Wedge prisms shifting the visual field laterally create a mismatch between the straight ahead position signalled by vision and that encoded by extraretinal and head-on-trunk proprioceptive information. Short-term adaptation to left-deviating prisms in normal subjects results in a visuomotor attentional bias towards the right-hand side (aftereffect). Prismatic adaptation (PA) is usually induced through a training consisting in repeated ballistic movements of the dominant arm towards visual targets, while participants are wearing prismatic goggles. The present study demonstrates that an original oculomotor PA procedure with leftward deviating prisms-without pointing movements and only consisting in repeated gaze shifts towards visual targets-can induce a rightward bias in normal subjects as assessed by visual straight ahead and line bisection tasks (Experiments 1 and 2). We show that oculomotor PA induces a bias in line bisection similar to that reported after visuomotor PA (Experiment 2). We suggest that a conflict between retinal, extraretinal and proprioceptive information about the straight ahead location causes the observed effects. In follow-up experiments 3, 4, and 5, we demonstrate that neither eye deviation without prisms nor shift of the visual field without eye deviation induces PA biases. We propose that an optimal integration model of visual and proprioceptive inputs can best account for the observed results.
Keywords: Line bisection from memory; Maximum Likelihood Estimation (MLE); Oculomotor prismatic training; Optimal integration model; Prismatic adaptation; Visual straight ahead.