Contextual influences on object recognition investigated by means of steady-state visual evoked potentials

本研究旨在探讨上下文信息如何影响目标识别。上下文信息可以通过风景背景或几个对象之间的语义关联来提供。这两种场景的特点都是对视觉输入的所谓“要点”的感知，即对场景一般意义的感知。到目前为止，这些背景影响对物体识别的神经元机制还没有得到很好的理解。造成这一缺陷的一个原因是方法论上的问题，即无法将上下文引起的大脑反应与与目标物体相关的皮层激活分开。更准确地说，传统的神经科学方法（例如BOLD信号）仅仅将混合信号反映到构成多刺激显示的所有元素上。我们打算利用稳态视觉诱发电位（SSVEP）来克服这一限制，这是脑电图（EEG）研究中的一种特殊技术。除了其他优点外，SSVEP方法还可以通过将每个刺激标记为单独的驱动频率（例如，以12 Hz呈现熟悉的物体，以9 Hz呈现相关的风景背景）来解开由同时呈现的刺激引起的神经元反应。接下来，将EEG信号转换为其频谱分量（在上面的例子中为12hz的目标响应和9hz的背景响应）。我们建议进行六个实验，从而利用SSVEP方法的主要优势。这些实验旨在回答以下两个主要的研究问题：(1)在语境中物体识别的神经元基础是什么？(2)哪些基本认知功能是情境对物体识别的影响（如特定记忆过程）的基础？综上所述，该研究项目将提供关于自然观看条件下（即与上下文信息）物体识别的神经元和认知机制的系统信息。因此，建议的实验将补充和扩展目前从孤立（即没有背景）研究物体识别中获得的知识。

The present research proposal aims at investigating how context information influences object recognition. Contextual information can be provided by means of a scenic background or by semantic associations between several objects. Both scenarios are characterized by the perception of the so-called "gist" of the visual input, i.e. the perception of the general meaning of a scene. Up-today, the neuronal mechanisms underlying these contextual influences on object recognition are not well understood. One reason for this shortcoming lies in the methodological problem to separate the brain response elicited by a context from the cortical activation related to a target object. More precisely, conventional neuroscientific methods (e.g. BOLD signals) solely mirror the intermingled signal to all constituting elements of a multi-stimulus display. We intend to overcome this limitation by making use of the steady-state visual evoked potential (SSVEP), a specific technique in electroencephalogram (EEG) research. Besides other advantages, the SSVEP method can disentangle neuronal responses elicited by simultaneously presented stimuli by means of tagging each stimulus with an individual driving frequency (for example by presenting a familiar object at 12 Hz and a related scenic background at 9 Hz). Next, the EEG signal is transformed to its spectral components (in the above example the 12 Hz object response and the 9 Hz background response). We suggest conducting six experiments thereby taking advantage of the SSVEP methods major strengths. These experiments aim at answering the following two main research questions: (1) What are the neuronal underpinnings of objects recognition in context? (2) Which basic cognitive functions underlie contextual influences on object recognition (e.g. specific memory processes)? To sum, the proposed research project will provide systematic information about neuronal and cognitive mechanisms of object recognition under natural viewing conditions (i.e. with contextual information). The suggested experiments will thereby complement and extend current knowledge derived from studying object recognition in isolation (i.e. without context).