ABSTRACT
Drivers are regularly engaged in self-motion heading discrimination tasks. When driving a remote-control vehicle through a closed-circuit TV, while physically travelling in another moving vehicle, drivers may be exposed to conflicting and nonconflicting visual and inertial heading cues. Modelling of how humans integrate visual and vestibular heading information to estimate their self-motion heading has been studied extensively (e.g., Fetsch et al. 2009; Butler et al. 2010; de Winkel et al. 2010). Heading discrimination thresholds (HDT) has been used as a quantifying measure of the strength of a heading cue. HDT has been defined as the smallest rightward deviations (degrees) from the straight ahead direction by a heading cue that can be correctly detected with an accuracy of 84%. A strong cue will have a small threshold while a weak cue will have a large threshold (i.e., larger deviation is needed to achieve 84% correct detection rate). The choice of 84% accuracy level was originated in Gu et al. (2008) when he proposed a Bayesian model to predict how humans integrate visual and vestibular heading information. This 84% for defining the HDT was adopted in Fetsch et al. 2009 and also in this study. Participants were repeatedly exposed to the heading stimuli of the same strength but with different heading directions ranging from −16◦ (left) to +16◦ (right). For each cue, participants were asked to judge whether the cue is to the left or to the right.
Figure 1. HDTs extraction from plots of percentages of rightward as functions of heading directions. Two hypothetical curves were plotted with HDTs of 1◦ and 15◦. Strong heading cues were used in stimuli 1 and weak cues
