Intermittent movement control emerges from information-based planning
Takagi, A.; Verdel, D.; Burdet, E.
Show abstract
Dominant models of reactive motion control in humans, based on optimal feedback control, predict smooth trajectories that reflect averaged behaviors. However, the observation of individual movements suggests that mammalian motor control is inherently discrete, with movement corrections occurring at a rate that depends on task demands and sensory information quality. To address these limitations, we introduce the Information Predictive Control (IPC) framework that integrates model predictive control with information theory, which triggers movement corrections only when unexpected deviations occur and corrections are likely to succeed. By quantifying "surprise" relative to anticipated internal and external states, IPC produces successful movements while robustly responding to sensorimotor noise, task constraints, and target variability. Simulations demonstrate IPCs ability to reproduce human-like responses to novel force fields during reaching, continuous target tracking, and adaptive planning under noise, while dynamically adjusting the planning horizon in complex, unpredictable environments.
Matching journals
The top 2 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Context-dependent extinction learning emerging from raw sensory inputs: A reinforcement learning approach 95%
- Intrinsic bursts facilitate learning of Levy flight movements in recurrent neural network models 95%
- Overcoming the limitations of motion sensor models by considering dendritic computations 95%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.