A global neuroscience collaboration has published a detailed cellular atlas of the basal ganglia, a group of deep brain structures involved in movement, decision-making, habit formation and emotion. The research package, released on November 12, brings together 26 studies and data from 25.5 million cells. Its importance is not simply the size of the dataset. Researchers have combined gene activity, epigenetic regulation, spatial location and physiological information to create a shared reference that laboratories can use to compare brain cell types across people and other mammals.
Why the basal ganglia became the first major target
The basal ganglia sit deep within the brain and form a network of interconnected structures that help select, initiate and regulate actions. They are central to voluntary movement, but their influence extends well beyond muscles. The same circuits contribute to habit formation, motivation, reward processing, decision-making and aspects of emotional behaviour. Because the network interacts with the cerebral cortex, thalamus and other brain regions, a small change in one cell population can affect a much wider circuit.
The region is also implicated in several neurological and psychiatric conditions. Parkinson’s disease is associated with the loss of dopamine-producing neurons that project into the basal ganglia. Huntington’s disease damages striatal neurons. Research into schizophrenia, bipolar disorder, addiction and autism has also pointed to cell types and circuits within this network. These conditions are not caused by one structure alone, and the new atlas does not claim to explain them. It provides a more precise map of the cells researchers can investigate.
The project was organised through the BRAIN Initiative Cell Atlas Network, or BICAN, part of the US National Institutes of Health’s broader effort to build a comprehensive reference for the human brain. The consortium involved dozens of laboratories and hundreds of scientists across nine countries and more than 50 institutions. The November 12 publication package includes 26 studies in journals within the Cell Press family, rather than a single paper from one laboratory.
What the 25.5 million cell profiles represent
The Allen Institute’s announcement reported profiles from 21 million human cells, 3 million cells from rhesus and cynomolgus macaques, and 1.5 million cells from marmosets. Those figures add to 25.5 million. The wider atlas work also includes cross-species comparisons involving mouse data, but the headline 25.5-million figure refers to the human and non-human primate cell profiles described in the announcement. Keeping the accounting clear matters because different studies in a large research package may use different samples and methods.
Single-cell and single-nucleus techniques allow researchers to measure which genes are active in individual cells. Instead of treating a piece of brain tissue as one blended sample, scientists can distinguish different cell populations and examine their molecular signatures. That makes it possible to identify cell types that look similar under a microscope but behave differently at the molecular level. It also helps researchers compare how similar cell populations are across species.
The atlas combines several layers of evidence. Gene-expression data show which genes are active. Epigenetic measurements examine chemical marks and regulatory features that influence whether genes are switched on or off. Spatial approaches preserve information about where cells sit within tissue. Other studies examined cell morphology and physiological properties, including how cells respond electrically. When those datasets are aligned, researchers can test whether a proposed cell type is supported by more than one method.
A common language for brain cell types
One of the biggest problems in neuroscience is that laboratories have historically used different names and classification systems for similar cell populations. A cell group described by one team might be divided into several subtypes by another. That makes it difficult to compare results across experiments, especially when samples come from different species, brain regions or technologies. A shared taxonomy can reduce this confusion by giving researchers a common set of labels and reference measurements.
The new atlas is intended to provide that common language for the basal ganglia. Researchers can use the reference to identify a cell population in a new sample, compare gene-expression patterns and determine whether a cell group has a counterpart in another species. This does not mean that a mouse cell and a human cell are interchangeable. It means that scientists have a more structured way to describe similarities and differences rather than relying on names that may not line up.
The cross-species component is especially important because many experiments that cannot be performed in living human brains are carried out in animal models. If researchers can map comparable cell populations across humans, macaques, marmosets and mice, they can ask which features are conserved and which appear to be more specialised. That can help them judge how well a model reflects human biology before drawing conclusions about disease mechanisms or potential treatments.
The striatum reveals hidden molecular zones
Several studies in the package focus on the striatum, the largest component of the basal ganglia. Broad Institute researchers reported six molecular zones within the human striatum, identified through patterns of gene expression in projection neurons. These zones were not obvious from conventional visual inspection alone. The researchers used spatial mapping methods to connect molecular signatures with the physical location of cells in tissue.
The finding is an example of what the researchers call hidden histology. Two groups of cells may look similar under a microscope but have distinct molecular profiles and different patterns of gene activity. Mapping those differences can reveal organisation that was previously difficult to see. It may help scientists understand why certain circuits are more involved in particular behaviours or disease processes, although the functional meaning of each zone still requires further investigation.
The Broad team also analysed more than four million cells from the striatum and prefrontal cortex of nearly 200 adult postmortem brains. It found that the proportions of cell types varied between individuals, especially among non-neuronal cells such as astrocytes and microglia. In one reported age comparison, the abundance of oligodendrocyte progenitor cells declined by approximately 40 percent between ages 30 and 80. That result describes the sampled tissue and cell population, not a universal measurement of brain decline for every person.
Age changes the map as well as the cells
The research suggests that ageing affects not only the number of some cell types but also the molecular boundaries that distinguish neighbouring zones. Broad researchers found that the six striatal zones became less sharply defined with increasing age, as patterns of gene expression grew more similar across some boundaries. This does not mean that the striatum simply loses its structure. It indicates that the molecular distinctions between certain populations change over the lifespan.
