Scientists Built the Largest Human Brain Cell Atlas Yet—Here’s What It Reveals About Alzheimer’s and More

The human brain contains tens of billions of cells, but two cells sitting only millimeters apart can be doing very different biological jobs.

That complexity is one reason disorders such as Alzheimer’s disease, Parkinson’s disease, schizophrenia and bipolar disorder remain so difficult to understand.

Now scientists have produced an extraordinary new map designed to expose some of that hidden cellular diversity.

Published in Nature on September 23, 2026, the PsychAD Consortium's research analyzed more than 6.3 million cell nuclei from the dorsolateral prefrontal cortex of 1,494 human donors, making it the largest comparative investigation of human brain disorders at single-cell resolution to date. The donors ranged from infancy to 108 years old and included neurologically typical people as well as people affected by eight major neurological and psychiatric disorders.

The result is not a new brain scan.

It is not a diagnostic test.

And it is certainly not a cure.

Instead, it is something more fundamental:

a molecular atlas showing which genes were active in millions of individual brain-cell nuclei—and how those patterns differed across ageing and disease.

Among the most intriguing findings, researchers found molecular similarities linking several neurodegenerative diseases, identified shifts in neurons, immune cells and vascular cells as Alzheimer’s became more severe, and connected particular neuronal populations with behavioral and psychiatric symptoms seen in dementia.

The atlas gives scientists something they have long lacked:

a common cellular reference map for comparing different brain disorders side by side.

And that could change the questions researchers ask next.

Scientists Have Built an Extraordinary Map of the Brain

What makes this atlas different

Scientists have mapped human brain cells before.

What makes the PsychAD project unusual is its scale, diversity and cross-disease design.

Earlier single-cell studies often involved relatively small numbers of donated brains, examined one disease at a time or relied on bulk tissue measurements that averaged signals from many different cell types.

That averaging can hide important biology.

Imagine analyzing an entire city by calculating the average activity of everyone inside it.

You might learn something about the city as a whole.

But you would lose the distinction between:

teachers,

doctors,

police officers,

children,

engineers,

shopkeepers,

and construction workers.

Brain tissue presents the same problem.

A sample contains neurons, immune cells, blood-vessel-associated cells and several types of glia. If their molecular signals are blended together, a major change occurring in a rare cell population can disappear inside the average.

Single-nucleus RNA sequencing allows researchers to separate much of that complexity.

The PsychAD atlas contains 6,320,459 nuclei from 1,494 donors.

That scale allowed scientists to compare cell types across individuals, diagnoses and disease severity with a level of statistical power that smaller studies cannot easily achieve.

The eight disorders included

The main cross-disorder analysis focused on eight conditions:

• Alzheimer’s disease
• diffuse Lewy body disease
• vascular dementia
• Parkinson’s disease
• tauopathy
• frontotemporal dementia
• schizophrenia
• bipolar disorder

The dataset also included neurologically typical control donors.

That combination is scientifically important.

Brain disorders are traditionally studied in separate research silos.

Alzheimer’s researchers study Alzheimer’s.

Parkinson’s researchers study Parkinson’s.

Psychiatric researchers study schizophrenia.

But patients do not always fit into such neat categories.

Symptoms can overlap.

Pathologies can coexist.

And some genetic risks influence more than one disorder.

By examining several conditions within the same brain region using the same laboratory and computational framework, researchers can look for biology that crosses conventional diagnostic boundaries.

How Scientists Map Gene Activity Cell by Cell

From donated brain tissue to millions of molecular profiles

The study depended on postmortem brain donations.

Scientists obtained preserved tissue from the dorsolateral prefrontal cortex, then isolated millions of cell nuclei from the samples.

Instead of sequencing each donor's complete genome to ask only what DNA they inherited, researchers used single-nucleus RNA sequencing, or snRNA-seq.

RNA offers a different kind of information.

Nearly every cell contains broadly the same genome.

But cells do not use every gene equally.

A neuron may activate one collection of genes.

An immune cell activates another.

An ageing cell may alter its gene activity compared with a younger one.

