Accommodation modulates the individual difference between objective and subjective measures of the final convergence step response.

Measuring vergence eye movements with dichoptic nonius lines (subjectively) usually leads to an overestimation of the vergence state after a step response: a subjective vergence overestimation (SVO). We tried to reduce this SVO by presenting a vergence stimulus that decoupled vergence and accommodation during the step response, i.e. reduced the degree of 'forced vergence'. In a mirror-stereoscope, we estimated convergence step responses with nonius lines presented at 1000 ms after a disparity step-stimulus and compared it to objective recordings (EyeLink II; n = 6). We presented a vertical line, a cross/rectangle stimulus and a difference-of-gaussians (DOG) pattern. For 180 min arc step stimuli, the subjective measures revealed a larger final vergence response than the objective measure; for the vertical line this SVO was 20 min arc, while it was significantly smaller for the DOG (12 min arc). For 60 min arc step-responses, no overestimation was observed. Additionally, we measured accommodation, which changed more for the DOG-pattern compared with the line-stimulus; this relative increase correlated with the corresponding relative change of SVO (r = 0.77). Both findings (i.e. no overestimation for small steps and a weaker one for the DOG-pattern) reflect lesser conflicting demand on accommodation and vergence under 'forced-vergence' viewing; consequently, sensory compensation is reduced and subjective and objective measures of vergence step responses tend to agree.