The road to understanding how your brain sees a square is not a scenic highway. It is more like a labeled airport transfer: wrong turns through "basic vision," a layover in "math," a detour through artificial intelligence, and everyone realizes the brain packed a geometry textbook.

Geometry: The Original Brain Theme Park

You can spot a square instantly. A rectangle? Easy. A wonky quadrilateral assembled during turbulence? Also easy, but in a mildly offended way. That ease is the mystery. Your retina does not receive "square" as a labeled package. It gets light, edges, contrast, angles, and a spreadsheet with no column headers.

Classic vision science says your brain builds objects from simpler parts: lines, contours, boundaries, surfaces, then whole forms. Gestalt psychology added that we group bits into tidy wholes because the brain loves order the way a theme park loves a queue rail.

The new eLife study asks where that geometric tidiness lives in the human brain, and how fast it shows up.

Scientists Put Shapes on the Brain Scanner Ride

The team used fMRI in adults and 6-year-old children, plus MEG in adults. fMRI tells you where activity changes, with the timing precision of "sometime this afternoon." MEG tells you when activity changes, with the spatial precision of "probably around there." Together, they make a decent buddy-cop duo.

Participants viewed triangles, quadrilaterals, hexagons, and categories like faces, houses, tools, numbers, words, and characters. The researchers also tested how well different models explained the brain signals. Could a convolutional neural network capture what humans were doing? Or did the brain use something more symbolic, like "parallel sides," "right angle," and "symmetry"?

Early Vision Sees Lines. Later Brain Says "Ah, Geometry."

Early visual activity looked fairly CNN-friendly. That makes sense. A neural network trained on images is good at edges, textures, contours, and visual features. It is the brain's early security line: shoes off, laptop out, edges detected.

But later signals, especially in dorsal parietal and prefrontal regions, were not well explained by ordinary CNN features. Those signals fit better with discrete geometric features: parallelism, symmetry, equal sides, right angles. Bigger vision models did better than smaller CNNs, but the study still points to something human-sounding: the brain may switch from pixels and contours to compressed abstract structure.

That word "compressed" matters. A square is four equal sides, four right angles, two pairs of parallel sides, and enough symmetry to make a graphic designer quietly nod. Instead of storing every visual detail, your brain can encode the rule. Basically, it zips the file.

The Math Network Crashes the Vision Party

Geometric shapes did not simply crank up the usual object-vision machinery. Compared with other visual categories, shapes produced lower activity in parts of the ventral visual stream, but higher activity in intraparietal and inferior temporal regions linked to mathematical processing.

That is deliciously weird. The brain sees a hexagon and parts of the math network perk up like someone mentioned group discounts. This does not mean your brain is doing Euclidean proofs every time you glance at a stop sign. Please do not ask your occipital cortex to show its work. But abstract shape regularities seem to tap systems beyond ordinary object recognition.

The child data make the story better. Six-year-olds showed related patterns, hinting this sensitivity is not just schooling wearing a tiny lab coat. Previous work found that humans detect odd shapes more easily when regularity is involved, such as spotting a distorted square among squares. The current paper adds the brain map to that puzzle.

Why This Matters Beyond Winning at Toddler Tangrams

If these results hold up, they help explain a bridge between vision, math, culture, and AI. Humans have covered caves, temples, tools, carpets, logos, and phone icons with regular shapes for thousands of years. Maybe regular geometry fits a shortcut our brains are eager to use.

For AI, the paper is a friendly shove. Systems can recognize images astonishingly well, yet still struggle with abstract rules unless trained at huge scale or built with the right biases. Human vision may be a roller coaster with multiple tracks: appearance, structure, symbolic regularity, and one that screams during tax season.

Clinically, this is not a diagnostic tool or treatment. But understanding how the brain extracts form could eventually inform education, developmental research, visual rehabilitation, and models of disorders where object recognition or spatial reasoning goes sideways.

Takeaway: when you see a square, your brain may not merely see a shape. It may recognize a compact little rule system hiding in plain sight. Geometry is not just something you survived in school. It is something your brain runs in the background, like a nerdy operating system update you never agreed to install.

References

1. Sablé-Meyer M, Benjamin L, Potier Watkins C, He C, Pajot M, Morfoisse T, Al Roumi F, Dehaene S. "A geometric shape regularity effect in the human brain." eLife. DOI: 10.7554/eLife.106464. PMID: 41524531.
2. Dehaene S, Al Roumi F, Lakretz Y, Planton S, Sablé-Meyer M. "Symbols and mental programs: a hypothesis about human singularity." Trends in Cognitive Sciences. 2022. DOI: 10.1016/j.tics.2022.06.010.
3. Sablé-Meyer M, Fagot J, Caparos S, van Kerkoerle T, Amalric M, Dehaene S. "Sensitivity to geometric shape regularity in humans and baboons: A putative signature of human singularity." PNAS. 2021. DOI: 10.1073/pnas.2023123118. PMCID: PMC8072260.
4. Fitousi D, Algom D. "The quest for psychological symmetry through figural goodness, randomness, and complexity: A selective review." i-Perception. 2024. DOI: 10.1177/20416695241226545.
5. Bertamini M, Rampone G, Makin ADJ, Jessop A. "Polygons have a small facilitatory effect on extraretinal symmetry perception." NeuroImage. 2024. DOI: 10.1016/j.neuroimage.2024.120894.

Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.