A diseased cell may activate stress, inflammatory or repair pathways.

RNA molecules provide a snapshot of which genes were being used when the tissue was collected.

In simple terms:

DNA tells scientists what instructions a cell possesses. RNA helps reveal which instructions the cell was using.

The researchers were therefore constructing an atlas of gene activity, not simply sequencing millions of complete genomes.

Why study nuclei instead of whole brain cells?

Brain tissue is technically challenging.

Long, delicate neurons do not survive tissue processing easily.

Cell nuclei are often more stable in frozen postmortem tissue.

That makes single-nucleus sequencing particularly valuable for large human brain banks where tissue may have been stored for years.

The approach still does not capture everything occurring inside the original living cell.

But it allows scientists to obtain detailed gene-expression profiles at enormous scale.

Why the dorsolateral prefrontal cortex?

The researchers focused on the dorsolateral prefrontal cortex, or DLPFC.

This region contributes to high-level cognitive abilities including:

planning,

working memory,

attention,

decision-making,

behavioral control,

and flexible problem-solving.

It is also implicated in several psychiatric and neurodegenerative conditions.

That makes it a particularly useful region for comparative research.

But it creates an important limitation.

The brain is not biologically uniform.

Parkinson’s disease strongly affects circuits involving the substantia nigra.

Alzheimer’s disease prominently affects regions including the hippocampus and temporal cortex.

Psychiatric disorders involve distributed networks.

So what scientists see in the prefrontal cortex cannot automatically be assumed to occur identically throughout the brain.

The atlas is an extraordinarily detailed map.

It is still a map of one major territory.

Turning Millions of Cells Into a Common Classification

To make comparisons possible, researchers organized the cells into a hierarchy.

The atlas resolves:

8 broad cell classes

27 subclasses

and

65 subtypes.

The broad classes included excitatory and inhibitory neurons as well as several supporting and non-neuronal populations such as:

astrocytes,

oligodendrocytes,

oligodendrocyte progenitor cells,

immune cells,

endothelial cells,

and mural cells associated with blood vessels.

This matters because “brain cell” is far too broad a category.

Neurons transmit electrical information.

Astrocytes regulate the environment around neurons and synapses.

Oligodendrocytes produce myelin around axons.

Microglia and related immune cells monitor and respond to injury.

Endothelial and mural cells help form and maintain the brain's vascular system.

A disease can affect each population differently.

What the Atlas Reveals About Eight Brain Disorders

Several diseases share basic molecular stress signals

One of the most striking discoveries was that the eight disorders were not molecularly isolated from one another.

Researchers found shared gene-expression signatures involving basic cellular functions, including RNA processing and protein localization.

That suggests cells facing very different diseases can sometimes enter overlapping biological states.

This does not necessarily mean the diseases begin in the same way.

Imagine several different emergencies affecting a city.

A flood.

An earthquake.

A major fire.

The initial causes are completely different.

Yet all three might activate:

emergency communications,

hospital systems,

road closures,

and rescue operations.

Shared cellular responses may work similarly.

Different diseases can trigger common stress, repair, immune or protein-management systems even when their initiating causes differ.

That distinction is critical.

A shared response does not automatically imply a shared cause.

Neurodegenerative Diseases Showed Particularly Strong Similarities

After accounting for the molecular patterns shared broadly across all eight disorders, researchers found especially strong genetic and transcriptomic similarities among:

Alzheimer’s disease

diffuse Lewy body disease

vascular dementia

and

Parkinson’s disease.

These similarities included alterations involving neuronal and synaptic biology.

The shared gene-expression vulnerability also corresponded with overlapping inherited genetic risk.

That matters because it suggests at least part of the molecular similarity may reflect common biology associated with genetic susceptibility rather than simply cells reacting similarly after becoming sick.

Still, researchers are nowhere near concluding that these diseases are one disorder.

Alzheimer’s pathology is not Parkinson’s pathology.

Vascular dementia is not Lewy body disease.

Their symptoms, pathology, progression and affected brain systems differ substantially.

The atlas instead suggests that conventional diagnostic labels may sit above some deeper biological pathways that overlap.

