Neural activity spreads like wildfire across the brain, touching far more regions than most scientists expected.
In a massive international project, researchers recorded from more than 600,000 neurons in mice and stitched together the most comprehensive map yet of how sensations, choices, actions, and rewards ripple through the brain. The work, published in Nature, is both a dataset and a challenge: what do all these overlapping signals really mean?
The project involved 139 mice performing a deceptively simple decision-making task: see a visual cue, spin a tiny wheel, earn water or get a burst of white noise. On the surface, it looks like mouse video game training. Underneath, it reveals the brain at full throttle, juggling stimulus, memory, expectation, and movement.
The scale was staggering. Across 12 laboratories, scientists inserted 699 Neuropixels probes, recording spikes from 279 distinct brain areas. That coverage included cortex, thalamus, midbrain, hindbrain, and even the cerebellum. For once, the entire brain was on the record, not just cherry-picked regions.
One of the most striking patterns was how little was siloed. Signals tied to movement and reward popped up almost everywhere. Sensory responses were more contained, but even vision bled into unexpected regions like the hindbrain. As one of the project leaders put it, “Neural responses correlated with impending motor action almost everywhere in the brain.” That is a sobering reminder for neuroscientists: good luck isolating a clean, local signal.
The real surprise came from choice encoding. Neurons reflecting which way a mouse would turn the wheel were not confined to cortical decision hubs. They emerged simultaneously in the thalamus, midbrain, medulla, and cerebellum, suggesting that even ancient brainstem structures join the deliberation. One researcher noted that “parts of the medulla, the pons and the cerebellum are all selectively responsive with similar timing to cortical areas.” In other words, the machinery of choice is deeply distributed.
If that makes brain function sound messy, it is. The dataset shows that many neurons are modulated by the task, but most of their variance remains unexplained. Much of the brain appears to be doing its own thing, tuned to movements, arousal, or hidden states outside the neat experimental script. Neuroscience has long lived with the comforting idea of localized function. This map suggests that model was only a partial truth.
The economic angle is worth noting. Building and maintaining such a resource requires not just scientific cooperation but also infrastructure that rivals big tech. Data are publicly available through an online browser and GitHub, meaning any lab with coding chops can dive in. This democratization raises stakes: the dataset will likely fuel years of reanalysis, from AI-driven decoding to disease modeling. Open access is both gift and gamble, since interpretation remains contentious.
For now, the map is best seen as a starting point. It will not tell us, on its own, why decisions falter in Parkinson’s, why reward circuits misfire in addiction, or how internal states like expectation are stitched into perception. But it provides the terrain where such questions can finally be asked without blind spots.
In the end, the mouse wheel task underscores a humbling truth: the brain rarely whispers in one corner. It shouts across the whole network, and our job is to make sense of the noise.
Journal: Nature. DOI: 10.1038/s41586-025-08176-3
Explainer
Imagine trying to understand how a city works by listening to only one block. You might hear cars, shopkeepers, and sirens, but you would miss the buses, parks, and factories elsewhere. Neuroscience has long faced that problem, recording from isolated brain areas and assuming they held the keys to perception or decision-making. The International Brain Laboratory solved it by recording from almost the whole mouse brain at once, while the animals performed a decision task. The dataset shows that signals about stimuli, choices, movement, and reward are far more widespread than expected, with even brainstem and cerebellar regions participating. The message is clear: cognition is not housed in one neighborhood but is broadcast across the city.
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.