Using sinewave speech to investigate the locus of informational interference during speech-in-noise perception
Speech-in-noise research typically distinguishes between energetic masking (EM: interference between target and masker at the periphery) and informational masking (IM: interference higher in the auditory pathway). IM can itself be broken down into low-level and high-level IM. We use the term “informational interference” (inf-int) to refer to high-level IM involving linguistic and cognitive factors, and which is influenced by long-term knowledge (e.g. familiarity with the language spoken by competing talkers).
Unlike EM, inf-int is poorly understood, in part because it is extremely difficult to manipulate inf-int without altering EM or lower-level IM. The current study aims to isolate one way in which inf-int may arise: awareness of a masker being speech as opposed to non-speech. This is achieved by taking advantage of the perceptual properties of sinewave speech (SWS). SWS is produced by extracting the first few speech formants of a natural utterance and replacing them with time-varying sinusoids reproducing their frequency and amplitude variations. SWS is not usually initially perceived as speech, but through training can be made at least partially intelligible.
Young normal-hearing listeners (N = 54) completed 3 tasks (pre-exposure, exposure, post-exposure). The pre- and post-exposure tasks were speech-in-noise tasks, with target sentences presented in a masker of SWS and amplitude-modulated white noise. During the intervening exposure phase, one listener group was trained to understand similar SWS/white noise stimuli. A second group listened to the same SWS stimuli, but were told that they were hearing randomly-generated machine noise, and performed a simple same/different task. As a result, all listeners heard the same stimuli (thus controlling for EM and lower-level IM), but only listeners in the trained group were aware of linguistic content in the masker, and only post-exposure. By comparing pre- and post-exposure performance across the two groups, it is therefore possible to isolate any effects on speech-in-noise perception arising specifically from an awareness of the masker being speech. Results will be presented and interpreted in light of the literature on informational masking.