Item

Abstract

Many well-learned and complex visuomotor tasks like grasping and bicycle riding are performed effortlessly despite little or no explicit knowledge of the required motor movements. Researchers disagree about the extent to which movement control depends on sensory feedback although it is often suggested that driving manoeuvres like lane changing or positional control can be completed without continuous visual feedback. Recently we showed that most drivers are unable to perform both phases of a lane change (pulling out and straightening the car), in the total absence of visual feedback (Wallis et al, 1997 Perception 26 Supplement, 100).

In a new series of experiments we sought to distinguish the mode of control for these two phases during a lane change. Subjects drove in a simulated, naturalistic environment projected on a 180 deg screen, using a force-feedback steering wheel. Subjects were required to complete the manoeuvre within a total distance of 80 m (velocity 72 km h-1, duration 4 s). In condition (a), vision was occluded 0.5 s after initiation of the manoeuvre for 1.5 s, preventing visual feedback throughout the first phase. In condition (b), vision was occluded after completion of the first phase until the end of the trial, preventing any further feedback.
In condition (a), temporary withdrawal of vision led to a production of significantly larger steering-wheel amplitudes, with a consequent overshoot in lateral position ( p < 0.001). Under condition (b), an exaggerated, complementary steering phase was observed. These results indicate that manoeuvres like lane correction or overtaking can be characterised by two distinct phases, each of which must be preceded by a critical period of visual feedback used to initiate the subsequent steering phase.