When I was a child, it felt like summer was never-ending.

My mom used to take us for long days at the beach. We also did bike rides, explored new places and it seemed like I was meeting new children to play with almost every day.

Entire weeks felt like they were packed with new, different and exciting experiences. Looking back, I'm grateful my parents gave us so much to see and do. I know it isn’t like that for every child.

Anyway, now it’s the end of June, although I swear it feels like it was March last week. And I know that I'll blink, and then suddenly it’ll be September…

What changed? Did time speed up?

Not exactly.

One compelling explanation is that summers seem shorter as we get older because of how our brains pay attention to and remember experiences.

1. The brain pays attention to what is new

You know the feeling of driving somewhere for the first time? How the way there seems to take forever. But the way back feels a lot shorter, even though it's the same road. Nothing changed except that the second time, your brain already knew the way.

If our brains recorded everything equally, we'd be overwhelmed. Instead, they prioritize information that is novel, surprising or meaningful.

For example, your first day at university when you knew nobody, your first solo trip abroad or the day you moved out of your parents house. These experiences require attention because they’re unfamiliar to us. The brain has to work harder to process what’s happening.

Novel or unexpected events draw our attention and are more easily remembered than predictable or familiar ones. Research suggests that novelty engages memory systems in the hippocampus, helping new experiences stand out from the background of everyday life.

One idea proposes that novelty detected by the hippocampus engages dopamine-producing neurons in the ventral tegmental area (VTA). The resulting dopamine release is thought to help strengthen memory formation, making novel experiences more likely to be remembered.

By contrast, familiar experiences don't drive this signal in the same way. Routine experiences require less conscious processing. Once something becomes familiar, neural responses often become more efficient and, in many brain regions, smaller. That can be useful of course, but it also means that fewer distinct memories may be formed.

2. Sense of time in the moment

The time you feel while something is happening and the time you judge looking back are not necessarily the same.

Novelty doesn't only influence memory. It can also influence how long an experience feels in the moment.

One demonstration of this comes from the "oddball effect." In these experiments, participants viewed a sequence of nearly identical images, with an occasional unusual image inserted into the sequence. Although every image is shown for exactly the same amount of time, people judge the unusual image as lasting longer.

In a 2019 study, researchers built a computational model that estimated duration without any built-in clock. Instead, the model measured changes in perceptual processing over time. The model reproduced several biases that are also seen in human duration judgments. For example, busy city scenes tended to be estimated as lasting longer than quieter/less busy scenes, such as a cafe or office.

It’s still being debated exactly how the brain generates its sense of duration. But studies like these suggest that novelty and change can influence not only what we remember, but also how we experience the passage of time itself.

The more important story for the summer question, though, may be what happens afterwards.

3. Sense of time in hindsight

Our sense of time is also influenced by how we remember a period afterward.

Most of us probably know how a dull meeting can feel like forever while we’re in it, but it can easily be forgotten a month later.

A trip somewhere new does the reverse. Time can race by while we’re there, yet it can feel substantial in hindsight because there’s so much to recall.

Research on time perception suggests that when we reflect on a period of life, our brains partly reconstruct how long it felt from the memories formed during that time.

Episodic memories are not stored as a continuous recording of our lives. Instead, they are built from discrete events. Periods that contain more distinct events may therefore leave behind a richer memory trace than periods dominated by routine. One hypothesis is that these memory markers contribute to our retrospective sense of duration.

A month spent following the same routines stores fewer distinct memories. And looking back, it can feel as though it passed in a blur. (The start of this summer has, honestly!).

So the sense that time is speeding up may be partly about the amount of distinct memories in a period of life. Moments, weeks and years with fewer novel experiences may leave behind fewer memorable landmarks, while the opposite can happen with those filled with novelty, surprise or change.

Subscribe now

Why did childhood summers feel so long?

Children are constantly encountering new things.

New places, skills, friendships and experiences. Much of the world is still being discovered for the first time.

Almost everything is a first. And those firsts are also more likely to become distinct memories rather than blending together with everything around them.

Adults, on the other hand, often settle into routines. We drive the same routes, visit the same places and follow the same schedules and routines week after week.

Life becomes more predictable and fewer experiences stand out from the background of everyday routine.

So the result isn't necessarily that time moves faster now, but that our perception of it changes. Fewer memorable landmarks may be created along the way.

All in all, there is evidence linking novelty to memory formation, and the idea that this explains why time seems to accelerate with age is plausible, although this area is still being researched and there might be more to the story.

How to make your summer feel longer or more expansive

We don’t necessarily need a four-week holiday or a dramatic life change.

The goal is to do novel things, and they can be small. That could mean:

• Taking a different route on your way to work or on your evening walk

• Visiting a place you’ve never been before

• Cooking a new recipe with ingredients you might not often use

• Trying a new hobby or sport

• Having a conversation with someone new

• Reading about a topic you know nothing about

• Saying yes to experiences that are slightly outside your usual routine or comfort zone

• Practicing micro-mindfulness by dedicating time to focus entirely on sensory details, like textures, sounds and scents

• Cultivating a general sense of curiosity

These moments don’t have to be extraordinary, just a little different from the usual.

If you want this summer to feel longer, focus on making it more memorable, not just busier.

Thank you for reading! If you try one small new thing this week, I'd love to hear what it is :)

-Eleni

Share

References and further reading

1. Ranganath, C. & Rainer, G. Neural mechanisms for detecting and remembering novel events. Nat. Rev. Neurosci. 4, 193–202 (2003).

2. Lisman, J. E. & Grace, A. A. The hippocampal-VTA loop: controlling the entry of information into long-term memory. Neuron 46, 703–713 (2005).

3. Roseboom, W. et al. Activity in perceptual classification networks as a basis for human subjective time perception. Nat. Commun. 10, 267 (2019).

4. Matthews, W. J. & Meck, W. H. Temporal cognition: Connecting subjective time to perception, attention, and memory. Psychol. Bull. 142, 865–907 (2016).

5. Block, R. A. & Zakay, D. Prospective and retrospective duration judgments: A meta-analytic review. Psychon. Bull. Rev. 4, 184–197 (1997).

You are probably familiar by now with the Lancet Commission’s report on dementia, which I have mentioned in quite a few posts (for anyone new: it's a group of experts who periodically put together the evidence on what raises and lowers the risk of dementia). In its 2024 update, the Lancet Commission estimated that hearing loss in midlife is among the most important potentially modifiable risk factors for dementia. And modifiable is a key word here, as we might be able to do something about it, unlike age or the genes we are born with.

Hearing loss is common, particularly as people age. Globally, it has been estimated that around 1.5 billion, or 20% of people, live with some degree of hearing loss.

Based on statistical models, the Commission’s report estimated that hearing loss accounts for around 7% of dementia cases at the population level. This doesn’t mean that hearing loss directly causes 7% of dementia cases, but rather that dementia rates might be lower if hearing loss could somehow be prevented or effectively treated. Of course, in reality you can’t remove hearing loss from a whole population, so this is a hypothetical estimate and built on a lot of assumptions. But it can still be informative.

So in today’s post we will look at what hearing loss actually is, what we know about its link to dementia, whether it’s a cause or an early sign and whether treating it helps.

What is hearing loss?

The most common kind of hearing loss is age-related, and it's called sensorineural hearing loss. There are thousands of tiny hair cells deep in the inner ear. These cells convert sound waves into electrical signals, which are then carried by the auditory nerve to the brain for interpretation. Over a lifetime, exposure to loud noise and aging gradually wear these cells down.

Hearing loss usually happens slowly, often without the person really noticing. The higher-pitched sounds tend to fade first and speech becomes harder to understand, especially in busy restaurants, meetings or other noisy environments. Because it can creep in gradually, people can live with it for years before doing anything about it.

But once these hair cells are lost, they don’t grow back. This is true not only for age-related hearing loss, but also for many forms of noise-induced hearing loss. Some hearing problems do have reversible causes, like earwax buildup or certain infections, but damage to the sensory cells of the inner ear is usually permanent. A hearing aid cannot rebuild what is already gone. It can only make better use of what is left. And that is part of why it’s importnat to catch hearing loss early.

How does hearing loss link to dementia?

A lot of research since 2010 has looked at whether people with worse hearing are more likely to develop dementia later on. Fairly consistently, the findings have suggested that worse hearing is associated with later risk of dementia. For example, one early study published in 2011 in Archives of Neurology, found that dementia risk increased progressively from mild to moderate to severe hearing loss, suggesting that the relationship follows a dose-response pattern.

More recently, a meta-analysis (a study that combines the results of many earlier ones) published in Ageing Research Reviews pooled 50 observational studies. It found that hearing loss was linked to a higher risk of developing dementia, mild cognitive impairment and Alzheimer's disease specifically. Each 10-decibel drop in hearing was associated with about a 16% higher risk of dementia.

Subscribe now

Is hearing loss a cause or an early sign of dementia?

An association is not proof of cause. The Commission's estimate I mentioned earlier, for instance, came from observational research, which can show that two things go together but can't prove that one causes the other. And Alzheimer's disease, for example, starts in the brain years, often decades, before any symptoms show up. That raises the possibility that in some people, hearing difficulties may be an early sign of the disease already underway, rather than something separate that is causing dementia.

Indeed, not all hearing problems start in the ear. Some begin in the brain itself. There is a difference between hearing loss caused by damage to the ear and difficulties that arise when the brain processes sound. Hearing, in other words, is not just about detecting sounds but about making sense of them. Because Alzheimer's can affect the brain regions that interpret speech and sound, a person may have relatively healthy ears yet still struggle to follow a conversation, because the brain is having trouble interpreting what it hears. From the outside that can look like age-related hearing loss, even though the underlying problem is different.

We don't yet know for sure which way the causal arrow points. Hearing loss might be increasing the risk of dementia. Also, early disease in the brain might be making sounds harder to process, well before memory is overtly affected. It may even run both ways at once, each adding to the other. We haven't been able to pull apart how much hearing loss actually raises dementia risk and how much it is an early sign of something already underway. Probably it is a bit of both.

But studies have tried to separate the two. In a large UK Biobank study, people who had trouble following speech in noise were considerably more likely to develop dementia over the next decade. If that were simply an early symptom of dementia, we would expect the link to be strongest for those diagnosed soon afterward and to weaken for those diagnosed many years later. Instead, the association was consistent across short and very long follow-ups. That doesn't prove hearing loss causes dementia, and it can't fully rule out a very slowly developing disease. But it pushes back against the idea that the association is only an early warning sign, and leans toward hearing being a genuine risk factor.

Why does hearing matter for the brain?

If hearing difficulties do contribute to cognitive decline, several pathways have been proposed.

One idea is that when speech becomes harder to hear, the brain has to work harder to decode what is being said. Then, the resources that would normally be available for memory, attention and other cognitive processes are instead diverted toward understanding speech. I have definitely experienced this type of cognitive load after trying to follow a conversation in a noisy restaurant. It can feel surprisingly mentally exhausting!

A second one is that hearing loss can also be isolating and it can gradually shrink a person's social world. Keeping up in conversations can get tiring, so often people opt out and some even begin avoiding social situations altogether. This is important because social isolation is itself a recognized risk factor for cognitive decline and dementia. Other psychosocial factors, like depression, have also been proposed as mechanisms.

A third possibility is that reduced auditory input leads to less stimulation of the brain, particularly the networks involved in processing sound and language. In one longitudinal study from 2014, older adults with hearing loss showed faster brain volume loss over the following years, especially in the temporal regions that are involved in sound and speech.

More broadly, a recent analysis from the Framingham study published in JAMA Network Open, showed that people with midlife hearing loss tended to have smaller brain volumes and more white-matter damage (small areas of injury that show up as bright spots on a brain scan), and worse performance on tests of planning and mental flexibility. The association appeared strongest among APOE ε4 carriers (the strongest common genetic risk factor for Alzheimer's disease), suggesting that hearing loss may matter most in people already vulnerable to the disease.

The exact mechanisms are still uncertain, though the proposed pathways are plausible. Overall, the findings suggest that hearing loss is tied to changes that reach beyond the brain's hearing areas and is linked to cognitive outcomes.

Do hearing aids mitigate risk?

If hearing loss raises dementia risk, the hope then is that treating it lowers it.

A 2023 meta-analysis published in JAMA Neurology included 8 studies and found that people with hearing loss who used hearing aids (or cochlear implants) had about a 19% lower rate of long-term cognitive decline than those who didn't. The benefit has tended to hold up even in studies with long follow-up, which makes it less likely to be explained by reverse causation. But people who get hearing aids might differ from those who don't in all sorts of ways, making it difficult to tease apart the contribution of hearing aids to lower risk.

So far, a trial called ACHIEVE is the largest randomized trial specifically designed to test whether treating hearing loss affects cognitive decline. This trial included around a thousand older adults aged 70 to 84, all with untreated hearing loss. The participants were randomly assigned either to a hearing intervention (hearing aids and counselling) or to a structured health education programme.

The main result of the trial was somewhat disappointing. Across the whole group, the hearing intervention didn’t slow down cognitive decline. However, part of the trial was made up of people who were recruited through a long-running heart study, who had started out older and at higher risk and were declining faster. In that group, hearing treatment slowed cognitive decline by about 48%. A likely reason is that the heart study participants were already declining faster cognitively, giving a better chance of detecting a benefit over three years. The large effect in this subgroup might also reflect hearing aids doing more than restoring hearing, perhaps supporting social contact, mood and mental stimulation, though that hasn't been shown directly.

This suggests that hearing aids may not have an impact on cognitive outcomes for everyone with a bit of hearing loss, but that they are likely the most beneficial for the people who are at an increased risk of dementia.

Summary

Hearing loss is common. It is one of the more strongly linked risk factors for dementia that we know of. The risk climbs with hearing loss severity and there are plausible ways in which it could increase risk. At the same time, we don’t definitely know it causes dementia.

However, hearing loss is one of the few risk factors that can often be identified and treated relatively easily. Whether hearing aids directly reduce dementia risk remains an active area of research. But protecting your hearing, getting it checked and treating hearing loss when it occurs are still very important.

In practical terms, if you, or the people around you, notice your hearing worsening, get it checked - the tests are quick and often free. The earlier hearing loss is caught, the more there is to work with. And where you can, protect your hearing from loud noise. These are low-risk interventions that can keep us in the conversation for longer, with a growing body of evidence suggesting that they likely benefit the brain as well.

References and further reading

1. Lin, F. R. et al. Hearing loss and incident dementia. Arch. Neurol. 68, 214–220 (2011).

2. Stevenson, J. S., Clifton, L., Kuźma, E. & Littlejohns, T. J. Speech-in-noise hearing impairment is associated with an increased risk of incident dementia in 82,039 UK Biobank participants. Alzheimers Dement. J. Alzheimers Assoc. 18, 445–456 (2022).

3. Cantuaria, M. L. et al. Hearing Loss, Hearing Aid Use, and Risk of Dementia in Older Adults. JAMA Otolaryngol.-- Head Neck Surg. 150, 157–164 (2024).

4. Yu, R.-C. et al. Adult-onset hearing loss and incident cognitive impairment and dementia - A systematic review and meta-analysis of cohort studies. Ageing Res. Rev. 98, 102346 (2024).

5. Kolo, F. B. et al. Hearing Loss, Brain Structure, Cognition, and Dementia Risk in the Framingham Heart Study. JAMA Netw. Open 8, e2539209 (2025).

6. Lin, F. R. et al. Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE): a multicentre, randomised controlled trial. Lancet 402, 786–797 (2023).

7. Yeo, B. S. Y. et al. Association of Hearing Aids and Cochlear Implants With Cognitive Decline and Dementia: A Systematic Review and Meta-analysis. JAMA Neurol. 80, 134–141 (2023).

8. Lin, F. R. et al. Association of Hearing Impairment with Brain Volume Changes in Older Adults. NeuroImage 90, 84–92 (2014).

