Institut für Psychologie, Universität Leipzig, Germany.
Brain Res. 2012 Feb 23;1439:73-81. doi: 10.1016/j.brainres.2011.12.044. Epub 2011 Dec 31.
This study addresses the processing of concurrent sounds based on inharmonicity and onset asynchrony cues. We used harmonic complex sounds with one component starting marginally (40 ms) or considerably (500 ms) earlier than the complex and being slightly (3%) or strongly (13%) inharmonic. To index sound segregation of concurrent events, we measured the object-related negativity (ORN) component of the event-related potential. We contrasted two hypotheses: According to the concurrent-segregation hypothesis, increased onset asynchrony is assumed to promote segregation of the leading partial from the harmonic complex, which should be reflected in increased ORN amplitudes. That is, even with large onset asynchronies concurrent events would be processed by a simultaneous sound segregation mechanism. According to the sequential-integration hypothesis, however, with increased onset asynchrony concurrent cues are assumed to be less considered by simultaneous grouping processes, which should be reflected in attenuated ORN amplitudes for long onset asynchronies. This assumption is based on the notion that due to sequential integration, a stable percept of the leading partial has been developed within ~350 ms after sound onset, thus less processing is required from scene analysis mechanisms based on concurrent cues. Indeed, with increased onset asynchrony ORN was found to decrease, which supports the sequential-integration hypothesis. In line with previous data, ORN was also found to increase with increased inharmonicity. The absence of an inharmonicity×onset asynchrony interaction further supports the assumption that both cues are used in different ways for simultaneous sound segregation.
本研究探讨了基于非谐和性和起始异步线索处理并发声音的问题。我们使用具有一个分量的谐波复合声音，该分量起始时间略有（40 毫秒）或明显较早（500 毫秒），并且略有（3％）或强烈（13％）的非谐性。为了指示并发事件的声音分离，我们测量了事件相关电位的与对象相关的负性（ORN）成分。我们对比了两个假设：根据并发分离假设，增加起始异步性被认为会促进领先部分与谐波复合的分离，这应该反映在 ORN 幅度的增加上。也就是说，即使起始异步性很大，并发事件也将通过同时的声音分离机制进行处理。然而，根据顺序整合假设，随着起始异步性的增加，并发线索被认为不太受同时分组过程的考虑，这应该反映在长起始异步性时 ORN 幅度的衰减上。这一假设基于这样一种观念，即由于顺序整合，在声音起始后约 350 毫秒内已经发展出了领先部分的稳定感知，因此需要基于并发线索的场景分析机制进行较少的处理。事实上，随着起始异步性的增加，ORN 被发现减少，这支持了顺序整合假设。与之前的数据一致，ORN 也随着非谐性的增加而增加。缺乏非谐性×起始异步性相互作用进一步支持这样一种假设，即这两个线索以不同的方式用于同时的声音分离。