Psychiatric and Neurodegenerative Disorders Were Still Distinguishable

The study did not collapse schizophrenia, bipolar disorder, Alzheimer’s and Parkinson’s into one molecular category.

Broad patterns of cell abundance and gene expression retained important disease-specific differences.

Psychiatric and neurodegenerative diseases generally showed distinguishable patterns even while some shared biological processes crossed diagnostic boundaries.

That is important because the atlas should not be interpreted as a replacement for clinical diagnosis.

A psychiatrist cannot sequence a few cells and declare that someone has schizophrenia.

A neurologist cannot currently diagnose Alzheimer’s from this atlas.

Diagnosis still depends on established clinical methods such as:

symptoms,

medical history,

cognitive testing,

neurological examination,

brain imaging,

laboratory biomarkers,

and, in some cases, genetic testing.

The atlas is a research framework.

Not a clinical classification system.

New Clues to Alzheimer’s Disease

Alzheimer’s disease received particularly detailed analysis because the PsychAD dataset contains rich neuropathological and clinical information across different disease stages.

And the resulting picture is much broader than a story about neurons alone.

Neuronal abundance fell as Alzheimer’s became more severe

Researchers found that more severe Alzheimer’s disease was associated with reduced neuronal abundance in the prefrontal cortex.

That finding is consistent with the neurodegenerative nature of Alzheimer’s.

But even here, interpretation requires care.

Postmortem data capture the state of a person's brain near the end of life.

They cannot directly show the exact chronological order in which every molecular event occurred.

And a lower proportion of neurons may reflect actual neuronal loss, changing proportions of other cells or both.

The atlas can identify associations.

It cannot observe the same brain decade after decade.

Immune and vascular cells increased

As Alzheimer’s severity increased, researchers also found rising representation of immune and vascular cell populations.

This contributes to a rapidly expanding view of Alzheimer’s disease.

For decades, public explanations often focused mainly on abnormal proteins and dying neurons.

Those remain crucial.

But researchers increasingly examine interactions involving:

neurons,

microglia and other immune cells,

astrocytes,

blood vessels,

the blood-brain barrier,

and metabolic support.

The PsychAD atlas strengthens the idea that Alzheimer’s is not simply a neuron-isolated disorder.

It affects an entire neuro-immune-vascular system.

That does not mean inflammation or blood-vessel dysfunction is “the cause” of Alzheimer’s.

It means the disease involves interacting cell systems that may change at different stages.

Early and Late Alzheimer’s May Involve Different Cellular Responses

The researchers used computational trajectory models to examine how gene-expression patterns varied with tau pathology and clinical dementia severity.

The analysis suggested that different cell types become prominent at different disease stages.

PsychAD's own summary describes an early phase involving immune activation followed later by damaging neurovascular alterations.

This could matter enormously for treatment development.

A therapy that helps during an early inflammatory response might be useless—or even harmful—during a later disease stage.

Conversely, a treatment targeting vascular dysfunction may arrive too late if the relevant process began years earlier.

This is one reason Alzheimer’s drug development is so difficult.

“Alzheimer’s disease” may not represent one biological state.

It may represent a moving sequence of states.

Some People Remain Cognitively Resilient Despite Pathology

Another major question within the PsychAD programme concerns cognitive resilience.

Some people accumulate substantial Alzheimer’s-related pathology in the brain yet retain better cognitive function than researchers might expect.

Why?

Related analyses using the PsychAD dataset identified molecular patterns associated with resilience, including neuronal, astrocytic and mitochondrial features.

Understanding resilience could be just as important as understanding vulnerability.

Scientists traditionally ask:

Why does this cell fail?

A resilience-focused approach asks:

Why did this other cell survive?

The answer could expose naturally protective mechanisms that future therapies might try to reproduce.

Deep-Layer Excitatory Neurons Were Linked to Behavioral Symptoms

Alzheimer’s disease affects more than memory.

Many patients experience neuropsychiatric symptoms such as:

agitation,

depression,

apathy,

anxiety,

delusions,

or other behavioral changes.

