Department of Communication Disorders, University of Arkansas, Fayetteville AR, USA.
Department of Audiology and Speech Pathology, University of Tennessee Health Science Center, Knoxville TN, USA.
Front Psychol. 2014 May 7;5:366. doi: 10.3389/fpsyg.2014.00366. eCollection 2014.
Oscillatory models of speech processing have proposed that rhythmic cortical oscillations in sensory and motor regions modulate speech sound processing from the bottom-up via phase reset at low frequencies (3-10 Hz) and from the top-down via the disinhibition of alpha/beta rhythms (8-30 Hz). To investigate how the proposed rhythms mediate perceptual performance, electroencephalographic (EEG) was recorded while participants passively listened to or actively identified speech and tone-sweeps in a two-force choice in noise discrimination task presented at high and low signal-to-noise ratios. EEG data were decomposed using independent component analysis and clustered across participants using principle component methods in EEGLAB. Left and right hemisphere sensorimotor and posterior temporal lobe clusters were identified. Alpha and beta suppression was associated with active tasks only in sensorimotor and temporal clusters. In posterior temporal clusters, increases in phase reset at low frequencies were driven by the quality of bottom-up acoustic information for speech and non-speech stimuli, whereas phase reset in sensorimotor clusters was associated with top-down active task demands. A comparison of correct discrimination trials to those identified at chance showed an earlier performance related effect for the left sensorimotor cluster relative to the left-temporal lobe cluster during the syllable discrimination task only. The right sensorimotor cluster was associated with performance related differences for tone-sweep stimuli only. Findings are consistent with internal model accounts suggesting that early efferent sensorimotor models transmitted along alpha and beta channels reflect a release from inhibition related to active attention to auditory discrimination. Results are discussed in the broader context of dynamic, oscillatory models of cognition proposing that top-down internally generated states interact with bottom-up sensory processing to enhance task performance.
言语处理的震荡模型提出，感觉和运动区域的节律性皮质震荡通过低频（3-10 Hz）的相位重置以及通过α/β节律（8-30 Hz）的去抑制从下向上和从上向下调节言语声音处理。为了研究所提出的节律如何调节感知表现，在高信噪比和低信噪比的噪声辨别任务中，参与者被动地聆听或主动地识别言语和声调扫描时，记录了脑电图（EEG）。使用独立成分分析对 EEG 数据进行分解，并使用 EEGLAB 中的主成分方法对参与者进行聚类。确定了左、右半球感觉运动和后颞叶集群。仅在感觉运动和颞叶集群中，α 和β 抑制与主动任务相关。在后颞叶集群中，低频相位重置的增加是由言语和非言语刺激的下传声音信息的质量驱动的，而感觉运动集群中的相位重置与自上而下的主动任务需求相关。将正确辨别试验与机会辨别试验进行比较，仅在音节辨别任务中，左感觉运动集群相对于左颞叶集群表现出与任务相关的更早的效应。仅对于声调扫描刺激，右感觉运动集群与与任务相关的差异相关。研究结果与内部模型理论一致，该理论表明，沿α 和β 通道传递的早期传出感觉运动模型反映了与听觉辨别主动注意相关的抑制释放。研究结果在提出自上而下的内部生成状态与下传感觉处理相互作用以增强任务表现的动态、震荡认知模型的更广泛背景下进行了讨论。
