Uncovering persistent biases in human path integration by separating left and right trials

Navigation ability strongly varies between humans. Careful analysis of errors occurring in navigation, or indeed any cognitive process, offers insight into the underlying mechanisms at different levels. While analyses at the individual level allow nuanced identification of error persistences across conditions and time, the population level facilitates generalisation but precludes conclusions about lower-level phenomena.

Regarding the critical navigation mechanism of path integration – the continuous tracking of navigated paths for self-localisation – previous studies have focused on populationlevel analyses, revealing systematic errors in estimating the travelled angles and distances. However, at the individual level, there are indications that people also possess left or right biases that are classically overlooked when pooling left and right trials. Therefore, we carefully investigate individual path integration errors in (1) a re-analysis of data from several influential human navigation studies, and (2) our own virtual reality path integration experiment. For both, we confirm time-persistent individual side biases, but find no evidence for consistent errors at the population level, suggesting that important aspects in human navigation performance might be overlooked by averaging across sides.