The new atlas found that a greater abundance of deep-layer excitatory neurons was associated with a broad range of neuropsychiatric symptoms in Alzheimer’s.

That does not mean those neurons directly cause agitation or depression.

The association could reflect:

disease progression,

compensatory changes,

network remodeling,

or another process correlated with both cell abundance and symptoms.

Replication and mechanistic experiments will be needed.

But the result offers a potential cellular bridge between the microscopic biology of disease and the behavioral symptoms families actually see.

What “Shared Vulnerability” Really Means

The phrase sounds dramatic.

But it needs careful interpretation.

A common response is not necessarily a common cause

Suppose Alzheimer’s and Parkinson’s both change genes involved in protein handling.

There are several possibilities.

The shared pathway might contribute directly to both diseases.

Or it might represent the brain's response to damage.

Or it might be protective.

Or it might become harmful only after prolonged activation.

The atlas tells scientists where to look.

It does not automatically tell them which direction the causal arrow points.

That is why follow-up experiments matter.

Genes Do Not Act Alone

Brain disorders are sometimes described in headlines as though scientists will eventually discover “the Alzheimer’s gene” or “the schizophrenia cell.”

Reality is much more complicated.

Gene-expression patterns depend on interactions among:

inherited DNA variants,

age,

cell type,

environment,

medication,

disease state,

stress,

and interactions with neighboring cells.

Expression is dynamic.

A gene may be highly active during development, quiet during midlife and active again during ageing.

The related PsychAD lifespan study found just such non-linear patterns, with rapid developmental remodeling, relative stability in adulthood and renewed molecular change later in life.

Complex brain disease therefore cannot be reduced to one gene turning on or off.

Rare Cell Types May Matter Disproportionately

One of the great promises of single-cell biology is the ability to study small populations that disappear inside bulk tissue averages.

A rare neuronal subtype might represent only a tiny fraction of a sample.

But if that subtype is especially vulnerable—or especially protective—it could have an outsized influence on symptoms.

Even a dataset containing 6.3 million nuclei may not perfectly capture every rare or transient cellular state.

Still, the atlas gives researchers an unprecedented place to begin searching.

Future experiments can focus on the populations that look most unusual.

Could the Atlas Lead to New Treatments?

Potentially.

But not quickly.

It can help researchers prioritize targets

Suppose one molecular pathway changes strongly in a specific cell type across several neurodegenerative diseases.

Scientists can investigate whether manipulating that pathway changes disease-related behavior in:

cultured cells,

brain organoids,

animal models,

or additional human tissue.

That process can turn an observational signal into a mechanistic hypothesis.

If the mechanism survives repeated testing, it may eventually become a therapeutic target.

But the phrase potential drug target should never be confused with medicine ready for patients.

Most candidate targets do not become successful drugs.

Shared mechanisms could inspire cross-disorder treatments

The overlap among Alzheimer’s, Parkinson’s, Lewy body disease and vascular dementia raises an especially interesting possibility.

Could a therapy aimed at a shared cellular vulnerability benefit people with more than one diagnosis?

Perhaps.

That would challenge the traditional model of developing completely separate treatment pipelines for every disorder.

But there are complications.

The same pathway may perform different functions at different stages.

It may be harmful in one cell but protective in another.

And basic processes such as RNA processing and protein localization occur throughout the body.

A drug that disrupts an essential cellular function in the brain might cause serious effects elsewhere.

Shared biology can reveal opportunities.

It can also reveal why targeting biology safely is difficult.

The Atlas Does Not Create a New Alzheimer’s Test

This point deserves emphasis because large biomedical studies often generate exaggerated headlines.

The PsychAD atlas does not currently provide:

a blood test for Alzheimer’s,

a genetic test that predicts schizophrenia,

a personalized drug recommendation,

or a way to diagnose Parkinson’s from one molecular measurement.

Moving from atlas to clinic requires several stages:

association,

replication,

mechanistic validation,

biomarker development,

drug discovery,

preclinical testing,

safety studies,

and clinical trials.

That can take years or decades.

