Schütz Alexander C, Souto David
Abteilung Allgemeine Psychologie, Justus-Liebig-Universität Gießen Gießen, Germany ; Allgemeine und Biologische Psychologie, Philipps-Universität Marburg Marburg, Germany.
Abteilung Allgemeine Psychologie, Justus-Liebig-Universität Gießen Gießen, Germany ; Department of Neuroscience, Psychology and Behaviour, University of Leicester Leicester, UK.
Front Hum Neurosci. 2015 Oct 20;9:566. doi: 10.3389/fnhum.2015.00566. eCollection 2015.
Adaptation of saccades can be induced by different error signals, such as retinal position errors, prediction errors, or reinforcement learning. Recently, we showed that a shift in the spatial goal of a perceptual task can induce saccadic adaptation, in the absence of a bottom-up position error. Here, we investigated whether this top-down effect is mediated by the visibility of the task-relevant object, by reinforcement due to the feedback about the perceptual judgment or by a target selection mechanism. Participants were asked to discriminate visual stimuli arranged in a vertical compound. To induce adaptation, the discrimination target was presented at eccentric locations in the compound. In the first experiment, we compared adaptation with an easy and difficult discrimination. In the second experiment, we compared adaptation when feedback about the perceptual task was valid and when feedback was provided but was unrelated to performance. In the third experiment, we compared adaptation with instructions to fixate one of the elements in the compound-target selection-to the perceptual task condition-target selection and discrimination. To control for a bottom-up stimulus effect, we ran a fourth experiment in which the only instruction was to look at the compound. The saccade amplitude data were fitted by a two-state model distinguishing between an immediate and a gradual error correction process. We replicated our finding that a perceptual task can drive adaptation of saccades. Adaptation showed no effect of feedback reliability, nor an effect of the perceptual task beyond target selection. Adaptation was induced by a top-down signal since it was absent when there was no target selection instruction and no perceptual task. The immediate error correction was larger for the difficult than for the easy condition, suggesting that task difficulty affects mainly voluntary saccade targeting. In addition, the repetition of experiments one week later increased the magnitude of the gradual error correction. The results dissociate two distinct components of adaptation: an immediate and a gradual error correction. We conclude that perceptual-task induced adaptation is most likely due to top-down target selection within a larger object.
扫视适应可由不同的误差信号诱发，如视网膜位置误差、预测误差或强化学习。最近，我们发现，在没有自下而上的位置误差的情况下，感知任务空间目标的改变也能诱发扫视适应。在此，我们研究了这种自上而下的效应是由与任务相关物体的可见性介导，还是由感知判断反馈的强化作用或目标选择机制介导。我们要求参与者辨别垂直排列的复合视觉刺激。为了诱发适应，辨别目标呈现在复合刺激的偏心位置。在第一个实验中，我们比较了简单辨别和困难辨别情况下的适应情况。在第二个实验中，我们比较了感知任务反馈有效时和反馈与表现无关时的适应情况。在第三个实验中，我们比较了在复合刺激中固定注视其中一个元素（目标选择）的指令下的适应情况与感知任务条件下的目标选择和辨别情况。为了控制自下而上的刺激效应，我们进行了第四个实验，其中唯一的指令是注视复合刺激。扫视幅度数据由一个区分即时和渐进误差校正过程的双状态模型拟合。我们重复了我们的发现，即感知任务可以驱动扫视适应。适应情况不受反馈可靠性的影响，除目标选择外也不受感知任务的影响。适应是由自上而下的信号诱发的，因为在没有目标选择指令和感知任务时就不存在适应。困难条件下的即时误差校正比简单条件下更大，这表明任务难度主要影响自愿性扫视目标定位。此外，一周后重复实验增加了渐进误差校正的幅度。结果区分了适应的两个不同组成部分：即时和渐进误差校正。我们得出结论，感知任务诱发的适应很可能是由于在更大物体内的自上而下的目标选择。
