Figure 2.
A behavioral measure of speech clarity reflects the effect of P. Subjects were presented with sequences of vocoded-original-vocoded speech snippets and were asked to identify the two noise-vocoded streams (NP and P stimuli) as standard or deviant presentations by comparing them with the original speech snippet. Responses consisted of a level of confidence from 1 (definitely a deviant) to 5 (definitely a standard). A, The response distributions (mean percentage occurrence ± SEM) confirm that subjects were more confident in detecting standard trials when P was available. B, The confidence level for two selected subjects. The result in the top panel shows that subject 5 improved in detecting both standard and deviant trials when P was available, which we interpret as evidence for an increase in perceptual clarity. In contrast, subject 12 (bottom panel) responded with higher values to P stimuli for both standard and deviant trials. In this case, the positive StP-StNP cannot be assumed to purely reflect an increase in perceived clarity, as deviants were not detected. C, The confidence level averaged across all subjects (mean ± SEM) is here reported for NP and P stimuli, and for both standard and deviant trials. The increase in confidence due to P is larger for standard than for deviant trials (*p < 0.05).