A couple of weeks ago I shared a note on Substack which read:

And one reader replied with this quote attributed to physicist John Wheeler: “As my island of knowledge grows, so does the shore of my ignorance.”

Honestly, it captures something that has surprised me throughout the years. When I started studying dementia I didn't expect how often the most honest answer would be "we don't know yet." Each answer seem to uncover three new questions. In fact, I think most of my published papers have generated more questions than answers.

Often, when you’re new to a topic, things look simple. As you learn more, you start seeing the limitations, unanswered questions, caveats and competing explanations. And you get more careful about certainty, because you start to see how rare it actually is. For Alzheimer’s disease, we understand some risk factors quite well, and we know that there are certain biological changes happening years before cognitive symptoms. But there are still enormous gaps in our understanding. And the more papers I've read, the harder it has become to tell neat stories.

And ironically, this might be part of why people sometimes struggle to trust experts. The more you know about something, the less certain you sound, because you can see all the exceptions and everything we still don't know. I notice it in myself all the time. Reality is messy. But simple, clean explanations are more satisfying. And there are reasons for that - and I'm not immune to them either.

We are much more likely to trust and believe whatever fits with what we already believe to be true and we cross-examine whatever doesn’t. Confirmation bias is basically the tendency to search out and trust information that supports what we already believe. A related tendency is directionally-motivated reasoning, which is less about what we already believe and more about what we want to be true. For example, I have a massive sweet tooth, so I'd be more likely to see a headline saying "sugar is beneficial" and think "great!" before I've spent even a second questioning it.

And the more often we hear something, the truer it can start to feel, whether or not it actually is. This is called the illusory truth effect. Repetition makes a claim easier to process and we can mistake that for accuracy. And this can happen even when we know better. One study from 2015 showed that repeating false information made people more likely to believe it, even when it contradicted facts that they already knew to be true.

Of course, uncertainty can also be really uncomfortable. It can be easier to believe the first confident answer available, a process called the need for closure, because we would rather have something deifinitive than sit with not knowing. A confident claim gives us that. “Vitamin D deficiency causes dementia” is simple, actionable and gives us something to do. But the honest answer is much less satisfying: vitamin D may matter, but dementia risk is shaped by genetics, vascular health, education, social factors, sleep, exercise, mental health and dozens of other interacting factors.

We also like the feeling that we can influence outcomes. This is called the illusion of control, the sense that we can sway outcomes more than we really do. A supplement or a morning routine offers something that uncertainty can’t. I suspect that a lot of health advice sells that feeling of control more than it sells any real effect, especially if it’s something that can be taken within seconds. Sometimes we're not buying the intervention itself. We're buying the feeling that we've tilted the odds slightly in our favour.

And emotion can amplify all of it. When a claim scares or reassures us, it can become harder to evaluate it objectively.

Personally, the more I keep learning, the less I think expertise is about having all the answers. It’s largely about becoming comfortable with uncertainty. And that may also be why experts can sometimes be difficult to trust. The people who know the most about a subject are often the least willing to give a simple answer. And tidy, simple answers are exactly what our brains are looking for…

This doesn’t mean that science is too complicated to be useful. It just means that the reality is usually more nuanced than the headlines or bold claims suggest.

We don’t have to choose between absolute certainty and knowing nothing. Most of science is in the middle. We accumulate evidence, reduce uncertainty and gradually get closer to clearer answers. The shoreline keeps growing, but so does the island.

A few months ago I wrote about the practical side of all of this, a toolkit if you will, on how to check a claim before believing it. Today’s post was more on the reasons why it might be needed.

What's a tidy answer you once wanted to be true?

-Eleni

Vitamin D is a fat-soluble vitamin, also known as the “sunshine vitamin” because our skin can produce it when exposed to sunlight. We can also obtain vitamin D through certain foods (like fatty fish, egg yolks and fortified foods), as well as through supplements.

Vitamin D is thought to have neuroprotective properties. It acts as a neurosteroid and plays important roles in many processes throughout the body, including brain function, mood, calcium homeostasis, bone health and immune regulation. It's also involved in regulating many genes that are important for brain function.

Vitamin D levels vary a lot between individuals and depend on things like where they live and how much sun they see, their age and their dieatary intake. It has been estimated that between 30% and 50% of people have low vitamin D levels.

You might have seen scary headlines like:

“Low vitamin D linked to dementia.”

“Deficiency doubles your risk of Alzheimer’s.”

These headlines are based on a large body of research, and the pattern is pretty consistent.

So let’s see what the evidence behind these headlines means.

There is a large body of longitudinal observational research looking at the association between vitamin D levels and cognitive function/dementia. In these studies, researchers follow large groups of people for years, measure their vitamin D at the start and then track them to see who develops dementia. There have been dozens of these studies, in different countries, with different populations. When the results are pooled together in meta-analyses, the pattern is fairly consistent. People who started out vitamin D deficient were more likely to develop dementia later on (around 40% more likely). Also, people with severe vitamin D deficiency have been shown to have more than double the risk of developing dementia.

That seems like a clear story: low vitamin D is associated with higher risk of dementia. So should we just top it up and protect the brain?

But when we look at studies designed to test that idea directly, things get less clear.

The rest of this post is for paid subscribers.

Below is the part of this post that discusses why the trials and observational studies disagree, and the things worth knowing and checking before you start taking vitamin D supplements.

Read more

1, 2, 3…5, 6, 7. Oh no, this song is faster than what I’m used to, I better keep up. My dance partner is raising my hand, that means I need to turn. I’m turning, one time, two times… how many more? This place is packed, I need to keep an eye on everyone dancing around me so I don’t bump into them. Wait, do I move my left or right leg now? My arms are tense again. Relax, relax, like my teacher keeps telling me.

Those are all real thoughts that are going through my mind when I’m dancing salsa. Simultaneoulsy. That’s a lot for the brain to do at once, on top of actually moving. So it's not surprising that dance keeps showing up as beneficial in studies of cognitive aging.

The evidence on dancing and cognition

Both observational and intervention studies have looked at dance in older adults. For example, a study from 2003 published in the New England Journal of Medicine followed nearly 500 adults over the age of 75 for five years. The study asked about participants’ leisure activities at baseline and tracked who went on to develop dementia. The findings showed that frequently engaging in activities like reading, board games and playing musical instruments was associated with lower dementia risk. Of all the physical activities they measured (cycling, swimming, walking, group exercise) only dancing (frequently versus rarely) was linked to a lower dementia risk. The hazard ratio for frequent dancers was 0.24, meaning that they had a 76% lower risk of developing dementia compared to those who reported dancing rarely.

Several meta-analyses have also examined whether dance interventions improve cognition in older adults. A very recent 2025 paper in Medicine & Science in Sports and Exercise combined 13 randomized controlled trials of older adults. They found that individuals who took part in a dance intervention showed improvements in cognition and depression levels. Similarly, an earlier 2021 review in Age and Ageing of older adults found that dance improved global cognition and executive function. Additionally, a 2023 review in BMC Geriatrics looked specifically at older adults with mild cognitive impairment, and found that dance therapy helped not just with global cognition but also with memory, attention, executive function and mental health.

So what is dance doing to the brain?

Dance engages many systems at once

Dance engages many brain systems at the same time. In a study from 2006, amateur tango dancers were scanned with PET while stepping to a beat. The study found that regions that are involved in timing, motor coordination and spatial processing were activated in participants during dancing, including parts of the cerebellum, basal ganglia and parietal cortex. But dancing may also require other things not measured in this study, including memory for sequences, coordination with a dancing partner and awareness of other people in a crowded room, all of which involve other brain systems.

Most forms of exercise challenge one or two systems at a time. Dance may uniquely challenge the brain because it involves multiple cognitive, motor, sensory and social demands simultaneously. Specifically, these can include aerobic exertion, coordination, rhythm, memory, prediction, error correction and often social interaction.

Dance is aerobic exercise but also more than that

Aerobic exercise alone is known to benefit the brain. A 2011 randomized trial of 120 older adults found that one year of aerobic training (compared to stretching) increased anterior hippocampus volume, reversing about one to two years of age-related volume loss and also improved spatial memory. These changes were linked to higher brain-derived neurotrophic factor (BDNF) levels, a protein that is involved in synaptic plasticity and neuronal health.

But dancing is more than just an aerobic exercise. A 2018 trial then compared dance directly with conventional fitness training matched for intensity. The dance group showed structural changes across a broader set of brain regions (the cingulate cortex, insula, corpus callosum, and sensorimotor cortex.) And only the dance group showed an increase in BDNF levels. Both groups improved similarly on the cognitive tests. So the dance advantage in this study was at the structural and BDNF level, which the authors suggested may translate into cognitive benefits later on. Notably, the dance protocol used constantly changing choreographies. So as novelty and motor skill learning are themselves associated with structural plasticity, they may be part of what makes dance more than just aerobic exercise.

The social part of dancing

The social aspect probably also matters. When you’re dancing with someone, you’re constantly reading the other person’s movements, predicting what they’ll do next, adjusting to what they actually do, and all of this while staying or trying to stay on the beat. In leader and follower styles like salsa, tango or swing, this becomes a continuous nonverbal exchange. One partner signals through pressure and movement, and the other one reads and responds in real time. Unlike running on a treadmill or even a solo dance routine, partnered dance is something that cannot be done on autopilot. The next step depends on what someone else just did. It's difficult to isolate the social component experimentally, but it very likely matters. And whether we are talking about partnered dance or dancing in a group, the social engagement in itself is likely protective, as it has been associated with a lower risk of cognitive decline.

In summary

There are a few caveats worth mentioning regarding these studies. The first one is that sample sizes in this research area are still pretty small. And the second one is self-selection. People who sign up for dance classes are probably already healthier than people who don’t, meaning that some of the observed benefits may partly reflect pre-existing differences between people who choose to dance and people who do not.

That being said, dance combines aerobic fitness, motor coordination, sequence learning, rhythm, and when partnered, social coordination, all within the same activity. The studies that compare it with conventional exercise consistently find it does as well or better. Which part is doing the most work, though, is still a bit of an open question. The aerobic component? The novelty of learning new sequences? The partner? Or maybe just the fact that all of it is happening at once. The brain rarely has to do so many things at once, which may be exactly the point.

Do you dance? What does it do for you?

References and further reading

1. Verghese, J. et al. Leisure Activities and the Risk of Dementia in the Elderly. N. Engl. J. Med. 348, 2508–2516 (2003).

2. Hewston, P. et al. Effects of dance on cognitive function in older adults: a systematic review and meta-analysis. Age Ageing 50, 1084–1092 (2021).

3. Jaldin, M. A. et al. Systematic Review and Meta-Analysis of the Effects of Dance on Cognition and Depression in Healthy Older Adults. Med. Sci. Sports Exerc. 57, 490–500 (2025).

4. Brown, S., Martinez, M. J. & Parsons, L. M. The neural basis of human dance. Cereb. Cortex 16, 1157–1167 (2006).

5. Erickson, K. I. et al. Exercise training increases size of hippocampus and improves memory. Proc. Natl. Acad. Sci. U. S. A. 108, 3017–3022 (2011).

6. Rehfeld, K. et al. Dance training is superior to repetitive physical exercise in inducing brain plasticity in the elderly. PLOS ONE 13, e0196636 (2018).

And if you're new to Alzheimer's basics like what amyloid and tau mean, my earlier post is a good starting point:

As you may know, around two-thirds of people living with Alzheimer’s are women. Another way to look at it is through lifetime risk. A woman who is 45 years old has roughly a one in five chance of developing Alzheimer’s dementia during her lifetime, compared to a man at the same age whose lifetime risk is about one in ten.

At first glance, the numbers suggest that women are simply at much higher risk of the disease itself. But it's more complicated.

For a long time, the explanation for these statistics seemed to be that it is because women live longer. As age is the biggest risk factor for dementia, it would make sense that there are more women with Alzheimer’s. And yes, the longevity answer may be partly true, but it isn’t the whole story.

Why isn’t the longevity answer the whole story?

Prevalence and incidence are two words that are easy to confuse, but they mean different things.

Prevalence means how many people currently have a disease. So when we say that roughly two-thirds of Alzheimer’s cases are women, we are talking about prevalence.

Incidence, on the other hand, means how many new cases develop over a period of time. Seeing more new Alzheimer's diagnoses in women might suggest that women are biologically at higher risk. But age complicates the picture. Alzheimer’s becomes much more common in the oldest age groups and women are more likely to survive into those ages. So if more women develop Alzheimer’s overall, part of the reason is simply that more women live long enough to develop it.

To get closer to the real question, we need to compare women and men at the same age. For example, does a 75 year old woman develop Alzheimer’s at a higher rate than a man of the same age?

And the answer to this is less clear. Most studies find little consistent sex difference in age-specific incidence, although some studies report a female excess at the oldest ages. But even these comparisons are imperfect, because men who survive into very old age may represent a healthier subset of people overall. This effect is called selective survival.

So longevity explains part of why more women are living with Alzheimer’s. Living longer adds years of exposure, but it doesn’t change what’s happening at any given age. And whether or not the incidence numbers show a clean sex difference, the underlying biology often does. For example, women show greater tau burden, a stronger APOE4 effect and different trajectories of cognitive decline - things I will discuss next.

Sex-specific differences in genetics and pathology

APOE4 is the strongest common genetic risk factor for late-onset Alzheimer’s, and it does not seem to have equal effects by sex. A large meta-analysis with nearly sixty thousand people found that among carriers of a single copy of APOE4, women had a substantially higher risk of developing Alzheimer’s than men between the years of 65 and 75.

Amyloid (the protein that is believed to accumulate first in Alzheimer’s) shows little consistent sex difference. However, tau, which correlates more closely with actual cognitive decline, does. Women tend to show higher tau burden, particularly among APOE4 carriers and people who are already amyloid-positive. So the genetic risk and the pathology line up, with women carrying more of both.

Cognitive resilience advantage in women

You would expect more tau and a risk gene that confers greater risk to show up as women showing worse cognition and earlier. But that’s not the case. At similar levels of brain pathology, women tend to perform better than men, especially on verbal and episodic memory. Most clearly this effect is seen in the earlier stages, with the advantage narrowing as pathology becomes more severe.

And once women cross into clinical territory (mild cognitive impairment, then dementia) they decline faster than men. The advantage doesn’t protect them so much as delay the point at which the decline becomes visible, and then the decline is steep.

Two things, that are not mutually exclusive, may drive this. One explanation is resilience that holds against pathology up to a threshold and then gives way, after which there is more to “lose”. The other explanation is a measurement problem. Standard memory tests were not built with women’s higher baseline in mind, so a woman can be developing the disease and still show to be in the “normal” range. Studies re-scoring the same people using cut-offs set separately for women show that roughly one in ten women who had been classed as healthy were actually impaired. So some women may not really be declining faster, but they are being caught later, at a more advanced stage.

Hormones

Estrogen appears to have multiple neuroprotective and vascular effects, and the menopausal transition is considered a plausible period of increased vulnerability.

What is not so clear is whether doing anything about it helps. The obvious idea of replacing estrogen to protect the brain has not really been supported. The largest and most rigorous recent synthesis, a 2025 review in Lancet Healthy Longevity, found no good evidence that menopause hormone therapy either raises or lowers dementia risk, and no evidence supporting the popular "critical window" idea where getting the timing right is the solution. So the situation in 2026 is that we have biological evidence but no clinical evidence (yet). We need more lognitudinal and high-quality research on it.

Subscribe now

The vascular part

This mechanism is important, because we can actually do something about it.

