Abstract
Various psychophysical techniques have been proven successful tools to render visual inputs subjectively invisible. For example, backward masking obstructs the awareness of a prior stimulus, and the masked stimulus remains invisible even with full attention at the correct spatiotemporal location. In contrast, observers would miss the presence of a strong stimulus if their attention were not directed to the stimulus. These two examples both demonstrate observers' failure to detect a stimulus. However, it remains unknown whether the nature of unawareness between those techniques is the same. Here we developed a method for classifying different types of psychophysical blinding techniques such as in current study, backward masking and attentional blink (AB). In experiment 1, we impaired the visibility of a target luminance blob by presenting a mask immediately after. The observers' detection accuracy deteriorated as a function of ISI between the target and the mask. In experiment 2, we used AB to impair the visibility of the letter ‘X’ embedded in a RSVP letter stream by manipulating the lag between letter ‘X’ and another attention-catching marker. In both experiments, the target was presented on half of the trials, and subjects were asked to report the presence or absence of the target together with their subjective confidence rating (high/med/low).
Our analysis showed that subjects were equally highly confident in reporting absence in ‘missed’ and ‘correct-rejection’ trials in backward masking, suggesting the unaware experience due to backward masking (miss) is subjectively similar to physical absence (correct-rejection). In AB, observers' confidence decreased together with the objective performance, implying that observers were aware of the transient attentional impairment. This distinct pattern in confidence rating supports the hypothesis that impairments in unconscious perception can be classified into sensory and attentional mechanisms, and this is in line with the view that perceptual (un)awareness involves multi-stage processing.
This study was supported by NSC 97-2410-H-002-188-MY3 to CT.