Abstract
It is generally accepted that humans use information from self-generated optic flow to guide steering towards a goal. However, visual direction can also serve as the control signal for steering, and Warren et al (2001) found comparable influences from these two factors in a paradigm that used a prism manipulation to influence visual direction, while preserving the location of the focus of expansion (FOE). However, the head's lateral oscillation during the gait cycle results in constant variability in the direction of the heading and the position of the FOE, which suggests optic flow's role in postural control rather than steering (Matthis et al, 2022). These lateral head movements result in a structured curl signal in the flow array that can be used to maintain stable posture and make momentary changes in gait to steer towards a goal. We repeated Warren et al’s (2001) experiment in a virtual environment, and in addition tracked body and gaze position. Our findings indicate that the interpretation of behavior in this paradigm is complex. One complexity is that the prism manipulation introduces an abnormal curl signal in the presence of ground texture. In addition, when subjects were instructed to look at the ground instead of at the goal, their trajectories became more curved, consistent with a greater influence of visual direction. This suggests that fixating the abnormal curl signal on the ground plane modifies the effect of optic flow in this paradigm. Investigation of the retinal flow patterns during both goal and floor fixations suggested subjects compensate for deviations from their expected curl signal by strategic placement of gaze, and that retinal flow and visual direction may be used in different ways in locomotor control.