Abstract Detail

Talk Sessions

Poster Sessions

All Presentations

Symposia

A behavioral dynamics approach to obstacle detection and avoidance by patients with tunnel vision

33.308, Sunday, May 12, 8:30 am - 12:30 pm, Royal Ballroom 6-8
Session: Perception and action: Complex actions, clinical

Adam W. Kiefer¹, Russell L. Woods², William H. Warren¹; ¹Department of Cognitive, Linguistic & Psychological Sciences, Brown University, ²Schepens Eye Research Institute and Massachusetts Eye and Ear, Harvard Medical School

Mobility issues are the typical sequelae of "tunnel vision" (severe peripheral visual field loss), and manifest in the difficulty patients have detecting and avoiding obstacles during locomotion. We examined the behavior of seven patients with tunnel vision (6˚-27˚ visual fields) and six normally sighted individuals as they walked toward a visible goal in a virtual environment with or without an obstacle present. Each obstacle was a pole 1.8m tall and either stationary or moving across the path of the participant. Stationary obstacles appeared at a distance of 4 or 6m and an angle of 1˚ from the path to the goal. The moving obstacles crossed the participant’s path at 4 or 6m while moving at a constant speed (50%, 70% or 90% of the participant’s preferred walking speed). All obstacles appeared after participants walked 1m, and participants pressed a button when they detected an obstacle. Based on detection rates for short 1.2m obstacles (Kiefer et al., 2012), the patients were divided into two subgroups: (1) a low-detection group and (2) a high-detection group. We simulated individual locomotor trajectories using Fajen & Warren’s (2003) steering dynamics model. RMSE between the model and human heading time series revealed no significant differences between the normally sighted (M = 0.28˚ +/- 0.01˚), low-detection (M = 0.36˚ +/- 0.04˚) and high-detection (M = 0.30˚ +/- 0.03˚) groups with stationary obstacles; simulations with moving obstacles are ongoing. There was a trend toward greater variability in RMSE for both patient groups compared to the normally sighted group. The results indicate that tunnel vision patients have different detection strategies (Kiefer et al.), but once an obstacle is detected their locomotor behavior is similar to normally sighted pedestrians. Patients appear to exhibit noisier trajectories, perhaps due to difficulty tracking the obstacle as they fixate and steer toward the goal.

< Back