus The experiment was conducted in a small gymnasium that afforded a 14 m long pathway and a moveable platform at its end. This length allowed ample number of strides to be viewed before the subject reached the platform. To increase experimental effect the platform was built to replicate the dimensions of a double-height curb (See Appendix B for pilot experiment using standard curb height). A curtain was set up 4.5 m from the starting point to block the subjects' view of the platform's placement. This was to be sure that the subject was in full motion when the platform placement was revealed, as well as providing the experimenter with a definitive time at which the subject was able to see the target. Gait changes could then be corresponded to this time. The curtain was quickly opened when the subject reached the 3 m mark of each trial. The curtain rod was attached to the wall at a height of 2.3 m and extended 2.5 m out from the 8 wall. This placement ensured that camera view would not be affected and the subject could walk by without interference by. Each subject was instructed to walk at a comfortable pace down the prepared path and, if the platform was present, to step up onto it only with his or her right foot and continue walking to the end of the path. Subjects had to start at the same starting point and use the same starting foot for each trial. To control speed, two sets of photocells were placed in the walkway at 4 m and 11m. The first set triggered a digital timer to start and the second set triggered the same timer to stop. By keeping the walking over this 7 m to within 0.5 sec of each subject's average comfortable walking speed, overall speed was kept relatively constant. Subject eye movements were monitored during the task using Applied Science Laboratories' (ASL) Series 4000 Eye Tracking System. Gaze data were collected using the mobile ASL 4000SU head mounted eye tracking system, a monocular corneal reflection system that measured eye-line-of-gaze with respect to the headset. An infrared light beam illuminated the eye and the optical system focuses the resulting eye image onto a solid state video sensor. The light source and image were reflected from a headset visor that is specially coated to be reflective to infrared light and transmissive to visible light. A camera lens captured the light beam that is reflected back from the retina and the pupil image was displayed on the sensor. The light-source, optics, solid-state sensor, relay lens and visor were mounted on a headband. The complete head assembly weighed 885 g. The camera control unit for the eye tracker weighed 900 g and was carried in a small fanny pack that was fastened about the subject's waist. The fanny pack and 9 m umbilical cord permitted normal mobility for the subjects. This unit collected information from the eye and scene cameras and relayed the information along the umbilical cord to the main control unit and accompanying PC computer for further processing and display (see Figure 1). Once the signal was processed the computer calculated pupil diameter and eye-line-of-gaze. From this information pupil and corneal reflections were displayed on the pupil monitor and eye-10 line-of-gaze with respect to head was displayed on a second monitor as a set of crosshairs superimposed on the scene camera image to show precise point of gaze. This image was duplicated on a third monitor where the image was recorded and later processed. Sampling and output rate for the eye tracker was 60 Hz. Total transport delay from input to output was 3 video fields, or 49.95 ms. Spatial error between true eye position and computer measurement was less than 1 degree, increasing to less than 2 degrees in the periphery of the visual field. Kinematic data were recorded during the entire experiment using the Peak Performance Video Analysis System (PEAK). Reflective markers were placed on each foot and each trial subsequently videotaped. The markers were made with 1-inch Styrofoam balls, which were cut in half and covered with 3M reflective tape. To record the experiment, three stationary video cameras were set up in series 7 m from and parallel to the prepared path and three 3 75-watt lamps were used to illuminate these markers. Two Panasonic WVD5100 VHS and one portable Panasonic AG195MP VHS movie cameras produced video signals that were recorded by Panasonic AG7300 and AG7400 VCR's. The portable camera was able to record its own signal. All signals were stored on Fuji ST 160 Super VHS videotapes. Horita TCG-1000 time code generators produced the audio LTC (longitudinal time code) and its associated video window and attached them to each video frame during the recording process. Although the eye movement videotapes did not require digitizing, a mixer was used to duplicate the video time code from the one of the cameras and place it on the eye movement videotapes. This would allow time synchronization of the kinematic data with the eye movement data. Figure 2 shows the schematic diagram of the experimental setup. 11 Figure 2: Schematic Diagram of Experimental Setup Peak PC-Event ^ Sync Time code^  generators Mixer Eye monitor and VCR Monitor Camera 3 Eye tracking equipment Camera 2 Camera 1 (portable) Photocells - setl Time Curtain Photocells - set2 4 Curb obstacle 12 Experimental Conditions There were 4 experimental conditions, 3 of which simulated various curb-to-subject distances and 1 control condition (condition 1) in which the platform was not present. Condition 2, the close condition, was equal to left foot step-up, meaning the platform was placed in a position where, if the subject maintained his/her normal walking pattern, he/she would have stepped up onto the platform with their left foot. But, since step-up was permitted only with the right foot, subjects were forced to make some adjustments to their gait in order to obey the set criteria. Condition 3, the normal condition, had the platform placed at right foot step-up, where the platform was placed in a position where, by maintaining normal walking, subjects would have stepped up on the platform with their right foot and easily followed the set criteria. Condition 4, the far condition, had the platform placed at next left foot step-up where the platform was placed much like it was for the close condition, only 1 stride further away. All subjects completed each of the four conditions. Procedure and Design Each subject attended one testing session that lasted approximately two hours. The session began with the experimenter providing the subject with a description of the experiment and obtaining informed consent. Anthropometric measures were then taken of lower limb segments for use in subsequent analysis. After putting on the eye tracker headset and "fanny-pack" the subject was permitted ten obstacle-free practice trials, not only to become comfortable with the equipment but also to determine which foot they would prefer to start with and to establish a comfortable walking pace. The experimenter used this time to determine the subject's step length as they walked down the travel path. This information was then used to determine the placement of the obstacle for conditions 2, 3 and 4. The obstacle was placed near the end of the travel path at a distance equal 13 to each individual subject's step length to