Reward-based improvements in motor control are driven by multiple error-reducing mechanisms
Reward-based improvements in motor control are driven by multiple error-reducing mechanisms
复制标题
基于奖励的电机控制改进是由多种误差减少机制驱动的
DOI：
10.1101/792598
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Codol O
中科院分区：
文献类型：
--
作者：
Codol O
奖励具有显著的激活运动行为的能力，使个体能够以更高的精度和速度选择和执行动作。然而，如果要在应用环境中利用奖励，例如康复，则需要彻底了解其潜在机制。在一系列实验中，我们首先证明了奖励同时改善了伸手动作的选择和执行成分。具体来说，奖励促进了在存在干扰物的情况下选择正确的行动，同时也通过提高速度和保持准确性来提高执行力。这些结果导致了选择和执行的速度-精度函数的变化。此外，惩罚对行动选择和执行也有类似的影响，尽管它在一个区块内的所有试验中都提高了执行性能，也就是说，它的影响与试验值无关。尽管奖赏驱动的运动执行增强已被提出是通过增强反馈控制发生的，但一种未经验证的可能性是，它也由增加的手臂刚度驱动，这是一个增强肢体稳定性的能量消耗过程。计算分析显示，奖励既增加了运动中间的反馈修正，又减少了目标附近的运动噪音。与我们的假设一致，我们提供了新的证据，证明这种噪音降低是由手臂刚度的奖励依赖性增加所驱动的。因此，奖励驱动多种减少错误的机制，使个人能够在不影响准确性的情况下活跃运动表现。虽然奖励能增强运动表现是众所周知的，但神经系统是如何产生这些改善的尚不清楚。尽管最近的研究表明，奖励会增强反馈控制，但一种未经验证的可能性是，它也会增加手臂僵硬度。我们证明了奖励同时改善了伸手动作的选择和执行部分。此外，我们表明惩罚对绩效也有类似的积极影响。重要的是，通过结合计算和生物力学方法，我们表明奖励既可以改善反馈纠正，也可以增加刚度。因此，奖励驱动多种减少错误的机制，使个人能够在不影响准确性的情况下激发绩效。这项工作表明，刚度控制在基于奖励的运动控制改进中起着至关重要的作用，但未被充分认识。
Reward has a remarkable ability to invigorate motor behavior, enabling individuals to select and execute actions with greater precision and speed. However, if reward is to be exploited in applied settings, such as rehabilitation, a thorough understanding of its underlying mechanisms is required. In a series of experiments, we first demonstrate that reward simultaneously improves the selection and execution components of a reaching movement. Specifically, reward promoted the selection of the correct action in the presence of distractors, while also improving execution through increased speed and maintenance of accuracy. These results led to a shift in the speed-accuracy functions for both selection and execution. In addition, punishment had a similar impact on action selection and execution, although it enhanced execution performance across all trials within a block, that is, its impact was noncontingent to trial value. Although the reward-driven enhancement of movement execution has been proposed to occur through enhanced feedback control, an untested possibility is that it is also driven by increased arm stiffness, an energy-consuming process that enhances limb stability. Computational analysis revealed that reward led to both an increase in feedback correction in the middle of the movement and a reduction in motor noise near the target. In line with our hypothesis, we provide novel evidence that this noise reduction is driven by a reward-dependent increase in arm stiffness. Therefore, reward drives multiple error-reduction mechanisms which enable individuals to invigorate motor performance without compromising accuracy.SIGNIFICANCE STATEMENTWhile reward is well-known for enhancing motor performance, how the nervous system generates these improvements is unclear. Despite recent work indicating that reward leads to enhanced feedback control, an untested possibility is that it also increases arm stiffness. We demonstrate that reward simultaneously improves the selection and execution components of a reaching movement. Furthermore, we show that punishment has a similar positive impact on performance. Importantly, by combining computational and biomechanical approaches, we show that reward leads to both improved feedback correction and an increase in stiffness. Therefore, reward drives multiple error-reduction mechanisms which enable individuals to invigorate performance without compromising accuracy. This work suggests that stiffness control plays a vital, and underappreciated, role in the reward-based imporvemenets in motor control.