Around four in five people with Alzheimer’s dementia also have some degree of cerebrovascular disease (small-vessel damage, white-matter injury or silent strokes that lower the threshold at which pathology turns into symptoms). And the vascular part is also gendered, partly connecting back to the hormones. Through most of adult life, estrogen gives women a vascular advantage. The steep estrogen drop across menopause is associated with that advantage eroding, and women’s vascular risk shifts upward relative to men’s after the transition. White-matter damage is more common in women, particularly post-menopause. And lifetime stroke risk is higher in women, and women tend to have worse functional outcomes after stroke. So menopause is not only a hormonal event for the brain, but also a vascular one.

Of course, we cannot choose our genes or chromosomes. But modifiable factors like vascular health, physical activity and the rest, are estimated to account for something like 40% of dementia cases, and these are distributed unequally by gender too. Women, for example, are less physically active than men across their lifespan. The encouraging part is that the lever works for both men and women. A major multidomain prevention trial that put older adults through a structured program of diet, exercise, cognitive training and vascular risk management, and found that the cognitive benefit did not differ significantly between women and men, despite their different starting risks.

Sex vs gender

And finally, some of this isn't about biology at all.

Education is one of the strongest protective factors against dementia that we know of. It builds what the field calls cognitive reserve. Put simply, it acts as a buffer that lets the brain “absorb damage” before it shows up as symptoms. Historically, women have had less access to education. The generations now in their 70s and 80s (the people in a lot of the older studies on dementia) had fewer years of schooling, less complex work and fewer of the things that build that buffer. As expected, where gender gaps in education have narrowed across birth cohorts, some of the sex difference in dementia risk appears to narrow with them.

And of course, this doesn’t stop at education. Women carry more of the caregiving burden and they have higher rates (or reports) of chronic stress, depression and anxiety. Each of these is independently linked to a higher dementia risk. And these are not cleanly separable from biology. For example, chronic stress has physiological consequences.

So why do women get more Alzheimer’s?

There isn't one single reason.

Women live longer, which increases prevalence.

Women may be more vulnerable to the effects of APOE4 and to developing tau pathology.

They tend to maintain cognitive performance longer despite pathology, but then decline more steeply later.

Hormonal changes around menopause probably matter, but we don't yet know what to do about it.

And some of the gap is less education, more caregiving and more chronic stress among women.

For decades, sex was treated in this research as a nuisance variable. Something that we should statistically remove so we could get at the “real” finding. Fortunately, the field is slowly but surely moving away from this idea. In a disease that affects women roughly twice as often, sex and gender are not just noise that should be removed from the story. They are a large part of it.

I hope you enjoyed this overview of sex and gender differences in Alzheimer's disease! If you would like me to cover any part in more detail, let me know.

And if you are looking for more science-filled information on this topic, I recommend this review that I co-authored, published in Alzheimer's and Dementia journal.

-Eleni

Subscribe now

References and further reading

1. Arenaza-Urquijo, E. M. et al. Sex and gender differences in cognitive resilience to aging and Alzheimer’s disease. Alzheimers Dement. J. Alzheimers Assoc. 20, 5695–5719 (2024).

2. Chêne, G. et al. Gender and incidence of dementia in the Framingham Heart Study from mid-adult life. Alzheimers Dement. J. Alzheimers Assoc. 11, 310–320 (2015).

3. Neu, S. C. et al. Apolipoprotein E Genotype and Sex Risk Factors for Alzheimer Disease: A Meta-analysis. JAMA Neurol. 74, 1178–1189 (2017).

4. Hohman TJ, Dumitrescu L, Barnes LL, et al. Sex-specific association of apolipoprotein E with cerebrospinal fluid levels of tau. JAMA Neurology. 2018;75(8):989–998.

5. Buckley, R. F. et al. Sex, amyloid, and APOE ε4 and risk of cognitive decline in preclinical Alzheimer’s disease: Findings from three well-characterized cohorts. Alzheimers Dement. J. Alzheimers Assoc. 14, 1193–1203 (2018).

6. Sundermann, E. E. et al. Female advantage in verbal memory: Evidence of sex-specific cognitive reserve. Neurology 87, 1916–1924 (2016).

7. Sundermann, E. E. et al. Sex-specific norms for verbal memory tests may improve diagnostic accuracy of amnestic MCI. Neurology 93, e1881–e1889 (2019).

8. Melville, M. et al. Menopause hormone therapy and risk of mild cognitive impairment or dementia: a systematic review and meta-analysis. Lancet Healthy Longev. 6, 100803 (2025).

9. Sindi, S. et al. Sex differences in dementia and response to a lifestyle intervention: Evidence from Nordic population-based studies and a prevention trial. Alzheimers Dement. J. Alzheimers Assoc. 17, 1166–1178 (2021).

10. Mielke, M. M. et al. Consideration of sex and gender in Alzheimer’s disease and related disorders from a global perspective. Alzheimers Dement. J. Alzheimers Assoc. 18, 2707–2724 (2022).

Being rejected is an inevitable part of life. Whether it's a text message that never gets answered, a failed job application or a rejected manuscript, most of us know the feeling. It feels uncomfortable and can make us question our value and worth. But why does rejection feel so bad and could there be something good to it?

As I have mentioned here before, I am a cognitive neuroscientist and a full-time researcher. I sometimes think that at least in my professional life, I have gotten very good at being rejected. As a researcher, you often get a “Thank you, but…” email monthly, and sometimes even weekly. Whether that’s about a paper that was submitted, a grant application or a collaboration request.

I started thinking about the psychology and neuroscience of rejection last week. I had gotten an email I had been hoping for, for a very long time. I had started working on a project in the second year of my PhD. That was four years ago now. As usual, the scientific project began with an idea, that then developed into hypotheses, then came the analyses (refined many times over), and eventually the findings were written up in the form of a manuscript ready to be sent to a journal. Sounds simple, right?

What many people outside academia might not know is that papers can take years to publish. This one took four. Between analyses, rewrites, feedback and multiple failed submissions, the project kept dragging on.

I had gotten the dreaded rejection email for this paper from four or five journals before it was finally accepted. Some papers are just more difficult to get published than others. It doesn’t necessarily mean that the research is bad. In my case, it was probably a combination of a relatively small sample size, submitting to journals where the paper didn’t quite fit their scope and the fact that the findings were somewhat original and unconventional.

I know I started this post by saying that I had gotten good at rejection. But when I read that acceptance email, I realised that I had been lying to myself a little. The four or five rejections along the way had stung a bit. Which got me thinking. What is going on in the brain when we get rejected?

The main explanation is evolutionary. For our ancestors, being part of a group was necessary. It often meant the difference between surviving and not. Getting socially excluded was a survival threat and the brain seems to have evolved to still somewhat treat it as one. Which may also be why a rejection email today can still feel much worse than it logically should.

One well-known study from 2003, published in Science, put people in a brain scanner and had them play a virtual ball-tossing game called Cyberball. After a few rounds, the other two “players” (that were actually controlled by the computer) stopped throwing the ball to the participant, essentially excluding the participant. Interestingly, they found similar results to studies looking at physical pain. The brain region that got activated during exclusion, the anterior cingulate cortex, is often asscociated with processing physical pain. Brain activation in this region also correlated with self-reported distress.

Another 2010 study published in Psychological Science had people take acetaminophen (paracetamol) every day for three weeks and then play Cyberball in the scanner. The acetaminophen group reported less social pain and showed less activity in those same pain-related brain regions. A common painkiller, appeared to dampen hurt feelings. Wild!

But to be fair, the neuroscience has been refined since. Newer meta-analyses suggest that physical and social pain don’t share the exact same neural code, even if they activate similar regions. This suggests that the neural correlates of social pain are more complex than we previously thought. More recent studies suggest that the brain may also treat rejection as a learning signal. A 2024 study in PNAS suggested that when we get rejected (socially), the brain is updating its sense of who values us and which relationships are worth investing in (relational value). And the interesting part is that the brain regions that were long called "social pain regions" appear to be tracking these updates to relational value. So this suggests that these regions may be involved less in registering “hurt” itself and more in updating how much we think others value us.

Most of this neuroscience of rejection is about social rejection, like being excluded or ignored by friends and peers. My example of a journal rejecting a paper, is not really that. The editor doesn’t know me and there is no group I am being kicked out of. But perhaps the perception of being valued less than we had hoped, whether the rejection comes from a person, a hiring panel or an editor, triggers a similar response.

Research also points towards ways to deal with rejection. Cognitive reappraisal, which basically means changing how you think about something to change how it makes you feel, engages parts of the prefrontal cortex involved in emotion regulation and has been shown to reduce negative emotional responses. And self-affirmation has research behind it too. Briefly reflecting on something you care about, like a piece of work you are proud of, has been shown to reduce cortisol and other stress responses to social-evaluative situations.

For me, the most useful question after a rejection has become “What is this telling me?”. For example, “Is this a good fit?” and “How can I improve the work?”, instead of “Am I good enough?”. And cliché or not, I have also come round to the idea that rejection is redirection.

Getting rejected time and time again after spending years on something can definitely feel overwhelming and discouraging, I’m not going to lie. But I don’t think that the goal is to stop rejection from hurting. Maybe the healthier goal is learning not to confuse rejection with a verdict about our worth.

Rejection hurts, but it can also be informative. Sometimes it may be a sign to improve. Other times it may be a sign of poor fit or just bad timing. But it probably isn’t as absolute as it feels in the moment.

- Eleni

Subscribe now

References and further reading

1. Williams, K. D. Ostracism. Annu. Rev. Psychol. 58, 425–452 (2007).

2. Eisenberger, N. I., Lieberman, M. D. & Williams, K. D. Does Rejection Hurt? An fMRI Study of Social Exclusion. Science 302, 290–292 (2003).

3. Dewall, C. N. et al. Acetaminophen reduces social pain: behavioral and neural evidence. Psychol. Sci. 21, 931–937 (2010).

4. Cacioppo, S. et al. A quantitative meta-analysis of functional imaging studies of social rejection. Sci. Rep. 3, 2027 (2013).

5. Babür, B. G., Leong, Y. C., Pan, C. X. & Hackel, L. M. Neural responses to social rejection reflect dissociable learning about relational value and reward. Proc. Natl. Acad. Sci. 121, e2400022121 (2024).

6. Buhle, J. T. et al. Cognitive reappraisal of emotion: a meta-analysis of human neuroimaging studies. Cereb. Cortex 24, 2981–2990 (2014).

7. Creswell, J. D. et al. Affirmation of personal values buffers neuroendocrine and psychological stress responses. Psychol. Sci. 16, 846–851 (2005).

One day last week, I was trying to finish revisions on a manuscript and my brain just wasn’t cooperating. I couldn’t focus and I felt tired. I’d already had three coffees and I considered getting a fourth (which, if I’m honest, I knew was a bad idea). So instead, I went for a walk. When I returned to my desk, I felt considerably more refreshed and my head was clearer. That raised a question: what works when it comes to short-term cognitive performance?

In a recent post on nicotine, I argued that nicotine is not the clean cognitive enhancer the internet has sometimes made it out to be. So what does work, then?

I’ll focus on what improves short-term cognitive performance and what doesn’t, despite the attention it gets.

By short-term cognitive performance, I mean the way the brain functions in the day to day. Things like your attention, processing speed, working memory and executive function (planning, organization, emotional control). In this post, I am not talking about long-term protection against dementia and broader brain health (although there is overlap in factors that affect both). I wrote about these longer-term factors here.

Here's what the evidence supports for short-term cognition and what it doesn't:

1. Sleep

Sleep is one of the biggest determinants of short-term cognitive performance.

A 2010 meta-analysis published in Psychological Bulletin, looked at 70 studies across 147 cognitive tests. It was found that short-term total sleep deprivation produces large impairments in simple attention. Effects were also found on complex attention, working memory and processing speed. A similar pattern has been shown repeatedly in studies since, including in a 2024 meta-analysis in Sleep Medicine Reviews. The study showed that even a single night of restricted sleep can increase subjective sleepiness and impair sustained attention, that is a cognitive function crucial for everyday tasks, like driving.

What this means in practice is that one of the most effective short-term cognitive intervention available to most people is sleeping a little bit more. Especially for people who are chronically getting less than seven hours.

Naps also deserve a brief mention. Short naps of around 10–20 minutes, taken in the early afternoon, can improve alertness and performance on attention tasks for several hours afterwards. Longer naps, instead, risk sleep inertia (the groggy feeling after waking) and they can interfere with night sleep.

You could say that sleep is the foundation. But most of us probably already know that. So what else helps?

Read more

Most of us have some memories that feel very vivid. We remember where we were when we heard about 9/11. Or the moment our parents told us someone had died. Or a phone call about a job or a diagnosis.

For me, one of those moments was hearing about the 2004 tsunami. I can still picture the room, who I was with, what I was doing and even what was on the TV.

These types of memories can feel almost photographic. But whether the details we remember are actually accurate is a different question.

A name for the phenomenon

In 1977, Brown and Kulik coined the term flashbulb memory to describe people’s recollections of where they were when they heard about surprising and emotionally significant public events such as the Kennedy assassination. These kinds of memories seemed to preserve the surrounding context (where you were, who you were with and what you were doing). They proposed that such events trigger a special encoding mechanism, sometimes called the “now print” mechanism, that captures the moment in unusual detail.

The vividness part of their proposal turned out to be well supported. But the idea that flashbulb memories rely on a distinct, special mechanism has not been supported by subsequent research.

What the research shows about flashbulb memories

In 1986, the day after the Challenger space shuttle exploded, psychologist Neisser interviewed studens asking them where they were, who they were with and what they were doing when they heard the news. Almost three years later, he asked the students again.

Surprisingly, the answers often did not match their original reports. After nearly three years, many participants gave accounts that were inconsistent with what they had written the day after the event, and sometimes in substantial ways. What stayed high, however, was their confidence: participants remained strongly convinced that their later memories were correct, even when those memories differed from their earlier reports.

Another direct comparison study came from Talarico and Rubin, who tested students on their memories of 9/11 and on a recent everyday event from the same week. Across follow-ups at one, six and 32 weeks, both kinds of memories declined in consistency at similar rates over time. What differed was, again, the students’ subjective experience. Ratings of vividness, recollection and belief in accuracy declined for everyday memories but stayed high for 9/11 memories. So flashbulb memories did not appear special in their actual accuracy, as previously claimed, but only in their perceived accuracy.

More evidence came from a study that followed more than 3,000 people across seven US cities after the September 11 attacks. Participants reported their memories one week, one year and three years after the event. Forgetting was rapid in the first year and then slowed. Importantly, stability of memories did not reflect greater accuracy. Confidence in accuracy stayed high, even as the details drifted from what people had originally reported.

The current consensus is that flashbulb memories do not require a special mechanism. They share most properties with ordinary autobiographical memory, including susceptibility to forgetting and distortion. What is distinctive about them is that they often feel accurate.

Subscribe now

What’s happening in the brain

The subjective vividness of these memories has a physiological basis, even when the factual accuracy doesn’t.

When something emotionally arousing happens, emotional arousal engages noradrenergic systems and stress hormones such as cortisol. Decades of work have shown that these systems engage the amygdala, particularly the basolateral region, which in turn modulates how the hippocampus and surrounding cortex consolidate the memory. Emotional events tend to be prioritised during consolidation relative to neutral ones. This is why a stranger’s face from a random Tuesday often fades, while a stranger’s face from a frightening moment can remain more memorable over time.

A 2007 fMRI study provided evidence consistent with this circuit in people recalling 9/11. Participants who had been in downtown Manhattan that day showed greater amygdala activation when remembering the attacks compared to a neutral event from the same period. Participants who had been in midtown, only a few miles further away, did not show the same pattern. The neural response tracked personal proximity to the event more than the event itself. So close personal experience seems to engage the emotional memory system more strongly, producing more vivid recollections.

But amygdala-mediated modulation of consolidation seems to enhance some aspects of memory more than others. Emotional arousal tends to sharpen memory for the central, emotionally charged details of an event while leaving peripheral details more vulnerable to distortion or loss. The sense of being there and the vividness of the recollection can be heightened, but the accuracy of surrounding details is often not.

