In addition to the ability to hear, accurate speech-in-noise perception also requires contributions from cognitive abilities. Cognitive abilities examined most often in the context of speech-in-noise (SiN) perception tend to be related to memory, with working memory (WM) the most commonly investigated aspect of memory in this context. According to Baddeley and Hitch (1974, doi:10.1016/S0079-7421(08)60452-1) WM combines two critical components: a storage component and a manipulation component. It is not well understood whether it is the storage or the manipulation component that has the highest predictive value, or whether both components have to combine for maximum predictive value. Also, it is unclear whether the predictive ability of different WM components differs either for different groups of listeners or for different speech-in-noise tasks.

One WM task commonly used in the context of speech perception is the reading span task (RST). The RST exists in many different versions, possibly differing on the emphasis they put on the manipulation and storage component. Here, we are investigating to what extent the setup of different RSTs favours either working-memory versus long-term memory, and which of these two components is more predictive of speech-in-noise perception in young and old English native speakers. Understanding the exact cognitive processes involved in various RSTs and their predictive value for speech-in-noise in various groups of listeners is the first step to use the RST in the clinic meaningfully as part of the hearing rehabilitation process.