correspond with left foot contact, right foot contact and the next left foot contact. Upon completion of the practice trials reflective markers were placed at the following anatomical landmarks: right heel, right lateral maleolus, right fifth metatarsal, left heel, left medial maleolus and left first metatarsal. When the subject was ready and all equipment at appropriate settings it was time to calibrate the eye tracker. To do this, nine calibration points were placed in grid fashion on a wall at an even spacing of 50 cm. The subject was seated 1.5m from the wall with their chin resting on a support post to stabilize the head. Once comfortable, the experimenter adjusted the optics module, visor and scene camera to the desired position. Locations of the target points were entered into the computer using the ASL software. To calibrate the optics module with the scene camera, the subject was asked to look at each point. When the crosshairs corresponded with the desired point the computer was triggered and information stored. The process was repeated for all nine points and the calibration was complete. The subject was positioned at the start point, all the VCPvS were set to record and the curtain was drawn across the path. In addition, head movements that naturally occur during locomotion made it necessary to check calibration after each trial to ensure accuracy of the results. Using a fixed starting point and leading with the previously chosen limb, the subject walked down the path at their selected pace toward the curtain 5 m away. Upon reaching the 3 m mark the curtain was quickly drawn away and the path was in full view. The subject continued without stopping to the end of the path and stepped up with their right foot leading if the obstacle was present. Subjects completed seven test trials in each of the four conditions over a period of 90 minutes and the order of these trials was randomly chosen for each subject. After each trial was completed, the experimenter checked calibration of the equipment to ensure the headset had not 14 moved. This was done by having the subject watch and follow a point of light emitted by an infrared pointer held by the experimenter. If the crosshairs on the scene monitor accurately followed the point of light, the equipment was considered calibrated. As mentioned, speed was monitored using a digital timer. Trials in which the subject did not cover the 7m in the established time, (+/- 0.5secs), or in which the subject stepped up onto the obstacle with their left foot were discarded. Trials not completed in the accepted time range were repeated. Data Analysis Obstacle accommodation was measured by stride length, stride frequency, stride speed and gaze position. These measures represented types of gait adaptations that were modified in previous studies of obstacle accommodation. All variables had the same start point and end point and were based on the presence of the obstacle and its distance from the subject. In addition, all analyses were based on information obtained by strides, specifically, the first stride and the last four strides. Figure 3 displays the layout of these strides. Note that the first stride refers to the first stride after the curtain has been pulled and that the last four strides have been numbered backward from the curb, with stride 1 being the stride from right heelstrike on the floor to right heelstrike on the curb. Kinematic Variables The videotapes were first digitized using PEAK to produce trajectory plots for the markers on each of the right and left feet. Heelstrike event codes were recorded for each trial during the digitizing process. The final raw data were edited for outliers and conditioned using a Butterworth filter with a cutoff frequency of 6 Hz. The next stage of analysis required calculation of linear displacement of the foot from the foot markers, again using Peak. Upon completing this step, further analysis was done using a spreadsheet program (Microsoft Excel) to 16 manipulate data. As described, there were three cameras set up in series to record each trial, thus one data file existed for each camera and for each trial. To analyze complete trials it was necessary to link all three linear displacement files for each corresponding trial. To do this, each partial file was appended to its partner by its start and/or ending frames to produce one complete trial data file of linear displacements, including heel strikes and total elapsed time. These appended files were then filtered, again using the Butterworth filter, but this time at 10 Hz. This frequency was chosen primarily for any minor extremes produced by peripheral camera views. This was a conservative cutoff frequency thus filtered out only the outliers produced by appending the data files. The rest of the data were not affected. Finally, an in-house program module was executed to compute and display the length, frequency and speed of each stride for all subjects and trials. Gaze Analysis Five gaze locations were identified by the line-of-gaze videotape and are presented in Figure 4. Area B refers to the top of the curb, area C refers to the front panel of the curb, area D refers to the floor between the subject and the obstacle, and area A refers to all other visible positions. Area E identified those gaze positions that were 'out of range' and was used when a subject's eye-line-of-gaze had moved beyond the visor system. This event occurred when subjects looked directly down at their feet, to their immediate right or left, or directly up without moving their heads to do so. The five gaze areas were exhaustive so all lines-of-gaze were coded. The total time for each gaze code was plotted for each stride as a function of % stride. The goal of this experiment was to determine when visual sampling appears to be used for obstacle accommodation strategies. To accomplish this goal a number of analyses were necessary. For the analysis, gaze times for 'E' (off-screen) were compared to the gaze times for the other locations. Because E represented gaze areas that were beyond the eye tracker's recording abilities, no relevant comparisons could be made at this stage of the analysis. As a 17 Figure 4: Gaze Areas Identified by Line-of-Gaze Videotapes A Surroundings / B \ A / Top of \ / platform \ A Surroundings / \ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ Front of platform Surroundings D Floor, between subject and platform Note: Gaze location 'E' represents areas that are not in this visual field 18 result, all other comparisons were completed without E. A number of manipulations were performed on the remaining data to extract as much useful information as possible. Because A and D represented the surrounding area about the curb, these were pooled to form AD. Using graphic representation, AD gaze time was then compared with B and C. Finally, B and C positions were pooled to form BC and the results compared with AD (see Figure 5). Figures 6A to 6D show a summary of the phases of analysis. Although all gaze data sets were observed, it is on this final set of data that the statistical analysis is based. Statistical Analysis All kinematic variables were analyzed using a 2-way, condition(4) x stride(4) repeated measures (RM) ANOVA on the last four strides, and a 1-way RM ANOVA on the first stride of each condition. Gaze variables AD, BC and E were analyzed using a 2-way, condition(3) x stride(4) RM ANOVA on the last four strides and a 2-way, condition(3) x gaze (2) on the first stride of each condition. The significance level for each analysis was set at p<.05. The control (CI) condition was not included in the gaze analysis. 