Can Brain Training Really Reduce Dementia Risk?

Can a computer-based brain-training program really reduce the risk of dementia?

In this episode of Brain Matters, host Neil Broomfield speaks with Dr Eddy Roccati about a recent headline claiming that specific cognitive training had an “astonishing” effect on dementia risk.

Together, they unpack the research behind the story — a long-term analysis of the ACTIVE study, which followed older adults for up to 20 years and found that one specific type of speed-of-processing training, when reinforced with booster sessions, was linked with a lower risk of diagnosed Alzheimer’s disease and related dementias. 

But does this mean all brain-training games prevent dementia? Not quite. Neil and Eddy slow down the headline, explore what kind of training was actually tested, discuss relative versus absolute risk, and consider what listeners should look for when judging brain-health claims in the news.

Based on:

New Scientist article: “Specific cognitive training has ‘astonishing’ effect on dementia risk”

Study: Impact of cognitive training on claims-based diagnosed dementia over 20 years: evidence from the ACTIVE study

Listener Reflection Questions

• Did hearing more about the study change how convincing the “brain training reduces dementia risk” headline felt to you — and why?
• Why does it matter whether a 25 per cent reduction is a relative reduction or an absolute reduction?
• What kinds of mentally stimulating activities do you already enjoy — and how do they fit into your wider brain-health habits?

Brought to you by the ISLAND Project and Wicking Dementia Research and Education Centre at the University of Tasmania. To continue the discussion and for more information visit https://island.mooc.utas.edu.au or email island@dementia.utas.edu.au

Transcript

Copy Transcript Copied

Speaker: 00:01
Neil Broomfield:

Welcome back to Brain Matters, the podcast where we go behind health headlines and take a closer look at what the research actually tells us, and just as importantly, what it doesn't. Of course, our objective here is to highlight the strengths of these research studies and to critique them rather than criticize them. In doing so, we hope that we encourage you to think more deeply about the research behind the headlines. I'm Neil Broomfield, recording on Palawa land. Today, we're looking at a headline that sounds pretty exciting, and I quote: "Specific cognitive training has astonishing effect on dementia risk." This is based on a long-running study known as the active trial. The headline suggests that a specific type of cognitive training was linked to about a 25% lower risk of dementia diagnosis over time. Many of us have seen claims about brain training before, things like apps, games, and puzzles. But do they really work? And can they actually reduce dementia risk? To help us examine that in detail, I'm joined today by Eddy Roccati from the Wiking Dementia Research and Education Centre. Eddie, thanks so much for being here.

Speaker 1: 01:19

Eddy Roccati:

Thanks, Neil. It's a pleasure to be with you today.

Speaker: 01:22
Neil Broomfield:

Would you like to tell us a bit about your expertise and interests?

Speaker 1: 01:25
Eddy Roccati:

My name's Eddie Roccati. I'm the Senior Research Fellow for ISLAND at the Wicking Centre. My primary focus is looking at how we can look at objective measures of cognitive function and other things, like blood-based biomarkers of dementia risk, and what we can do in our life that can reduce our risk of those objective measures.

Speaker: 01:44
Neil Broomfield: 

So maybe it's time to look at the headline, shall we? The headline is based on research from the so-called Active Study, as I said before. This is an acronym for advanced cognitive training for independent and vital elderly. Bit of a mouthful. At its core, this study asks a big question. If older adults do different types of cognitive training, does it alter their chances of being diagnosed with dementia later in life? And importantly, not just in the short term, but over 20 years. Why is this important?

Speaker 1: 02:22
Eddy Roccati: 

It is really important. First of all, I love the ACTIVE acronym. As you say, it is a bit of a mouthful to explain, but we know with with ISLAND, we've got a really good acronym, and I think ACTIVE is an equally good acronym for this particular research study. What's important here is that these researchers essentially wanted to know something pretty simple but quite hard to test. If you train older adults' thinking and memory skills, or we can call their cognition, for a few weeks, does that training actually protect them from developing dementia decades later? This study actually happens in the late 90s and early follow-ups a few years ago from the same trial, which which I should say is based out of the the US,  showed that cognitive training improved performance on tests, and it even helped with managing everyday activities, like managing finances and medications, and that persisted up to 10 years after the initial training. But nobody really checked the ultimate real-world outcome, which in this case is an actual dementia diagnosis over a full 20 years. We're now two decades later from the initial intervention, and this study linked the original trial participants to their Medicare records over two decades later, to see who eventually got diagnosed with Alzheimer's disease or another related dementia, and if that differed between whether they'd be in in one training group or another. We should emphasize here is that there's a real difference between short-term and long-term outcomes. In this study, it was a long-term outcome. It was actual diagnosis of Alzheimer's disease.