The role of social context

There is also a very important social piece. Flashbulb memories are often replayed in media, talked about, retold and rehearsed far more than ordinary memories. They become part of how groups remember themselves. Each retelling is itself a memory event, shaped by the audience, the conversational context and the versions of the story currently in circulation. In large longitudinal studies of memory for the September 11 attacks, people’s recollections became more similar over time, consistent with the idea that shared media exposure and discussion gradually align how events are remembered. Whether rehearsal stabilises or distorts a memory depends on the study and the time scale, but it almost certainly strengthens confidence.

What to take from this

Flashbulb memories are not as Brown and Kulik first suggested, photographic snapshots stored by a dedicated mechanism. They are vivid, emotionally consolidated, repeatedly rehearsed, socially reinforced autobiographical memories. And this is enough to make them feel categorically different, even though, on most measures of accuracy, they aren’t.

The implication is that the memories we feel most certain about are not necessarily the ones most worth trusting. Vividness and confidence track emotional salience and rehearsal more reliably than they track accuracy. The brain produces the sense of certainty, but that doesn’t mean that the memory itself is accurate.

This is important because we often rely on our most vivid memories to make sense of the past. But those memories are often precisely the ones most likely to have been reshaped.

References & further reading

1. Brown, R. & Kulik, J. Flashbulb memories. Cognition 5, 73–99 (1977).

2. Neisser, U. & Harsch, N. Phantom flashbulbs: False recollections of hearing the news about Challenger. in Affect and accuracy in recall: Studies of ‘flashbulb’ memories 9–31 (Cambridge University Press, New York, NY, US, 1992). doi:10.1017/CBO9780511664069.003.

3. Talarico, J. M. & Rubin, D. C. Confidence, not consistency, characterizes flashbulb memories. Psychol. Sci. 14, 455–461 (2003).

4. Hirst, W. et al. Long-term memory for the terrorist attack of September 11: Flashbulb memories, event memories, and the factors that influence their retention. J. Exp. Psychol. Gen. 138, 161–176 (2009).

5. Hirst, W. & Phelps, E. A. Flashbulb memories. Curr. Dir. Psychol. Sci. 25, 36–41 (2016).

6. Sharot, T., Martorella, E. A., Delgado, M. R. & Phelps, E. A. How personal experience modulates the neural circuitry of memories of September 11. Proc. Natl. Acad. Sci. 104, 389–394 (2007).

7. Kensinger, E. A. Remembering the Details: Effects of Emotion. Emot. Rev. J. Int. Soc. Res. Emot. 1, 99–113 (2009).

8. McGaugh, J. L. Making lasting memories: remembering the significant. Proc. Natl. Acad. Sci. U. S. A. 110 Suppl 2, 10402–10407 (2013).

9. Hirst, W. et al. A ten-year follow-up of a study of memory for the attack of September 11, 2001: Flashbulb memories and memories for flashbulb events. J. Exp. Psychol. Gen. 144, 604–623 (2015).

I think everyone would agree that cigarette smoking is harmful. But somewhere in the last couple of years, the conversation around nicotine has shifted.

The social media posts that I keep seeing are telling me that nicotine is actually a nootropic, good for focus and fine if you’re careful about the delivery system.

It’s a compelling idea, but not as straightforward as it sounds.

So let’s unpack what the evidence actually says.

How nicotine affects the brain

To understand why nicotine is framed as a cognitive enhancer, it helps to start with how it acts in the brain.

Nicotine binds to a family of receptors called nicotinic acetylcholine receptors, or nAChRs. Acetylcholine is a key neurotransmitter involved in attention, memory and arousal. Nicotine acts as an agonist at these receptors, mimicking some of acetylcholine’s effects. This is true regardless of how nicotine is delivered. Whether it comes from cigarettes, vapes, pouches or patches, it acts on the same receptor systems in the brain.

Two subtypes are especially important for understanding nicotine’s effects:

α4β2 is one of the most abundant high-affinity subtypes in the thalamus and cortex and is closely tied to reward-related circuitry. When nicotine activates these receptors, it increases dopaminergic firing in the ventral tegmental area, leading to dopamine release in downstream regions like the nucleus accumbens. This is the same circuit engaged by food, sex and most drugs of abuse, which helps explain nicotine’s addictive potential. It also contributes to nicotine’s short-term effects on focus, as α4β2 activation can transiently sharpen attention and processing speed. In part, this reflects an improvement in signal-to-noise in attention networks, making relevant inputs stand out more clearly.

α7 is widely expressed in the hippocampus and cortex and is also found on immune cells like microglia. It’s involved in memory and synaptic plasticity, and in what’s known as the cholinergic anti-inflammatory pathway. This is the aspect of nicotine that has been explored in preclinical and early-stage research on neuroprotection.

But two things about these receptors are important to note:

First, chronic nicotine exposure upregulates them. Smokers have more α4β2 receptors on their neurons than non-smokers, not fewer. The brain builds more of them in response to persistent stimulation. This contributes to dependence and makes quitting more difficult.

And second, nAChRs desensitise quickly. Within minutes, they become less responsive. So with chronic nicotine use, you end up with more receptors, but a system that is not fully responsive. Which is part of why it has been hard to translate this into anything therapeutic.

The “nicotine as nootropic” claim

One of the foundational citations for nicotine improving cognition is a meta-analysis published in Psychopharmacology in 2010. This included 41 placebo-controlled studies showing small improvements in attention, motor performance and short-term memory. This study is frequently cited in discussions of nicotine as a cognitive enhancer.

But it’s important to note that the effects were small, more like subtle shifts in performance than changes that meaningfully alter day-to-day functioning.

More importantly, a 2020 systematic review in Substance Abuse looked at 32 randomized controlled trials and found no consistent evidence of cognitive benefit of nicotine. Positive, mixed and null results were all found. At the same time, a substantial proportion of authors had prior tobacco industry funding and many of these ties were not disclosed. The review did not find a clear link between funding and outcomes, but the influence of industry funding may be underestimated.

This matters because the tobacco industry has influenced how nicotine research is conducted and presented. One example is the Foundation for a Smoke-Free World, which was funded with a $1 billion pledge from Philip Morris International in 2017 and later rebranded in 2024. The World Health Organization has advised against engaging with the foundation.

So, nicotine can have acute cognitive effects, but the way the “cognitive enhancer” narrative has been built is more complex than it appears.

Subscribe now

The clinical trials in neurodegenerative diseases

For decades, the cleanest version of the case for nicotine has been that if acetylcholine signalling declines in Alzheimer’s disease and nicotine activates acetylcholine receptors, maybe nicotine can slow cognitive decline. A 2012 pilot trial tested transdermal nicotine patches in people with mild cognitive impairment. The trial was promising enough to motivate a much larger follow-up, the MIND study, with 348 participants over two years of treatment.

The results of the MIND study have not yet been formally published, but they were presented in December 2025 at the CTAD conference in San Diego.

Transdermal nicotine did not slow cognitive decline in people with mild cognitive impairment. The researchers’ interpretation was that nicotine may produce short-term cognitive effects, but these do not appear to be sustained over two years. The patch was safe and well tolerated in non-smokers, but it did not slow cognitive decline over that period.

This was one of the most rigorous tests of nicotine as a cognitive intervention, and so far, the answer appears to be no.

Parkinson’s disease initially looked like a similar story. Epidemiologically, smokers show a lower risk of Parkinson’s. This has been interpreted as nicotine possibly having neuroprotective effects on dopamine neurons.

But the NIC-PD trial, published in NEJM Evidence in 2023, tested transdermal nicotine against placebo in early Parkinson’s patients for a year. Nicotine did not slow disease progression. If anything, outcomes were numerically slightly better in the placebo group. This suggests that whatever protects smokers from Parkinson’s is unlikely to be explained by nicotine alone.

To this, we can also add the α7 nicotinic agonists that pharmaceutical companies have tried to develop as cognitive drugs (encenicline, ABT-126, TC-5619). These showed early promise but ultimately failed to translate into approved treatments.

The pattern seems consistent. If nicotine were an effective medicine for long-term outcomes, we would likely know by now.

What about smoking and neurodegenerative diseases?

Parkinson’s and Alzheimer’s are both neurodegenerative diseases and involve neurotransmitter systems nicotine interacts with. But the epidemiological evidence for cigarette smoking points in opposite directions.

For Parkinson’s, smokers develop PD at substantially lower rates across epidemiological studies. The question is what explains it. Nicotine has been the obvious candidate, but after the NIC-PD trial I mentioned before, that explanation looks unlikely to be sufficient on its own.

For Alzheimer’s, the association goes the other way. Smokers have higher Alzheimer’s risk. This is supported by prospective cohort meta-analyses, with smokers showing increased risk. Also the 2024 The Lancet Commission on dementia prevention, intervention and care lists smoking as one of fourteen modifiable risk factors.

There is an older literature claiming smokers had lower Alzheimer’s risk. However, a 2010 analysis published in Journal of Alzheimer’s Disease re-examined these studies using the UCSF Truth Tobacco industry documents library. The apparent protective effect was largely confined to case-control studies with tobacco-industry-affiliated authors. Independent prospective cohort studies showed increased Alzheimer’s risk. So, once the industry-affiliation was taken into account, the “protection” disappeared.

So to summarise: smokers have lower Parkinson’s risk, but the protection is unlikely to be explained by nicotine alone, if at all. And for Alzheimer’s, the association goes in the opposite direction, with smokers showing higher risk.

The delivery system problem

Nicotine gets delivered in very different ways and this matters.

A cigarette is essentially a combustion system that happens to deliver nicotine. The thousands of other compounds in tobacco smoke (tars, oxidants, carbon monoxide, carcinogens) are what can increase risk of cancer, cardiovascular disease and stroke. Smoking is one of the most harmful exposures for long-term health, and quitting is associated with measurable cardiovascular benefits within weeks.

E-cigarettes remove combustion but introduce new uncertainties. We don’t yet have decades of cohort data that we have for cigarette smoking. What we do have is evidence suggesting cardiovascular effects, including endothelial dysfunction, increases in heart rate and blood pressure and emerging evidence of adverse pulmonary effects. The American Heart Association 2023 scientific statement provides a clear summary.

Nicotine pouches are the newest category. They don’t involve smoke or aerosol. They deliver nicotine through the oral mucosa, more slowly than a cigarette but faster than a patch. The U.S. Food and Drug Administration authorised multiple products in January 2025 on the basis that they are likely to be less harmful than cigarettes for adult smokers who completely switch.

This framing may be appropriate for a 50-year-old smoking a pack a day. But it is much less clear for a 19-year-old who has never smoked and is now using pouches “for focus” during exams. Recent data suggest that youth use of nicotine pouches has increased. These users are a new population of nicotine users, many of whom were not previously smokers.

The cardiovascular pharmacology of nicotine itself, independent of combustion, is also not negligible. It increases heart rate and causes vasoconstriction. The long-term effects in never-smokers using pouches are not yet well characterised, because the population hasn’t existed long enough to study them.

Nicotine and the vulnerable period of adolescence

The adolescent brain continues developing into early adulthood, particularly in the prefrontal cortex. This is the part of the brain that supports executive function, impulse control and decision-making. Nicotinic receptors contribute to synaptic and circuit refinement that happens during this period.

When rodent brains are exposed to nicotine in adolescence, they show lasting changes in prefrontal synaptic plasticity that are not seen to the same extent with adult exposure. And human imaging studies (correlational) show associations with differences in brain structure and function, including in regions such as the hippocampus in adolescents who use nicotine regularly.

There’s a real asymmetry. A 70-year-old with potential age-related changes in cholinergic function and a 17-year-old with a still-developing prefrontal cortex are likely not experiencing the same biological effects of nicotine.

So is nicotine good or bad?

Nicotine isn’t the clean, low-risk cognitive enhancer it’s often framed as online.

Acutely and temporarily, it can affect cognitive functioning, but the effects are modest.

Over time, it is addictive.

As a medicine for neurodegenerative conditions, nicotine has not translated into effective treatments.

And the neurologically vulnerable population, adolescents, are also the fastest-growing group adopting nicotine use.

So the biohacker framing of nicotine as a clean cognitive tool is, in my view, largely misleading. What’s left is a mildly stimulating, addictive substance with complex pharmacology, and a cultural story that has outpaced the evidence.

References & further reading

1. Heishman, S. J., Kleykamp, B. A. & Singleton, E. G. Meta-analysis of the acute effects of nicotine and smoking on human performance. Psychopharmacology (Berl.) 210, 453–469 (2010).

2. Pasetes, S. V., Ling, P. M. & Apollonio, D. E. Cognitive performance effects of nicotine and industry affiliation: a systematic review. Subst. Abuse Res. Treat. 14, 1178221820926545 (2020).

3. Newhouse, P. et al. Nicotine treatment of mild cognitive impairment: a 6-month double-blind pilot clinical trial. Neurology 78, 91–101 (2012).

4. Oertel, W. H. et al. Transdermal Nicotine Treatment and Progression of Early Parkinson’s Disease. NEJM Evid. 2, EVIDoa2200311 (2023).

5. Cataldo, J. K., Prochaska, J. J. & Glantz, S. A. Cigarette smoking is a risk factor for Alzheimer’s Disease: an analysis controlling for tobacco industry affiliation. J. Alzheimers Dis. JAD 19, 465–480 (2010).

6. Zhong, G., Wang, Y., Zhang, Y., Guo, J. J. & Zhao, Y. Smoking Is Associated with an Increased Risk of Dementia: A Meta-Analysis of Prospective Cohort Studies with Investigation of Potential Effect Modifiers. PLOS ONE 10, e0118333 (2015).

7. Rose, J. J. et al. Cardiopulmonary Impact of Electronic Cigarettes and Vaping Products: A Scientific Statement From the American Heart Association. Circulation 148, 703–728 (2023).

8. Yuan, M., Cross, S. J., Loughlin, S. E. & Leslie, F. M. Nicotine and the adolescent brain. J. Physiol. 593, 3397–3412 (2015).

You get the promotion you’ve been working hard towards.

You buy your dream house that you’ve wanted for so long.

And yet…something feels off.

Flat. Dull. Like something is missing, except you don’t know what.

I felt this after my PhD defence. Years of hard work and anticipation. And when it was finally done, I expected relief and joy.

And logically, I knew I was relieved. But it felt flat and empty. And the stress didn’t lift for months.

That feeling is easy to misread. It can feel like going backwards. Like the work didn’t work.

But in many cases, it’s the opposite.

The brain is keeping score differently than you may think

The brain doesn’t evaluate how good life is in purely absolute terms. It tracks the gap between what it expects and what actually happens.

One study tracked the moment-to-moment happiness of over 18,000 people via smartphone. Happiness wasn't explained by how much people were winning overall. It was explained primarily by recent prediction errors. Outcomes that matched expectations produced little to no change, regardless of how objectively good they were.

Another study found that what drives mood isn't reward itself but the brain's learning signal (the moment of updating expectations). In stable, predictable environments where there's nothing left to learn, that signal shrinks toward zero and there’s less driving mood upward. Part of why comes down to how the dopamine system itself is structured, as it doesn’t have a fixed sensitivity level. It adjusts continuously, turning up its responsiveness in uncertain, volatile environments where new information matters, and turning it back down in stable, predictable ones where everything is already known. Evidence suggests that in a stable situation, the dopamine system reads the environment as low-variability and recalibrates accordingly. Expected good things produce almost no response because there’s nothing left to track.

And there's a further layer. Neuroscience distinguishes between two systems most people conflate: wanting (the dopaminergic drive to pursue) and liking (the actual hedonic experience of having). The two are neurologically distinct. You can strongly want something and feel very little when you get it. And you can enjoy something you didn’t particularly seek.

