Task-modulated Sensitivity to Vocal Pitch in the Dorsal Premotor Cortex during Multitalker Speech Recognition.
Task-modulated Sensitivity to Vocal Pitch in the Dorsal Premotor Cortex during Multitalker Speech Recognition.
复制标题
多说话者语音识别过程中背侧前运动皮层对音高的任务调制敏感性。
DOI：
10.1162/jocn_a_01907
复制
发表时间：
2022
影响因子：
3.2
通讯作者：
Yi,Alex
中科院分区：
文献类型：
--
作者：
Venezia,JonathanH;Herrera,Christian;Whittle,Nicole;Leek,MarjorieR;Barnes,Samuel;Holshouser,Barbara;Yi,Alex
人们早就知道，听语言会激活额下运动区（前运动区）和背侧运动前区（dPM）。最近的研究表明，dPM位于喉部运动皮层附近，除了音素类别等高级语音线索外，还对包括音调和语音包络在内的低级语音线索做出反应。一种新出现的假设是，dPM是听觉引导的喉部控制回路的一部分，在产生语言和其他自主的听觉-发声行为中起作用。我们最近报道了一项研究，在一项退化的语音识别任务中，dPM对音高有反应，但只有当语音被评为难以理解时；可理解言语和不可理解言语之间的分类差异对dPM的调节更为稳健。与一般的听觉-声音假说相反，这表明可理解的言语是dPM的主要驱动因素。然而，在音调敏感的听觉皮层中也观察到相同的结果。至关重要的是，音高与可理解性判断任务无关，这可能以牺牲音高线索为代价促进了语音信息的处理。目前的fMRI研究（n= 25）验证了这样一种假设，即对于多说话者任务，强调音调以分离说话者，左dPM和音调敏感的听觉区域将对音调做出反应，而不管整体的语音清晰度如何。这表明音调处理确实是该回路的主要关注点，只有在任务需要时才会在感知过程中显现出来。在两种情况下（Competing, Unison），仅一种情况下需要基于音高的隔离（Competing），利用光谱时间调制失真来独立调制两种说话者（男性/女性）话语的音高和语音内容。贝叶斯分层漂移-扩散模型用于预测语音识别性能，从光谱时间畸变模式施加在每次试验上。该模型的漂移率参数与《Competing》的音高密切相关，但与《Unison》无关。使用第二个贝叶斯层次模型，我们确定了dPM中与行为相关的声学特征与fMRI激活相关的区域。我们对分层漂移-扩散模型的后验预测进行了回归，反映了每次试验中与行为相关的声学特征的相对存在或不存在，而不是试验激活幅度。在两种情况下，在左侧dPM和双侧听觉皮层观察到总体漂移率显著正相关，反映了与总体可理解性相关的音调和语音线索。在左侧dPM中观察到与“音调受限”漂移率的显著正相关，仅反映与行为相关的音调线索的相对存在或不存在，而不管语音内容的存在或不存在（可理解性），但仅在竞争条件下。有趣的是，在两种情况下，双侧听觉皮层也观察到同样的效果。事后中介分析排除了决策负荷是负责观察到的音高效应的可能性。这些发现表明，音高的处理是听觉皮层- dPM回路的主要关注点，尽管在感知核心音高的过程中，处理是由听觉皮层进行的，听觉皮层受到dPM的潜在调节影响。
It has long been known that listening to speech activates inferior frontal (pre-)motor regions in addition to a more dorsal premotor site (dPM). Recent work shows that dPM, located adjacent to laryngeal motor cortex, responds to low-level acoustic speech cues including vocal pitch, and the speech envelope, in addition to higher-level cues such as phoneme categories. An emerging hypothesis is that dPM is part of a general auditory-guided laryngeal control circuit that plays a role in producing speech and other voluntary auditory–vocal behaviors. We recently reported a study in which dPM responded to vocal pitch during a degraded speech recognition task, but only when speech was rated as unintelligible; dPM was more robustly modulated by the categorical difference between intelligible and unintelligible speech. Contrary to the general auditory–vocal hypothesis, this suggests intelligible speech is the primary driver of dPM. However, the same pattern of results was observed in pitch-sensitive auditory cortex. Crucially, vocal pitch was not relevant to the intelligibility judgment task, which may have facilitated processing of phonetic information at the expense of vocal pitch cues. The present fMRI study (n= 25) tests the hypothesis that, for a multitalker task that emphasizes pitch for talker segregation, left dPM and pitch-sensitive auditory regions will respond to vocal pitch regardless of overall speech intelligibility. This would suggest that pitch processing is indeed a primary concern of this circuit, apparent during perception only when the task demands it. Spectrotemporal modulation distortion was used to independently modulate vocal pitch and phonetic content in two-talker (male/female) utterances across two conditions (Competing, Unison), only one of which required pitch-based segregation (Competing). A Bayesian hierarchical drift-diffusion model was used to predict speech recognition performance from patterns of spectrotemporal distortion imposed on each trial. The model's drift rate parameter, ad′-like measure of performance, was strongly associated with vocal pitch for Competing but not Unison. Using a second Bayesian hierarchical model, we identified regions where behaviorally relevant acoustic features were related to fMRI activation in dPM. We regressed the hierarchical drift-diffusion model's posterior predictions of trial-wise drift rate, reflecting the relative presence or absence of behaviorally relevant acoustic features from trial to trial, against trial-wise activation amplitude. A significant positive association with overall drift rate, reflecting vocal pitch and phonetic cues related to overall intelligibility, was observed in left dPM and bilateral auditory cortex in both conditions. A significant positive association with “pitch-restricted” drift rate, reflecting only the relative presence or absence of behaviorally relevant pitch cues, regardless of the presence or absence of phonetic content (intelligibility), was observed in left dPM, but only in the Competing condition. Interestingly, the same effect was observed in bilateral auditory cortex but in both conditions. A post hoc mediation analysis ruled out the possibility that decision load was responsible for the observed pitch effects. These findings suggest that processing of vocal pitch is a primary concern of the auditory-cortex–dPM circuit, although during perception core pitch, processing is carried out by auditory cortex with a potential modulatory influence from dPM.