McPeek R M, Keller E L
The Smith-Kettlewell Eye Research Institute, 2318 Fillmore St., San Francisco, CA, USA.
Vision Res. 2001 Mar;41(6):785-800. doi: 10.1016/s0042-6989(00)00287-x.
In human subjects, two mechanisms for improving the efficiency of saccades in visual search have recently been described: color priming and concurrent processing of two saccades. Since the monkey provides an important model for understanding the neural underpinnings of target selection in visual search, we sought to explore the degree to which the saccadic system of monkeys uses these same mechanisms. Therefore, we recorded the eye movements of rhesus monkeys performing a simple color-oddity pop-out search task, similar to that used previously with human subjects. The monkeys were rewarded for making a saccade to the odd-colored target, which was presented with an array of three distractors. The target and distractors were randomly chosen to be red or green in each trial. Similar to what was previously observed for humans, we found that monkeys show the influence of a cumulative, short-term priming mechanism which facilitates saccades when the color of the search target happens to repeat from trial to trial. Furthermore, we found that like humans, when monkeys make an erroneous initial saccade to a distractor, they are capable of executing a second saccade to the target after a very brief inter-saccadic interval, suggesting that the two saccades have been programmed concurrently (i.e. in parallel). These results demonstrate a close similarity between human and monkey performance. We also made a new observation: we found that when monkeys make such two-saccade responses, the trajectory of the initial saccade tends to curve toward the goal of the subsequent saccade. This provides evidence that the two saccade goals are simultaneously represented on a common motor map, supporting the idea that the movements are processed concurrently. It also indicates that concurrent processing is not limited to brain areas involved in higher-level planning; rather, such parallel programming apparently occurs at a low enough level in the saccadic system that it can affect saccade trajectory.
在人类受试者中，最近描述了两种提高视觉搜索中扫视效率的机制：颜色启动和两个扫视的并行处理。由于猴子为理解视觉搜索中目标选择的神经基础提供了重要模型，我们试图探究猴子的扫视系统在多大程度上使用这些相同机制。因此，我们记录了恒河猴执行简单颜色奇异性弹出式搜索任务时的眼动，该任务类似于之前用于人类受试者的任务。猴子向奇色目标进行扫视会得到奖励，该目标与三个干扰物组成的阵列一起呈现。在每次试验中，目标和干扰物被随机选择为红色或绿色。与之前对人类的观察结果相似，我们发现猴子表现出一种累积的短期启动机制的影响，当搜索目标的颜色在试验间碰巧重复时，这种机制会促进扫视。此外，我们发现与人类一样，当猴子对干扰物进行错误的初始扫视时，它们能够在非常短的扫视间隔后向目标执行第二次扫视，这表明这两个扫视是并行编程的（即同时进行）。这些结果证明了人类和猴子表现之间的密切相似性。我们还进行了一项新的观察：我们发现当猴子做出这种双扫视反应时，初始扫视的轨迹往往会朝着后续扫视的目标弯曲。这提供了证据，表明两个扫视目标在一个共同的运动图谱上同时被表征，支持了这些运动是并行处理的观点。这也表明并行处理并不局限于参与高级规划的脑区；相反，这种并行编程显然发生在扫视系统中足够低的水平，以至于它可以影响扫视轨迹。
