Abstract
While it has been well-established that optostatic and optokinetic cues contribute to the perception of vertical, it is unclear how the brain processes their combined presence with the non-visual vestibular cues. Using a psychometric approach, we examined the percept of vertical in human participants (n=17) with their body and head upright, presented with a visual frame tilted at one of eight orientations (between ±45°, steps of 11.25º) or no frame, surrounded by an optokinetic roll-stimulus (velocity = ±30°/s or stationary). Both cues demonstrate relatively independent biases on vertical perception, with a sinusoidal modulation by frame orientation of about 4º and a general shift of about 1-2º in the rotation direction of the optic flow. Variability was unaffected by frame orientation, but was higher with than without optokinetic rotation. An optimal-observer model in which vestibular, optostatic, and optokinetic cues provide independent sources to vertical perception was unable to explain these data. In contrast, a model in which the optokinetic signal biases the internal representation of gravity, which is then optimally integrated with the optostatic cue, provided a good account, at the individual participant level. We conclude that optostatic and optokinetic cues interact differently with vestibular cues in the neural computations for vertical perception.
