A1 neural basis of frequency discrimination learning

描述（由申请人提供）：本申请概述了听觉学习机制的直接研究。动物有能力在积极强化因素、之前的感觉环境和导致它们的运动行为之间建立新的联系。该研究将使用多个微电极植入技术，在动物获得新关联的一段时间内同时记录大量的听觉皮层神经元。具体的目的是观察听觉皮层对感官环境的表征是如何被学习改变的。这些猴子将被训练来辨别短音调。当他们开始执行任务以获得奖励时，他们的大脑学习机制就被激活了，并且不可避免地导致了声音表现方式的变化。数据将被并行收集，包括动物的感知能力和行为任务内外听觉领域神经元的反应特性。在过去十年中，神经系统植入物的使用取得了很大进展；这项研究将把这项技术应用于研究学习。在训练的第一阶段，动物会对频率略高于固定对比音调的短音调做出反应。初步研究表明，如果目标刺激来自一个固定的范围，那么该频率范围的表征就会增长和增强。在这个初始阶段之后，动物将使用随机频率进行比较和目标音调点的相同频率识别任务。对照实验将在只是被动听者的幼稚动物身上重复第一阶段训练的试验。这些研究有助于将皮质场的塑性特性定义为动态学习神经网络。这种网络最容易理解的是它们的初始连接集和学习规则。这项研究将分离并测试不同的学习规则假设，并将导致对大脑皮层在工具性学习和经典条件反射中的适应性作用的更好理解。

DESCRIPTION (provided by applicant): This application outlines a direct study of the mechanisms of auditory learning. Animals are capable of making new associations between positive reinforcers, the sensory context that preceded them, and the motor acts that cause them. The proposed studies will use multiple microelectrode implant technology to record simultaneously from a large number of neurons in auditory cortex throughout a period in which an animal acquires a new association. The specific intent will be to observe how the auditory cortex representation of the sensory context is altered by the learning. The monkeys will be trained to discriminate short sound tones. As they begin to perform the task to receive rewards, their brain learning machinery is engaged and causes unavoidable changes in how sounds are represented. Data will be collected in parallel, in terms of the animal's perceptual capabilities and the response properties of neurons in auditory fields within and outside the behavioral tasks. The use of implants in neural systems has advanced much in the last decade; this study will apply that technology to study learning. In the first phase of training, animals will respond to short tones that are just higher in frequency than fixed comparison tones. Preliminary studies have shown that if the target stimulus comes from a fixed range, the representation of that frequency range grows and intensifies. After this initial phase, the animals will perform the same frequency discrimination task using randomized frequencies for comparison and target tone pips. A control experiment will replicate the trials of the first phase of training in naive animals that are only passive listeners. These studies should contribute to defining the plastic properties of cortical fields as dynamic learning neural networks. Such networks are most easily understood in terms of their initial sets of connections and their learning rules. This study will isolate and test different hypotheses for learning rules, and will lead to a greater understanding of the adaptive role of the cerebral cortex in general, both in instrumental learning and in classical conditioning.