THE ROLE OF SPATIAL SEPARATION ON SELECTIVE AND DISTRIBUTED ATTENTION TO SPEECH

Processing speech in multi-speaker environments poses substantial challenges to the human perceptual and attention system. Moreover, different contexts may require employing different listening strategies. For instance, in some cases individuals pay attention Selectively to one speaker and attempt to ignore all other task-irrelevant sounds, whereas other contexts may require listeners to Distribute their attention among several speakers. Spatial and spectral acoustic cues both play an important role in assisting listeners to segregate concurrent speakers. However, how these cues interact with varying demands for allocating top-down attention is less clear. In the current study, we test and compare how spatial cues are utilized to benefit performance on these different types of attentional tasks.
To this end, participants listened to a concoction of two or four speakers, presented either as emanating from different locations in space or with no spatial separation. In separate trials, participants were required to employ different listening strategies, and detect a target-word spoken either by one pre-defined speaker (Selective Attention) or spoken by any of the speakers (Distributed Attention).
Results indicate that the presence of spatial cues improved performance, particularly in the two-speaker condition, which is in line with the important role of spatial cues in stream segregation. However, spatial cues provided similar benefits to performance under Selective and Distributed attention. This pattern suggests the despite the advantage of spatial cues for stream segregation, they were nonetheless insufficient for directing a more focused ‘attentional spotlight’ towards the location of a designated speaker in the Selective attention condition.

在多说话人环境中处理语音对人类的感知和注意力系统构成了重大挑战。此外，不同的语境可能需要采用不同的听力策略。例如，在某些情况下，个体会选择性地关注一位说话人，并试图忽略所有其他与任务无关的声音，而在其他语境中，听众可能需要将注意力分散在多位说话人之间。空间和频谱声学线索在帮助听众区分同时出现的说话人中发挥着重要作用。然而，这些线索如何与分配注意力时不断变化的需求相互作用尚不明确。在当前研究中，我们测试并比较了空间线索如何被利用以提升不同类型注意力任务的表现。为此，参与者聆听了两位或四位说话人的混合语音，这些语音要么来自空间中的不同位置，要么没有空间分隔。在单独的试验中，参与者需要采用不同的听力策略，并检测一位预先定义的说话人所说的目标词（选择性注意力）或任何说话人所说的目标词（分散注意力）。结果表明，空间线索的存在改善了表现，尤其是在两位说话人的条件下，这与空间线索在流分离中的重要作用相一致。然而，空间线索在选择性注意力和分散注意力下对表现提供了相似的好处。这一模式表明，尽管空间线索在流分离方面具有优势，但它们仍然不足以在选择性注意力条件下将更集中的‘注意力焦点’指向指定说话人的位置。