Products claiming that the new atlas already enables personalized treatment should therefore be treated with skepticism.

How Representative Is the Brain Atlas?

Scale is impressive.

Representation matters too.

More than 30% of donors had non-European genetic ancestry

The full PsychAD cohort included 509 donors—more than 30% of the total—with non-European genetic ancestry.

That is important because biomedical datasets have historically overrepresented people of European ancestry.

Broader diversity improves the chances of identifying biology that generalizes across populations and of discovering genetic associations that would otherwise remain hidden.

The related genetic studies included European, African, American and smaller East and South Asian ancestry groups.

Still, diversity can improve further.

Nearly 1,500 donated brains are enormous for a single-cell study.

They are tiny compared with humanity.

Future atlases will benefit from broader geographic, socioeconomic and ancestral representation.

A lifespan study built from snapshots

The full PsychAD cohort spans people from infancy to 108 years old.

That creates a remarkable cross-sectional view of human brain biology across the lifespan.

But the study did not follow the same person from infancy to 108.

Each donated brain represents a different person at one point in life.

This distinction matters.

Researchers can compare younger and older donors and model trajectories.

But they cannot directly observe exactly how one individual's cells changed over 80 years.

Differences between age groups can also reflect:

generation,

environment,

medication,

life history,

disease,

and other factors.

Postmortem tissue has unavoidable limitations

Brain donation makes research of this scale possible.

It also introduces challenges.

Gene activity can change around the time of death.

Tissue preservation differs.

The interval between death and preservation can matter.

Clinical information may vary in completeness between brain banks.

Medication histories differ.

Some donors have multiple diagnoses.

And a postmortem sample primarily captures the endpoint of a biological story.

It cannot show every intermediate event directly.

Researchers use quality control and statistical methods to reduce these problems.

They cannot make them disappear completely.

What the Atlas Cannot Tell Us Yet

One brain region is not the whole brain

The dorsolateral prefrontal cortex is important.

But Alzheimer’s, Parkinson’s and psychiatric illness affect distributed neural networks.

Researchers now need similarly deep maps of other regions, including areas such as:

the hippocampus,

temporal cortex,

substantia nigra,

amygdala,

and additional cortical and subcortical structures.

A cell subtype may respond one way in the prefrontal cortex and another way elsewhere.

An atlas of one region can reveal important biology.

It cannot substitute for a whole-brain map.

RNA is only one layer of biology

The atlas measures RNA transcripts.

That is extremely informative.

But RNA is not the entire cell.

To understand disease fully, researchers also need information about:

proteins,

DNA regulation,

epigenetic changes,

cell shape,

metabolism,

electrical activity,

synaptic connections,

and spatial organization.

Two cells can express similar genes yet behave differently because their proteins are modified differently or because they occupy different local environments.

That is why future brain atlases increasingly combine multiple “omics” layers.

Spatial information is especially important

Ordinary single-nucleus sequencing separates nuclei from their original tissue.

That gives researchers excellent molecular profiles but weakens information about where each cell originally sat.

Yet location matters enormously.

A neuron beside one vascular structure may behave differently from an otherwise similar neuron in another cortical layer.

Cells communicate with neighboring cells.

Immune cells respond to local damage.

Astrocytes interact with nearby synapses.

Blood vessels create local metabolic environments.

The PsychAD programme includes spatial-transcriptomic resources, and future research will increasingly combine cell identity with physical location.

Association Is Not Causation

This may be the single most important caveat.

Suppose a gene is more active in severe Alzheimer’s disease.

What does that mean?

Possibility one:

it is helping cause damage.

Possibility two:

it is responding to damage.

Possibility three:

it is trying to protect the brain.

Possibility four:

another process drives both the disease and the gene-expression change.

The atlas cannot distinguish these possibilities by itself.

Likewise, computational trajectories showing “early” and “late” disease patterns are models built from different donors.

They are not time-lapse recordings of the same patient's brain.

The models generate hypotheses.

Biological experiments must test them.

The Atlas Is an Open Resource for the Next Generation of Brain Research

One of the strongest features of the PsychAD project is that it is not intended to remain locked inside a single laboratory.

