This review discusses the computational neuroanatomy of motor control, emphasizing the interplay between lesion studies and theoretical models. It explores how impairments in motor control, motor learning, and higher-order motor control in patients with lesions of the corticospinal tract, cerebellum, parietal cortex, basal ganglia, and medial temporal lobe can be explained through computational concepts such as state estimation, optimization, prediction, cost, and reward. The cerebellum is proposed to function in system identification, building internal models to predict sensory outcomes of motor commands and correct them through internal feedback. The parietal cortex integrates predicted proprioceptive and visual outcomes with sensory feedback to form a belief about body and environmental states. The basal ganglia are related to optimal control, learning costs and rewards associated with sensory states, and estimating "cost-to-go" during motor tasks. The primary and premotor cortices implement optimal control policies by transforming beliefs about proprioceptive and visual states into motor commands. The review outlines a computational framework for motor control, including the problem of inferring function from lesions, the steps involved in making a movement, and the effects of focal and distributed lesions on these steps. It discusses the computational problem in reaching, including the costs and rewards of the task, the best way to maximize rewards and minimize costs, and the types of computations required to achieve this goal. The review also presents examples of how these computational principles apply to motor control, including the role of the cerebellum in predicting sensory consequences of motor commands and the construction of internal models. It highlights the importance of understanding the relationship between learning better sensory predictions and learning better motor commands, as well as the role of amnesia in learning tool use. The review concludes with a discussion of the relationship between the cerebellum and adaptation, emphasizing the importance of forward models in predicting sensory consequences of motor commands.This review discusses the computational neuroanatomy of motor control, emphasizing the interplay between lesion studies and theoretical models. It explores how impairments in motor control, motor learning, and higher-order motor control in patients with lesions of the corticospinal tract, cerebellum, parietal cortex, basal ganglia, and medial temporal lobe can be explained through computational concepts such as state estimation, optimization, prediction, cost, and reward. The cerebellum is proposed to function in system identification, building internal models to predict sensory outcomes of motor commands and correct them through internal feedback. The parietal cortex integrates predicted proprioceptive and visual outcomes with sensory feedback to form a belief about body and environmental states. The basal ganglia are related to optimal control, learning costs and rewards associated with sensory states, and estimating "cost-to-go" during motor tasks. The primary and premotor cortices implement optimal control policies by transforming beliefs about proprioceptive and visual states into motor commands. The review outlines a computational framework for motor control, including the problem of inferring function from lesions, the steps involved in making a movement, and the effects of focal and distributed lesions on these steps. It discusses the computational problem in reaching, including the costs and rewards of the task, the best way to maximize rewards and minimize costs, and the types of computations required to achieve this goal. The review also presents examples of how these computational principles apply to motor control, including the role of the cerebellum in predicting sensory consequences of motor commands and the construction of internal models. It highlights the importance of understanding the relationship between learning better sensory predictions and learning better motor commands, as well as the role of amnesia in learning tool use. The review concludes with a discussion of the relationship between the cerebellum and adaptation, emphasizing the importance of forward models in predicting sensory consequences of motor commands.