The systems run in parallel, and arrival tends to quieten wanting without necessarily activating liking. While you were working toward the thing, both were fully engaged. The day it became your reality, the wanting signal moved to the next gap.

Subscribe now

The lottery problem

In a well-known study from 1978, Brickman and colleagues compared lottery winners to controls and found that winners were no happier, reporting less pleasure even from everyday things. It became foundational evidence for the “hedonic treadmill.”

That study had only 22 lottery winners and no baseline measure, but the topic has been studied since:

A study from 2020, tracked Swedish lottery winners for up to 22 years. Life satisfaction gains from a substantial win persisted over a decade with no sign of decay. The hedonic treadmill is not as total as the study by Brickman suggested. But the nuance is where it gets interesting. Life satisfaction, referring to how you evaluate your life when you stop and think, improved. But happiness and mental health (affective well-being) showed weaker and less consistent effects.

This distinction is crucial and underappreciated. You may genuinely judge your life as better, and still not feel better in any given moment. Because the evaluation is a deliberate cognitive act and the feeling is downstream of prediction errors. And prediction errors normalise fast.

When the nervous system forgets what calm feels like

Maybe for a long time, the baseline has included urgency, pressure, constant input and even stress. All of that can become familiar —it did for me. And the brain, as a prediction machine, builds a model of what normal looks like and navigates everything against it.

When the environment changes, when things get calmer, more stable, the brain doesn’t immediately register that as better. It registers it as different.

A 2020 mathematical modelling study showed why stress recovery takes so long at the level of tissue. Under prolonged stress, the adrenal glands physically enlarge. When the stressor is removed, that tissue has to shrink back, and that process unfolds over weeks to months, with hormonal dysregulation persisting well after things appear to have normalised. The body cannot reset its stress machinery overnight.

After sustained stress, the nervous system may adapt in a way that makes calm feel exposing rather than relieving. Being relaxed before something bad happens means the emotional drop can hit harder. In people who've spent a long time under pressure, this dynamic can become entrenched: calm can stop feeling like rest and start feeling like vulnerability.

The lag

When your life improves but it doesn’t feel like it, several things are likely happening at once.

The reward signal quieted because the outcome became predictable. What was once a goal is now expected and expected outcomes generate less signal. Your evaluation of life may have genuinely improved, and that can last, but moment-to-moment feeling tracks change, not stability.

The biological machinery of stress takes longer still. The nervous system trained on sustained activation doesn’t immediately read the drop in demand as relief. And that can take weeks to months, not days.

The external situation changed. The internal one may still be catching up. That took me longer than I expected and longer than I gave myself credit for needing.

Has there been a moment in your life where something good happened and you couldn't feel it? What was it?

References and further reading

1. Rutledge, R. B., Skandali, N., Dayan, P. & Dolan, R. J. A computational and neural model of momentary subjective well-being. Proc. Natl. Acad. Sci. U. S. A. 111, 12252–12257 (2014).

2. Blain, B. & Rutledge, R. B. Momentary subjective well-being depends on learning and not reward. eLife 9, e57977 (2020).

3. Diederen, K. M. J., Spencer, T., Vestergaard, M. D., Fletcher, P. C. & Schultz, W. Adaptive Prediction Error Coding in the Human Midbrain and Striatum Facilitates Behavioral Adaptation and Learning Efficiency. Neuron 90, 1127–1138 (2016).

4. Lindqvist, E., Östling, R. & Cesarini, D. Long-Run Effects of Lottery Wealth on Psychological Well-Being. Rev. Econ. Stud. 87, 2703–2726 (2020).

5. Berridge, K. C. & Kringelbach, M. L. Pleasure systems in the brain. Neuron 86, 646–664 (2015).

6. Karin, O. et al. A new model for the HPA axis explains dysregulation of stress hormones on the timescale of weeks. Mol. Syst. Biol. 16, e9510 (2020).

7. Newman, M. G. & Llera, S. J. A novel theory of experiential avoidance in generalized anxiety disorder: a review and synthesis of research supporting a contrast avoidance model of worry. Clin. Psychol. Rev. 31, 371–382 (2011).

A lot of brain health advice is built on weak evidence.
Sometimes, the things that get the most attention are the least supported.
And the factors with the strongest causal evidence are rarely the ones being marketed.

I read brain health literature almost daily. This paid post outlines how the evidence actually ranks.

What’s inside:

• Which factors have the strongest evidence for lowering risk of cognitive decline and dementia-related outcomes

• Which widely promoted strategies fail in trials

• Why supplements underperform and the few exceptions worth watching

• What actually matters most in practice (based on the strongest available evidence)

If you only change a few things based on this, this is where the data suggests you should focus.

A note on scope

“Brain health” is a broad and somewhat vague term. In this post I draw primarily on research into cognitive aging and dementia prevention, because that’s where some of the most rigorous long-term evidence is available.

Presenting factors individually is a necessary simplification here and I'll cover specific factors in more depth in future posts. And it’s also important to note, that the structural factors I discussed in the last post don't disappear from biology just because they're hard to put in a trial.

So here, I’m ranking lifestyle factors by how confidently we can say they matter for cognitive aging and dementia prevention:

• Strong evidence: includes factors supported by randomized controlled trials (RCTs), Mendelian randomization studies or both.

• Moderate evidence: includes factors from large, well-conducted observational studies with plausible mechanisms.

• Limited or mixed evidence: includes factors from trials that have been disappointing, factors with unclear causal direction and factors from observational research that don’t survive rigorous testing.

• Preliminary evidence: early-stage, biologically plausible factors but not yet replicated at scale.

Strong evidence, what probably moves the needle most

Read more

Brain health advice is quite consistent.

“Sleep eight hours”, “exercise regularly”, “eat well”, “stay socially connected”, “challenge your mind”. These are real recommendations, and they are supported by real and substantial evidence.

But they’re often presented as a checklist, as if each habit works independently and brain health is simply their sum.

Brain health isn’t the sum of isolated habits.
It emerges from interacting systems, operating across an entire lifetime.

The checklist model and why it’s incomplete

The checklist approach is intuitive because it mirrors the logic of simple interventions: if something is missing, add it. Not enough sleep? Sleep more. Not enough social connection? Add social connection. Each item is discrete, actionable and checkable. It’s a reassuring way to feel in control of something that can otherwise feel abstract and threatening.

The problem isn’t that the items on the list are necessarily wrong. Many of them are right. The problem is more the underlying model. The assumption that brain health is the sum of independent parts, each contributing its own fixed amount, each modifiable in isolation, working the same way across individuals and contexts.

Brain health is also often framed as the absence of disease.
But a more useful definition is the brain’s capacity to function, adapt and resist damage across the life course, encompassing cognitive, vascular, psychological and social dimensions simultaneously.

By that definition, brain health is a systems concept. And the checklist, by design, is not well suited to capturing a system.

Brain health unfolds across a lifetime, not a single moment

Much of what we know about brain health comes from studying dementia risk, because it reflects long-term changes in how the brain functions and ages.

One of the most widely cited scientific syntheses in the dementia field is the Lancet Commission on Dementia Prevention. The 2024 update published in The Lancet organizes its findings by when in life different risk factors may matter the most for dementia risk.

Education and early cognitive stimulation in childhood and early adulthood build what we call cognitive reserve, a kind of accumulated capacity that offers resilience against later damage.

In midlife, physical inactivity, hypertension, obesity and hearing loss (among others) appear to carry the greatest burden. In later life, social isolation, vision loss and air pollution may be most consequential.

This temporal organization reflects the brain accumulating exposure. Risk isn’t a snapshot of current habits, but an accumulation of what the brain has experienced, and when.

The brain has sensitive periods, windows where experience can leave a more lasting imprint and shape the trajectory that follows. But sensitivity to timing doesn't end in childhood. Midlife can also be understood as a critical window: in Alzheimer’s disease (the most common form of dementia) amyloid and tau proteins begin accumulating decades before any symptom appears, meaning that vascular and metabolic risk factors in 40s and 50s act on a brain that may already be undergoing pathological change.

The same exposure can have different effects depending on when it occurs.

And a checklist collapses time.

Why risks converge

One landmark prevention trial in this field, the Finnish FINGER study, published in The Lancet in 2015 (and since expanded into a global network of trials), tested what happens when you address multiple risks simultaneously. They combined dietary guidance, exercise, cognitive training and vascular risk monitoring over two years. Relative to the control group, the intervention group showed improvements in cognitive functioning. The finding supports the idea that addressing multiple domains simultaneously may be more effective than targeting single factors in isolation. This fits a broader pattern in the literature: risk factors tend to accumulate and interact rather than operate independently.

This is biologically plausible, though the exact mechanisms are still being worked out. Most dementia risk factors don’t operate through entirely separate mechanisms. Many of them appear to converge on overlapping pathways. For example, hypertension, diabetes and smoking all damage the vascular system. Disrupted sleep has been associated with impaired clearance of amyloid. Chronic psychological stress dysregulates cortisol and amplifies inflammatory signalling. Depression and social isolation can compound these effects through both behavioral and biological pathways. What emerges downstream (chronic inflammation, impaired cerebrovascular function, metabolic stress or reduced synaptic resilience) is often the product of multiple converging pressures, not a single cause. Risk factors may appear additive in epidemiological studies, without being truly independent in biology.

The term allostatic load is used to describe a cumulative burden, the total physiological cost the body and the brain carry from chronic exposure to multiple stressors across cardiovascular, metabolic, neuroendocrine and immune systems. It’s not the cost of one bad thing, but the weight of many things, carried across time.

A 2026 study from the Berlin Aging Study II, in Communications Medicine, compared 16 biological aging markers in over 1,000 older adults across seven years and found allostatic load to be among the most consistent markers associated with age-related health outcomes, though the field continues to debate how best to measure biological aging.

If risk factors tend to accumulate and converge on shared biological pathways, the question becomes whether some processes sit upstream of multiple risks at once. That's a question I'll return to.

What about genetics?

Genetics matter. The APOE ε4 allele, which is the strongest known genetic risk factor for late-onset Alzheimer’s disease, increases risk and roughly one in four people carries at least one copy. Rare variants in genes like PSEN1 and APP cause forms of dementia that are largely deterministic regardless of lifestyle. But only around 1 per cent of dementia cases are attributed to these rare genetic variants.

And so genetics rarely writes a fixed outcome, at least when it comes to dementia. It shapes the gradients of risk, without determining where you end up. Even APOE ε4 carriers show substantially different trajectories depending on vascular health, education, sleep and stress exposure across their lives. The FINGER trial found that its multidomain intervention produced cognitive benefits in APOE ε4 carriers who would otherwise have been at highest risk. This suggests that genetic predisposition doesn’t eliminate room for intervention. These findings are consistent with what gene-environment interaction research more broadly suggests.

Genetics sets the parameters while the life course fills them in. The same genetic profile can lead to different outcomes in different environments, at different time points and in different social contexts.

Subscribe now

An important layer most checklists leave out

There’s a dimension the checklist model rarely addresses, because it sits upstream of individual choice. And that is the conditions in which people live their lives.

Brain aging doesn’t happen in a social vacuum and the evidence is beginning to make this harder to overlook.

For example, a 2025 study published in Nature Aging examined over 2,000 adults across Latin America and the United States. Greater structural inequality was associated with reduced brain volume and altered connectivity in regions critical for memory and executive function. These associations held after adjusting for age, sex, education and individual cognitive ability, and were more pronounced in people with Alzheimer’s disease. One interpretation is that macrosocial factors become biologically embedded over the life course. It’s not that inequality acts directly on the brain overnight, but that cumulative exposure to environments shaped by structural disadvantage can leave measurable traces over time.

A growing body of research also suggests that neighborhood disadvantage independently associates with smaller gray matter volumes, greater white matter lesion burden and accelerated cognitive decline in otherwise healthy older adults. Histories of lower socioeconomic status are associated with differences in brain structure detectable decades later. And as a 2022 life-course review in Stroke documents, declining dementia incidence in recent cohorts in some high-income countries appears to be partly attributable to population-level investments. These include increased access to education, mass public health programs such as vaccination, and improved access to healthcare, rather than to shifts in individual lifestyle choices alone. Brain health at the population level is also a policy outcome.

None of this dissolves individual agency, and I’m not arguing it does. What it suggests is that individual choices may act differently depending on conditions: which neighborhood, which country, which decade and which starting conditions. A checklist that ignores this isn’t wrong exactly. It’s just describing a much smaller and incomplete part of the picture.

A note on causality

Showing causality in dementia prevention is genuinely hard. Risk accumulates across decades, pathology develops long before any symptoms appear and for ethical and practical reasons you cannot randomize people to have hypertension for twenty years. This means that observational evidence has limits, and some associations that appear causal may reflect reverse causality, shared pathways, or confounding.

A 2025 systematic review in the European Journal of Neurology examined Mendelian randomization studies. This method uses genetic variation to test causal relationships, and is often considered less prone to confounding than standard observational studies. They found that nearly three-quarters of evaluable analyses provided what the authors classified as insufficient evidence for a causal association. Only some factors, most consistently education and type 2 diabetes–related dysfunction, showed stronger and more reproducible causal support.

The authors themselves note that null findings should be interpreted cautiously given measurement limitations. And this doesn’t mean the other risk factors don’t matter. It may mean that when risks share biology and converge on common mechanisms, disentangling individual causal contributions was always going to be methodologically difficult.

It’s complex because it is complex.

What this means, in my opinion

None of this is an argument for fear or for the conclusion that nothing we as individuals do matters. It’s an argument for asking better questions.

“Am I doing enough for my brain health?” is a difficult question to answer because it’s built on the checklist frame as if there’s a threshold to cross or a list to complete. The more useful question is: what actually matters most, when does it matter most, and for whom?

The evidence suggests that not everything matters equally. And the factors that matter most aren’t always the ones receiving the most attention.

That’s what I’ll explore next.

-Eleni

References and further reading

1. Livingston, G. et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. The Lancet 404, 572–628 (2024).

2. Ngandu, T. et al. A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring versus control to prevent cognitive decline in at-risk elderly people (FINGER): a randomised controlled trial. The Lancet 385, 2255–2263 (2015).

3. Legaz, A. et al. Structural inequality linked to brain volume and network dynamics in aging and dementia across the Americas. Nat. Aging 5, 259–274 (2025).

4. Vetter, V. M. et al. Comprehensive cross-sectional and longitudinal comparison of sixteen markers of biological aging from the Berlin Aging Study II. Commun. Med. 6, 168 (2026).

5. Hilal, S. & Brayne, C. Epidemiologic Trends, Social Determinants, and Brain Health: The Role of Life Course Inequalities. Stroke 53, 437–443 (2022).

6. Solomon, A. et al. Effect of the Apolipoprotein E Genotype on Cognitive Change During a Multidomain Lifestyle Intervention. JAMA Neurol. 75, 462–469 (2018).

7. Mostert, C. M. et al. Broadening dementia risk models: building on the 2024 Lancet Commission report for a more inclusive global framework. eBioMedicine 120, (2025).

8. Desai, R. et al. Evidence for Causal Links Between Known Modifiable Risk Factors and Dementia: A Systematic Review of Mendelian Randomisation Studies. Eur. J. Neurol. 32, e70458 (2025).

The Brain Brief is four months old and I wanted to share a small change.

If you’ve been reading, commenting or sharing the posts with someone, I want to thank you. I really appreciate it.

I want to keep writing The Brain Brief in a way that stays thoughtful, consistent and sustainable.

Starting next month, one post per month will be for paid subscribers. The other three will remain free for now.

So I’ve added a paid option for those who find value in this work and want to support it. And, of course, get access to everything I post.

Writing The Brain Brief takes real time each week: reading the research, making sense of it and turning it into something clear and honest without oversimplifying or overclaiming.

If that’s been useful to you, this is a way to support it.

You can subscribe below for $6/month or $49/year.

The price may increase in the future, but your rate will stay the same if you subscribe now.

