Mechanisms of Visuomotor Interception

Background/Objectives: Visuomotor interception requires aligning action with the future state of moving targets under sensory and motor delays. This constraint provides a tractable framework to examine how predictive and feedback-driven processes interact. This narrative review evaluates theoretical and empirical accounts of interception, with emphasis on how prediction and online control are integrated across behavioral and neural levels. Methods: We conducted a narrative synthesis of behavioral, eye-tracking, computational, and neurophysiological studies on visuomotor interception. Literature was identified through searches of PubMed, Web of Science, and Google Scholar using search terms including “visuomotor interception,” “predictive motor control,” “eye–hand coordination,” “time-to-contact,” “sensorimotor delay,” and related combinations. Studies published between 1986 and 2026 were considered, with emphasis on peer-reviewed empirical and theoretical work. Preprints were included only when directly relevant and are identified as such. The review compares internal model, ecological, and hybrid frameworks, and organizes evidence around spatial (“where”) and temporal (“when”) components of control. Results: Across paradigms, interception behavior is not well accounted for by purely predictive or reactive mechanisms. Instead, trajectories reflect a continuous interaction between anticipatory guidance and online correction. Spatial and temporal components show partial dissociation across tasks and manipulations. Available evidence supports the involvement of distributed circuits, including parietal, frontal, cerebellar, and subcortical systems, while indicating that eye movements play an active role in both information sampling and motor planning. Conclusions: Interception is best understood as the product of interacting biological, environmental, and learned constraints. Similar behavioral signatures can arise from distinct mechanisms, arguing against a unitary account. Progress requires integrating behavioral analyses with model-based and neural approaches to dissociate underlying computations.