01 — Attention and gaze

Attention and eye movements are tightly coupled.

Attention can move without the eyes. A saccade — the rapid movement between two gaze locations — is nevertheless usually preceded by an attentional shift toward its target. Eye movements therefore make part of attentional orientation observable.

Eye tracking records these movements. A fixation indicates where the fovea points; the fovea is the central retinal region that provides the sharpest vision. Interpretation then connects the signal to the task, stimulus and an explicit cognitive hypothesis.

02 — Saliency and goals

Gaze combines image properties with the observer’s objectives.

Exogenous attention is captured by physical scene properties such as contrast, color, orientation or shape. This stimulus-driven influence is called bottom-up processing. Saliency models attempt to represent the visual priority it creates.

Task and prior knowledge exert a goal-driven influence called top-down processing. Semantic context, faces, age, expertise and personality also guide the next saccade. Modern priority maps combine both families of factors. The same painting produces different scanpaths — ordered gaze trajectories — when viewed for memorization or to locate a detail.

03 — Two complementary modes

First grasp the whole, then inspect the detail.

Ambient mode is classically associated with fixations below 180 ms followed by saccades above 5 degrees. It quickly captures spatial organization and the global impression of a scene, sometimes called its gist. Focal mode shows the reverse pattern: longer fixations, smaller saccades and deeper local processing.

Exploration often begins with an ambient-to-focal transition, after which both modes alternate as needs change. The K coefficient combines fixation duration and saccade amplitude to place this dynamic on a continuum without relying only on fixed thresholds.

04 — What this changes

A heatmap summarizes space and reduces the temporal dimension.

Interface analysis combines the most fixated regions with their viewing time, order and task instruction. A heavily viewed area may be important, ambiguous or simply difficult to decode.

Ambient and focal modes have been linked to dorsal and ventral brain pathways. This proposal remains a hypothesis about their neural signatures and provides an interpretive framework to test with dedicated neurological measures.