Neural And Cognitive Mechanisms Affecting Perceptual Adaptation To Distorted Speech

The majority of everyday communication occurs in the presence of distortions, such as background noise, yet the human ability to understand speech in adverse listening conditions is remarkably robust. Past research has investigated perceptual adaptation to different speech conditions, however, our knowledge of the individual differences and the associated cognitive and neural mechanisms affecting perceptual adaptation is still limited. The work described in this thesis therefore aimed to advance our understanding of this research area, with specific focus first on determining the extent to which adaptation to one distortion generalises to another, second, determining the underlying cognitive mechanisms of this adaptation process and finally determining what role, if any, the left ventral premotor cortex plays in adaptation. This thesis presents results from eight experiments, two behavioural and six using Transcranial Magnetic Stimulation (TMS) as the primary research tool. Results from experiments 1 and 2 (behavioural) show that measures of verbal intelligence, specifically vocabulary knowledge, working memory and general cognitive functioning underpin the perceptual learning process, providing support for statistical learning to occur and assist adaptation to distorted speech. Additionally, the results suggest participants possess a general skill that enables generalisation of learning from one adverse listening condition to another. Experiments 3 to 8 used TMS to modulate perception of speech in noise in a bilateral superior temporal region. However, no effect of using this protocol was found when applied to the left ventral premotor cortex whilst participants adapted to time-compressed speech. The results of the experiments described in this thesis are considered in the context of our current understanding of the cognitive and neural mechanisms associated with perceptual adaptation to distorted speech. It is believed that the results will contribute significantly to existing knowledge due to use of novel research methodologies e.g., use of multiple distortions, multiple speakers and TMS.