Speaker: 04:02
Neil Broomfield:

Can I just interject here and go off at a bit of a tangent? You talked about using Medicare records. Now, in America, their Medicare system is very different from ours. Are there a good spread of people in the Medicare system there, do you know? Does it represent a broad scope of socioeconomic groups, for example?

Speaker 1: 04:24
Eddy Roccati:

As with any sort of study, there's always inherent biases in how we can actually measure an outcome. And yes, we have an actual diagnosed cases of of dementia. We have someone going to a hospital or a GP or a clinic and actually getting a diagnosis of dementia, which is a really objective outcome measure. But in that objective outcome measure, there's all sorts of things that can play a role. We know that Medicare records are a very good standard of evidence, but there are some things that can impart a bias there. For example, socioeconomic status. This particular one found a difference between people who are on a program called Medicare Advantage, which we can think of a bit more like a private health aspect of a Medicare program. And again, in that, some people don't present to healthcare professionals with issues. There's all sorts of different reasons why that might not be the case. So, yes, it's avery good standard of evidence, but there's also things we need to take into consideration when we consider a diagnosis as an outcome measure.

Speaker: 05:28
Neil Broomfield: 

Thank you for clarifying that, Eddy. Let's now take more of a closer look at what kind of study this was. One reason it has attracted attention is the design of it. It wasn't just observational, it was a randomized, controlled trial. This means, I think, that participants were randomly assigned to different groups, which helps make these groups comparable and reduces some bias about which you spoke a moment ago. And then years later, researchers linked those participants in this one to health records to see who have been diagnosed with dementia. Isn't that right, Eddy?

Speaker 1: 06:06
Eddy Roccati:

That's correct, Neil. To use another acronym, a randomized controlled trial or an RCT, is a real gold standard in research methods because it accounts for a lot of the biases in group selection. And because we randomize group assignment, we randomize what intervention or a control group that we put a person into, it allows us to make claims about causality over a correlation, which we might find in an observational trial. RCTs are often considered strong evidence because if people had chosen their own level of training, we might see that healthier or more motivated people might pick the intervention group over other less health-minded people. Random assignment of group cancels out that. It cancels out  the factors that we know about and also the factors that we don't know about. When we look at the gold standard of of research methods, RCTs are right at the top. However, they still don't remove all levels of uncertainty.

Speaker: 07:07
Neil Broomfield: 

Okay. So getting back to this particular research, years later, researchers linked these participants to health records to see who have been diagnosed with dementia, I believe.

Speaker 1: 07:19
Eddy Roccati:

Correct. So this is a really strong paper and a strong research design. But with any interpretation of any scientific article or news presentation, we should always be cautious. For this reason, in this in this study, we know that only 70% of the original sample from the 90s made it into the analysis. This was due to a lot of different things, which some and some not are due to biases. So some people weren't matched with Medicare records, some people were on this Medicare Advantage program and some people passed away in that period. So there's only 70% of the people in that study. We can see that the diagnosis that happened two decades later, is to do with a lot of different things like health care access, being prone to engage with healthcare professionals, having family around you that might notice symptoms and suggest that you go and see someone about those and education, just in general healthcare education. So all these sort of factors could shape who gets diagnosed, separate from who actually has the underlying disease.

Speaker: 08:23
Neil Broomfield: 

That all points to the fact that we've got here a really strong study design, I think. But obviously, and as always, we need to look closely, still, at what was actually tested. So let's examine what the brain training in fact was. Because brain training can mean a lot of different things. In this study, the result wasn't about brain training in general, it was about a very specific type, that is, computer-based speed of processing training. Another mouthful. How might these compare with memory and recent training, Eddy?

