{"slug": "reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves", "title": "Reading Minds Almost: Scientists Just Reconstructed Vision from Brain Waves", "summary": "Researchers at Princeton University, led by Professor Ken Norman, have developed a system that reconstructs images from brain activity measured by fMRI in real time, producing reconstructions about 15 seconds after viewing. The study, published as 'Real-time Reconstruction of Human Visual Perception from fMRI,' uses generative AI to translate brain signals into visual approximations, outperforming chance on image similarity metrics. This marks a significant speed improvement over prior methods that took hours or days.", "body_md": "By Frederick d’Oleire Uquillas, Science Communications Fellow for the AI Lab\n\nImagine lying in an MRI scanner while looking at a picture of a dog. A computer analyzes your brain activity and, seconds later, generates a rough image of that dog.\n\nNot by reading the picture on the screen, but by reading your brain.\n\nThat’s the premise behind a new study, “Real-time Reconstruction of Human Visual Perception from fMRI,” in which researchers from Princeton University, supervised by Professor Ken Norman, demonstrate that it’s possible to reconstruct images someone is looking at using only brain activity measured in an MRI scanner – and to do it within seconds.\n\nIt’s not perfect mind reading, but it’s close enough to make you do a double take.\n\n**The big idea: Translating brain activity into pictures**\n\nWhen you look at an image, say a lighthouse, a person skiing, or a plate of food, your visual cortex produces a unique pattern of activity. Functional MRI (fMRI) measures this activity indirectly through changes in blood oxygen levels across the brain.\n\nFor decades, neuroscientists have tried to decode these patterns. The question has always been simple to ask and very hard to answer: *If we know the brain activity pattern, can we figure out what someone is seeing?*\n\nRecent advances in artificial intelligence have made that question much more tractable. Instead of trying to map brain activity directly to pixels, modern approaches translate brain signals into a high-level representation of the image. Essentially a semantic description of its visual features. Once you have that representation, generative AI can turn it back into an image.\n\nThis new study shows that such decoding can now happen in real time.\n\n**The experiment**\n\nThe experiment itself is surprisingly straightforward. A participant laid inside a 3-Tesla MRI scanner while looking at hundreds of natural images – things like animals, people, and outdoor scenes. Each image appeared for a few seconds while the scanner measured activity across visual regions of the brain.\n\nThe researchers first collected about an hour of training data from the participant. During that phase, the AI system learned how that individual’s brain responded to different images. Once trained, the model could be used in a second scanning session to reconstruct what the participant was seeing.\n\n**What the reconstructions look like**\n\nThe reconstructions aren’t photographic copies of the original images. Instead, they look more like AI-generated approximations.\n\nIf the participant saw a skier wearing a red jacket on a snowy mountain, the reconstruction might show a person-shaped figure in red surrounded by snowy terrain. If the participant saw animals grazing in a field, the reconstruction might produce a scene containing animal-like shapes in a grassy landscape.\n\nThey’re fuzzy and imperfect, but unmistakably related to the original image. The system captures the broad structure and meaning of the scene even when the details are not exact.\n\nQuantitatively, the reconstructions perform significantly better than chance according to multiple image similarity metrics used in the study.\n\n**The real breakthrough: speed**\n\nEarlier methods for reconstructing images from fMRI data could take hours, or even days.\n\nThis system works much, much faster.\n\nIn the fastest configuration, the researchers were able to produce image reconstructions about 15 seconds after the image was viewed.\n\nThat might not sound instantaneous, but for fMRI it is remarkably fast. The delay mostly comes from biology rather than computing power: The blood-oxygen signal measured by fMRI naturally peaks several seconds after a stimulus appears and can take up to 20 seconds to fully dissipate.\n\nEven with that constraint, the system processes the brain data and runs the AI models in just a few seconds once the signal arrives.\n\n**Why this is technically difficult**\n\nTwo fundamental challenges make real-time brain decoding hard.