A novel diffusion-based model of choice reaching experiments
A novel diffusion-based model of choice reaching experiments
复制标题
一种新颖的基于扩散的选择到达实验模型
DOI：
10.1167/jov.20.11.776
复制
发表时间：
2020
影响因子：
1.8
通讯作者：
Heinke D
中科院分区：
文献类型：
--
作者：
Heinke D
在长期的研究中，先前的研究表明，伸手动作可能会受到注意力选择过程的影响。这种效应的关键证据来自于颜色奇怪任务中的到达轨迹。这些实验表明，到达曲率的调节与颜色启动相关，即，与目标颜色切换的试验相比，颜色重复导致更小的曲率（例如，Song & Nakayama，2008；Moher & Song，2016）。根据这一证据，Heinke 及其同事开发了一种受神经学启发的颜色启动机器人模型（Strauss 等，2015）。重要的是，该模型表明，在神经结构中，注意力选择过程很容易泄漏到运动系统中，从而导致曲率效应。在这里，我们提出了一个替代的、更简单的模型，它允许我们对泄漏效应进行定量研究（例如模型拟合）。这种新颖的模型采用了两个扩散过程 (DP)。（通常 DP 用于模拟感知决策。）这里假设一个 DP 捕获奇数颜色目标的选择（“注意”阶段）并“泄漏”到第二个 DP（“运动”阶段）。这种运动 DP 可以被视为描述了到达轨迹的嘈杂空间进展，并且仅在知觉阶段达到阈值时才开始，该阈值允许我们对到达延迟进行建模。重要的是，我们的模拟研究还表明，这个新模型可以轻松捕获曲率效应。进一步的模拟研究发现，该模型往往表现出到达时间和曲率之间的相关性。然而，这一结果与实证结果相矛盾。有趣的是，对模型进行简单修改，从运动 DP 到感知 DP 的额外反馈，使我们能够解决这个问题。这种反馈可以解释为通过到达轨迹影响视觉注意力的新颖预测。此外，我们将展示贝叶斯模型拟合的结果，强调该模型的潜在用途。
In long-standing research, previous studies have shown that reaching movements can be influenced by attentional selection processes. Critical evidence for this effect stems from reach trajectories in colour-oddity tasks. These experiments showed that the modulation of the reaching curvature is linked to colour priming ie, colour repetitions lead to smaller curvatures compared to trials where target colour switches (eg, Song & Nakayama, 2008; Moher & Song, 2016). Following this evidence, Heinke and colleagues developed a neurologically inspired robotics model for colour priming (Strauss et al., 2015). Critically, the model shows that in neural structures the attentional selection process easily leak into the motor system causing the curvature effect. Here, we present an alternative, simpler model which allows us to conduct quantitative investigations (eg, model fitting) into the leakage effect. This novel model employs two diffusion processes (DP).(Normally DPs are used to model perceptual decision making.) Here one DP is assumed to capture the selection of the odd colour target (“attentional” stage) and “leaks” into the second DP (“motor” stage). This motor DP can be seen to describe the noisy spatial progression of reaching trajectories and only begins when the perceptual stage reaches a threshold allowing us to model reaching latencies. Importantly, our simulation studies also showed that the curvature effect can be easily captured by this new model. Further simulation studies found that the model tends to exhibit correlation between reaching time and curvature. However, this result contradicts empirical findings. Interestingly, a simple modification of the model, additional feedback from motor DP to perceptual DP, allowed us to fix this problem. This feedback can be interpreted as a novel prediction for an influence of visual attention through reaching trajectories. In addition, we will present results from Bayesian model fitting underlining the potential usefulness of the model.