19 Figure 5: Gaze Areas Identified by Line-of-Gaze Videotapes AD Off- target AD Off-target BC On-target BC On-target AD Off-target AD Off-target Note: Gaze location 'E' represents areas that are not in this visual field ^ s WD o « .S € -2 •o « o. S a, «a a _ 8 j 8 i c o << ffl U Q W _ m M ca n o 5 3 S ^ g © *< ^ S vo CZJ O M S3 M E ""7 ¥ * Q H ON © 00 © t-© © © © © © <S iH aaraxs % 21 ? ? © <*> DC u o a • MM TJ a s "3. Q< o (surroi (top of (front (floor) < CQ U c • • m IHIliiiifltlsSltl VO Q 2 H © © © © 00 © © vo © I T ) © © © aaraxs % "3 © O H a «< V© JJ-H o •a W o H CQ u VO -9 flit ft lljltl 4 If H o o o 00 o o vo o o o © © a a r a i s % a a r a i s % RESULTS 24 Right Stride Length To directly test the hypothesis that no step adjustments would be observed between the first step after the curtain and last step before the obstacle, each analysis was based on individual stride variables for significant difference across four obstacle conditions. A 4 x 4 RM ANOVA performed on right stride length for the last four steps indicated significant main effects of stride (see Table 1 for ANOVA results of kinematic variables). A Scheffe post-hoc test indicated stride lengths for strides 1 and 2 were significantly shorter than they were for strides 3 and 4, particularly for stride 1 (Figure 7). All other effects were not significant (p>.05). Stride length appeared to increase slightly from stride 4 to stride 3, then decreased significantly from stride 3 to stride 2 for all conditions. The control (CI) and close C2) conditions show a similar pattern for stride 2 to stride 1 in that stride length decreases again, while for the normal (C3) condition stride 1 leveled off. The far (C4) condition increased in stride length from stride 2 to stride 1. Factor means and standard deviations for all dependent variables are presented in Appendix A. Right Stride Frequency Right stride frequency data showed a non-significant trend for stride main effects approaching significance (p=.056) (Table 1). In particular, stride frequency remained constant throughout strides 2, 3 and 4 but increased for stride 1 (Figure 8). The overall trend presented no change in stride frequency from stride 4 to stride 3, a small increase from stride 3 to stride 2, and a large increase for stride 2 to stride 1. In addition, significant interaction effects were observed for condition by stride, indicating stride frequency patterns did not remain the same across conditions. The control, normal and far conditions showed the steady increase in stride frequency 25 Table 1: ANOVA results for all kinematic dependent variables (last four strides). All main effects and interaction effects are presented. Effect F-test p-value p<.05 Right Stride Length - last 4 strides Condition F(3,33)= 0.94 P = .433 Stride F(3,33)= 7.43 P = .001 * Condition x Stride F(9,99) = 0.77 P = .641 Right Stride Frequency - last 4 strides Condition F(3,33)= 0.88 P = .463 Stride F(3,33) = 2.79 P = .056 Condition x Stride F(9,99) = 2.49 P = .013 * Right Stride Speed- last 4 strides Condition F(3,33) = 0.56 P = .648 Stride F(3,33) = 2.35 P = .091 Condition x Stride F(9,99)= 0.43 P = .917 Left Stride Length - last 4 strides Condition F(3,24) = 1.36 P = .278 Stride F(3,24) = 13.08 P = .000 * Condition x Stride F(9,72)= 1.44 P = .187 Left Stride Frequency - last 4 strides Condition F(3,24) = 2.25 P = .108 Stride F(3,24) = 2.68 P = .069 Condition x Stride F(9,72)= 1.47 P = .176 Left Stride Speed - last 4 strides Condition F(3,24) = 1.65 P = .204 Stride F(3,24) = 8.00 P = .001 * Condition x Stride F(9,72)= 1.11 P = .367 26 XJ •i—< _. ^1 W D ^  •-»< '« +-> i •+•> W) -«3 if) 3t XJ •i—< : _ : © IT) o o OS (ui) q;§u3i apiJis >>\ >*\ 00 § u S3 0D 3 -S3 , DJD 2 i i i i i i i i i i i i i i i i i i i i i i i i I i i i i i i i i i i i i i i O O o o o c o ON oo r f^> n • • • • • T H l - H T H I - H O (aas/sapujjs) ^ Duanbaij apujs described above, with the far (C4) condition having the lowest increase from stride 4 to stride 2 and finishing with a large increase for stride 1. The control (CI) and normal (C3) conditions were nearly identical in pattern with the control 1 frequency only slightly more overall. In contrast, the close (C2) condition began with the same stride frequency as the other conditions, then steadily decreased from stride 4 to stride 2. A Scheffe post hoc test for interactions indicated that the difference between conditions was greatest at stride 1. In addition, the post hoc test revealed that the greatest difference between strides occurred in condition 1. These effects are indicative of the main effects observed for the stride length variable. With the close condition (C2) showing a decrease for both stride length and stride frequency at stride 2, adaptations in gait are evident. All other effects were not significant (p>.05). Right Stride Speed Although no significant effects were observed for right stride speed, the general trend showed strides 3 and 4 remained the same, slowing down for stride 2 then increasing speed again for stride 1, except in the case of condition 1 where speed decreased slightly for stride 1. The greatest decrease in speed was observed during condition 2 from stride 3 to stride 2 (Figure 9). Left Stride Length Results of a 4 x 4 RM ANOVA on left-foot stride length indicated significant main effects for stride (Table 1). Three test subjects did not have a fourth stride to anlayze, therefore the statistical analysis was based on 9 subjects rather than 12. Although there were no main effects for condition, average stride lengths were shorter for stride 4 in comparison to strides 1, 2 and 3, and this was especially true for the close (C2) condition. The general trend saw stride length increase from stride 4 to stride 3, than gradually decrease as the subject approached the curb (Figure 10). A Scheffe post hoc test verified a significantly shorter stride length for stride 4 29 a to E S W) 2 O -S O C8 U U Z h u t i i i i i i i i i i i i i i i i i i i i i i i i i i i i i S CZ5 OS 00 IT) (oas/ui) paads apu;s 30 lit! 31 in comparison to all other strides. All other effects were not significant (p>.05). Left Stride Frequency Although there were no significant effects, a similar pattern was seen for stride 4 as was identified for left stride length (Figure 11). Strides 1, 2 and 3 show no differences between each other except for a mild increase in stride frequency for stride 1, condition 1, and a mild decrease for stride 1 during the close (C2) condition. Left Stride Speed Significant main effects for stride were also observed for the dependent variable stride speed. As with left stride length, chart patterns for left stride speed (Figure 12) indicated a slow speed was used for stride 4, after which speed increased for stride 3, then gradually decreased as the subject approached the curb. A 4 X 4 RM ANOVA verified a significant main effect of stride and Scheffe post hoc test indicated that the means of strides 1, 2 and 3 were significantly greater than the mean of stride 4. All other effects for left stride speed were not significant (p>.05). Gaze Time - Preliminary Findings Statistical analysis for this experiment was performed on gaze data that was presented as AD (off-target) and BC (on-target), however the original data collected for the eye-line-of-gaze portion of this experiment consisted of five gaze measures: 2 on-target measures, 2 off-target measures, and one error measure "E". Data were displayed and analyzed as a function of time and stride for each subject and each condition. Due to the variety of subjects who participated in this experiment the stride times varied considerably and the results were difficult to interpret thus gaze data were first normalized and presented as time as a function of % stride. 