Nature Human Behaviour, 2026 active vision measured association

The first look has a map

In the same scene, some people send their eyes to faces, others to text. A new study links that habit to how distinctly the visual brain encodes the same categories.

The study shows associations between gaze, visual representations and performance. The causal direction between experience, brain organization and looking habits remains open.

102 adults in eye-tracking

700 complex scenes

61 participants with

4.7x range in first fixations to text

r=.38/.44 face and text correlations

Tap the first thing you noticed

when you're ready

The scene appears after this button. Choose the first thing that pulls your gaze, without hunting for a correct answer.

Choose directly:

readout

Start the scene.

Press the button in the image, then choose the first thing that pulls your gaze. In the lab, a tendency is estimated from many scenes and fixations.

r .38 faces: first fixation and

r .44 text: first fixation and

from scene to scanpath

The same image can draw different gazes.

In the experiment, participants saw hundreds of scenes for 3 seconds each. The tendency to fixate faces or text repeated strongly across different halves of the image set.

Faces were a stable target.

For first fixations to faces, consistency across image splits was very high: r = .93 in the eye-tracking sample.

Text separated participants even more.

The tendency to first fixate text had a 4.7x maximum-minimum range. That makes text a useful window onto individual differences.

Competition matters.

When faces and text appear together, they compete for the first fixation. Direct competition makes the anticorrelation stronger.

face first

text first

from scanpath to

Gaze matches the individual visual map.

In a separate session, participants viewed centered stimuli in the scanner. Faces and words produced measurable patterns in .

Distinctiveness is calculated by comparing how similar responses are within one category against responses across different categories. In the study, faces and words were tested separately from free scene viewing.

what the evidence supports

The strong result is the measured link.

The evidence chain connects spontaneous gaze to individual visual representations. The large developmental question remains: the brain may guide gaze, gaze may train the brain, or both processes may become matched over time.

supported

Stable gaze tendencies

Faces and text drew first fixations differently across participants, and those differences repeated across separate image sets.

for first fixations to faces; for first fixations to text.

supported

Match with representations

Participants who more often fixated a category had more distinctive representations for that category in .

for faces in ; for text in .

partial

Links to performance

Evidence is clearer for face recognition. For reading, the extended-sample results are weaker.

CFMT: with face distinctiveness; pseudowords: in a subsample.

open

Causal direction

The authors treat the main analysis as correlational. The next step would test how gaze and representations change during development or training.

Developmental hypothesis.

how the map might form

Visual experience leaves measurable traces.

The new study fits a broader literature on development, reading, faces and visual expertise. The shared line is simple: categories we see often and use heavily can receive special treatment in visual cortex.

Faces

Early face exposure matters for the formation of face domains in visual cortex.

Reading

Learning to read reorganizes visual and language circuits, including the visual word form region.

Expertise

Objects heavily used in childhood, including Pokémon in the Gomez study, can show distinct cortical responses in adults.

Gaze

The new study adds spontaneous looking: what we choose in the first fraction of a scene matches what the brain encodes more distinctly.

sources

A verifiable map for the first look.

The interaction above uses synthetic scenes. The cited data come from published studies, with different samples, methods and limits.

primary study

Active vision is linked to category selectivity in the individual brain

Supports the link between face/text fixation tendencies and the distinctiveness of representations.

Nature Human Behaviour

data and code

OSF: Kollenda et al.

The authors state that anonymized data and code used for the figures and results are available.

OSF

individual gaze

Individual differences in visual salience vary along semantic dimensions

The basis for the idea that individual salience can vary along semantic categories such as faces and text.

PNAS

development

Longitudinal development of category representations

Links the development of word and face representations to reading and recognition performance.

Nature Communications

faces

Seeing faces is necessary for face-domain formation

Shows, in non-human primates, the role of face exposure in forming cortical face domains.

Nature Neuroscience

visual expertise

Extensive childhood experience with Pokémon

A memorable example of childhood visual experience appearing in adult visual cortex organization.

Nature Human Behaviour