How the brain rapidly learns to interpret ambiguous images
· Medical Xpressby Ingrid Fadelli, Medical Xpress
Ingrid Fadelli
Author
Meet our staff & contributors
Learn about our editorial standards
edited by Sadie Harley, reviewed by Robert Egan
Sadie Harley
Scientific Editor
Meet our editorial team
Behind our editorial process
Robert Egan
Senior Editor
Meet our editorial team
Behind our editorial process Editors' notes
This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:
fact-checked
peer-reviewed publication
trusted source
proofread
The GIST Add to Preferred Sources
Both humans and image recognition algorithms can take time to identify the objects and scenes depicted in some images. Emblematic examples of this phenomenon are Mooney images, pictures that are intentionally reduced to patches of black and white, with no shades of gray.
A lack of gray shades can make recognizing meaningful objects or scenes in Mooney images more challenging. When humans view them, they often struggle to interpret them at first, yet they can find that the images suddenly make sense after they see the grayscale originals.
Researchers at Harvard Medical School, MIT's Center for Brains, Minds and Machines and other institutions recently tried to better understand the neural underpinnings of this rapid shift in perception. The results of their experiments, published in Nature Neuroscience, showed that after participants learned to recognize Mooney images, neurons in two brain regions responded to these images more like they responded to the grayscale originals.
"We were inspired by a striking feature of human intelligence, our ability to learn so quickly that one brief glimpse can transform how we see things," Marcelo Armendariz, a co-first author of the paper, told Medical Xpress.
"Sometimes we look at something that seems to make no sense, and then suddenly a single piece of information from our environment completely changes its meaning. For that to happen, the patterns of activity of the neurons in our brains must change to reshape our perception."
Armendariz and his colleagues wanted to better understand how neurons in the human brain rapidly encode new information. This could also potentially help to improve artificial intelligence (AI) algorithms, many of which are trained to recognize objects in images by analyzing a large pool of examples.
"The recognition of 'Mooney' images (impoverished black-and-white images that are hard to interpret initially) is a paradigmatic example of rapid learning," said Gabriel Kreiman, senior author of the paper.
"The effect is basically all or none. People cannot recognize these images initially and they learn to interpret them in one shot after exposure to the grayscale counterpart. A prominent example is the famous textbook 'Dalmatian dog illusion.' I was struck by these examples when I saw them for the first time in college."
Turning black-and-white patches into recognizable images
To investigate the mechanisms responsible for the speed with which the human brain learns to interpret ambiguous images, the researchers performed experiments involving 13 individuals with drug-resistant epilepsy. These people had electrodes implanted in their brains to monitor their seizures, and the researchers used those electrodes to record the activity of individual neurons.
"While the participants performed a rapid learning and image recognition task, we recorded the activity of more than 1,000 neurons in their brains," explained Armendariz. "We then used machine learning methods to track what information was represented in their patterns of neuronal activity, and where and when this occurred."
The task completed by the participants required them to look at Mooney images and try to interpret them. Initially, the participants reported recognizing the objects in only 4% of trials.
"We then showed participants the grayscale counterparts of those images," said Kreiman. "When we showed the black-and-white Mooney images again, participants immediately recognized the objects in the image. To study the mechanisms, we recorded the neuronal responses from visual and memory areas in the brain using the electrodes implanted in the patients' brains."
After one or a few learning sequences, during which participants saw a grayscale image followed by its Mooney counterpart, participants recognized the black-and-white images in 83% of trials, on average.
Moreover, once participants had seen the grayscale original, the patterns of neural activity elicited by the black-and-white Mooney image became more similar to those evoked by that original. These similar responses were observed in the occipital cortex, which processes visual information, and in parts of the medial temporal lobe, including the hippocampus, which plays a role in memory.
"We found that neurons in the visual cortex and in memory structures such as the hippocampus changed their patterns of activity after learning," said Armendariz. "The visual cortex needed extra time to process the newly acquired information. This delay was then observed in memory structures, suggesting that the neocortex must first resolve the learned images before passing the information on."
Possible implications for AI and clinical research
Based on their findings, the researchers hypothesized that after rapid learning, recognizing Mooney images may involve signals passing back and forth within the neocortex, possibly with feedback from higher-order cortical areas. Medial temporal lobe responses appeared later, suggesting that feedback from this region may not drive the initial recognition of Mooney images.
"What we found is a remarkable transformation in the neuronal representation of visual information accompanying rapid learning," said Kreiman.
"In other words, neurons in the human brain change how they react to images in a single shot, reflecting the fast-learning process. The mechanisms involved could inspire novel AI algorithms to learn fast. Clinically, the insights can help in the future design of treatments for learning and memory disorders."
The recent work by Armendariz, Kreiman and their colleagues could potentially contribute to the understanding of how humans learn to interpret ambiguous images. The insight they gathered could in turn inform both efforts to advance AI algorithms for image recognition and the study of disorders associated with disruptions in visual learning.
"In future work, we want to understand how higher-order regions of the neocortex might provide feedback to the visual cortex to support rapid learning and the encoding of new information," said Armendariz.
"We are also implementing learning and memory AI algorithms inspired by these findings to endow machines with the capability of rapid learning," added Kreiman.
Written for you by our author Ingrid Fadelli, edited by Sadie Harley, and fact-checked and reviewed by Robert Egan—this article is the result of careful human work. We rely on readers like you to keep independent science journalism alive. If this reporting matters to you, please consider a donation (especially monthly). You'll get an ad-free account as a thank-you.
Publication details
Marcelo Armendariz et al., Single-neuron encoding of rapidly learned visual information reshapes human perception, Nature Neuroscience (2026). DOI: 10.1038/s41593-026-02496-6.
Journal information: Nature Neuroscience
Key medical concepts
Clinical categories
NeurologyPsychology & Mental health Who's behind this story?
Ingrid Fadelli
Freelance journalist with BSc Psychology and MA International Journalism. Covers AI, robotics, neuroscience, and astrophysics since 2018. Full profile →
Sadie Harley
BSc Life Sciences & Ecology. Microbiology lab background with pharmaceutical news experience in oil, gas, and renewable industries. Full profile →
Robert Egan
Bachelor's in mathematical biology, Master's in creative writing. Well-traveled with unique perspectives on science and language. Full profile →
© 2026 Science X Network
Citation: How the brain rapidly learns to interpret ambiguous images (2026, November 13) retrieved 13 November 2026 from https://medicalxpress.com/news/2026-11-brain-rapidly-ambiguous-images.html This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.