32 33 34 The first step in the analysis was to observe the error measure, E, and note any trends. E was coded whenever cursor display was lost. This phenomenon occurred when the eye camera lost pupil/corneal fixation and was attributed to the subject gazing outside the range of the eye tracker, for example, if looking to the far right or left. It was noted that the presence of E increased when the subject approached the curb, at about the same time when the curb became partially out of the subject's view. Based on cursor activity observed on the videotapes, for most of E subjects were looking down at his/her feet, the floor or at the platform. The first stage of data observed that the greatest incidence of E appeared in the last 2 strides (strides 1 and 2), and that the least incidence appeared in strides 4, 5 and 6. In fact, for stride 1 all but 2 subjects displayed this trend. For each of the platform conditions, most (8 of the 12) of the subjects displayed this trend. Two subjects presented with virtually no E gaze time. Most of E was distributed normally across strides for the control (CI) condition in comparison with the platform conditions, with a variation of approximately 10%. Condition 3 was set up to demand the least effort by the subject and, therefore, should have presented with the least incidence of E. This was not the case, in fact, 6 of the 12 subjects in condition 2 presented with less error when compared to condition 3, while 3 subjects presented with more. Most of these observations took place during strides 1 and 2, closer to the curb. One subject demonstrated no E observations whatsoever. In summary, the experimenter concluded from these notes that E was greatest for steps 1 and 2 for all conditions and slightly more for condition 4. Otherwise, E was distributed normally. With the observations of E complete, the data were set aside and the remaining measures were observed. (See Appendix A for gaze times for E represented as % stride.) For the next step in the analysis, times for gaze areas A, B, C and D (see Figure 4) were compared and observed for trends. A and D observations were found to be evenly distributed within conditions and the majority of D observations occurred at the far end of the path in the 35 area of the curb. Incidence of D was least for strides 1 and 2, possibly because there was less area of D to observe. D presented the most during the first couple of steps after the curtain opened for all but 2 subjects, while for one subject, D was distributed evenly across all conditions. For six of the subjects, observations of D decreased as they neared the curb, while for 2 others D remained the same across all strides. For one subject, condition 3 had the greatest incidence of D while for one other it was condition 2. In short, D was greatest at stride 4 and decreased as the subjects approached the curb, possibly because area D became smaller as the subject drew nearer to the obstacle. Based on a visual analysis only there were no other consistencies across subjects, conditions or trials. Because the focal point of this experiment was on the platform, A and D were next pooled to form one gaze time for area AD (see Figure 5) then compared with gaze times for areas B and C. As was expected, condition 1 distributed evenly across strides and was therefore not required for this or the remainder of the analyses. In comparing gaze time across conditions and strides, the main observation was that AD did not play a large part in gaze time. For at least 6 of the 12 subjects, C made up a large part of the stride time, and for 3 other subjects, B made up the majority of the stride time. For the remaining subjects there did not appear to be a pattern to any condition or stride. Although C appears to be prominent in all steps, there does not appear to be a common trend across conditions. C was especially prevalent in the first step when the curtain had just cleared the path. In fact, during the data recording it was observed that even while the curtain obstructed their view, subjects often looked in the direction of the obstacle. For the next stage of data manipulation, gaze measures B and C were merged and the new data, deemed ' B C , was compared and contrasted with gaze AD (originally gaze A and D). Gaze BC was present in all but one step and appeared to make up more of the stride than gaze AD. From the accumulated observations the experimenter has concluded that subjects target the object throughout the task and look very little in areas other than the curb. Thus, visual sampling 36 was occurring for the entire task. All of these observations were strictly preliminary and provided the experimenter with a glimpse of the real results demonstrated in the final statistical analysis. Due to the conversion of time to % stride it was necessary to confirm prior to statistical analysis that the final data were normally distributed. Normal distribution was confirmed using a logarithmic comparison. Where necessary, Huynh-Feldt Epsilon was used for adjusted degrees of freedom. Gaze Time To gain more information about the effects observed for gaze time, further statistical analysis was performed on gaze areas (Table 2). A condition (3) x stride (4) RM ANOVA for gaze time BC indicated a significant main effect for stride, and a Scheffe post hoc test indicated that the means for strides 1 and 2 were significantly less than the means for strides 3 and 4 (Figure 13). In addition, a significant condition by stride interaction effect showed the normal position (C3) and far position (C4) decreased from stride 3 to stride 2 and stride 1. But for the close condition (C2) gaze times do not change nearly as much. General trends indicated no change in gaze times for the normal (C3) and far (C4) conditions for strides 3 and 4, while gaze times decreased significantly for strides 1 and 2, especially for the normal (C3) condition. The close (C2) condition began with the same gaze time as the other conditions, decreased slightly, then leveled off. In general, the results indicate that for the obstacles that were further away, subjects required less visual information because they knew they would be able to step up with their right foot without problem. For the normal platform placement, it appeared subjects knew by stride 3 that they would be able to step up without much difficulty. The far condition required a little more gaze time, but by stride 2 subjects were more confident. The close condition placed a little more demand on the subject, being that the platform-to-subject