| Description |
Perceiving and understanding speech is a complex and dynamic task. Speech is often poorly articulated and is rarely heard in isolation. It is unsurprising, then, that adaptive speech processing methods, such as relying on context, are common. While making predictions about spoken words can facilitate real-time speech processing, these predictions may be relied upon too heavily, occasionally resulting in a misperception of the speech signal. This can lead to false hearing, where listeners confidently report hearing a word that was not spoken. Most models of false hearing assume that cognitive control, defined as the ability to inhibit automatic responses, is central; false hearing occurs because the listener fails to appropriately inhibit the top-down, predictive signal in light of the contradictory bottom-up auditory signal. When the speech signal does not match the predicted word, incongruency exists, and when the signal itself is unclear (such as being masked by background noise), the correct word present in the signal is activated less strongly than the predicted but ultimately incorrect word. However, recent research has suggested that cognitive control may not be so crucial. To further test its role, the current study employed a within-subjects dual-level cognitive demand task, in which 39 undergraduate participants were asked to drive in a simulator under two conditions, one easy and one difficult. In the simulator, participants also completed a verbal response task in which they were presented with a series of sentences and asked to identify the last word of the sentence they heard. Three sentence-final conditions were presented with each sentence context: a congruent/predictable word, which was found in previous norming studies to be highly predictable based on the sentence context; an incongruent baseline, which was semantically incongruent and phonologically dissimilar from the target; and a phonological lure, which was also incongruent with the sentence but phonologically similar to the predictable word. The results of the study suggested that, while the hard driving condition was indeed more cognitively demanding than the easy condition, rates of false hearing were unaffected by the driving condition. This suggests that cognitive control specifically is not central to false hearing, challenging current models. These findings may have implications for the design of auditory systems, such as hearing aids or alarms, especially for those in which the speech signal has been obscured by background noise. |
