The Role of Vision during Path Integration in Darkness

Path integration in the dark has shown that previous visual experience may not be recruited and requires only interoceptive information. One study showed there was no difference in a path integration task between blindfolded and sighted subjects [1]. Another showed no improvement in performance when subjects were shown a visual preview of the environment before completing the path integration task [2].
Using a triangle completion paradigm, we examined the contribution of visual cues to the accumulation of translational or rotational information during path integration in a fully immersive virtual environment. We later extended these experiments to show that previous exposure to a virtual environment with a mismatch between visual and interoceptive rotational information does affect subsequent performance in a path integration task carried out wholly in the dark. We also argue that visual imagery must be recruited during path integration in complete darkness.
Our goals - to determine whether path integration in the dark is influenced by previous visual experience during a triangle completion task. ii) Determine whether path integration in the dark requires the use of visual imagery, despite the explicit lack of visual information in the scene.
In the first two experiments, subjects experienced a rotational or translational visual gain whilst being led on a path in a rich and textured virtual environment viewed in a head-mounted display (see Figure 1 for the environment). At this point, the environment became dark and subjects made their own way back to the origin. If subjects rely entirely on interoceptive information during self-location, they should return accurately to the origin. If instead, they use vision, the perception of their own location and subsequent response will be biased by the gain in the altered environment. In the next two experiments, subjects performed a tangential task, whilst experiencing a non-detectable constant rotational gain in the virtual environment. The task required them to search for a football in the virtual scene, walk towards it and report their approximate location relative to the scene. Control trials were used to check that subjects had become adapted to the mismatch between visual and interoceptive cues. Subsequently, subjects were led entirely in the dark along a path and asked to return to the origin also in the dark. If subjects used visual imagery to self-localise, their return to origin should be biased in accordance with the manipulation they experienced.
In the first two experiments subjects responses were consistent with the visual bias they experienced whilst being led. In the last two experiments, subject did adapt to the mismatch, as shown by the increased fractional turn after adaptation to a gain of 0.7 (See figure 2) and responses in complete darkness were consistent with the use of visual imagery.
Vision is integral to path integration in the dark when a mismatch between vision and other cues exist. Even in the absence of explicit visual information, vision in the form of visual imagery is recruited enabling subjects to self localize in darkness.
Last updated: Friday, 05.10.2012