Dynamics of Auditory Reflective Attention in Speech-in-noise Perception
T.M. Vanessa Chan · TSpace (University of Toronto) · 2020
The ability to listen to speech in noisy environments is supported not only by mechanisms of sound segregation and auditory streaming, but also by higher-order cognitive functions such as semantics, attention, and memory. Major strides have been made to understand how each of these cognitive functions contributes to comprehension of incoming speech signals, leading to the emergence of predictive models of speech comprehension in noise. Whereas most studies have focused on the effects of prior contextual information on speech-in-noise (SIN) processing, less is understood about the mechanisms involved in the effects of subsequent contextual information, in which a noisy signal is followed by context. This dissertation characterizes the involvement of subsequent semantic contextual information across three studies, informed by an auditory attention to memory (AtoM) approach. Study 1 first directly compared the effects of prior and subsequent semantic context on identification of a word in noise, and found that a subsequent cue word that was related to a target word in noise could boost identification of the target, although not as well as presenting it prior to the target. Study 2 built on Study 1 by comparing the neural mechanisms underlying prior and subsequent semantic context using electroencephalography (EEG), with results suggesting different temporal dynamics underlying performance. Finally, Study 3 incorporated multiple words in noise as targets and manipulated the participants’ instruction of cue usage, finding dissociable effects of instruction, relatedness among the targets, and relatedness of a subsequent cue to targets. These findings are discussed in the context of models of SIN processing, semantic priming, and AtoM.