Attention and Auditory Scene Analysis

点击翻译按钮获取中文摘要

Auditory scene analysis (ASA) is an area of study in cognitive neuroscience that seeks to understand how we perceive and identify sound objects in the environment from the soundscape (i.e., a mixture of sound waveforms) that reaches our ears. To date, ASA research has primarily focused on low-level feature-driven processes using a wide range of pure tone configurations, whereas the influence of prior knowledge on ASA has received relatively little attention. This leaves a large gap in our understanding of ASA because, while we may suspect that schema-driven processes play a major role in perception of everyday auditory environments, we do not yet understand how prior knowledge (i.e., auditory schema) impacts the grouping of sound features. Under what conditions do schema-driven processes influence feature-driven processes? How does the brain deal with sound objects that fall outside of attention, and does this depend on prior experience or expectation? These are just a few of the questions that, when answered, will advance theories and training programs that strive to improve listening skills in adverse listening conditions. This comprehensive research program includes a series of experiments on auditory attention and ASA with an emphasis on how feature- and schema-driven ASA interact during soundscape perception. My working hypothesis is that schema-driven rather than feature-driven ASA is the primary determinant of soundscape perception. The specific aims of this research program are to: (i)characterize the interaction between feature- and schema-driven ASA; (ii)determine under which conditions we use local vs. global properties of a soundscape when we perceive, notice changes, and recognize previously heard soundscapes; (iii)examine the effect of implicit learning on attention and ASA; (iv)assess how involuntary attentional capture influences feature- and schema-driven ASA. All studies will include adults with normal hearing and measurements of neuroelectric brain activity while participants are presented with a virtual auditory environment. We will use familiar and unfamiliar material and will manipulate task instructions and acoustic parameters such as timbre and location to influence the separation and identification of distinct sound objects. The proposed research will advance theories of attention and ASA by revealing under what conditions schema-driven ASA influences how spectral and temporal properties of sounds are bound together into sound object representations. Determining the balance between feature- and schema-driven ASA will help to develop training programs targeted towards improving listening skills in normal hearing as well as individuals with brain damage, hearing loss, and other medical conditions such as deficits in attentional regulation and executive functions. Lastly, this research will be transformative by demonstrating that soundscape perception is first and foremost guided by top-down processes.