\n\nThe first challenge is that fMRI signals are noisy and indirect. Instead of measuring neurons directly, the scanner measures blood flow changes associated with neural activity. That signal unfolds *slowly* and can overlap across multiple stimuli.\n\nThe second challenge is computational. The AI models used to reconstruct images are large and complex, with hundreds of millions of parameters. Running them quickly enough to keep up with incoming brain data requires careful engineering and substantial computing resources.\n\nThe researchers addressed this by combining a cloud-based real-time fMRI framework with a streamlined version of the MindEye2 architecture, allowing the system to analyze brain activity and generate reconstructions within seconds.\n\n**Why we should care**\n\nBeyond the “mind reading” headlines, the real significance lies in what this technology enables.\n\nReal-time decoding opens the door to closed-loop neuroscience experiments, where brain activity can be analyzed and fed back to the participant while the experiment is still happening. For example, researchers could monitor how the brain represents a visual stimulus and then adjust the stimulus based on that representation.\n\nThere are also potential clinical applications. Real-time decoding might eventually help researchers understand how people with depression, anxiety, or other conditions interpret emotional images or stimuli. (A [2021 study](https://doi.org/10.1016/j.bpsc.2020.10.006) from a team that included Princeton researchers found that people with depression get better at ignoring faces with negative expressions when they’re given real-time feedback on their brain activity.) In principle, the reconstructed image could reveal subtle biases in how someone perceives the world.\n\n### **A quick reality check**\n\nBefore we start worrying about someone reading our thoughts on the subway, it’s worth emphasizing the limitations.\n\nThe system requires a participant to lie inside a large and very expensive MRI scanner. It also requires personalized training data for each individual before decoding can work reliably. And most importantly, it reconstructs visual perception, not arbitrary thoughts or memories.\n\nSo this isn’t a universal mind reader. It’s more like a sophisticated translator between brain activity and visual representations.\n\nStill, it’s an impressive step.\n\n**The bigger picture**\n\nFor decades, neuroscientists have dreamed of decoding the contents of perception from brain activity. What makes this moment different is the convergence of two technological revolutions: Large-scale neuroimaging datasets, and powerful generative AI models.\n\nTogether, they allow scientists to map brain activity into rich visual feature spaces that AI already understands.\n\nOnce that mapping exists, generating an image becomes surprisingly straightforward.\n\n**The takeaway**\n\nThe images reconstructed in this study aren’t perfect. But the fact that they exist at all, and can appear within seconds of seeing a stimulus, is remarkable.\n\nIt suggests that the gap between brain signals and subjective experience may be smaller than we once thought.\n\nWe’re not reading minds yet.\n\nBut we’re starting to sketch them.\n\nThe research was funded by Princeton’s Office of the Dean for Research Innovation Fund for New Industrial Collaborations, in partnership with [Sophont](https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fsophontai.com%2F&data=05%7C02%7Ckristine.lloyd%40princeton.edu%7Cd3ee1bb2349c4999fdaf08dec302e725%7C2ff601167431425db5af077d7791bda4%7C0%7C0%7C639162614438552345%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=JL1gX5l3LIgrM7EjS6W%2Fcztq14IJ%2F9pdJlnrfZ51ii8%3D&reserved=0). Early stages of the research were supported by the National Institute of Mental Health.\n\n**Curious to learn more?**\n\nCheck out the [full paper here](https://arxiv.org/abs/2607.22753).", "url": "https://wpnews.pro/news/reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves", "canonical_source": "https://blog.ai.princeton.edu/2026/07/31/reading-minds-almost-scientists-just-reconstructed-what-someone-sees-from-their-brain/", "published_at": "2026-08-01 17:05:35+00:00", "updated_at": "2026-08-01 17:22:41.345102+00:00", "lang": "en", "topics": ["artificial-intelligence", "generative-ai", "computer-vision", "neural-networks"], "entities": ["Princeton University", "Ken Norman", "fMRI"], "alternates": {"html": "https://wpnews.pro/news/reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves", "markdown": "https://wpnews.pro/news/reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves.md", "text": "https://wpnews.pro/news/reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves.txt", "jsonld": "https://wpnews.pro/news/reading-minds-almost-scientists-just-reconstructed-vision-from-brain-waves.jsonld"}}