A simple computational principle predicts vocal adaptation dynamics across age and error size.
A simple computational principle predicts vocal adaptation dynamics across age and error size.
复制标题
DOI：
10.3389/fnint.2014.00075
复制
发表时间：
2014
影响因子：
3.5
通讯作者：
Sober SJ
中科院分区：
文献类型：
--
作者：
Kelly CW;Sober SJ
大脑利用感官反馈来纠正行为错误。鸣禽和人类通过模仿成人发出的声音来获得发声行为，并在一生中依靠听觉反馈来纠正发声错误。在鸟类和人类中，在获得发声行为后，声音变异性随着年龄的增长而稳步下降。先前对成人的研究表明，虽然在声音变异性范围内的感觉错误会引起强烈的运动纠正，但更大的错误不会引起学习。尽管这些结果表明，年轻的动物具有更大的声音变异性，可能比年老的动物更容易纠正大的错误，但尚不清楚年龄依赖性的变异性变化是否伴随着声音错误纠正的速度或幅度的变化。我们测试了这样一个假设：听觉错误会在年轻的动物中引起更大的声音变化，一个共同的计算决定了感觉信息如何驱动不同年龄和错误大小的运动学习。与我们的假设一致，我们发现在鸣禽中，错误纠正的速度和程度随着年龄的增长而急剧变化，并且通过一个模型预测了年龄依赖性学习差异，其中感觉错误和先前感觉反馈分布之间的重叠决定了适应的动态。我们的研究结果表明，大脑采用了一种简单而强大的计算原理来校准声音适应的速度和幅度，这些速度和幅度跨越了行为表现的年龄依赖性变化和对不同感觉错误的反应。
The brain uses sensory feedback to correct errors in behavior. Songbirds and humans acquire vocal behaviors by imitating the sounds produced by adults and rely on auditory feedback to correct vocal errors throughout their lifetimes. In both birds and humans, acoustic variability decreases steadily with age following the acquisition of vocal behavior. Prior studies in adults have shown that while sensory errors that fall within the limits of vocal variability evoke robust motor corrections, larger errors do not induce learning. Although such results suggest that younger animals, which have greater vocal variability, might correct large errors more readily than older individuals, it is unknown whether age-dependent changes in variability are accompanied by changes in the speed or magnitude of vocal error correction. We tested the hypothesis that auditory errors evoke greater vocal changes in younger animals and that a common computation determines how sensory information drives motor learning across different ages and error sizes. Consistent with our hypothesis, we found that in songbirds the speed and extent of error correction changes dramatically with age and that age-dependent differences in learning were predicted by a model in which the overlap between sensory errors and the distribution of prior sensory feedback determines the dynamics of adaptation. Our results suggest that the brain employs a simple and robust computational principle to calibrate the rate and magnitude of vocal adaptation across age-dependent changes in behavioral performance and in response to different sensory errors.