Problem difficulty and waiting time shape the level of detail and temporal organization of visual strategies in human planning

A new study by researchers Mattia Eluchans and Giovanni Pezzulo, published on arXiv, provides a quantitative look at how humans visually plan and execute actions under different constraints. The team designed a multi-target problem-solving task on a grid and manipulated two key variables: problem difficulty, using a novel 'misleadingness' metric, and the waiting time allowed before participants could act. By recording both eye movements and cursor trajectories, they captured the intricate relationship between visual strategy and physical execution.

Their findings reveal distinct patterns. Increased problem difficulty significantly reduced success rates and required more corrections, but it also prompted a more thorough visual inspection of the plan before movement began. This pre-movement coverage of the to-be-executed path was higher for harder problems. The availability of planning time was equally critical. When participants were forced to act immediately, they executed with less consolidated plans, leading to more pauses and backtracking. However, this condition resulted in more precise real-time alignment of gaze and cursor, suggesting a compensatory mechanism of improved online control.

Conversely, when given time to wait and plan, participants used that period to build a more coherent strategy. For difficult problems, this led to a stronger temporal alignment between the initial visual scan and the subsequent cursor movement during execution. The research suggests that visual planning is not a monolithic process but a flexible one: difficulty pushes us to gather more visual information upfront, while time availability determines whether we rely on a solid pre-formed plan or adapt through superior online control during the action itself.