Top-Down Control of Attention

描述（由申请人提供）：心理物理学和生理学实验已经证明，在实际发生之前关注预期的刺激可以增强其感觉处理。控制预期注意力的神经机制的破坏是许多精神和神经疾病的症状，包括精神分裂症和癫痫。神经影像学研究已经确定了对预期注意力有贡献的大脑区域，并提出了一种自上而下的控制机制。目前尚不清楚的是实现这种控制的神经生理信号。本项目有三个目标。首先，我们希望建立α范围（10 Hz）的活动，在当地的场电位和表面脑电图的信号，调解不同方面的自上而下的注意力控制在视觉皮层。其次，我们希望确定颞下皮层可能作为一个中转站，将自上而下的控制信号从前额皮质引导到较低的视觉区域。第三，通过比较刺激出现前后的注意效应，我们希望检验自上而下控制和自下而上加工之间的相互作用。这些目标将通过利用一个独特的资源- 120 GB的高质量多电极局部场电位，表面脑电图和多单元活动数据记录从猴子执行联运选择性注意任务来解决。实现这些目标的一个关键组成部分是一套信号处理方法，用于分析PI及其同事在过去十年中开发和验证的多元神经记录。初步应用的分析方法的数据已经产生了一些见解，以支持我们的研究目标，奠定了基础，成功执行拟议的项目。

DESCRIPTION (provided by applicant): Psychophysical and physiological experiments have demonstrated that attending to an anticipated stimulus before its actual onset can enhance its sensory processing. Disruption of the neural mechanism that controls anticipatory attention is symptomatic of a number of psychiatric and neurological disorders including schizophrenia and epilepsy. Neuroimaging studies have identified the brain areas that contribute to anticipatory attention and suggested a top-down control mechanism. What remains not well understood is the neurophysiological signal that implements such control. The present project has three objectives. First, we wish to establish that alpha range (10 Hz) activities in local field potentials and in surface EEGs are the signals that mediate different aspects of top-down attentional control in the visual cortex. Second, we wish to establish that the inferotemporal cortex may serve as a way station that directs the top-down control signal from the prefrontal cortex to lower visual areas. Third, by comparing the attentional effect before and after stimulus onset, we wish to examine the interplay between top-down control and bottom-up processing. These objectives will be addressed by utilizing a unique resource - 120 GB of high quality multielectrode local field potential, surface EEG and multiunit activity data recorded from monkeys performing an intermodal selective attention task. A key enabling component for accomplishing these objectives is a set of signal processing methods for analyzing multivariate neural recordings developed and validated over the past decade by the PI and his colleagues. Initial applications of the analysis methods to the data have already yielded a number of insights in support of our research aims which lay the foundation for the successful execution of the proposed project.