Perceptual And Motor Decisions In The Primate Brain

认知神经科学的一个突出问题是我们如何能够在不移动眼睛的情况下将视觉注意力集中在特定的物体和位置上。为此，我们一直在研究前额视野（FEF）在视觉感知中的作用。FEF位于前额叶皮层，参与将视觉信息转化为移动眼睛的命令。多种证据表明，源自FEF的动眼肌指令的发展介导了视觉空间注意。我们通过记录在没有眼球运动的情况下执行隐蔽视觉搜索任务的猴子FEF中的视觉和扫视相关神经元来验证这一假设。我们发现视觉反应神经元是FEF中注意信号的来源，而不是与眼跳相关的神经元。通过分析和比较这种认知信号在FEF输入和输出中的时间过程，我们发现，在FEF中，视觉场景的空间选择性表征是由非空间选择性输入重新生成的（参考文献3）。这意味着FEF是调节自上而下的视觉注意力的重要信号来源。在参考文献4中，我们证明了猴子FEF中对目标位置的准确感知表征并不总是能预测猴子对目标位置的准确行为报告。这一发现很重要，因为它表明知觉过程和运动过程可以在某种程度上独立运行，并对理解报告知觉决策时公开行为的可靠性具有启示意义。使用一个简单的扩散模型，我们表明与感知加工无关的噪声是决策的重要因素。

One of the outstanding questions of cognitive neuroscience is how we are able to focus visual attention on specific objects and locations without moving our eyes. To this end, we have been investigating the role of the frontal eye field (FEF) in visual perception. The FEF is located in the prefrontal cortex and participates in the transformation of visual information into commands to move the eyes. Multiple lines of evidence suggest that developing oculomotor commands originating from the FEF mediate visual spatial attention. We tested this hypothesis by recording from visual and saccade related neurons in the FEF of monkeys performing covert visual search tasks without eye movements. We found that the visually responsive neurons, and not the saccade-related neurons, are the source of attention signals in FEF. By analyzing and comparing the time course of this cognitive signal in the inputs and the outputs of FEF, we found that a spatially selective representation of the visual scene is generated de novo in FEF from non-spatially selective inputs (ref. 3). This means that FEF is an importante source of signals that mediate top-down focused visual attention. In ref. 4 we show that an accurate perceptual representation of target location in a monkeys' FEF does not always predict an accurate behavioral report of target location by the monkey. This finding is important because it demonstrates that perceptual and motor processes can run somewhat independently and has implications for understanding the reliability of overt behavior in reporting a perceptual decision. Using a simple diffusion model, we show that noise unrelated to perceptual processing is an important factor in decision-making. In another study (ref. 5), we addressed the question of how cognitive processes control spatial attention in the absence of visual input. We recorded frontal eye field activity in monkeys trained to perform a difficult discrimination task in which the monkeys attended the locations of the visual stimuli to be discriminated before they actually appeared. We found that most FEF neurons exhibited elevated activity when a cue informed the monkey that a stimulus would appear. This anticipatory attention-related activity in FEF occurred without any visual stimulation and was not related to motor processes. Together, these studies demonstrate that stimulus-driven and cognitively-driven spatial attention signals are present in FEF and are independent of saccade command signals. Therefore, FEF probably serves an important role in controlling visual spatial attention in addition to its well known role in saccade production. We also examined the neural basis of how we adjust our behavior when confronted with unexpected events that render our current actions inappropriate (ref. 1). We recorded saccade command signals in the FEF of monkeys performing a visual search task in which the target of a search unexpectedly changed position with one of the distractors. Consequently, some saccades were directed in error to the original target location and were followed frequently by unrewarded corrective saccades to the final target location. We found that saccade-related activity producing corrective saccades on error trials begins before visual afferent signals and before error recognition signals could respond to the changed image following the execution of the error saccade. These observations provide direct evidence for a rapid error correction mechanism before an errant behavior can be detected by the visual system or by the error monitoring system. In ref. 2 we compared human and non-human primate behavior during this task in which a saccade target unexpectedly changes locations. We show that human and monkey behavior are indistinguishable and can be modeled by a race between three independent stochastic processes, thereby providing a computational account of saccade production when the image changes unexpectedly. These studies have extended our understanding about the frontal eye field far beyond its familiar role in controlling eye movements. With this knowledge we can design experiments to investigate the flow of sensory information through the brain as it is transformed into perception and action. This work helps us understand the mechanisms of how the brain focuses attention to make perceptual decisions and guide behavior, which is necessary to be able to understand and treat attention-related disorders in humans.