【作者】 黄炫；

【导师】 周逸峰；

【作者基本信息】 中国科学技术大学 ， 生物物理学， 2007， 博士

【摘要】 知觉学习指依赖于训练，由感觉经验引起的在基本知觉任务上表现的改变。此种改变一般需要训练达到一定强度才会发生，且特异于训练的基本设定(如任务、训练方向、频率、位置等)，并可保持较长时间。知觉学习的特异性一直是研究的热点之一，较早期的知觉学习特异性研究认为特异性反映了知觉学习的低级中枢加工机制，认为可以藉特异性来探究知觉学习的发生位点，但越来越多互相矛盾的实验结果对知觉学习的低级中枢加工机制提出了挑战。Ahissar和Hochstein认为任务难易度是决定知觉学习特异性的关键因素，而Liu和Weinshall甚至认为知觉学习的传递具有普遍意义。本研究主要是从视知觉学习特异性(传递特性)方面出发，分别从任务角度和刺激特征角度研究两种不同形式运动点的知觉学习。包括不同随机运动点任务的知觉学习特异性(实验一)和径向运动点运动速度辨别任务的知觉学习特异性的研究(实验二)。实验一比较了随机运动点知觉学习中两种不同任务的学习特异性。被试在训练前后需要进行两种任务(信号点运动识别任务和信号点运动方向辨别任务)的心理物理曲线的测量，通过比较前后结果来衡量不同训练任务的学习特异性。实验将被试随机分为两组进行训练：一组进行信号点识别任务的训练即识别训练(12人)；另外一组进行信号点运动方向辨别任务的训练即辨别训练(11人)。结果发现：识别训练组的被试在训练后运动识别任务成绩有显著性的提高，而运动方向辨别任务成绩没有明显变化，即信号点运动识别任务的知觉学习具有任务特异性；不同的是辨别训练组的被试在运动识别任务和运动方向辨别任务中的成绩都有显著的提高，即信号点运动方向辨别任务的知觉学习并不特异于该任务。由于精巧的实验设计，信号点的具体运动方向对于运动识别任务来说是一个无关信息，而信号点的识别则对于两个任务来说都是相关的信息。因此，结果提示在随机运动点知觉学习过程中可能只有与任务相关的信息才能被学会，这为研究知觉学习的本质即知觉学习到底学习的是什么，提供了有力的证据。实验二则比较详细地研究了径向运动点运动速度辨别任务的知觉学习特异性。被试需要完成运动点运动速度辨别任务，根据不同研究目的，实验分成两个独立的部分：第一部分主要研究运动点速度辨别知觉学习的运动方向特异性；第二部分主要研究运动点速度辨别知觉学习在不同速度差之间的特异性。第一部分中被试被随机分配到两个不同的训练组中：一组训练所呈现的刺激是径向向内运动的点即向内运动训练(5人)；另一组训练所呈现的刺激是径向向外运动的点即向外运动训练(5人)。还有5人作为对照组，没有给予任何训练。在训练前后分别测量向内运动方向和向外运动方向上的心理物理曲线。结果发现：与对照组相比不管是向内运动训练组还是向外运动训练组，训练后被试在各自的训练方向上(向内运动方向或向外运动方向)速度辨别任务成绩都有显著的提高；同时两组被试在没有训练的方向上成绩也都有显著的提高，说明这种径向运动点运动速度辨别的知觉学习可以传递到相反的方向。第二部分的实验被试(12人)在一个中等的观察距离(114cm)进行训练，在训练前后分别测试被试在三个不同观察距离(5.7cm，114cm，228cm)的心理物理曲线，比较训练前后的结果发现：速度辨别任务成绩的提高不仅仅发生在训练距离，而且还发生在另外两个非训练距离。由于与训练距离(114cm)相比，在57cm运动点运动速度差加倍而在228cm速度差减半，这说明径向运动点运动速度辨别的知觉学习在不同速度差之间发生传递。因此径向运动点运动速度辨别任务的知觉学习具有普遍传递的性质。综合以上结果：随机运动点的知觉学习过程中可能只有任务相关的信息才能被学会，学习具有任务特异性；而径向运动点速度辨别的知觉学习没有方向特异性和明显的速度差特异性。更多还原

【Abstract】 Perceptual learning is defined as a change of performance as a result of repeated perceptual task training. The change of performance tends to persist over weeks and months. Perceptual learning is often quite specific for the exact task and stimulus attribute. Specificity is always a central topic in the perceptual field. The specificity of learning was used to support the hypothesis that perceptual learning embodies neuronal modifications in the brain’s early cortical areas and specificity was supposed to reflect the location of perceptual learning. However, different studies suggested huge variability of specificity on the same stimulus attribute. Now the low-level mechanism is facing a big challenge. Ahissar & Hochstein suggested that specificity of perceptual learning depends on task difficulty. Liu & Weinshalll found that generalization is a common phenomenon in perceptual learning.The present research studied specificity of perceptual learning on dots motion from task and stimulus attributes aspects. The study included two experiments: 1. Task transfer of perceptual learning on random dots motion. 2. The specificity of perceptual learning on speed discrimination.Experiment 1 compared task specificity of motion perceptual learning between detection and discrimination. Observers either decided which of the two stimuli contained a coherent motion signal (detection task), or whether the coherent motion direction was clockwise or counter-clockwise relative to a reference direction (discrimination task). Psychometric functions were measured before and after training. Twenty-three observers were randomly assigned to detection training group and discrimination training group. We found that the detection training improved only their detection but not discrimination sensitivity, whereas the discrimination trainees improved both. Whereas the exact direction of the coherent motion signal was irrelevant to the detection task, detection of the coherent motion signal was necessary for the discrimination task. Therefore the importance of task-relevance was demonstrated in both detection and discrimination learning. This suggests that only task-relevant information is learned in motion perceptual learning. This is an evidence that ’What is learned?’ during perceptual learning.Experiment 2 focused on specificity of perceptual learning on discriminating motion speed of radial moving dots. Observers were asked to perform a speed discrimination task. Two parts of study were included because of the different research purpose. Part 1 studied motion specificity of speed discrimination learning. Part 2 studied the specificity of learning under three speed differences. In part 1, ten observers were randomly assigned to inward training group and outward training group. Inward training group were presented with inward radial motion stimuli and outward training group were presented with outward radial motion stimuli. Also another five observers who went through no training were as controls. We found both groups improved their performance at the training direction (inward or outward), compared with control group. Moreover this improvement of performance happened at non-training direction. This means that perceptual learning of speed discrimination on radial motion dots was not specific to direction. In Part 2, twelve observers were trained at the viewing distance of 114cm. Psychometric functions were measured at three viewing distance of 57cm, 114cm and 228cm. We found that the improvement of performance happened not only at 114cm, but also at 57cm and 228cm. The speed difference doubled at 57 cm and halved at 228cm. We suggested that perceptual learning transferred between three speed differences. In summary, generalization of learning on speed discrimination was a common character.In summary, we found two main results. First, we suggested that only task-relevant information was learned in motion perceptual learning. Perceptual learning of motion was task specific. Second, perceptual learning of speed discrimination were not direction specific and speed difference specific.更多还原