Abstract
One recent study reported that when ongoing pre-stimulus fMRI activity in the dorsal attention network was high, the hit rate in an auditory detection task was surprisingly low. This result is puzzling because pre-stimulus activity in the dorsal attention network presumably reflects the subjects' attentional state, and high attention is supposed to improve perception, not impair it. However, it is important to distinguish between the capacity and decision/criterion aspects of perception. Using signal detection theoretic analysis, we provide empirical evidence to show that spatial attention can lead to conservation bias in detection, although it boosts detection capacity. In behavioral experiments we confirmed the prediction, derived from signal detection theory, that this conservative bias in detection is coupled with lowered confidence ratings in a discrimination task. Based on these results, we then used fMRI to test the hypothesis that low pre-stimulus ongoing activity in the dorsal attention network predicts high confidence rating in a visual motion discrimination task. We confirmed this counter-intuitive hypothesis, and also found that functional connectivity (i.e. correlation) between areas within the dorsal attention network negatively predicts confidence rating.
Taken together, these results support the notion that attention may have a negative impact on the decision/criterion aspects of perception. This negative relationship may explain why under the lack of attention, we may have an inflated sense of subjective experience: e.g. the vividness of peripheral vision; and the overconfidence in na&ıuml;ve subjects in inattentional blindness and change blindness experiments despite their poor performance capacity.