Researchers have released data and tools so other scientists can explore and challenge the findings.

Scientists can access the data through the AD Knowledge Portal

Raw and processed single-nucleus data are available through the AD Knowledge Portal under controlled-access conditions designed to protect participant privacy.

Researchers can use the resource to:

test new hypotheses,

compare future samples,

investigate specific genes,

examine particular cell populations,

or develop new computational methods.

CELLxGENE allows interactive exploration

The consortium also provides interactive visualization through CELLxGENE, allowing researchers to explore cell types, metadata and gene-expression patterns in a web browser.

That dramatically increases the value of the project.

The atlas is not only one paper.

It is infrastructure for many future studies.

Why Human Molecular Data Still Need Privacy Protections

“Open science” does not mean every raw human dataset should be freely downloadable without safeguards.

Genomic and detailed molecular information can be sensitive.

Even when obvious identifiers such as names are removed, combinations of genetic and clinical information create privacy risks.

That is why some PsychAD data require:

registration,

data-use agreements,

and controlled access.

The scientific goal is to make the dataset useful while protecting donor anonymity.

Behind every one of those 6.3 million molecular profiles is a person who donated their brain—or whose family helped make that donation possible.

The atlas exists because of that contribution.

What Researchers Should Build Next

The PsychAD atlas answers many questions.

It creates even more.

The next generation of work will need:

additional brain regions,

larger globally representative cohorts,

more spatial measurements,

protein-level information,

epigenetic maps,

genetic integration,

whole-cell morphology,

and biomarkers that can be measured repeatedly in living people.

The final point is especially important.

Postmortem tissue gives scientists extraordinary cellular resolution.

Clinical medicine needs ways to follow disease while a patient is alive.

The ideal future will connect detailed postmortem maps with:

blood biomarkers,

cerebrospinal-fluid measurements,

brain imaging,

cognitive testing,

and longitudinal clinical data.

That could allow researchers to connect microscopic cellular changes with the progression doctors actually observe over time.

Why This Atlas Matters Even If It Produces No Drug Tomorrow

Scientific breakthroughs are often judged by an impatient question:

When will this cure something?

But foundational maps have their own value.

The Human Genome Project did not instantly create treatments for every genetic disease.

Detailed anatomical maps did not immediately cure neurological illness.

Resources change research by giving scientists better coordinates.

The PsychAD atlas does something similar.

Instead of saying:

“Alzheimer’s changes the brain,”

researchers can increasingly ask:

Which neuronal subtype changes?

At what disease stage?

Which genes change with it?

What are neighboring immune cells doing?

Does the same pathway change in Parkinson’s?

Is the signal associated with genetic risk?

Is it present in cognitively resilient people?

Those are much sharper questions.

And better questions are often the beginning of better treatments.

The Biggest Insight May Be That Brain Disease Is a Systems Problem

For decades, neuroscience often searched for the critical malfunction.

One protein.

One neurotransmitter.

One brain region.

One cell.

Those discoveries remain enormously important.

But the new atlas reinforces a more complicated reality.

Brain disorders emerge from systems.

Neurons interact with glia.

Glia interact with immune pathways.

Immune cells interact with blood vessels.

Blood vessels influence metabolism.

Genetics influences every layer.

Age changes how those cells respond.

And disease can push the entire network into new states.

The PsychAD findings repeatedly point toward interactions involving neuronal, immune and vascular systems, especially in Alzheimer’s disease.

That systems view may eventually prove more important than any individual gene identified in the atlas.

A Map, Not Yet a Cure

The scale alone is difficult to comprehend.

6,320,459 nuclei.

1,494 human donors.

Eight major brain disorders.

Eight broad cell classes.

Twenty-seven subclasses.

Sixty-five subtypes.

Ages ranging from infancy to 108 years.

And millions of measurements showing which genes were active in which cells.

The atlas reveals both unity and difference.

Alzheimer’s, Parkinson’s, Lewy body disease and vascular dementia share parts of their molecular vulnerability.

But they remain distinct diseases.