(If you’re reading this in the Substack app, you may need to open it in your browser to subscribe).

If not, nothing changes. You’ll keep getting three posts a month.

Thank you for reading.

- Eleni

What if the problem isn’t what happens to you,
but how long your brain keeps running it?

I’m not anxious all the time.
But I’ve had nights where I lie awake at 2am replaying conversations from six months ago or running through future scenarios.

Most people will probably recognize some version of this.

A loop that runs without permission.

That doesn’t respond to “stop it.”

Many of the topics I write about focus on things that can actually be changed.

And today’s topic is modifiable too. I will cover:

• What is repetitive negative thinking (RNT)

• How it may relate to long-term brain health

• What we know about reducing it

What is repetitive negative thinking

RNT is the umbrella term for two very familiar cognitive habits:

• Rumination: replaying the past. What happened, why it happened, what it means about me.

• Worry: rehearsing the future. What might go wrong, what I’d do, how bad it would be.

Researchers have traditionally studied these separately. Rumination was a depression thing and worry was an anxiety thing. But over the past decade or so, the field has shifted. We now know that rumination and worry are highly correlated, load onto a common factor and predict and maintain a whole spectrum of mental health conditions.

A comprehensive 2025 review in Nature Reviews Psychology argues that RNT is best understood as a transdiagnostic cognitive process. In other words, not as a disorder-specific symptom, but as a shared underlying feature that cuts across diagnostic boundaries.

The content and context changes, but the process is hypothesized to be the same. Same loop, just running different material through it.

The cognitive debt framework

Before getting into the empirical data, there’s a conceptual idea worth understanding.

In 2015, University College London researchers proposed what they called the Cognitive Debt hypothesis.

In a way, it could be seen as the theoretical opposite of cognitive reserve, which refers to factors that build resilience over a lifetime. Things like education, mental stimulation and social engagement.

Cognitive debt, in contrast, reflects the gradual accumulation of factors that erode that resilience over time, increasing vulnerability to cognitive decline.

Depression, chronic sleep disruption, PTSD, anxiety symptoms and high neuroticism have all been linked to increased risk of dementia or Alzheimer’s disease in observational research. Why?

The cognitive debt framework argues that they share a common underlying mechanism: RNT. It suggests that these risk factors increase vulnerability in part because they are associated with chronic, persistent patterns of negative thinking, which can keep the brain’s stress systems engaged over time.

This framework gives a coherent mechanistic story for something that otherwise looks like a puzzling list of unrelated risk factors. And it makes a testable prediction: if RNT is the common pathway, then measuring RNT directly should predict AD pathology better than measuring depression or anxiety as broad clinical syndromes.

And that prediction was tested in an influential study.

Association between RNT and Alzheimer’s pathology

In 2020, the UCL team published a paper where they followed cognitively normal older adults across two cohorts (one in Canada, one in France) and measured their levels of RNT using a validated questionnaire. Participants also underwent PET brain imaging to measure amyloid-β and tau protein deposition, the two hallmark pathological features of Alzheimer’s disease.

The results were striking.

Higher RNT was associated with greater amyloid deposition, replicated across both cohorts. It was also associated with greater tau accumulation in the entorhinal cortex, which is one of the earliest regions affected by tau pathology. And it predicted decline in memory and global cognition over four years.

All of this held after controlling for age, sex, education and APOE-ε4 status.

This is what I find interesting: symptoms of depression and anxiety predicted cognitive decline too. But they were not associated with Alzheimer’s pathology. In this study, only RNT showed associations with amyloid and tau, which may suggest that it connects to Alzheimer’s biology in a more specific way than broader emotional distress.

To my knowledge, this was among the first studies to directly link a cognitive pattern to the biology of Alzheimer’s disease in humans.

Why the brain? The anatomy of RNT

Why might a thinking pattern relate to processes involved in Alzheimer’s pathology?

The default mode network (DMN) is a set of interconnected brain regions that become most active when we’re not focused on the outside world, but when we’re daydreaming, remembering the past, imagining the future or thinking about ourselves. It’s like the brain’s introspective circuit.

It’s also one of the networks where amyloid accumulation is thought to begin.

The regions showing early and heavy amyloid deposition include the posterior cingulate cortex, precuneus and medial prefrontal cortex. These regions overlap closely with core DMN hubs, which also have some of the highest baseline metabolic activity in the brain.

RNT appears to engage this same network.

A 2019 fMRI study published in Translational Psychiatry measured momentary sadness, rumination and resting-state functional connectivity. They found that the relationship between sadness and RNT in daily life was modulated by connectivity between the DMN, fronto-parietal and salience network. Individuals who reported higher rumination showed patterns consistent with impaired disengagement. In other words, difficulty shifting away from internally focused thought.

A 2021 systematic review in Psychiatry Research: Neuroimaging, synthesizing 24 neuroimaging studies found that RNT has been linked to structural differences in prefrontal and cingulate regions, as well as altered microstructure of the superior longitudinal fasciculus (a white matter tract connecting frontal and parietal regions involved in cognitive control.)

However, findings across studies were heterogeneous and no single brain region has been consistently implicated. It may be that RNT reflects altered dynamics across distributed large-scale networks, rather than localized structural changes.

What’s important is that the systems repeatedly engaged during RNT overlap with some of the ones most vulnerable in Alzheimer’s disease. What that means over time is still somewhat unclear.

Does RNT cause anything?

All the human evidence linking RNT to amyloid, tau and cognitive decline is observational. We cannot say, based on existing data, that rumination causes Alzheimer’s pathology.

The causal arrow could run the other way too: early, preclinical pathology might itself generate anxiety, low mood and ruminative thinking as early behavioral signatures of the disease. There’s evidence for this direction too.

There’s also the possibility that a third variable, like genetics, personality or chronic stress exposure drives both RNT and AD risk without one causing the other.

What we have instead is a pattern that shows up repeatedly, is biologically plausible and holds after accounting for important confounders. But it’s not proof of causation. We already know that RNT can be reduced. What we don’t yet know is whether changing this pattern changes what happens in the brain over time. Intervention studies will be needed to answer that question.

So what’s the possible mechanism?

If RNT does affect brain health, how might it work? There are several plausible, non-mutually-exclusive pathways.

1. Cortisol. This is one of the best-supported pathways. A large meta-analysis showed that perseverative cognition (a process closely related to RNT) is associated with sustained cortisol elevation. Chronic cortisol exposure has well-established effects on brain structure and function.

Evidence linking cortisol to amyloid comes primarily from animal models, where stress can increase amyloid-β levels. In humans, higher midlife cortisol has been associated with greater amyloid burden years later in some longitudinal studies.

Taken together, the association between RNT and prolonged cortisol activation is supported. The link between cortisol and amyloid in humans is still emerging, with some longitudinal evidence but not yet consistent.

2. Neuroinflammation. This one has experimental evidence at the human level. In a randomized study, women who were instructed to ruminate after a speech stressor showed more prolonged CRP and cortisol responses, whereas distraction was associated with faster recovery. Chronic neuroinflammation is strongly implicated in neurodegeneration. The chain from “rumination sustains inflammation” to “rumination causes brain damage” still has multiple steps that have not been established. But the first link, that repetitive thinking can prolong inflammatory responses, is supported.

3. Sleep and the glymphatic system. This pathway is speculative. What we know is that during sleep, the brain’s waste-clearance system (the glymphatic pathway) facilitates the removal of metabolic byproducts, including amyloid. We also know that RNT may disrupt sleep. What we don’t yet have is evidence in humans that the full chain operates:
RNT → disrupted sleep → impaired clearance → amyloid accumulation.

But I find it worth mentioning because it illustrates how a cognitive pattern could, in principle, translate into biological consequences.

4. Vascular pathways. Perseverative cognition has been linked to a pattern of sustained cardiovascular activation, including increased peripheral resistance and slower blood pressure recovery after stress. Vascular health is one of the most well-established modifiable risk factors for dementia. Again, however, the full chain linking these processes has not been demonstrated.

Beyond stress: what predicts brain aging

In 2021, a neuroimaging study published in Neurobiology of Aging used machine learning to estimate “brain age”. They measured how structurally old the brain appears relative to chronological age.

They then asked whether psychological factors predicted accelerated aging.

Worry and rumination did, but perceived stress did not.

In other words, it wasn’t how stressed people felt, but how much their minds replayed and rehearsed negative content that predicted older brain age.

These findings are consistent with the idea that patterns like worry and rumination, key components of repetitive negative thinking, may relate to accelerated brain aging rather than stress itself.

This suggests that it may not be what happens to you, but how long your brain keeps running it.

Subscribe now

A different side of the story

I want to tell you about a finding from my work that I think speaks directly to the other side of this story.

For the past few years, I have mostly studied cognitively unimpaired older adults with significant Alzheimer’s risk. Many with elevated amyloid or tau in their brains and biologically at-risk, but showing no cognitive symptoms. For one project, we looked at whether how people cope with stress predicts how their cognition evolves over time.

In this study, published in Alzheimer’s & Dementia last year, we showed that higher use of adaptive coping strategies (including active coping, positive reframing, planning, acceptance) was associated with better cognitive trajectories over five years of follow-up, independent of AD pathology. More strikingly, we also found that adaptive coping buffered the impact of tau on cognitive trajectories. So participants with high tau burden but high adaptive coping maintained better function than those with the same amount of pathology but less use of adaptive coping strategies.

I find this finding hopeful because it suggests that the link between brain pathology and cognitive symptoms is not fixed. How we respond to stress may shape the cognitive consequences of biology that’s already there.

RNT, in a sense, could be seen as the opposite of adaptive coping. The mind refusing to disengage, reframe or accept. The evidence suggests that this pattern of response, not just the stressor itself, may be what matters for the brain over the long term.

What the research supports doing about RNT

A 2023 meta-analysis of 61 randomized controlled trials in Journal of Affective Disorders found that mindfulness-based interventions produce a moderate reduction in ruminative thinking.

Similarly, a 2025 transdiagnostic meta-analysis in Psychological Medicine showed that CBT broadly reduces RNT with moderate effect sizes, with RNT-specific therapies being significantly more efficacious. The takeaway is that traditional CBT works, but interventions designed to change how someone relates to their thinking, appear to be the most powerful tools.

Further, a 2015 PNAS study found that a 90-minute walk in nature reduced both self-reported rumination and neural activity in the subgenual prefrontal cortex (a region implicated in depressive self-referential processing).

What we can say is that RNT is modifiable and the biology connecting it to brain health is plausible enough that reducing it seems, at minimum, not a bad idea.

What to take from all of this

Most of us have repetitive negative thoughts. The brain is built to return to unresolved things.

What may matter is not whether the loop appears, but how long it keeps running and whether it becomes the default rather than something that passes.

We don’t yet have clinical trials showing that reducing RNT changes long-term brain pathology or dementia risk.

But across different studies, methods and populations, the pattern shows up.

Depression, sleep, exercise and social connection are all well-recognized.

Repetitive negative thinking is less so.

And yet it may be a common pathway through which many other risk factors operate.

And it’s something we can learn to interrupt.

Share

References and further reading

1. Moulds, M. L. & McEvoy, P. M. Repetitive negative thinking as a transdiagnostic cognitive process. Nat. Rev. Psychol. 4, 127–141 (2025).

2. Marchant, N. L. & Howard, R. J. Cognitive Debt and Alzheimer’s Disease. J. Alzheimer’s Dis. 44, 755–770 (2015).

3. Marchant, N. L. et al. Repetitive negative thinking is associated with amyloid, tau, and cognitive decline. Alzheimers Dement. J. Alzheimers Assoc. 16, 1054–1064 (2020).

4. Lydon-Staley, D. M. et al. Repetitive negative thinking in daily life and functional connectivity among default mode, fronto-parietal, and salience networks. Transl. Psychiatry 9, 234 (2019).

5. Demnitz-King, H., Göehre, I. & Marchant, N. L. The neuroanatomical correlates of repetitive negative thinking: A systematic review. Psychiatry Res. Neuroimaging 316, 111353 (2021).

6. Karim, H. T. et al. Aging faster: worry and rumination in late life are associated with greater brain age. Neurobiol. Aging 101, 13–21 (2021).

7. Palpatzis, E. et al. Adaptive stress coping is associated with cognitive resilience in at-risk cognitively unimpaired older adults. Alzheimers Dement. 21, e70713 (2025).

8. Stenzel, K. L., Keller, J., Kirchner, L., Rief, W. & Berg, M. Efficacy of cognitive behavioral therapy in treating repetitive negative thinking, rumination, and worry – a transdiagnostic meta-analysis. Psychol. Med. 55, e31 (2025).

9. Bratman, G. N., Hamilton, J. P., Hahn, K. S., Daily, G. C. & Gross, J. J. Nature experience reduces rumination and subgenual prefrontal cortex activation. Proc. Natl. Acad. Sci. 112, 8567–8572 (2015).

10. Ottaviani, C. et al. Physiological concomitants of perseverative cognition: A systematic review and meta-analysis. Psychol. Bull. 142, 231–259 (2016)

11. Zoccola, P. M., Figueroa, W. S., Rabideau, E. M., Woody, A. & Benencia, F. Differential effects of post-stress rumination and distraction on cortisol and C-reactive protein. Health Psychol. 33, 1606–1609 (2014).

In Part 1, I wrote about what happens when the brain is under pressure: working memory becomes constrained, behaviour can shift toward habit and the future can start to feel less relevant.

All of this can push decision-making in a similar direction. Not necessarily toward worse outcomes in every case, but toward faster, simpler and more immediate ones.

In Part 2, the question is: what can we do about it?

It’s not (just) about willpower

Under pressure, the brain isn’t choosing to perform poorly, but is operating under different constraints.

Advice like ”just relax” or “think more clearly” under pressure or stress is not very useful.

A more useful approach is to adjust the conditions under which the decision is being made and to work with the constraints. Here are a few ways:

1. Recognise the state you’re in

This sounds obvious, but often it isn’t.

The same systems that help us monitor our own thinking are among the first affected by stress. So stress can partially impair our ability to notice that stress is impairing us.

It probably doesn’t feel like “my working memory is compromised” or “my time horizon is collapsing.”

It often feels more like:

• something physical: tension in the body, a faster heartbeat or maybe a sense of urgency that feels disproportionate

• something cognitive: difficulty holding more than one option in mind or the sense that the decision has to happen now

• something temporal: the future version of you may feel abstract, what matters is resolving the present moment

Even briefly naming the state and the reason for it can help. For example, “I’m overwhelmed by too many options”, “this feels urgent” or “I am anxious about disappointing others” can help create distance from the decision at hand, and even provide some clarity.

Noticing these things doesn’t fix everything. But it can be the first step.

2. Reduce cognitive load before deciding

Working memory takes an early hit under stress. One useful response can be to reduce how much needs to be held in mind.

In practice, that might mean:

• writing options down instead of holding them mentally

• separating what you actually know from what you’re assuming

• narrowing the decision to the one or two things that matter most

• identifying whether the decision is reversible or irreversible

The goal isn’t to oversimplify the problem but to offload the cognitive demand of navigating it, so that whatever capacity remains can be used more effectively.

3. Check whether you’re in habit mode

Under pressure, habitual responses or familiar patterns can feel like automatic decisions.

Before committing, it’s worth asking:

Is this what I would actually choose or just what I always do?

You don’t have to abandon the familiar option, sometimes it’s the right one.

But asking the question can create a brief interruption to re-engage deliberate thinking before you commit.

4. Re-anchor the decision to the outcome

Even if you interrupt the default response, you still need a direction.

One of the reasons decision-making can shift under pressure is that the outcome itself can become harder to keep in mind.

Goal-directed thinking depends on being able to keep the outcome actively in mind. Under stress, that becomes harder. The decision can become less about where this leads and more about what feels actionable right now.