distance was less and that 37 38 less strides were available for the subject to incorporate the necessary changes to step up with the right foot. As a result, subjects needed to obtain more visual information. A 2-way, 3x4 RM ANOVA for gaze AD data indicated no significant effects (Figure 14) for gaze time in position AD. Gaze time remained constant across all strides and conditions with only a marginal increase during the normal (C3) condition for strides 1 and 2. A 3 x 4 RM ANOVA performed on gaze E data identified significant main effects for stride. A Scheffe* post hoc test identified strides 1 and 2 as having more overall error gaze time than strides 3 and 4 (Figure 15). Trend lines indicated no difference between strides 3 and 4 overall, and no difference between conditions for stride 2. The close (C2) condition again is identified with a slight deviation from the normal (C3) and far (C4) conditions, appearing to level off for stride 1 while the normal (C3) and far (C4) conditions increase. Non significant variables Each of the dependent variables was tested again for effects on the first stride (after the curtain was clear). Seven 1-way RM ANOVAs were performed on the 'first stride' data, all had no significant effects (Table 3). No significant effects were seen in the kinematic data for the first stride, nor were there any significant differences between the gaze times for each of the gaze areas. Note that subjects do, however, look at the obstacle in the first stride. These results are consistent with the pilot data and the results of Lee, Lishman and Thomson (1982) that subjects first gather information then make changes when we get closer to the target. Table 2: A N O V A results for all gaze data. A l l main effects and interaction effects are presented. Effect F-test p-value p<.05 Gaze Time A (2-way) - last 4 strides Condition F(2,22)= 2.23 p = .131 Stride F(3,33) = 0.04 p = .988 Condition x Stride F(6,66) = 0.92 p = .489 Gaze Time B (2-way) — last 4 strides Condition F(2,22) = 2.63 p = .095 Stride F(3,33) = 8.25 p = .000 * Condition x Stride F(6,66) = 2.54 p = .028 * Gaze Time E (2-way) — last 4 strides Condition F(2,22)= 0.78 p = .471 Stride F(3,33)= 8.89 p = .000 * Condition x Stride F(6,66)= 1.46 p = .205 Gaze Time -first stride Condition F(2,22)= 0.27 p = .765 Stride F ( l , l l ) = 0.93 p = .355 Condition x Stride F(2,22) = 0.03 p = .974 Table 3: A N O V A results for all kinematic dependent variables (first stride). A l l main effects and interaction effects are presented. Effect F-test P" value p<.05 Right Stride Length —first stride Condition F(3,33) = 0.71 P = .554 Right Stride Frequency -first stride Condition F(3,33) = .058 P = .635 Right Stride Speed -first stride Condition F(3,33) = 0.27 P = .847 Left Stride Length - first stride Condition F(3,33) = 1.65 P = .197 Left Stride Frequency —first stride Condition F(3,33) = 2.10 P = .118 Left Stride Speed-first stride Condition F(3,33) = 1.83 P = .160 40 \ i J I I L I I I o o o (apu;s o / o ) 3UIi X 41 DISCUSSION 42 The purpose of the experiment was to examine the role that vision plays in the guidance of locomotion, specifically, to monitor gait adaptations and the corresponding visual sample patterns in response to an obstacle's distance from the observer. Measurable effects on visual sampling were evident and gait kinematics when subjects were required to negotiate a raised platform during normal walking. The effects were seen as changes in stride length, stride frequency, stride speed and gaze time. No termination criteria were set (i.e. subjects were instructed to step up onto the platform, continue to the end and step off). The success rate of obstacle clearance was 100% for all of the conditions and the ability of the subjects to clear the obstacle was not affected by obstacle placement. From the opening of the curtain, subjects first gathered visual information. This initial sample provided the subject with details about their position relative to the platform. Subjects were then able to decide whether a response was necessary and how immediate a response would be required. These observations were in keeping with the findings by Patla and Rietdyk (1993) who found that the adaptive strategies that were chosen were based on the characteristics of the obstacle. Subjects were able to accurately determine the dimensions of obstacles and to actively change their limb trajectory to provide for a small safety margin when going over the obstacle. Patla (1991) had also found that subjects were able to incorporate changes to their gait within one step cycle. In the present experiment, visual sampling was evident for each stride. No stride adaptations were made for any of the platform placement conditions when the curtain was pulled, thus, with Patla's findings in mind, it was concluded that the subject had decided, based on the visual information just gathered, the platform was still far enough away that reaction was not yet necessary. 43 Visual information from the platform was taken for at least 50% of each of the first few strides, and reduced to approximately 35% for the last 2 strides. This information was not taken continuously by the subject, but was intermittent within each stride. From the findings of Thomson (1980, 1983, 1986), Elliott (1986,1987) and Steenhuis and Goodale (1988), the subjects in this experiment gathered information, remembered what they saw and looked away while continuing on the same course. The subjects then traveled based on what they had just seen and looked back at the platform when they felt more information was needed. An example of such information included how much time they had left before they would come into contact with the obstacle. Information about time-to-contact is obtained by optical flow (Lee and Thomson, 1982), a visual flow which runs in the opposite direction of the subject's direction of movement. The rate of dilation of the platform on the retina as the subject moves closer provides the subject with direction and velocity information, and therefore time-to-contact. If after a few visual samples had been taken and little change in terrain was seen, the subjects realized that the terrain was not difficult and did not require as much visual monitoring as was first performed. This was especially true for the normal and far conditions. The greatest change in on-target gaze time occurred when the platform was in its normal placement. In this position, subjects did not need to make large gait adjustments to obey the right-foot step-up criteria. Subjects were also observed to spend more time looking outside the range of the eye tracking visor as they approached the platform. This did not necessarily mean they were not looking at the platform. In fact, for much of the out of range gaze time, subjects were looking down and this was implied by the activity of the eye-line-of gaze indicator. It could not be determined exactly where the subject was looking, just that they were looking down in the direction of the platform, the floor and their feet. Subjects looked longer at the platform during the close and far conditions, especially in the last 2 strides. When right-foot step-up was awkward, subjects required more visual 44 information about the platform to determine appropriate adaptations. This was