Perception of visual direction was investigated by requiring subjects repeatedly to adjust a single small light, in an otherwise darkened room, to perceived ‘straight ahead’. This task presumably requires comparing concurrent extra-retinal information (either proprioception or an efference copy) with an internally stored ‘standard’ of comparison. Moment-to-moment precision in that performance is remarkably good, with median threshold (standard deviation) of 47 arc min. Nevertheless, the responses often involved a monotonic shift of direction over a few minutes during a test session in this reduced visual environment. These trends led to final settings that were immediately recognised as grossly erroneous when the room was relit, implying that the presumptive internal standard of comparison, while unstable, can be rapidly updated in a full visual environment. There are clear similarities between this phenomenon and the sudden ‘visual capture’ that occurs in a re-illuminated room, following distortions of visual direction that arose in a similarly reduced setting for subjects whose extraocular muscles were partially paralysed (Matin et al, 1982 Science216 198 – 201). In both cases, the visual stimuli that underlie rapid recalibration are unknown. Among the several possibilities that can be imagined, the strongest candidate hypothesis for this calibration of the straight-ahead direction is that, during fixation in a lit room, one utilises the directional distribution of image motion that arises because of microscale drift of the eye, as it moves toward its equilibrium orientation, much as a moving observer can use optic flow to evaluate ‘heading’ (the dynamic analogue of ‘straight ahead’).
No dynamic visual capture for self-translation minimum audible angle
