Abstract
Is conscious change detection produced by the implicit registering of a sensory mismatch drawing attention to the change? Electrophysiological correlates for early, implicit stages in this process have been claimed in the auditory domain (Näätänen, 1992), but evidence is mixed in vision. The strongest evidence for early implicit detection would be significant event-related potentials (ERPs) indicating a sensory mismatch representation when psychophysical performance on conscious change detection is close to chance. In the present experiment subjects had to detect an orientation shift in one of six heterogeneously oriented D6 pattern elements. To maximise sensory mismatch, while keeping subject performance poor, a 90 increment miscued condition was included. For such large element changes, performance, even in an uncued (‘change blindness’) condition was substantially above threshold (83%) but for the miscued condition; performance remained close to chance (56%). ERPs were recorded from 20 electrodes from 9 subjects in a miscuing paradigm (60% correctly cued, 20% uncued, 20% miscued trials). In the uncued condition, there was a larger posterior negativity (ca. 200–300ms, N2, p<0.001) and a larger positivity (ca. 350–600ms, the P3, p<0.02) to increments, but when miscued these differences were absent. There was also no evidence of the putative sensory mismatch component (Tales et al, 1999) which should be evident in the miscued condition. Hence, in a paradigm designed to optimize the detection of implicit components we do not find evidence of implicit change detection. This is consistent with a recent psychophysical study (Mitroff et al, 2002).