especially the case for the close platform condition because subjects had less time and strides to incorporate adjustments. Because the platform was closer, the subject had less time to decide what to do and incorporate the necessary changes. As a result, kinematic changes in gait were greater in magnitude and required more visual feedback. Finally, to adapt to the platform, subjects made the initial stride adjustment at approximately mid-path, particularly for the close and far platform placement positions. Subjects slowed down when they came to within 4 strides of the platform. At this time, they felt their stride position was incorrect, checked their stride, made the appropriate adjustments then resumed their regular stride pattern. These initial changes were achieved only by adjusting stride length. This is in keeping with the findings by Patla et al. (1991) and Patla, (1996) that for most obstacle accommodations strategies, change in stride length is the most common form of adaptation. Like the long-jumpers of Lee, Lishman and Thomson's study, when subjects neared their goal, or in this case were within 2 strides of the platform, they fine-tuned their approach to ensure correct foot placement. Again, stride length was the primary method of adaptation. CONCLUSION 45 Based on the pilot experiment and on the literature, hypotheses were made at the outset of the study stating that multiple visual sampling would not be necessary for gait adaptations, that visual sampling would occur in the first and second last step before the obstacle, and that no step adjustment would be observed between the first and last steps before the obstacle. The data presented in this study showed visual sampling of the platform for the majority of each stride and for every condition, thus the hypothesis was maintained for sampling in the first and last strides. Furthermore, stride adjustments were not only observed in the last two strides, but also for the fourth stride before the platform (mid-path), and therefore the hypothesis was rejected that no step adjustments would be observed between the first and second last strides. The main goal of the experiment was to provide some insight into the role of vision in obstacle negotiation. In general, intermittent changes in stride length were observed throughout the task, particularly in the middle of the path and slightly more when the subject had almost reached the obstacle. Moreover, except to step up onto the curb, variation in stride frequency was not a method of adaptation used by the subjects to negotiate the obstacle that was provided. Any observed changes in speed were a result of the change in stride length, therefore subjects were observed to slow down in the middle of the path, resume speed, then slow down (slightly) again two steps before the obstacle in preparation for stepping up. Most changes happen during the last two strides before the obstacle. It was concluded that for subjects to make large gait adjustments four steps before the obstacle, then again to make small adjustments two steps before, and continue to do so until they have cleared the obstacle. In summary, subjects looked at the obstacle during most of the stride, from the time the curtain was opened to the time they stepped up on the curb. For the majority of conditions, gaze time on the obstacle decreased as the subjects approached the obstacle, however, for obstacles that are close to the subject, gaze time remains constant. This means that for obstacles within close range the subject is required to be more attentive to the task and make adaptations much faster than for the longer distances. 47 REFERENCES Bardy, B. G., Baumberger, B., Fluckiger, M., & Laurent, M. (1992). On the role of global and local visual information in goal directed walking. Acta Psychologica, 81.199-210. Berg, W. P., Wade, M. G. & Greer, N. L. (1994). Visual regulation of gait in bipedal locomotion: revisting Lee, Lishman, and Thomson (1982). Journal of Experimental Psychology: Human Perception and Performance, 20, 854-863. Brown, B., Brabyn, J. A. (1987). Mobility and low vision: a review. Clinical Experiences in Optometry. 70. 96-101. Corlett, J. (1992). The role of vision in the planning and guidance of locomotion through the environment. In L. Proteau & D. Elliott (Eds.), Vision and Motor Control (pp 375-397). New York: Elsevier Science. Elliott, D. (1986). Continuous visual information may be important after all: a failure to replicate Thomson (1983). Journal of Experimental Psychology: Human Perception and Performance. 12. 388-391. Elliott, D. (1987). The influence of walking speed and prior practice on distance estimation. Journal of Motor Behavior. 19. 476-485. Eng, J. J., Winter, D. A., & Patla, A. E. (1994). Neuromuscular strategies for recovery from a trip in early and late swing during human walking. Experimental Brain Research, 102. 339-349. Forssberg, H. (1982). Spinal Locomotor Functions and Descending Control. In B. Sjolund & A. Bjorklund (Eds.), Brain Stem Control of Spinal Mechanisms (pp. 253-271). New York: Elsevier Biomedical Press. Genesky, S. M . , Berry, S. H., Bikson, T. H., Bikson, T. K. (1979). Visual environmental adaptation problems of the partially sighted, final report. Santa Monica medical center: center for the partially sighted. Gibson, J. J. (1958). Visually controlled locomotion and visual orientation in animals. British Journal of Psychology, 49,182-194 Gibson, J. J. (1966). The senses considered as perceptual systems. Boston: Houghton Mifflin. Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mifflin. Georgopoulous, A. P. & Grillner, S. (1989). Visuomotor coordination in reaching and locomotion. Science. 245.1209-1210. Grillner, S. (1985). Neurobiological bases of rhythmic motor acts in vertebrates. Science, 228, 143-149. Hay, J. G. (1988). Approach strategies in the long jump. International Journal of Sport Biomechanics. 4. 372-392. 48 Lee, D. N., & Lishman, R. (1977). Visual control of locomotion. Scandinavian Journal of Psychology. 18,224-230. Lee, D. N., Lishman, J. R., & Thompson, J. A. (1982). Regulation of gait in long jumping. Journal of Experimental Psychology: Human Perception and Performance. 8,448-459. Llewellyn, M., Yang, J.F., & Prochazka, A., (1990). Human H-reflexes are smaller in difficult beam walking than in normal treadmill walking. Experimental Brain Research, 83, 22-8. Nashner, L. M. (1980). Balance adjustments of humans perturbed while walking. Journal of neurophysiology, 44, 650-664. Patla, A. E. (1986). Adaptation of postural responses to voluntary arm raises during locomotion in humans. Neuroscience Letters, 68, 334-338. Patla, A. E. (1989a). In search of laws for the visual control of locomotion: some observations. Journal of Experimental Psychology: Human Perception and Performance, 15, 624-628. Patla, A. E. (1991). Visual control of human locomotion. In A. E. Patla (Ed.), Adaptability of Human Gait: Implications for the Control of Locomotion (pp. 37-99). Amsterdam: North-Holland. Patla, A. E., & Rietdyk, S. (1993). Visual control of limb trajectory over obstacles during locomotion: effect of height and width. Gait and posture, 1, 45-60. Patla, A. E., Calvert, T. W., & Stein, R. B. (1985). Model of a pattern generator for locomotion in mammals. American Journal of Physiology, 248, R484-R494. Patla, A. E., Prentice, S. D., Armand, M., & Huissoon, J. P. (1994). The role of effector system dynamics in the control of limb trajectory over obstacles during locomotion: empirical and modeling approaches. In K. Taguchi, M. Igarashi & S. Mori (eds.), Vestibular and Neural Front (XHth Symposium on Posture and Gait) (pp.33-336). North-Holland: Elsevier. Patla, A. E., Prentice, S.D., Robinson, C. & Neufeld, J. (1991). Visual control of locomotion: strategies for changing direction and going over obstacles. Journal of Experimental Psychology: Human Perception and Performance, 17, 603-634. Patla, A. E., Rietdyk, S., Martin, C, & Prentice, S. (1996). Locomotor patterns of the leading and trailing limb while going over solid and fragile obstacles: some insights into the role of vision during locomotion. Journal of Motor Behavior, 28. 