Alzheimer’s involves changes not only in neurons but in immune and vascular cells.

Yet those changes do not reveal a single cause.

Specific neuronal populations correlate with behavioral symptoms.

But correlation is not prediction.

Genetic diversity improves the atlas.

But the world's populations are still not represented equally.

The resource is enormous.

But it still covers one major brain region.

That combination of power and limitation is exactly what good science looks like.

The PsychAD atlas does not tell us that scientists have solved Alzheimer’s disease.

It tells us that they can now study it at a level of cellular resolution that would have been extraordinarily difficult only a decade ago.

And perhaps the most important lesson is that there may never be one simple answer.

Brain disease appears to involve conversations among many kinds of cells.

Those conversations change with age.

They change with genetics.

They change as disease progresses.

Now scientists have a much better map of those conversations.

The map is not the destination.

But if researchers hope to understand where disease begins, why some people remain resilient and where future treatments should intervene, having better coordinates is a very good place to start.

Frequently Asked Questions

What is the new human brain-cell atlas?

It is a large molecular map created by the PsychAD Consortium using single-nucleus RNA sequencing of the human dorsolateral prefrontal cortex. It allows researchers to compare gene activity across millions of cells from people with and without major brain disorders.

How many cells were included?

The final cross-disorder dataset contained 6,320,459 cell nuclei.

How many people donated brain tissue?

The atlas includes tissue from 1,494 unique donors.

How old were the donors?

Across the full PsychAD cohort, donors ranged from infancy to 108 years old.

Which brain disorders were studied?

The eight main disorders were Alzheimer’s disease, diffuse Lewy body disease, vascular dementia, Parkinson’s disease, tauopathy, frontotemporal dementia, schizophrenia and bipolar disorder.

What part of the brain did scientists study?

Researchers studied the dorsolateral prefrontal cortex, a region involved in working memory, planning, decision-making and other executive functions.

What is single-nucleus RNA sequencing?

Single-nucleus RNA sequencing measures RNA from individual cell nuclei, allowing scientists to estimate which genes were active in different cell types.

Why use nuclei instead of whole cells?

Nuclei can be recovered reliably from frozen postmortem brain tissue, making the technique particularly useful for large human brain-bank studies.

Is RNA sequencing the same as sequencing the entire genome?

No. Genome sequencing examines inherited DNA. RNA sequencing measures gene-expression activity—the genes a cell was using at the time the tissue was collected.

How many cell types did the researchers classify?

The atlas organized nuclei into eight major cell classes, 27 subclasses and 65 subtypes.

What did the atlas reveal about Alzheimer’s disease?

More severe Alzheimer’s was associated with reduced neuronal abundance and increased immune- and vascular-cell populations. Researchers also identified stage-related gene-expression patterns and neuronal populations associated with neuropsychiatric symptoms.

Did scientists find similarities between Alzheimer’s and Parkinson’s?

Yes. After accounting for broadly shared disease signatures, Alzheimer’s, Parkinson’s, diffuse Lewy body disease and vascular dementia showed particularly strong genetic and transcriptomic similarities.

Does that mean Alzheimer’s and Parkinson’s are the same disease?

No. They remain clinically and pathologically distinct disorders. Shared molecular pathways can exist without the diseases having the same cause or progression.

What are glial cells?

Glia are non-neuronal cells that support and regulate the nervous system. They include astrocytes, oligodendrocytes and immune-related cells such as microglia.

Why do immune cells matter in Alzheimer’s disease?

Immune cells may participate in responses to injury, abnormal proteins and inflammation. The atlas found immune-cell changes associated with Alzheimer’s severity, but it does not establish inflammation as the single cause of the disease.

Why do vascular cells matter?

The brain depends on a tightly regulated blood supply. Changes involving endothelial and other vascular cells could affect oxygen delivery, metabolic support and blood-brain-barrier function.

What are neuropsychiatric symptoms in Alzheimer’s disease?

They can include symptoms such as depression, agitation, apathy, anxiety or psychosis. Researchers found associations between a deep-layer excitatory-neuron population and a range of such symptoms.

