Greater response conflict from weaker visual signals

Abstract

It is generally thought that strong conflicting signals have greater negative impacts on task performance than relatively weaker conflicting signal. Here we show that this is not necessarily the case for visual-motor contingencies. We used a novel procedure where subjects judged the color (blue or green) of a visual stimulus by making a color-contingent leftward or rightward motion. In the background, a certain percent of coherently moving dots was presented with leftward or rightward direction, but which was irrelevant to the task. In half the trials, the direction of the “correct” motor response was the same as the direction of motion-stimulus (ie congruent trials), whereas in the other half of trials, the correct response direction was opposite to the coherent motion direction (ie incongruent trials). The coherent motion percentage was randomly chosen for each trial and ranged from 0–100%. Interestingly, we found a non-monotonic relationship with a significant dip in performance at around 25% motion coherence with recoveries of performance on either side. Reaction times trended higher at 25% coherence, opposite to the prediction of a reaction-time/accuracy trade-off. The generality of this effect was confirmed with tasks using different motor responses, such as saccadic eye movements and steering wheel operations. These results are at odds with the model that stronger irrelevant signals more severely interfere with a task performance. One possible explanation is that visual signals that are too weak to be “noticed” and suppressed can directly interfere with a motor response as conflicting motor signals.

Footnotes

 NSF grant (BCS-9905914), NIH grant (RO1EY015980-01), Human Frontier Research grant (RGP18/2004), NSF (CELEST), ERATO JST