35-47. Patla, A. E., Robinson, C , Samways, M., & Armstrong, C. J. (1989). Visual control of step length during overground locomotion: task-specific modulation of the locomotor synergy. Journal of Experimental Psychology: Human Perception and Performance, 15, 605-615. Patla, A. E. (1995 in press). Neurobiomechanical bases for the control of human locomotion. In A. Bronstein, T. H. Brandt, & M. Woollacott, (Eds.), Clinical Aspects of Balance and Gait Disorders. City, UK: Edward Arnold. 49 Pearson, K. G. (1987). Central pattern generation: a concept under scrutiny. In H. McLennan, J. R. Ledsome, C. H. S. Mcintosh & D. R. Jones (Eds.), Advances in Physiological Research (pp. 167-185). New York: Plenum Press. Senders, J. W., Kristofferson, A. B., Levison, W. H., Dietrich, C. W., & Ward, J. L. (1967). The attentional demand of automobile driving. Highway Research Record, 195,15-32. Steenhuis, R. E., & Goodale, M. A. (1988). The effects of time and distance on the accuracy of target directed locomotion: does an accurate short term memory for spatial location exist? Journal of Motor Behavior. 20, 399-415. Thomson, J. A. (1983). Is continuous visual monitoring necessary in visually guided locomotion? Journal of Experimental Psychology: Human Perception and Performance, 12, 427-443. Thomson, J. A. (1986). Intermittent versus continual control: a reply to Elliott. Journal of Experimental Psychology: Human Perception and Performance, 9, 392-393. Warren, W. H., Jr. (1984). Perceiving affordances: visual guidance of stair climbing. Journal of Experimental Psychology: Human Perception and Performance, 10, 683-703. Warren, W. H., Jr., & Whang, S. (1987). Visual guidance of walking through apertures: body scaled information for affordances. Journal of Experimental Psychology: Human Perception and Performance, 13, 371-383. Warren, W. H., Jr., Young, D. S., & Lee, D. N. (1986). Visual control of step length during running over irregular terrain. Journal of Experimental Psychology: Human Perception and Performance, 12, 259-266. Watanbe, K., & Miyakawa, T. (1994). Gait analysis during stepping over the different height of obstacles in aged persons. In K. Taguchi, M. Igarasha & S. Mori (Eds.), Vestibular and Neural Front (pp. 195-198). North-Holland: Elsevier Science. Winter, D. A., Patla, A. E., Frank, J. S. (1990). Biomechanical walking pattern changes in the fit and healthy elderly. Physical Therapy, 70. 340-347. Appendix A Mean and Standard Deviations for all Dependent Variables Left Stride Length (m/stride) Mean and standard deviation for each of the 2 factors, 16 conditions.* Stride 1 Stride 2 Stride 3 Stride 4 CI 1.46 1.47 1.58. 1.52 1.51 (0.14) (0.15) (0.15) (0.15) (0.05) C2 1.41 1.53 1.52 1.11 1.39 (0.22) (0.14) (0.13) (0.67) (0.20) C3 1.46 1.55 1.53 1.43 1.50 (0.17) (0.15) (0.13) (0.48) (0.06) C4 1.50 1.53 1.56 1.37 1.49 (0.20) (0.14) (0.14) (0.46) (0.08) 1.46 1.52 1.55 1.36 (0.04) (0.03) (0.03) (0.18) Mean and standard deviation for each of the 4 conditions.* CI C2 C3 C4 First stride 1.56 (0.16) 1.53 (0.12) 1.57 (0.15) 1.54 (0.14) 1.55 (0.02) Left Stride Frequency (strides/sec) Mean and standard deviation for each of the 2 factors, 16 conditons.* Stride 1 Stride 2 Stride 3 Stride 4 CI 1.14 1.11 1.12 1.09 1.12 (0.08) (0.09) (0.10) (0.08) (0.02) C2 1.07 1.11 1.10 0.84 1.03 (0.09) (0.09) (0.09) (0.51) (0.13) C3 1.11 1.12 1.10 1.03 1.09 (0.11) (0.08) (0.09) (0.33) (0.04) C4 1.10 1.10 1.10 1.00 1.08 (0.09) (0.08) (0.09) (0.33) (0.05) 1.11 1.11 1.10 0.99 (0.03) (0.01) (0.01) (0.11) * Standard deviations appear in parentheses. C refers to condition. Left Stride Frequency (strides/sec), cont'd Mean and standard deviation for each of the 4 conditions. CI C2 C3 C4 First stride 1.11 (0.08) 1.10 (0.09) 1.12 (0.09) 1.09 (0.10) 1.11 (0.01) Left Stride Speed (m/sec) Mean and standard deviation for each of the 2 factors, 16 conditons.* Stride 1 Stride 2 Stride 3 Stride 4 CI 1.68 1.65 1.78 1.66 1.69 (0.19) (0.19) (0.26) (0.17) (0.06) C2 1.52 1.70 1.66 1.24 1.53 (0.29) (0.17) (0.16) (0.76) (0.21) C3 1.64 1.73 1.67 1.61 1.66 (0.27) (0.20) (0.17) (0.54) (0.05) C4 1.66 1.69 1.71 1.51 1.64 (0.25) (0.16) (0.18) (0.53) (0.09) 1.62 1.69 1.71 1.51 (0.07) (0.04) (0.05) (0.19) Mean and standard deviation for each of the 4 conditions.* CI C2 C3 C4 First stride 1.74 (0.20) 1.69 (0.17) 1.75 (0.19) 1.68 (0.19) 1.71 (0.04) Right Stride Length (m/stride) Mean and standard deviation for each of the 2 factors, 16 conditions.* Stride 1 Stride 2 Stride 3 Stride 4 CI 1.48 1.52 1.58 1.58 1.54 (0.14) (0.15) (0.13) (0-13) (0.05) C2 1.48 1.49 1.57 1.56 1.52 (0.21) (0.18) (0.12) (0.14) (0.05) C3 1.54 1.54 1.59 1.57 1.56 (0.19) (0.18) (0.14) (0.16) (0.02) C4 1.56 1.53 1.59 1.55 1.56 (0.19) (0.16) (0.13) (0.18) (0.02) 1.51 1.52 1.58 1.57 (0.04) (0.02) (0.01) (0.01) * Standard deviations appear in parentheses. C refers to condition. Right Stride Length (m/stride), cont'd Mean and standard deviation for each of the 4 conditions.* CI C2 C3 C4 First stride 1.58 (0.13) 1.57 (0.13) 1.58 (0.14) 1.58 (0.15) 1.57 (0.01) Right Stride Frequency (strides/sec) Mean and standard deviation for each of the 2 factors, 16 conditions Stride 1 Stride 2 Stride 3 Stride 4 CI 1.14 1.12 1.11 1.10 1.12 (0.08) (0.09) (0.09) (0.08) (0.01) C2 1.12 1.09 1.10 1.10 1.10 (0.13) (0.08) (0.08) (0.09) (0.01) C3 1.14 1.11 1.10 1.10 1.11 (0.10) (0.10) (0.09) (0.09) (0.02) C4 1.13 1.10 1.10 1.10 1.11 (0.10) (0.08) (0.09) (0.10) (0.02) 1.13 1.10 1.10 1.10 (0.01) (0.01) (0.00) (0.00) Mean and standard deviation for each of the 4 conditions.* CI C2 C3 C4 First stride 1.11 (0.08) 1.11 (0.09) 1.11 (0.08) 1.10 (0.09) 1.11 (0.00) Right Stride Speed (m/sec) Mean and standard deviation for each of the 2 factors, 16 conditions Stride 1 Stride 2 Stride 3 Stride 4 CI 1.68 1.70 1.75 1.74 1.72 (0.17) (0.19) (0.18) (0.15) (0.03) C2 1.68 1.62 1.72 1.73 1.69 (0.36) (0.20) (0.15) (0.19) (0.05) C3 1.76 1.72 1.75 1.73 1.74 (0.31) (0.25) (0.17) (0.20) (0.02) C4 1.77 1.69 1.74 1.71 1.73 (0.29) (0.19) (0.15) (0.26) (0.04) 1.72 1.68 1.74 1.73 (0.05) (0.04) (0.01) (0.01) * Standard deviations appear in parentheses. C refers to condition. Right Stride Speed (m/sec) Mean and standard deviation for each of the 4 conditions.