Examining the neural network dynamics underlying the temporal integration window of crossmodal perception

长期以来，认知神经科学一直试图解释这样一种观察，即意识知觉显然是由刺激开始前后发生的神经过程塑造的。研究表明，刺激前的神经活动会影响我们的知觉。此外，在刺激后几百毫秒内呈现的信息仍然会影响对刺激的感知。我们用后记这个词来描述这种倒退的知觉现象。总而言之，这些发现表明，知觉是由对刺激周围的时间窗口中呈现的信息进行综合处理而形成的。在神经层面上，许多关于后显现象的理论解释都假设反馈和前馈加工之间存在相互作用。此外，有证据表明，同步的大脑活动，如神经振荡所表现的，是影响感知的神经机制。在这个项目中，我们将测试反馈和前馈过程，分别反映在β频段(13-30赫兹)和伽马频段(&gt；30赫兹)振荡中，以及诱发的大脑反应如何有助于后判断性跨模式知觉。我们将结合贝叶斯因果参考(BCI)和脑电(EEG)数据分析来研究两种交叉模式的神经机制：幻觉视听(AV)兔和隐形AV兔。在这些范式中，要么虚幻的视觉闪光是由随后的视听刺激诱导的(幻觉AV兔)，要么是对真实闪光的感知被抑制(看不见的AV兔)。选择这些范式有两个原因：(1)跨峰过程以前与神经振荡有关，它们需要综合的网络处理。(Ii)这些范式唤起了相反的知觉，即诱导或抑制闪光，这些范式背后的神经机制可以直接比较。我们将研究神经网络动力学在振荡中的作用，并从诱发的大脑反应模式中解码交叉模式后行为的特征。使用脑-机接口，我们将检验这样的假设，即如果观察者推断出闪光和声音的共同原因，就会出现后遗症。我们将考察先前的因果预期和可能性，即刺激在观察者身上诱导特定知觉的感觉证据，如何影响这两种范式中的知觉和神经加工。这个联合的脑电建模项目将提高我们对时间整合窗口下的神经网络动力学如何塑造跨模式知觉的理解。

Cognitive neuroscience has long sought to explain the observation that conscious perception is apparently shaped by neural processes occurring both before and after stimulus onset. Studies have revealed that neural activity preceding a stimulus influence our perception. Additionally, information that is presented within a few hundreds of milliseconds after a stimulus can still affect perception of the stimulus. We use the term postdiction to describe such backwards-oriented perceptual phenomena. Together, these findings suggest that perception is shaped by the integrative processing of information presented in a temporal window surrounding a stimulus. At the neural level, many theoretical accounts of postdiction phenomena postulate an interplay between feedback and feedforward processing. Furthermore, there is evidence that synchronized brain activity, as expressed in neural oscillations, serves as a neural mechanism that affects perception. To date, it is not well understood how neural oscillations in the temporal integration window contribute to postdiction phenomena.In this project, we will test how feedback and feedforward processes, as reflected in beta-band (13-30 Hz) and gamma-band (> 30 Hz) oscillations, respectively, as well as evoked brain responses, contribute to postdictive crossmodal perception. We will combine Bayesian causal interreference (BCI) and electroencephalography (EEG) data analysis to study the neural mechanisms underlying two crossmodal paradigms: the illusory audiovisual (AV) rabbit and the invisible AV rabbit. In these paradigms, either an illusory visual flash is induced by a subsequent audiovisual stimulus (illusory AV rabbit), or the perception of a real flash is suppressed (invisible AV rabbit). The paradigms were selected for two reasons: (i) Crossmodal processes have been previously linked to neural oscillations and they require integrative network processing. (ii) The paradigms evoke opposite percepts, i.e., inducing or suppressing a flash, and the neural mechanisms underlying these paradigms can be directly compared. We will investigate the role of neural network dynamics in oscillations and decode signatures of crossmodal postdiction from patterns of evoked brain responses. Using BCI, we will test the assumption that postdiction can arise if observers infer a common cause of the flashes and sounds. We will examine how prior causal expectations and the likelihood, i.e., the sensory evidence that a stimulus induces a particular percept in an observer, influence perception and neural processing in the two paradigms. This combined EEG-modelling project will improve our understanding about how neural network dynamics underlying the temporal integration window shape crossmodal perception.