Decoding parametric attributes of auditory working memories from human brain activity

从人脑活动解码听觉工作记忆的参数属性 听觉工作记忆（WM）是指我们保持和操纵相关声音信息的能力， 支持沟通和解决问题。听觉WM功能障碍在以下个体中得到证实： 言语感知障碍、语言和阅读障碍，以及涉及言语功能障碍的其他障碍。 听觉认知，如中枢听觉处理障碍（CAPD）。更好地理解听觉工作记忆 将有助于开发更精确的生物标志物和针对这些缺陷的有针对性的干预措施。证据 在实验室动物和非实验动物中已经发现了与听觉WM模式相关的神经元激活， 侵入性人类神经成像研究，包括听觉皮层（AC）和大脑其他部位。最近 人类功能磁共振成像数据也提供了一些证据，在AC和额叶皮层的项目特定的激活模式。 然而，听觉工作记忆内容在大脑中的具体存储位置，以及更重要的是， 信息表示仍不清楚。 本研究旨在通过尝试推断人类听觉工作记忆， 它的内容从大脑：使用国家的最先进的非侵入性神经成像和先进的信号分析 方法，我们将寻找皮质激活模式，将预测项目特定的WM信息。我们将 用磁和脑电图（MEG/EEG）非侵入性地检查脑功能，经颅 磁刺激（TMS）和功能性MRI（fMRI），并使用颅内EEG（iEEG）验证结果 术前患者的测量。最近的研究表明，不是持续的激活模式， WM由短期突触促进支持，即，“活动沉默”的人口级机制。我们 假设这些活动沉默的表征可以通过分析通用的后效来解码， 听觉“脉冲刺激”或TMS，用于在记忆过程中“ping”底层皮层网络 上维护我们还假设，虽然听觉工作记忆可能涉及多个大脑的合作， 参与发音运动功能、感知分类或语义处理的区域， 诸如频谱时间调制模式的听觉-感觉属性的保留关键地取决于 进化人为了研究WM的这种听觉属性，我们将使用具有动态涟漪声音刺激的任务， 与人类的发声类似，但抵制非听觉处理策略。 该项目的主要意义在于，它将增加对听觉工作记忆的理解，这是一个至关重要的 认知功能的神经基础仍然难以捉摸。这些结果也可能有助于开发更多 精确的工具来表征听觉WM功能障碍的听力和沟通缺陷，以及 例如阅读障碍、注意力缺陷/多动障碍（ADHD）和精神分裂症。

Decoding parametric attributes of auditory working memories from human brain activity Auditory working memory (WM) refers to our capacity to maintain and manipulate relevant sound information to support communication and problem solving. Auditory WM dysfunctions are evidenced in individuals with speech perception disorders, language and reading impairments, and other disorders involving dysfunction of auditory cognition, such as central auditory processing disorders (CAPD). Better understanding of auditory WM would help develop more precise biomarkers and targeted interventions for these deficits. Evidence for neuronal activation related to auditory WM patterns have been found in laboratory animals as well as in non- invasive human neuroimaging studies, both in auditory cortices (AC) and elsewhere in the brain. Recent human fMRI data also provide some evidence on item-specific activation patterns in ACs and frontal cortices. However, exactly where in the brain auditory WM content is stored and, more importantly, how the memorized information is represented is still unclear. This research program pursues a better understanding of human auditory WM by attempting to infer its contents from the brain: Using state-of-the-art non-invasive neuroimaging and advanced signal-analysis methods we will search for cortical activation patterns that would predict item-specific WM information. We will examine brain function non-invasively with magneto- and electroencephalography (MEG/EEG), transcranial magnetic stimulation (TMS), and functional MRI (fMRI), and validate the findings using intracranial EEG (iEEG) measurements in presurgical patients. Recent studies suggest that, instead of persistent activation patterns, WM is supported by short-term synaptic facilitation, i.e., "activity-silent" population-level mechanisms. We hypothesize that these activity-silent representations could be decoded by analyzing the aftereffects of generic auditory "impulse stimuli" or TMS, which are utilized to "ping" the underlying cortical network during memory maintenance. We also hypothesize that although auditory WM likely involves co-operation of multiple brain regions, which are involved in articulatory-motor functions, perceptual categorization, or semantic processing, the retention of auditory-sensory attributes such as spectrotemporal modulation patterns critically depends on ACs. To examine WM of such auditory attributes, we will use tasks with dynamic ripple sound stimuli, which are spectrotemporally similar to human vocalizations but resist non-auditory processing strategies. The major significance of this project is that it will increase the understanding of auditory WM, a crucial cognitive function whose neuronal bases have remained elusive. The results may also help develop more precise tools for characterizing auditory WM dysfunctions in hearing and communication deficits as well as in disorders such as dyslexia, attention deficit/hyperactivity disorder (ADHD), and schizophrenia.