Age-related shifts are difficult to interpret because they can reflect several processes. Cells may alter their gene activity in response to inflammation, changes in connectivity, metabolic stress or other biological factors. The proportions of cell types in a sample can also vary because of differences between individuals and the health histories of the donors. A reference atlas can help researchers separate ordinary variation from patterns that are associated with particular conditions, but it cannot establish cause and effect on its own.
Another value of the atlas is that it gives researchers a baseline against which to compare diseased tissue. If a future study finds that a cell population is altered in Parkinson’s disease or schizophrenia, investigators can ask whether the difference is also present in healthy ageing, whether it is specific to one molecular zone and whether it appears in the same cell type across more than one dataset. That is a more precise starting point than treating the basal ganglia as a single uniform region.
A rare neuron draws attention in schizophrenia research
One of the findings highlighted by National Geographic concerns a rare subtype of neuron in the striatum that expresses the serotonin receptor HTR7. Researchers compared cell-level molecular data with large genetic studies of psychiatric disorders and found that genetic risk for schizophrenia was enriched in this cell group. The result gives scientists a candidate population for further study, but it does not show that the neuron causes schizophrenia or that its presence can diagnose an individual.
That distinction is essential. Genetic risk is distributed across many variants and biological pathways. A statistical association between risk signals and a cell type can suggest where some of the relevant biology may be occurring, but the association must be tested in additional datasets and experiments. Researchers still need to understand how the cell behaves, how it connects to other neurons and whether its activity differs in people with schizophrenia.
The finding may eventually help scientists design more targeted experiments. If a particular neuron subtype is involved in a circuit that contributes to symptoms, researchers could investigate its receptors, signalling pathways and interactions with neighbouring cells. The atlas makes it easier to identify and label the population. It does not provide an immediate treatment, a screening test or a reason to change clinical care.
Open data could matter as much as the discovery
The consortium has made major datasets and workflows available to the research community. Open access is important because a resource of this size cannot be fully explored by the teams that created it. Independent laboratories can test alternative classifications, compare the atlas with their own samples and look for patterns that were not the focus of the original studies. A common data framework can also make it easier to reproduce analyses and build on them.
The atlas is a milestone, not a completed map of the human brain. The basal ganglia are only one network within an organ containing billions of neurons and many more supporting cells. Other brain regions have different circuits, cell types and developmental histories. The BRAIN Initiative’s long-term ambition is to create a comprehensive catalogue of cell types across the whole human brain, and the basal ganglia project provides a set of methods and standards that can be extended to other regions.
Open data also makes it possible to ask questions that require larger samples than one laboratory can assemble. Researchers can investigate variation across age groups, compare cell populations between people, and examine how different molecular layers fit together. The value will depend on how well the datasets are documented, how accessible the tools remain and whether other teams can reproduce the findings. A large data release becomes more useful when it is treated as shared scientific infrastructure rather than a one-time headline.
What the atlas cannot tell us yet
Most of the human tissue used in this kind of research comes from postmortem samples. That provides access to cell types and molecular information that would be difficult to collect from living brains, but it also limits what can be observed. A static sample cannot show how a neuron changes its firing pattern while a person makes a decision or learns a new habit. Researchers must combine atlas data with imaging, electrophysiology, animal experiments and clinical studies to understand how cell populations behave over time.
Cell classification is also an evolving process. New technologies may reveal subtypes that are not visible with current methods, while different samples may show additional variation. A standard taxonomy is useful because it allows comparison, but it should not be mistaken for a final list that can never change. Scientists will need to update the reference as larger datasets, new spatial techniques and better functional measurements become available.
For patients and families affected by neurological or psychiatric disorders, the near-term message is cautious. The atlas is not a new diagnostic test and it does not produce a treatment that is ready for clinical use. Its contribution is upstream: it gives researchers a clearer map of the cells and molecular systems that future studies can examine. That work may take years, but it can help move neuroscience from broad regional descriptions toward more precise questions about which cells, in which circuits, are involved in disease.
Frequently asked questions
What is the basal ganglia?
The basal ganglia are a group of interconnected structures deep within the brain. They help regulate movement, habits, reward, decision-making and aspects of emotion through circuits that interact with the cortex and other brain regions.
Does the 25.5-million-cell atlas identify a cure for Parkinson’s or schizophrenia?
No. It is a research reference that helps scientists identify cell types and molecular patterns. It may guide future work on disease mechanisms, but it is not a treatment or diagnostic test.
Why did researchers compare humans with other mammals?
Cross-species comparisons help scientists identify cell types and molecular features that are shared or different across mammals. This can improve the interpretation of animal models while highlighting where human brain biology may be distinct.
Sources and reporting notes
- Allen Institute, “Scientists create landmark cellular atlas of one of the brain’s most important regions,” November 12, 2026.
- Broad Institute, “Researchers uncover distinct zones and age-associated changes in a key brain area,” November 12, 2026.
- EL PAÍS, report on the cross-species basal ganglia atlas, November 12, 2026.
- National Geographic, “Scientists Discover a Neuron That May Play a Role in Schizophrenia,” November 12, 2026.
Updated November 13, 2026. Cell counts and age-related findings are reported as described by the research teams; associations are not presented as proof of causation.
