Max Planck Research Group "Auditory Cognition", Max Planck Institute for Human Cognitive and Brain Sciences, Stephanstraβe 1A, 04103 Leipzig, Germany.
Research Group "Subcortical contributions to comprehension", Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany; School of Psychological Sciences, The University of Manchester, Manchester, UK.
Neuroimage. 2014 Aug 15;97:387-95. doi: 10.1016/j.neuroimage.2014.04.005. Epub 2014 Apr 18.
Slow neural oscillations (1-15 Hz) are thought to orchestrate the neural processes of spoken language comprehension. However, functional subdivisions within this broad range of frequencies are disputed, with most studies hypothesizing only about single frequency bands. The present study utilizes an established paradigm of spoken word recognition (lexical decision) to test the hypothesis that within the slow neural oscillatory frequency range, distinct functional signatures and cortical networks can be identified at least for theta- (3-7 Hz) and alpha-frequencies (~8-12 Hz). Listeners performed an auditory lexical decision task on a set of items that formed a word-pseudoword continuum: ranging from (1) real words over (2) ambiguous pseudowords (deviating from real words only in one vowel; comparable to natural mispronunciations in speech) to (3) pseudowords (clearly deviating from real words by randomized syllables). By means of time-frequency analysis and spatial filtering, we observed a dissociation into distinct but simultaneous patterns of alpha power suppression and theta power enhancement. Alpha exhibited a parametric suppression as items increasingly matched real words, in line with lowered functional inhibition in a left-dominant lexical processing network for more word-like input. Simultaneously, theta power in a bilateral fronto-temporal network was selectively enhanced for ambiguous pseudowords only. Thus, enhanced alpha power can neurally 'gate' lexical integration, while enhanced theta power might index functionally more specific ambiguity-resolution processes. To this end, a joint analysis of both frequency bands provides neural evidence for parallel processes in achieving spoken word recognition.
慢神经振荡（1-15Hz）被认为协调口语理解的神经过程。然而，该广泛频率范围内的功能细分存在争议，大多数研究仅假设单一的频带。本研究利用已建立的口语识别范式（词汇决策）来检验以下假设：在慢神经振荡频率范围内，至少在θ波（3-7Hz）和α波（~8-12Hz）频率范围内，可以识别出不同的功能特征和皮质网络。听众对一组形成单词-伪单词连续体的项目执行听觉词汇决策任务：从（1）真实单词到（2）模棱两可的伪单词（仅在一个元音上与真实单词不同；与言语中的自然发音错误相当）再到（3）伪单词（通过随机音节明显偏离真实单词）。通过时频分析和空间滤波，我们观察到一种独特但同时发生的α功率抑制和θ功率增强模式的分离。随着项目越来越符合真实单词，α表现出参数抑制，这与更类似于单词的输入的左优势词汇处理网络中的功能抑制降低一致。同时，仅在模棱两可的伪单词中，双侧额颞网络中的θ功率选择性增强。因此，增强的α功率可以在神经上“门控”词汇整合，而增强的θ功率可能指数值更具体的歧义解决过程。为此，两个频带的联合分析为实现口语识别的并行过程提供了神经证据。
