Does moving your eyes make a face preview more useful?
Seeing an upright face in the periphery sped up judgements about that face only when people actually moved their eyes to it, not when the same images simply appeared in the centre of their gaze.
Source
Saccade execution increases the preview effect with faces: An EEG and eye-tracking coregistration study
Study at a glance
- Design
- Human experiment — 2 x 2 within-subject design: preview valid (upright face) vs invalid (inverted face) crossed with saccade vs no-saccade blocks, where the no-saccade block replayed each person's own saccade timing while gaze stayed central.
- N
- N=26 · 26 adults with concurrent EEG and eye-tracking; 320 trials per viewing condition.
- Population
- Healthy adult volunteers with normal or corrected vision tested at the University of Trento.
- Outcome
- Response time and errors judging the tilt of the foveal target face, fixation-related N170 amplitude (165-250 ms), saccade latency and first fixation duration.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
Adults looked at a central cross while a face appeared to the left, either upright (valid preview) or upside-down (invalid preview). In saccade blocks they moved their eyes to it and the face was swapped mid-movement for an upright, slightly tilted target; in no-saccade blocks they kept still and the target appeared at the centre after a delay matched to their own eye-movement timing. They reported the tilt while EEG and eye position were recorded.
What they found
Tilt judgements were faster after a valid preview in the saccade condition (579 vs 590 ms), but identical after valid and invalid previews without a saccade. For the face-sensitive N170 window, the preview effect was larger with saccades than without (a significant interaction), though within the saccade condition alone it did not reach significance and without saccades it went slightly in the opposite direction. The N170 was much larger overall when there was no saccade, and first fixations on the target were shorter after an inverted preview.
The limits
What it doesn't show
The EEG preview effect was weaker than in the authors' earlier studies and not significant within the saccade condition, so the neural evidence that saccades are necessary is ambiguous. The design used only upright targets, faces only in the left visual field and a task-irrelevant preview, which may have shrunk the effect and limits generality. Residual eye-movement artefacts remained in the EEG after ICA, and the large N170 difference between conditions may partly reflect eyeball rotation rather than face processing. With 26 participants, small effects are hard to pin down.
Key terms
- Saccade
- A fast jump of the eyes from one point to another, which brings a new object into central (foveal) vision.
- Preview effect
- The change in processing of an object after it was already glimpsed in peripheral vision before being looked at directly; valid previews usually help.
- N170
- An EEG brain response peaking about 170 ms after a face appears, over the back of the head, commonly treated as a marker of face detection.
- Active vision
- The view that perception relies on the eye-movement system to predict what will land on the fovea, rather than passively receiving images.
- Fixation-related potential
- EEG activity time-locked to the moment the eyes land on something, the free-viewing version of a stimulus-locked ERP.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What was manipulated to make a preview valid or invalid?
Common questions
Why did the no-saccade condition use randomly drawn delays?
To match the timing of the saccade condition, the authors drew a latency plus duration for each trial from each participant's own practice-block eye movements, so the only difference was whether the eyes moved.
Does this prove eye movements are necessary for any preview effect?
For the behavioural effect it points that way, since it vanished without saccades, but a small (reversed) EEG preview effect remained without saccades, so some peripheral-to-central integration may still happen.
More on Face processing