Can the atlas predict who will develop Alzheimer’s?

No. It is not currently a predictive clinical test.

Can the atlas diagnose schizophrenia or bipolar disorder?

No. The atlas is a research resource, not an approved diagnostic system.

Did the study identify a single cause of Alzheimer’s?

No. The findings point toward complex interactions among neuronal, immune, vascular and genetic processes.

What does “shared molecular vulnerability” mean?

It means that different diseases can show overlapping changes in gene activity or cellular pathways. Those similarities may reflect shared genetic risk, shared disease mechanisms or common responses to injury.

Does shared vulnerability mean the diseases could use the same treatment?

Possibly in some cases, but that remains speculative. Researchers must determine whether a shared pathway is causal, harmful, protective or simply a consequence of disease.

What is cognitive resilience in Alzheimer’s research?

Cognitive resilience describes people who maintain better thinking abilities than expected despite substantial Alzheimer’s-related pathology.

Why are cognitively resilient people important to study?

They may reveal naturally protective cellular or molecular mechanisms that could inspire future therapeutic research.

Can the atlas lead to new medicines?

It can help identify potential genes, pathways and cell types for further investigation. Turning those findings into medicines requires extensive laboratory validation and clinical testing.

How long could that process take?

Potentially many years. Most biological targets discovered in basic research never become approved drugs.

How diverse was the donor group?

More than 30% of donors in the full cohort were classified as having non-European genetic ancestry.

Why does ancestry diversity matter?

Broader representation helps researchers identify biology that generalizes across populations and reduces reliance on datasets dominated by European ancestry.

Did researchers follow people from childhood into old age?

No. The atlas is primarily cross-sectional. Different donors represent different ages rather than the same individuals being followed throughout life.

Why is postmortem tissue useful?

It allows researchers to directly examine human brain cells at molecular resolution, something that usually cannot be done safely in living people.

What are the limitations of postmortem tissue?

Gene expression can be influenced by death, tissue preservation, medication, disease severity and other factors. Postmortem tissue also captures a late snapshot rather than the full history of disease progression.

Does the atlas cover the entire brain?

No. The principal dataset focuses on the dorsolateral prefrontal cortex.

Why is that a limitation?

Cell populations and disease effects can vary greatly among brain regions. Alzheimer’s, Parkinson’s and psychiatric disorders involve distributed brain networks.

What additional regions should scientists study?

Important targets include the hippocampus, temporal cortex, substantia nigra and other cortical and subcortical regions relevant to specific diseases.

Does gene expression tell scientists everything about a cell?

No. RNA is only one biological layer. Protein abundance, epigenetics, metabolism, spatial position, electrical activity and connectivity also matter.

What is spatial transcriptomics?

Spatial transcriptomics measures gene activity while preserving information about where cells or gene-expression signals are located within tissue.

Why is spatial information important?

Cells communicate with nearby cells. Their location can influence metabolism, immune activity, vascular support and neuronal connectivity.

Can researchers access the PsychAD data?

Yes. PsychAD makes research data available through resources including the AD Knowledge Portal and provides interactive single-cell visualization through CELLxGENE.

Is all of the data completely open without restrictions?

No. Some molecular and genetic data require controlled access and data-use agreements to protect donor privacy.

Who funded the work?

The PsychAD programme has received support from the U.S. National Institute on Aging and other NIH resources.

When was the major atlas published?

The main Nature papers were published online on September 23, 2026, with the journal issue dated September 24.

Is this the largest brain-cell atlas ever made?

Nature described it as the largest-known map of gene expression in the human prefrontal cortex, while NIH characterized the project as the largest comparative investigation of human brain disorders at single-cell resolution.

What is the most important takeaway?

The atlas suggests that complex brain disorders cannot be understood by studying neurons alone.

Alzheimer’s and related diseases involve interacting networks of neuronal, immune, vascular and supporting cells, while several disorders also share molecular vulnerabilities that cross traditional diagnostic boundaries.

The atlas does not provide a cure.

It gives researchers something that may be just as important at this stage:

a far better map of where to look next.