Abstract
Human and animals are able to navigate to a previously visited place based on visual landmarks or path integration, a process that keeps track of location and orientation during self-motion. Path integration is often assumed to function automatically, continuously running in the background (e.g., May & Klatzky, 2000), whereas other studies have shown that it is overshadowed by salient visual landmarks and may serve as a back-up system when landmarks fail (e.g., Foo et al., 2005; Zhao & Warren, Psychonomics, 2010). We tested whether path integration in navigation with visual landmarks is a continuous process or is a back-up system modulated by landmark reliability. Participants performed a triangle completion task in an open virtual environment. Three set of trials were tested sequentially: landmark trials (landmarks were always present); catch trials (landmarks disappeared at response in 10% of trials), and path integration trials (no landmarks). If path integration is continuously running in the background, performance in catch trials should be equivalent to that in path integration trials. However if it is modulated by landmark stability, performance in catch trials should be worse than in path integration trials. Participants showed significantly greater constant and variable errors in catch trials than in path integration trials (p < 0.05), suggesting that path integration is “dialed down” in the presence of stable landmarks. This effect was primarily due to the first catch trial (p < 0.01), while the rest showed no difference from path integration trials, indicating that path integration is quickly “dialed up” when landmarks fail. These results are consistent with the idea that human navigation is dominated by salient visual landmarks, while path integration serves as a back-up system rather than running automatically and continuously.
