Auditory scene analysis and attentional focusing in speech perception under complex dynamic listening conditions in younger and older adults

在不利的（鸡尾酒会）听力条件下的言语理解随着年龄的增长而恶化。据推测，除了真正的听力缺陷，年龄相关的认知能力下降有助于这些困难。在该项目的第一阶段，我们证明了注意力和抑制控制的下降会影响言语感知，特别是在高度动态的听觉场景中和发生扬声器变化后。在行为表现中观察到的缺陷与年龄相关的变化，电生理措施，反映了增加分心，灵活的注意力分配，减少准备的相关语音信息的处理在老年人。然而，以前的研究也表明，老年人受益于提示听觉场景和扬声器设置的变化，其程度与年轻人相似。一种非常自然的提示类型是由说话人的面部提供的视觉语音信息。众所周知，视觉语音可以大大提高（听觉）语音理解。这种双模态视听（AV）促进是基于这样一个事实，即视觉语音携带一个特定的预测值的听觉话语。听觉和视觉语音信息的整合降低了信号的不确定性，并支持相关语音信息的提取，特别是在不利的收听条件下。有趣的是，事件相关电位（ERP）的分析显示，老年人从双峰语音信息中受益更多。在高度动态的鸡尾酒会情况下，视听语音信息可以在多大程度上增强语音理解，目前还不清楚。然而，关于我们以前的结果，可以预期，视觉语音也有积极的影响时，在听觉场景和扬声器设置发生变化的语音理解。在该项目的第二阶段，应在四个实验中对此进行调查，使用真实的听力情景和结合的行为和电生理措施。通过将听觉语音与（a）一致的视觉语音，（B）产生非（语音）特定嘴唇运动的面部，以及（c）静态面部相结合，应在年轻人和老年人中测试动态听力条件下AV语音的具体益处。实验的重点是改变扬声器设置和切换成本与变化相关的注意力切换。将通过ERP方法分析潜在的皮层过程。在这里，分析将集中在ERP被证明是敏感的扬声器的变化（N400diff和LPCdiff），注意力转换（N2ac和LPCpc），和年龄（N2和CNV）在第一阶段的项目。总之，新实验的目的是更多地了解视听言语加工的神经生理机制及其与不利听力下言语理解益处的关系。

Speech comprehension under adverse (cocktailparty) listening conditions deteriorates with age. It is assumed that, in addition to genuine hearing deficits, age-related declines in cognitive abilities contribute to these difficulties. In the first phase of the project, we demonstrated that declines in attentional and inhibitory control affect speech perception especially in highly dynamic auditory sceneries and after the occurrence of speaker changes. Deficits observed in behavioural performance were associated with age-related changes in electrophysiological measures, reflecting increased distraction, less flexible attention allocation, and reduced preparation of the processing of relevant speech information in older adults. However, previous studies also showed that older adults benefit from cues indicating changes in the auditory scenery and in speaker settings to a similar degree as younger adults. A quite natural type of cue is visual speech information provided by the speakers face. It is long known that visual speech can substantially improve (auditory) speech comprehension. This bimodal audio-visual (AV) facilitation is based on the fact that visual speech carries a specific predictive value for the auditory utterance. The integration of auditory and visual speech information reduces signal uncertainty and supports the extraction of relevant speech information especially under adverse listening conditions. Interestingly, analyses of eventrelated potentials (ERPs) revealed that older adults benefit even more from bimodal speech information than younger ones. The degree to which audio-visual speech information may enhance speech comprehension also in highly dynamic cocktailparty situations is yet unclear. Regarding our previous results, however, it can be expected that visual speech also has a positive effect on speech comprehension when changes in the auditory scenery and in speaker settings occur. In the second phase of the project, this should be investigated in four experiments, using realistic listening scenarios and combined behavioural and electro-physiological measures. By combining auditory speech with (a) congruent visual speech, (b) faces that produce non(speech)specific lip movements, and (c) static faces, the specific benefits of AV speech under dynamic listening conditions should be tested in younger and older adults. The experiments focus on changing speaker settings and on switch costs associated with change-related attentional switching. The underlying cortical processes will be analysed by ERP methodology. Here, the analysis will focus on ERPs that proved to be sensitive to speaker changes (N400diff und LPCdiff), to attentional switching (N2ac and LPCpc), and to age (N2 and CNV) in the first phase of the project. In sum, the aim of the new experiments is to learn more about the neurophysiological mechanisms of audio-visual speech processing and their relation to benefits in speech comprehension under adverse listening.