Pushing the boundaries of a physical approach for the study of sensorimotor control
Pushing the boundaries of a physical approach for the study of sensorimotor control
复制标题
突破物理方法研究感觉运动控制的界限
DOI：
10.1016/j.plrev.2021.02.002
复制
发表时间：
2021
影响因子：
11.7
通讯作者：
Santello, Marco
中科院分区：
文献类型：
--
作者：
Santello, Marco
Mark Latash 的评论文章 [1] 对感觉运动控制研究的局限性和机遇进行了深入的讨论。作者的观点和他广泛的研究工作是以基于“自然法则”的自下而上的方法为基础的。这个框架引出了一个极其重要的概念：与无生命的物体不同，生物系统的运动变化是通过改变参数而不是变量来产生的。这个命题是 lambda (λ) 模型及其更新的表述——参考构型 (RC) 假设的基础。因此，理论上，生物系统与环境之间的相互作用可以通过控制一个时变参数 λ（代表牵张反射阈值）来调节。在支持 RC 假设的各种实验结果的背景下回顾了这一理论框架。尽管这篇评论文章提供了多种有趣的反思和讨论途径，但在这里我将重点关注实现基于 RC 的感觉运动控制所需的“构建模块”以及识别潜在神经机制的未来潜在研究途径。 RC 假设的一个吸引人的特点是其明显的简单性：神经系统不需要执行任何类型的“计算”，正如其他运动控制理论似乎暗示的那样，控制特定变量或最小化成本函数，如[1]。然而，其中一些理论也强调了参考系变换的重要性[例如，[2, 3]；在[4, 5]]中进行了回顾，即修改在感觉、受体特异性参考系（例如，视觉输入的视网膜专题参考系）中接收的感觉输入的过程，以便能够在基于效应器的参考系（例如，肌肉或关节）中传递运动命令。这个概念也出现在 RC 框架中（图 2）。作者还利用丰度原理解决了RC空间数量不可避免的爆炸，为系统提供了多种解决方案以确保性能的稳定性。这些考虑因素与作者审查的实验证据一起构成了支持 RC 假设的令人信服的案例。然而，这里有人想知道是否可以通过实验探测神经系统的较高级别（例如，初级和感觉皮层、前运动皮层、后顶叶皮层），以有效地识别通过开发或修改来构建给定 RC 的过程。
Mark Latash’s review article [1] provides a thorough discussion about limitations and opportunities for sensorimotor control research. The author’s views, and his extensive research work, are anchored in a bottom-up approach based on “laws of nature”. This framework leads to a critically important concept: Unlike inanimate objects, changes in motion of biological systems are generated by changing parameters, rather than variables. This proposition is the very foundation of the lambda (λ) model and its more recent formulation, the reference configuration (RC) hypothesis. Consequently, interactions between biological systems and the environment could theoretically be regulated by controlling one time-varying parameter, λ, representing the stretch reflex threshold. This theoretical framework is reviewed in the context of a wide variety of experimental findings supporting the RC-hypothesis. Although the review article provides multiple intriguing avenues for reflection and discussion, here I will focus on the ‘building blocks’ that would be required to implement RC-based sensorimotor control and potential future research avenues to identify the underlying neural mechanisms.One of the appealing features of the RC-hypothesis is its apparent simplicity: The nervous system does not need to perform any type of ‘computation’, as other motor control theories seem to imply, to control specific variables or minimize cost functions, as pointed out in [1]. However, some of these theories have also highlighted the importance of frame of reference transformations [eg,[2, 3]; reviewed in [4, 5]], ie, processes that would modify sensory inputs–received in the sensory, receptor-specific frame of reference (eg, retinotopic frame of reference for visual inputs)–to enable delivery of motor commands in an effector-based frame of reference (eg, muscles or joints). This concept is also present in the RC-framework (Fig. 2). The author also addresses the inevitable explosion in the number of RC spaces using the principle of abundance, which affords the system with multiple solutions to ensure stability of performance. These considerations, together with the experimental evidence reviewed by the author, make for a compelling case in support of the RC-hypothesis. However, here one is left wondering if the higher-levels of the nervous system (eg, primary and sensory cortices; premotor cortex; posterior parietal cortex) can be experimentally probed to effectively identify the processes through which a given RC might be built through development or modified