Simple ways to make the goal explicit:

• writing down the outcome you are actually trying to move toward

• briefly shifting perspective: what would this decision look and feel like in 10 minutes, 10 months or 10 years?

This doesn’t require solving the entire problem, it just reintroduces a reference point.

5. Delay decision-making when possible, even briefly

We might feel a pull toward immediate resolution when we are under pressure. The future can feel abstract, while the present may feel urgent.

Even short delays can counteract this.

A pause creates distance between the physiological state and the decision. It gives stress-related signals time to begin settling and prefrontal systems a better chance to re-engage.

In practice, this might look like:

• stepping away briefly before committing

• separating your need for time from others' urgency. Practice saying: “Can I get back to you on this?”

• revisiting the decision later the same day

• avoiding finalising anything at the peak of emotional activation

The thinking and the committing don’t have to happen at the same time. Even though pressure makes it feel like they should.

Asking for more time is often possible, it just may not feel that way in the moment. Sometimes, the urgency is artificial.

That said, not everyone responds to pressure the same way. In some cases, the opposite can happen. Instead of rushing, people can become stuck or overanalyze. In those situations, setting a simple time boundary can help shift the decision forward.

6. Pre-commit outside of pressure

One of the most reliable ways to make better decisions under pressure is to make fewer decisions in that state.

When things are calm we’re better able to evaluate trade-offs, think long-term and consider what we actually want.

Decisions made in those conditions can be carried forward into more pressured contexts.

In practice:

• set rules or boundaries in advance

• decide what “good enough” looks like ahead of time

• define your non-negotiables before you’re in the situation

Having decided in advance means fewer things to figure out in the moment of pressure.

None of this eliminates the effects of pressure or stress on the brain. But these tools can help with noticing the state we’re in before it drives the decision, keeping the goal in view when everything feels urgent and creating space between the pressure and the response.

Which of these do you think would actually change how you decide under pressure? I'd love to hear in the comments!

This post draws on the same literature cited in Part 1.

Most of us have probably made decisions under pressure that we later regretted. For example, saying something you immediately regretted in the middle of an argument. Or saying yes to a “limited time offer” simply because it felt like a good deal in the moment.

It’s easy to blame yourself in these situations: “I’m not a quick thinker.” “I panic.”

But what’s actually happening is more specific than that, and often less about personality.

I’ve written before about what the stress response looks like physiologically. If you haven’t read this post, it’s a useful foundation.

But today I want to go further, discussing not just that stress impairs decision-making, but how it does so.

First, working memory takes a hit

Before we get to the larger patterns that can emerge in decision-making under stress, it’s worth starting with something more immediate. One of the most consistently observed cognitive effects of acute stress involves working memory.

Working memory is the brain’s ability to hold information in mind while using it. It’s what allows us to track an argument as it unfolds, compare multiple options or keep the broader context in mind while dealing with a specific detail. And it depends heavily on the prefrontal cortex. Specifically, on the stability of neural circuits that allow information to remain actively represented for a few seconds at a time.

Studies show that even mild, uncontrollable stress can rapidly impair these circuits. The mechanism involves a surge of catecholamines, particularly noradrenaline and dopamine. At moderate levels, these neurotransmitters help prefrontal networks function. But at high levels, which can occur during stress, they weaken the recurrent signaling that keeps prefrontal networks stable.

The effect is dose-dependent and can happen quickly, sometimes before we have fully recognized how stressed we are.

Put simply, the practical consequence is this: under pressure, we can hold less in mind at once.

We can lose the thread of complex reasoning, comparing multiple options becomes harder and details that would normally fit into a coherent picture start competing with each other.

So the brain simplifies. The system responsible for managing complexity has temporarily lost bandwidth.

Consistent with this idea, experimental studies have found that acute stress can impair decision-making particularly when decisions become more complex and require integrating multiple pieces of information.

And this matters for everything that follows. Because once working memory is compromised, the brain has to compensate somehow. In the literature, two common patterns tend to emerge.

The brain shifts from flexible thinking to habit

Under stress or pressure, particularly when the stressor is uncontrollable, the way behaviour is controlled can shift. Decision-making can move away from flexible, goal-directed reasoning and toward more habitual responses.

Goal-directed behaviour is forward-looking: it involves considering the outcome we want and choosing actions that are likely to produce it.

Habitual behaviour is more automatic: we repeat what has worked before, even when the current situation is not exactly the same.

Both systems operate in parallel and behaviour is normally influenced by a dynamic balance between them. However, they rely on somewhat different neural circuits. Goal-directed behaviour depends heavily on prefrontal regions interacting with the dorsomedial striatum, while habitual behaviour relies more on sensorimotor circuits involving the dorsolateral striatum.

Under stress, the balance between these systems can shift.

A well-known study published in Translational Psychiatry demonstrated this pattern in humans. Individuals experiencing high levels of chronic stress showed less flexible decision-making: their behaviour became less sensitive to changes in outcome value. This means that they were more likely to keep repeating previously learned actions even when those actions were no longer optimal.

When the same participants were reassessed several weeks later, after the stress had passed, the pattern reversed.

In everyday life, this shows up in familiar ways. Under pressure, people fall back on the same coping strategies, even the ones they may know don’t work very well. Arguments can follow old scripts. The first solution that comes to mind tends to dominate and switching approaches becomes harder.

Part of this is habit. Part of it is also a reduction in cognitive flexibility: the ability to update strategies when conditions change. Stress makes it harder for the brain to abandon an approach once it has started down that path.

In other words, when resources are constrained, the brain prioritizes efficiency over flexibility.

That said, the literature here isn’t entirely uniform. Some studies have found weaker effects, and individual differences in baseline stress levels or working memory capacity may moderate the shift. However, the overall pattern of stress pushing behaviour toward habit and away from flexible control has been observed across multiple paradigms.

Pressure makes the future feel further away

The second pattern is subtler and, in many ways, more consequential.

Under stress, people sometimes show what behavioural economists call increased temporal discounting. This means a stronger preference for smaller rewards available immediately (e.g. $20 now) rather than larger rewards available later (e.g. $40 in a month).

Put simply: when we’re under pressure, the future may shrink.

There may also be a physiological mechanism behind this shift. In one double-blind experiment published in Psychoneuroendocrinology, researchers administered hydrocortisone to healthy participants before presenting them with intertemporal choices (decisions between smaller rewards available sooner and larger rewards available later.)

Shortly after administration, participants showed a stronger preference for immediate rewards. Several hours later, when cortisol levels had stabilized, the effect disappeared.

This suggests that acute stress hormones can temporarily bias decision-making toward more immediate rewards.

That said, the evidence overall is mixed. A 2024 meta-analysis found no consistent effect of acute stress on monetary delay discounting across eleven studies. Individual differences, particularly in cortisol reactivity, appear to play an important role. One study found the effect only emerged in participants who showed a measurable cortisol response to stress.

In other words, the pattern is real in some contexts, but it doesn’t appear to be universal.

From an evolutionary perspective, the logic is plausible. If you are genuinely under threat, prioritizing immediate outcomes can sometimes be adaptive. Taking what is available now and leaving quickly may be safer than waiting for a better option.

But under modern forms of pressure, like work deadlines, financial decisions and interpersonal conflict, that same bias can push behaviour in directions we would not normally choose.

The everyday version of this is subtle:

You may send an email before thinking it through because you want the situation resolved. You choose the quickest option rather than the best one. Or you agree to something in the moment that tomorrow’s version of you would probably decline.

Then, there are more high-stakes situations.

For example, if you're buying a house and the agent tells you another offer came in overnight. Or when you’re negotiating a job offer and HR says they need an answer by end of day. Or car salespeople disappearing to "check with the manager".

Often, none of that is accidental.

Artificial deadlines, manufactured scarcity and time pressure are classic sales tactics, in part because they can exploit pressure-driven shifts in decision-making. They collapse the time horizon and make the immediate option feel far more compelling than it may actually be.

When we’re in that state, the future version of us who has to live with the decision can feel abstract and distant. The discomfort of the present moment can feel very real.

So pressure doesn’t just make us think worse, but it can change what the brain is optimizing for.

The goal can shift from “what leads to the best outcome?” to “what ends this feeling fastest?”

So what does this mean in practice?

Under pressure, as cognitive resources shrink, the brain may rely more on familiar routines and habits. As fewer alternatives can be held in mind at once, the first available option can become disproportionately attractive. And as the future feels more distant, immediate relief can win.

Of course, these effects don’t look exactly the same for everyone or in every situation. But taken together, they can push decision-making in a similar direction: toward speed, simplicity and short-term resolution.

The question isn’t whether stress affects decision-making. We know it does.

The real question is whether we can recognize that we’re in a high-pressure situation, before the decision is made.

In Part 2, I’ll introduce a practical framework for spotting these moments and creating conditions to think more clearly, even when pressure or stress are unavoidable.

Before then, I’m curious: when was the last time you noticed your thinking change under pressure?

If a situation comes to mind, feel free to share it! I may draw on examples when writing Part 2.

References & further reading

1. Arnsten, A.F.T. (2009). Stress signalling pathways that impair prefrontal cortex structure and function. Nature Reviews Neuroscience, 10, 410–422.

2. Arnsten, A.F.T. (2015). Stress weakens prefrontal networks: molecular insults to higher cognition. Nature Neuroscience, 18(10), 1376–1385.

3. Soares, J.M., et al. (2012). Stress-induced changes in human decision-making are reversible. Translational Psychiatry, 2, e131.

4. Yin, H.H., & Knowlton, B.J. (2006). The role of the basal ganglia in habit formation. Nature Reviews Neuroscience, 7(6), 464–476.

5. Schwabe, L., & Wolf, O.T. (2011). Stress-induced modulation of instrumental behavior: from goal-directed to habitual control of action. Behavioural Brain Research, 219(2), 321–328.

6. Riis-Vestergaard, M.I., et al. (2018). The effect of hydrocortisone administration on intertemporal choice. Psychoneuroendocrinology, 88, 173–182.

7. Kimura, K., et al. (2013). The biological effects of acute psychosocial stress on delay discounting. Psychoneuroendocrinology, 38(10), 2300–2308.

8. Forbes, P.A.G., et al. (2024). No effects of acute stress on monetary delay discounting: a systematic literature review and meta-analysis. Neurobiology of Stress, 31, 100653.

9. Doroc, K., Yadav, N., & Murawski, C. (2025). Acute stress impairs decision-making at varying levels of decision complexity. Communications Psychology, 3, 179

Many of us drink. Maybe not a lot, but enough that it’s worth asking: what is alcohol actually doing to the brain, especially long-term?

What one drink does to your brain

Alcohol affects brain function by interacting with multiple neurotransmitter systems, which is why its effects can feel so layered: relaxing and disinhibiting and mood-altering all at once.

GABA goes up, glutamate goes down

GABA is the brain’s main inhibitory neurotransmitter, it calms things down. Glutamate, on the other hand, is the main excitatory one. It increases neural firing and keeps circuits active.

Alcohol enhances GABA signalling while simultaneously inhibiting glutamate receptors (specifically, NMDA receptors). The effect is a shift toward inhibition, with neural activity becoming less excitable overall. This is why alcohol can initially feel calming or anxiolytic and, at higher doses, lead to slowed cognition and impaired coordination.

That NMDA inhibition has a particularly important consequence in the hippocampus, an area crucial for forming new memories and one of the earliest regions damaged in Alzheimer’s. This NMDA inhibition interferes with long-term potentiation, a cellular process that is thought to underlie memory formation. At moderate alcohol doses, this can lead to fragmented recall. At higher doses, however, memory encoding can fail altogether, resulting in blackouts where events are not encoded into long-term memory.

The reward system activates

Alcohol also activates the brain's reward pathway, indirectly increasing dopamine release, particularly in a brain region called the nucleus accumbens (a key hub in the brain’s reward circuitry).

Rodent studies show that moderate ethanol doses can increase accumbens dopamine by roughly 40–100%, depending on dose and prior exposure. Human PET imaging also shows dopamine release in the brain’s reward system after drinking. That signal can contribute to wanting the next drink.

Alcohol also triggers endogenous opioid release in the nucleus accumbens and orbitofrontal cortex. In the nucleus accumbens, this release has been correlated with feelings of pleasure across all drinkers. In the orbitofrontal cortex, the opioid release response has been found to be larger in heavy drinkers, suggesting that differences in endogenous opioid function in this region contribute to excessive alcohol consumption. Together, this opioid surge contributes to feelings of pleasure and it’s why naltrexone (an opioid blocker) is used as a treatment for alcohol use disorder.

Adenosine builds up

Alcohol also blocks a transporter that normally clears adenosine from the space between neurons. Adenosine is the same molecule that builds up during the day and makes you sleepy and it's what caffeine blocks (I wrote about it here). So alcohol taps into your sleep-pressure system, which is partly why a nightcap can feel like it helps you sleep (but it doesn't really).

Although alcohol can shorten the time it takes to fall asleep, sleep quality deteriorates later in the night. Meta-analyses show that even low doses of alcohol (around two drinks) delay the onset of rapid eye movement (REM) sleep and reduce its duration. Higher doses shorten sleep onset latency further but produce larger disruptions to REM sleep later in the night. REM sleep is thought to play an important role in emotional regulation and memory consolidation.

Downstream consequence: the brain’s control system weakens

Even at relatively low levels of intoxication, activity in the prefrontal cortex begins to decline. This region is critical for judgement, impulse control and decision-making. As its regulatory influence weakens, behaviour becomes more disinhibited, which is one reason why drinking can lead to riskier choices.

And alcohol does not affect the brain uniformly. Functional imaging studies show that it alters communication between large-scale brain networks, particularly between prefrontal control networks and subcortical reward circuits. As top-down regulatory signalling weakens, behaviour becomes more influenced by immediate reward signals rather than long-term goals. In other words, alcohol shifts the balance of the brain from control toward impulse.

Subscribe now

What happens when this becomes chronic

The acute effects largely reverse as alcohol clears the system. But the brain adapts to repeated exposure and those adaptations can carry a cost.

Neurotransmitter balance shifts

With regular drinking, GABA receptors downregulate and glutamate receptors upregulate. More alcohol is needed to achieve the same effect, and without it, the brain is left in a state of neural hyperexcitability which can induce restlessness, anxiety and disrupted sleep.

Gray matter volume differences

A UK Biobank imaging study punlished in Nature Communications, studied over 36,000 participants’ intake in UK alcohol units (one unit ≈ half a standard beer). Regression models showed that going from one to two units per day was associated with lower brain volume equivalent to approximately 2 additional years of age. Whereas the increase from two to three daily units corresponded to lower brain volume equivalent to 3.5 additional years.

A separate UK Biobank analysis published in NeuroImage: Clinical followed over 25,000 participants whose alcohol intake had been recorded approximately ten years earlier. It was found that alcohol was associated more strongly with lower grey matter volume than smoking, BMI or hypertension within that statistical model.

White matter and blood vessels suffer too

A meta-analysis in Translational Psychiatry found consistent white matter damage to the corpus callosum and cingulum in heavy drinkers (regions important for processing speed and executive function).

And alcohol has been associated with increased blood pressure in a dose-dependent manner. Over time, this can contribute to small vessel disease, white matter hyperintensities and microbleeds (areas of vascular damage visible on brain scans). For the ageing brain, this vascular burden can be particularly harmful.

Neuroinflammation adds up

Chronic alcohol activates microglia, the brain’s resident immune cells. A recent postmortem study of the orbitofrontal cortex found that people with alcohol use disorder showed a distinct reactive microglial phenotype, accompanied by reactive astrocytes, oxidative DNA damage and neuron loss. Mediation analyses suggested a cascade: reactive microglia drive astrocyte activation, which in turn contributes to neuronal damage.

Alcohol also weakens the gut barrier, allowing bacterial toxins into the bloodstream that can cross into the brain and fuel the same inflammatory cycle.

