Conjoint effects of decision making and stimulus probability (oddball-effect) on pupil dilation

In the physical environment as well as in many experiments, we need to decide whether an incoming stimulus is relevant to us. Especially outside laboratories, stimuli have uneven probabilities to occur. Both decision making and stimulus probability are reflected in pupil size, with stronger dilation for targets than distractors (decision making) and for rare than frequent stimuli (oddball-effect). However, conjoint investigations into both factors are still scarce.
In two experiments, we sequentially presented letters that either matched (target) or mismatched (distractor) a given word. Moreover, the ratio of target to distractor letters was varied blockwise from 25:75 over 50:50 to 75:25. Experiment 1 suggested that the effect of decision making was much larger than the effect of stimulus probability. Further, decision making and stimulus probability interacted in a way that only target letters were affected by differential stimulus probability. In Experiment 2, we tested whether stimulus probability alone would affect pupil dilation by adding a condition during which only letters of differential probability but no target word were given. In this condition without goal relevance, no oddball effect was found; further, results showed that only goal relevant letters (not irrelevant or only task relevant letters) were affected by differential probabilities.
On this basis, we propose a framework integrating theories on decision making and stimulus probability, suggesting that incoming stimuli are first checked for their goal-relevance. Second and only if relevant, stimuli are evaluated regarding their relative value in a temporarily subsequent step.