Your brain predicts the shape of unseen things by filling in, from context and experience, what your eyes cannot directly see. When an object is partly hidden, higher visual areas represent its completed shape, and even parts of early visual cortex that receive no direct input carry information about the surrounding scene. Expectations also make visual processing more efficient. The process is constant, automatic, and mostly invisible to you.
Seeing Whole Objects From Partial Views
Look at a mug half hidden behind a book, and you still perceive a whole mug. Scientists call this amodal completion. In an fMRI study of nine people, researchers showed images that evoke completion and then tested which brain regions responded to the visible shape and which to the completed one.
Early visual cortex, presumably V1, tracked the local contours of what was actually visible. Regions in the inferior temporal cortex responded to the completed shape instead, even though the two shapes shared three-quarters of their area, and they did not preserve the original contours. The authors favor a multi-step account in which visible features are represented first and replaced by the completed object at higher levels.
Early Visual Cortex Knows What Surrounds the Gap
The brain also uses context to fill gaps in the earliest stages of vision. In a study of nonstimulated visual cortex, researchers examined a quarter of the visual field that received no direct input while varying the surrounding image. Activity patterns in V1 and V2 could still discriminate the surrounding context, and the effect was driven mainly by V1.
Those patterns were related to the ones evoked by direct stimulation. The authors suggest the information arrives through lateral and feedback connections, not from the thalamus, which is the relay station that carries direct visual input. In plain terms, a region with no view of a scene still holds a hint of what belongs there.
Why Expectation Makes Seeing Easier
Prediction does more than fill gaps. In a Neuron study from the Donders Institute, participants viewed images that were expected or unexpected while a scanner recorded their brain activity. Visual regions were less active for expected images, yet a decoding algorithm identified those images more accurately. The authors concluded that expectations make visual processing less intensive but more efficient.
This fits predictive coding, the idea that perception is built by comparing sensory input with expectations. A profile of cognitive neuroscientist Peter Kok describes how this view treats the brain as an active forecaster rather than a passive camera.
When Predictions Go Wrong
The same machinery that completes a partly hidden mug can complete it incorrectly. The Kok profile notes that expectations can bias interpretation and produce errors, illusions, or even hallucinations, and that when images are ambiguous, people reported seeing dots move in the direction they expected. It’s reasonable to see this as the cost of a fast system: speed comes from guessing, and guesses are sometimes wrong.
Test Your Own Predictions
You can notice this process with a few simple experiments:
- Cover part of a photo and sketch what you expect underneath before revealing it.
- Glance at a partly hidden object and describe its hidden side, then check.
- Pay attention to moments of surprise, since they mark where a prediction failed.
Treat these as observation exercises, not tests. The research above did not examine whether practicing them improves perception.
Prediction, Focus, and Where Nootropics Fit
Noticing your own predictions takes attention and a rested brain. Curious readers sometimes ask about nootropics, supplements marketed for focus and memory. Evidence varies widely by ingredient, and no capsule replaces curiosity or sleep.
Frequently Asked Questions
How Does the Brain Fill In Hidden Parts of Objects?
Early visual areas track the visible contours, while higher areas in the inferior temporal cortex represent the completed shape. Context from the surrounding scene also reaches early visual cortex through feedback and lateral connections.
Does What I Expect Change What I See?
Yes. In one study, visual regions were less active for expected images yet represented them more accurately, and expectations can bias how ambiguous images are interpreted.
Can Predictions Make Me See Things That Are Not There?
They can bias interpretation, especially for ambiguous images. In one study, people reported seeing dots move in the direction they expected.