What happened to “a glass of wine is good for you”?

For decades, studies showed that moderate drinkers were healthier than abstainers. This finding has been substantially challenged in recent years.

The problem was that most studies lumped together lifelong non-drinkers with people who had quit. When studies corrected for this and examined only high-quality studies, the apparent benefit of moderate drinking disappeared entirely.

Mendelian randomisation studies, which use genetic variants to separate the effects of alcohol from the other lifestyle factors that tend to go along with drinking, also agree. Multiple large genetic analyses found that alcohol uniformly increases health risks with no protective threshold. And a very recent study published in BMJ Evidence-Based Medicine, the largest combined genetic study on alcohol and dementia to date, including over 2.4 million people, found that dementia risk rises steadily with no evidence of a protective effect at low levels.

The old J-curve was most probably an artefact of reverse causation: people who would later develop dementia had been reducing their drinking for years before diagnosis.

Is the damage reversible?

This is where the news is good. The brain is remarkably plastic.

In people with heavy alcohol use, gray matter volume begins to recover in early abstinence, with roughly half of total volume recovery occurring in the first month. After about seven months, cortical thickness has been found to return to control-equivalent levels in the majority of brain regions.

White matter takes longer: months to years, but abstainers show steady improvement.

Cognitively, a systematic review of longitudinal studies in PLoS ONE found that attention, executive function and memory largely recover within six to twelve months. After several years of sustained abstinence, many individuals approach control-equivalent performance across most cognitive domains, particularly when abstinence is sustained and comorbid vascular risk is managed.

The brain’s capacity to heal, even after years of heavy use, is one of the most encouraging findings in this literature.

So should we never drink again?

The best current evidence suggests that there is no completely risk-free level of alcohol consumption for long-term brain health. That's what the data shows. At the same time, the effect sizes at low consumption levels are small and risk is not destiny. Many factors shape long-term brain health and alcohol is one of them. The dose-response relationship appears non-linear, meaning that the greatest benefit comes from cutting back if you drink heavily. The difference between four drinks a day and two is vastly more consequential than the difference between one and zero.

And personally? I believe in moderation.

If a glass of wine with dinner brings you pleasure, connection and a moment of calm, I’m not going to tell you the science says you can’t have it. What I will say is that it’s worth knowing what alcohol does to the brain, especially if consumed regularly and long-term, so the choice is an informed one rather than one based on an outdated assumption that it’s somehow good for you.

I'm Eleni, a neuroscientist studying cognition and brain ageing. The Brain Brief translates complex neuroscience into something you can actually use. If you found this useful, consider subscribing or sharing it with someone who'd appreciate it.

Subscribe now

References & further reading

1. Mitchell, J. M. et al. Alcohol consumption induces endogenous opioid release in the human orbitofrontal cortex and nucleus accumbens. Sci. Transl. Med. 4, 116ra6 (2012).

2. Gardiner, C. et al. The effect of alcohol on subsequent sleep in healthy adults: A systematic review and meta-analysis. Sleep Med. Rev. 80, 102030 (2025).

3. Daviet, R. et al. Associations between alcohol consumption and gray and white matter volumes in the UK Biobank. Nat. Commun. 13, 1175 (2022).

4. Topiwala, A., Ebmeier, K. P., Maullin-Sapey, T. & Nichols, T. E. Alcohol consumption and MRI markers of brain structure and function: Cohort study of 25,378 UK Biobank participants. NeuroImage Clin. 35, 103066 (2022).

5. Spindler, C., Mallien, L., Trautmann, S., Alexander, N. & Muehlhan, M. A coordinate-based meta-analysis of white matter alterations in patients with alcohol use disorder. Transl. Psychiatry 12, 40 (2022).

6. Crews, F. T., Qin, L., Coleman, L., Vidrascu, E. & Vetreno, R. Cortical reactive microglia activate astrocytes, increasing neurodegeneration in human alcohol use disorder. Brain. Behav. Immun. 131, 106156 (2026).

7. Stockwell, T. et al. Why Do Only Some Cohort Studies Find Health Benefits From Low-Volume Alcohol Use? A Systematic Review and Meta-Analysis of Study Characteristics That May Bias Mortality Risk Estimates. J. Stud. Alcohol Drugs 85, 441–452 (2024).

8. Topiwala, A. et al. Alcohol use and risk of dementia in diverse populations: evidence from cohort, case–control and Mendelian randomisation approaches. BMJ Evid.-Based Med. 31, 13–22 (2026).

9. Durazzo, T. C., Stephens, L. H. & Meyerhoff, D. J. Regional cortical thickness recovery with extended abstinence after treatment in those with alcohol use disorder. Alcohol 114, 51–60 (2024).

10. Powell, A., Sumnall, H., Smith, J., Kuiper, R. & Montgomery, C. Recovery of neuropsychological function following abstinence from alcohol in adults diagnosed with an alcohol use disorder: Systematic review of longitudinal studies. PLOS ONE 19, e0296043 (2024).

I started this newsletter almost three months ago, from scratch. Today, more than 300 of you are here. Hello to all of you, and welcome!

And thank you, truly. It means a lot that you’ve chosen to spend your time here.

I thought this was the right time to properly introduce myself and get to know you a little bit better. So today’s newsletter will be slightly different.

My story

I began with a BSc in Applied Psychology because I was curious about how the mind works. During my undergraduate studies, I became especially fascinated by memory and cognition. When we later covered dementia, something clicked. I became determined to understand why some brains remain resilient as they age and why others become vulnerable.

So, I went on to complete an MSc in dementia and other neurodegenerative diseases, and later a PhD in biomedicine, where I studied prevention, mental health, resilience and vulnerability in the context of Alzheimer’s disease. My thesis work examined how psychiatric symptoms, stress and biological changes interact years before dementia becomes clinically visible.

Today, I’m a postdoctoral researcher at a university hospital. I now study post-stroke recovery and dementia. I examine how biological signals (such as blood-based biomarkers or network-level damage) and resilience factors like education and lifelong enrichment shape cognitive recovery trajectories.

Across all of it, the thread has remained the same:

What protects the brain and what makes it vulnerable? In what contexts?
And how early can we detect those changes?

Whether we’re talking about aging, mental health, stroke, dementia or everyday stress, the underlying theme could be summarized as resilience under pressure.

Why The Brain Brief exists

Honestly? I started this newsletter because I noticed a gap between the way neuroscience is discussed in research settings and how it’s presented in everyday life.

Oversimplified headlines and brain myths travel fast. But nuanced, mechanism-based neuroscience rarely does. And I wanted to change that. I wanted to build a space where the science is taken seriously without losing clarity. A space where complex ideas can be explored carefully, and not just consumed quickly. So, here we are.

Everything I write is grounded in peer-reviewed research and I try to be honest about what the science actually says, including when the answer is 'we don't know yet.'

A few things about me outside the lab

I’m Finnish-Greek, which means that I take food and sauna culture very seriously (and yes, there’s actual research on the latter, I wrote about it).

Outside work, you’ll usually find me doing pilates, cooking something new or deep in an audiobook. I recently moved to Switzerland and I’m still adjusting to the fact that everything here runs exactly on time. I do love the chocolate and cheese, though!

Subscribe now

Now I’d love to hear from you

I'd love to get to know you a bit better too and make sure I'm writing what's actually useful to you. There are 4 poll questions (anonymous) and 2 open questions for the comments, and I'd really appreciate your answers, whenever you get to them. You can answer the polls and comment even if a lot of time has passed since this post first came out.

And I’d love to hear in the comments:

What brought you to The Brain Brief? Was it a specific post, a topic you were searching for or something else?

And lastly, what is one brain-related question you’ve been thinking about recently?

Thank you for being here, I hope you have a lovely rest of the week!

- Eleni

This is one of the questions I spent years exploring during my PhD and it’s still actively debated in my field. The conversation (at least in the media) often swings between two extremes: depression will give you dementia or mood changes are just part of normal aging. The evidence is more nuanced and the answer depends on timing, biology and which brain circuits are affected long before cognition is.

In a recent post, I explored why the pathological changes of Alzheimer’s often start decades before a diagnosis. Depression sits squarely in that story, because when it appears, why it appears and which symptoms dominate all influence its connection with dementia.

What we know: depression and Alzheimer’s disease are associated

Depression is common. It has been estimated that about one in five people experience at least one episode in their lifetime (although regional differences exist). Multiple large studies have shown that a history of depression is linked to a higher risk of developing Alzheimer’s. A recent study of 1.4 million people published in JAMA Neurology found that depression roughly doubles the incidence of dementia across 20 years of follow-up. Also, the Lancet Commission’s 2024 dementia report estimated that mid‑life depression could contribute up to about 3 % of dementia cases worldwide.

But association doesn’t equal causation. The fact that these two conditions are linked doesn’t tell us which one is driving the other or whether something else entirely is going on.

So let’s look at what might be happening.

Possibility one: depression changes the brain in ways that make it more vulnerable

One explanation is that depression, especially when it’s recurrent or happens during early or midlife, gradually makes the brain more vulnerable.

The most studied pathway involves the stress system. Chronic depression often hyperactivates the HPA axis (the body’s central stress response), driving prolonged cortisol release. Over time, this can affect areas like the hippocampus and prefrontal cortex, increase neuroinflammation, and, at least in animal models, promote the build-up of amyloid-beta and tau, the hallmark proteins that define Alzheimer’s dementia.

There’s also a behavioural side. Depression often disrupts sleep, reduces physical activity and can lead to social withdrawal. All of these have independently been linked to a higher dementia risk.

This “risk factor” interpretation has strong epidemiological support. Early- and midlife depression has been consistently associated with roughly a twofold increase in Alzheimer’s risk. Stronger studies have tried to rule out reverse causation by excluding people who developed dementia within the first several years after their depression was assessed. The association still holds, suggesting it’s not simply that people were already on the path to dementia when their depression was recorded.

A recent study published in The Lancet Psychiatry added an interesting layer to the story. The researchers followed over 5,800 middle-aged adults for 23 years and found that midlife depression was associated with a 27% higher risk of dementia, but in adults under 60, this increase was driven entirely by six specific symptoms. These included loss of self-confidence, difficulty facing problems and reduced social warmth. Each of these carried about a 50% increased risk over the follow-up period. Other symptoms we might expect to matter, like sleep disturbance or low mood, showed no significant link.

In other words, which depressive symptoms you experience in your 40s and 50s may matter more than the diagnosis itself.

Possibility two: depression is the brain’s earliest response to pathology that’s already there

Alzheimer’s pathology begins accumulating in the brain decades before any noticeable memory problems. And some of the earliest regions affected are circuits that regulate mood and emotion.

This means that in some people, depression or depressive symptoms may be a consequence of pathology and neurodegeneration that’s already quietly unfolding.

This idea has gained a lot of evidence too. Longitudinal studies in cognitively healthy individuals have shown that higher amyloid and tau burden (but particularly tau) are associated with worsening mood symptoms over time.

A recent tau‑PET study showed that tau burden in amygdala (one of the brain regions associated with emotional processing) was linked to both worsening depressive symptoms and declining memory performance, with the strongest associations in APOE ε4 carriers (the strongest genetic risk factor for Alzheimer’s). Entorhinal and hippocampal tau remained stronger predictors of memory decline overall, while amygdala tau uniquely captured mood‑related change.

Our own research supports this too. In an analysis of 681 participants from the Alzheimer’s Disease Neuroimaging Initiative, we found that higher neuropsychiatric symptoms, including depressive and anxiety symptoms, were associated with greater tau burden up to two years later, suggesting that neuropsychiatric symptoms may be among the earliest signals of underlying pathology.

Research indicates that clinically significant depression occurs in roughly 15–40 % of individuals with Alzheimer’s dementia, while anxiety disorders occur in about 9–39 %, often emerging before cognitive symptoms appear.

The genetics are complicated and inconclusive

I won’t go too deep into the genetics here, but it’s worth noting. Depression and Alzheimer’s share some genetic overlap. A 2021 study estimated around 98 shared causal variants between the two, with a key locus in TMEM106B, a gene involved in cellular waste clearance.

But shared genetics doesn’t mean one causes the other. A large 2025 genetic analysis tested whether depression causally increases Alzheimer’s risk and found no significant effect in either direction. Importantly, the authors showed that previous studies suggesting genetic overlap had been inflated by the inclusion of “proxy” cases (people classified as having Alzheimer’s based on family history rather than a clinical diagnosis). When only clinically diagnosed cases were used, most of the genetic overlap disappeared with one exception: the TMEM106B locus, which showed colocalization regardless of how Alzheimer's cases were defined.

Overall, the genetic data hint at some shared biology, but the epidemiological link between depression and Alzheimer's likely runs through stress, inflammation or behaviour.

So which is it?

Likely both, depending on when depression or depressive symptoms show up.

Depression appearing decades before cognitive symptoms in your 30s, 40s or 50s, more likely represents a genuine risk factor. It may contribute to brain vulnerability through chronic stress, inflammation and the behavioural changes that come with it.

Depression emerging for the first time in later life, especially without a prior history, is more suspicious for prodromal disease. It may be the brain’s earliest response to pathology that hasn’t yet crossed the threshold for cognitive symptoms. A 2025 study showed that people who developed depression at age 75 or older had significantly higher Alzheimer’s risk and faster progression compared to those with earlier-onset depression, even after accounting for age.

What this means

Having depression doesn’t mean you’ll develop dementia. The association is real but statistical. Most people with depression will never develop Alzheimer’s. Depression is one of many modifiable factors that may shape long-term brain health.

Treating depression matters, though not only for dementia prevention. Evidence showing that treating depression reduces dementia risk is still limited. But effective treatment may help with sleep, social engagement and physical activity, all of which independently protect the brain.

Midlife symptoms deserve a closer look. Research suggests that specific symptoms like loss of self-confidence, difficulty coping and reduced warmth toward others may matter the most in midlife.

Late-life mood changes may not be “just ageing.” New mood changes in later life, especially with apathy or cognitive complaints should prompt assessment.

Mood and memory share circuitry. The tau imaging studies show the brain regions processing emotion and those supporting memory overlap and are vulnerable to the same pathology. Protecting one often means protecting the other.

This is an evolving area of research, and there's still a lot we don't fully understand. But paying attention to how we feel may be just as important as worrying about how much we remember.

References & further reading

• Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet. 2024.

• Elser H, Horváth-Puhó E, Gradus JL, et al. Association of early-, middle-, and late-life depression with incident dementia in a Danish cohort. JAMA Neurol. 2023.

• Frank P, Singh-Manoux A, Pentti J, et al. Specific midlife depressive symptoms and long-term dementia risk: a 23-year UK prospective cohort study. Lancet Psychiatry. 2025.

• Gilchrist L, Spargo TP, Green RE, et al. Depression symptom-specific genetic associations in clinically diagnosed and proxy case Alzheimer’s disease. Nat Mental Health. 2025.

• Monereo-Sánchez J, et al. Genetic overlap between Alzheimer’s disease and depression mapped onto the brain. Front Neurosci. 2021.

• Markova TZ, Fonseca CS, Ciampa CJ, et al. Defining the contributions of tau pathology in the amygdala to increasing depressive symptoms in aging. Alzheimers Dement. 2025.

• Aguilar-Dominguez P, Palpatzis E, Akinci M, et al. Temporal associations of neuropsychiatric symptoms, demographics and amyloid with subsequent tau burden in older adults. J Prev Alzheimers Dis. 2025.

• Jang YJ, Kim M-J, Moon YK, et al. Changes in dementia risk along with onset age of depression: a longitudinal cohort study of elderly depressed patients. BMC Psychiatry. 2025.

• Ismail Z, Smith EE, Geda Y, et al. Neuropsychiatric symptoms as early manifestations of emergent dementia: provisional diagnostic criteria for mild behavioral impairment. Alzheimers Dement. 2016.