* CI C2 C3 C4 First stride 1.74 (0.17) 1.73 (0.18) 1.74 (0.17) 1.74 (0.18) 1.74 (0.00) Gaze Time A (% stride) Mean and standard deviation for each of the 2 factors, 12 conditions. Stride 1 Stride 2 Stride 3 Stride 4 C2 31.54 30.26 35.19 32.71 32.42 (18.82) (20.65) (18.02) (14.37) (2.09) C3 38.24 34.59 31.32 33.40 34.39 (23.79) (21.01) (16.80) (16.54) (2.90) C4 28.61 29.10 30.64 33.40 30.44 (20.56) (18.57) (20.34) (18.40) (2.16) 32.80 31.32 32.38 33.17 (21.06) (20.08) (18.39) (16.44) Gaze Time B (% stride) Mean and standard deviation for each of the 2 factors, 12 conditions. Stride 1 Stride 2 Stride 3 Stride 4 C2 46.29 49.81 51.91 57.75 51.44 (23.31) (25.63) (22.70) (17.05) (4.80) C3 31.12 44.43 56.55 57.74 47.46 (20.17) (28.94) (24.37) (16.62) (12.44) C4 39.70 52.46 58.91 55.94 51.75 (19.25) (25.44) (27.67) (20.75) (8.45) 39.04 48.90 55.79 57.14 (20.91) (26.67) (24.91) (18.14) * Standard deviations appear in parentheses. C refers to condition. 54 Gaze Time E (% stride) Mean and standard deviation for each of the 2 factors, 12 conditions.* Stride 1 Stride 2 Stride 3 Stride 4 C2 22.17 19.93 12.91 9.54 16.14 (16.23) (15.26) (11.94) (10.29) (5.91) C3 30.64 20.98 12.12 8.86 18.15 (21.20) (18.19) (11.90) (7.08) (9.77) C4 31.68 18.44 10.45 10.65 17.81 (18.00) (14.25) (14.46) (8.14) (9.97) 28.16 19.78 11.83 9.69 (18.47) (15.90) (12.77) (8.50) Mean and standard deviation for each of the 2 conditions.* First stride Gaze A GazeB 42.23 (18.95) 52.26 (21.80) * Standard deviations appear in parentheses. C refers to condition. i 55 Appendix B Vision and Gait: The Role of Visual Sampling in Obstacle Compensation Pilot Study #3 Purpose The focus of this investigation is to twofold: first is to determine if there are any observable changes in limb kinematics that may occur prior to an obstacle that are associated with accommodation; second to flush out possible difficulties in attempting testing of this nature. Methods Subjects Five healthy individuals age 22-35 years from Vancouver area. Setup and Protocol The experiment took place in a small gymnasium in which a 17 meter long path was set up. Photocells were set at 4 meters and 11 meters from the start, providing a 7 m space to monitor speed. Near the end of the path, an obstacle (2 wooden boxes, placed flush one behind the other, 16.5cm x 66cm x 66cm each) of standard curb dimensions was placed. The neutral obstacle position corresponded to the location where the right foot would naturally be placed if there had been no obstacle. This position was predetermined using the control condition. The 'close' condition referred to the obstacle being placed one step ahead of the neutral condition while the 'far' condition referred to the obstacle being placed one step further than the neutral. These last 2 conditions would require the subject to naturally step up with their left foot, however tests require the subject to step up only with the right foot. Prior to testing, subjects were given several practice trials without the obstacle to determine natural walking pace. Reflective markers were placed on the shoulder, hip, knee and ankle of the ipsilateral limb. Markers were also placed at intervals along the wall in order to 56 reference movement spatially. A video-based data acquisition system (Peak Performance Technologies software) was used to collect the kinematic data. Forty trials were collected beginning with ten trials of normal unobstructed gait and ten trials each for the 3 platform placement conditions. The platform positions were chosen at random with a limit often for each. Subjects were instructed to walk at a comfortable pace, consistently starting each trial with the same foot and clearing the 'curb' only with the right foot. Data Analysis Stride duration was taken from ipsilateral heel strike to ipsilateral heel strike. During this time, hip and knee velocities were extracted from recorded markers for each frame. Velocity was determined using the filtered displacement data. Results When compared to the control trial, the paths of all the markers of the obstructed conditions differ in shape. Maximal vertical displacement of all markers increases when the subject steps up on the obstacle. Quantitative results showed no differences except where the ipsilateral limb stepped up onto the curb. Stride length, stride frequency and stride duration all showed no differences when viewed across strides and across conditions. Figure 1 displays stride duration as taken from shoulder joint data for each trial. The average standard deviation between trials when compared remained the same across all conditions, thus showing no difference except at the curb where step-up was required. The same results occurred when comparing angular displacement at the hip and knee across strides and across conditions. Figure 2 shows the angular displacement at the hip and knee for the control (condition 1). All strides remain consistent throughout each trial. One might expect differences to appear when the curb was introduced, and indeed there were obvious changes during the stride that cleared the curb; nonetheless there was no observable change during the previous approaching strides. In the 57 normal condition (condition 2) where the curb was placed in a natural right-foot placement position, we did not expect to see any changes (Figure 3), however the same observations were noted for both the close and far conditions. Each stride, when compared across conditions, also showed no differences. Figure 4 shows the average hip and knee angles of all trials for stride #3, which occurs approximately mid-way along the prepared path. All other strides, with the exception of the step-up stride, appear the same. 59 00 4» S <U S W P O WD « 3 E S v© ID p« (apujs/SDas) uopejnQ 60 61 00 ID •a •c (apuis/soas) UOIIBJIIQ 62 s q« <U «S a ^ o "3c *"3 << o w DX 2 fS J. 3 (3D 00 o 9i "2 "C o o © O o © o © 00 © © VO (•Sap) 9[3uy 63 33 •*•» xn s «* - § < s a 61) U • • PQ s ox • p"« OO o VO 4> •c o o © ON © 00 o o I T ) o © O o © © ON (•Sap) aiSuy 0> •d •c PQ -5S § s & _ u - g « «s &. h o "wo & « c <3i © OX S3 -tS hi - B < <u S DJ3 O 00 VO -a •c 0S O o o fN O fN fN O fN O © fN O © 00 o © v o (•Sap) ajSuy 65 66 O © o o o o o o o o o (•8ap) a[Suy (•Sap) afSuy 68 Discussion In this study subjects were required to mount a single obstacle placed on an otherwise normal, even path at various distances from the start position. The foci of the discussion were to identify whether there are specific changes in gait patterns and if there are, to determine what they were what are they are the foci of this discussion. Although no changes in any of the dependent variables could be seen, one must take into account that the results were averaged across subjects and not within. Consequently, some important observations may have been averaged out of the results. Most tripping accidents occur when the center of mass lies outside the base of support, i.e. during the single support phase, and are an effect of limb trajectory during the swing phase. It is therefore important to look more closely at limb kinematics during each trial. As the ipsilateral limb is the one clearing the curb first, it may be important to look at the kinematics when the ipsilateral limb is over the obstacle together with the ensuing foot contact. Additionally, movement outcomes are generally a direct result of what has been perceived by the visual sense. A closer look at the involvement of vision in negotiating obstacles would also provide more information as to its role in the act of locomotion. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items