"By 2025, a quarter of Europeans will be over 65 years old. In everyday life, older adults need to communicate efficiently, yet they often find it more difficult than younger adults to attend to target speech when other people are talking, specifically when operating in their second language. This could be the outcome of age-related cognitive and/or sensory changes. To address these issues, the current research proposal highlights three themes which contrast the performance of younger and older, native and non-native speakers, in two types of noise, speech-spectrum and a multi-talker babble: 1. The time-line needed to segregate target speech from noise. Listeners are asked to repeat spoken words presented in different delays between the onset of a spoken word and a noise masker. Four signal-to-noise ratios (SNR) are employed to determine the SNR for 50% thresholds in each delay. Functions relating thresholds to onset-delay describe the time-line for stream segregation. 2. Real-time lexical processing of spoken words using eye-tracking. Listeners follow spoken instructions referring to depicted objects, e.g., ""Look at the candle"". Eye movements capture listeners' ability to differentiate the target (candle) from a phonological (candy) or semantic alternative (lamp) as the word unfolds in time. 3. Identification of emotions in speech. Lexical sentences, associated with one of five emotions (anger, fear, sadness, happiness and neutral) are recorded spoken in these five distinct prosodies, creating all possible combinations of emotions across both dimensions. Listeners are asked to rate the degree to which each token expressed each emotion in the lexical content, prosody or both, revealing the interplay between the two dimensions, and the ability to selectively attend to one of them. Finally, each theme generates contrasting predictions for cognitive and sensory aging theories, and suggests possible strategies that can mitigate the outlined communication difficulties."