Speaker 1: 09:00
Eddy Roccati:
This study was looking at speed training and what you can do to visualize what the participants went through. You can picture a computer screen, and what comes up in flashes in the centre of that computer screen, you have to spot it, identify it within your peripheral vision and also in your central vision.

Neil Broomfield:
Can you give us an example of what they were looking at?

Eddy Roccati:
If you imagine a computer screen and an object flashes in the centre of that screen.

Speaker: 09:28
Neil Broomfield:
Any old random object?

Speaker 1: 09:30
Eddy Roccati: 

Any old random object, a chair, a couch, a milkshake. That can happen in your central vision or peripheral. So you need to quickly identify and respond to that stimulus. This relies on quickly taking in visual information, responding under time pressure, and keeping up as the tasks become more challenging. I'll talk about this in a little bit, but the tasks got harder and harder. That's a really important part of this study and and the findings.

Speaker: 09:59
Neil Broomfield:

I'm rather interested in this because peripheral vision varies an awful lot from individual to individual.

Speaker 1: 10:06
Eddy Roccati:

Yes, it does.

Speaker: 10:08
Neil Broomfield:

And did they, do you know, try and select in or out people with good or bad peripheral vision in order to equal the playing field, so to speak?

Speaker 1: 10:18
Eddy Roccati:
I don't know that in particular. I'm not sure if it was part of their inclusion or exclusion criteria, but because it was an RCT, that's one of the things you can say, would have been accounted for as a known confounding factor. We know that if it's a visual task, that there's going to be your visual system.  If you wear glasses,  those sort of things are going to have an impact on whether or not the intervention has an effect. And in this study, because it's an RCT, we can assume that that known confounding factor was adjusted for in the analyses.

Speaker: 10:49
Neil Broomfield: 
We've talked about confounding factors in previous Brain Masters episodes. So I imagine our listeners are getting kind of accustomed to the term. Could we explain speed of processing as you just have in everyday language?

Speaker 1: 11:03
Eddy Roccati:
In that task, we can consider it against something like a memory or a reason training, if we want to use those kind of terms. Memory training gives people techniques like mnemonics. Remember mnemonics from high school or primary school? You know, Roygiv gives you the color of the rainbow. Every good boy deserves fruit. Remember that from learning piano. I remember that one. I'll never forget that one. That's from learning piano, it teaches you the notes on a scale. That uses strategies to encode and recall lists, things that aren't necessarily intuitive. Reasoning training is more like methods for spotting patterns and solving problems with a logical sequence, like figuring out what comes next in a series If you're doing the seven times table, seven, fourteen, twenty-one, trying to figure out what the real solution is there. So those trainings are more about teaching a trick or a strategy that's going to optimize your response. Whereas speed training, in contrast, there's no real strategy to speed training. It's very proactive and it's very much about not how you do it, but how quickly you can do it. In this particular study, this is structured and adaptive training. And what I mean by adaptive is that it got harder as people got better at it. That's really important in any sort of study of cognitive function when you're trying to measure someone's cognitive capacity. If you make a task and it's very easy, all you'll find is that people all just score perfect, 100%, 10 out of 10, A plus. There's no real value in learning that. What we need to do with studying cognitive function is have tests get harder and harder. So we figure out where someone's threshold is at, where their where their peak cognitive capacity is at. When you multiply that by, thousands, hundreds of thousands of people that are doing a task, that's when you start to see a nice kind of bell curve of cognitive function and capacity.

Speaker: 13:08
Neil Broomfield:
I'm now, Eddy, going to look at what the participants actually did. They completed about five to six weeks of training with sessions of about an hour. And then importantly, some people went on to do booster sessions, months or years later. It turns out that these booster sessions are really important. Is that right, Eddy?

Speaker 1: 13:30
Eddy Roccati:

They are. That was really the key finding of this study, that addition of the booster sessions was what actually resulted in the reduced risk of dementia over time. They did their 10 initial sessions of training over six weeks. It was only participants who had done eight out of those ten sessions who are randomized again to the boosters. The boosters happened at about 11 months and 35 months after the initial training. They always consisted of up to four 75-minute sessions in the same style as the original training. So this again imparts an element of bias. You've only got people in the randomized booster group who have been part of the study for 80%. That already shows high level of engagement with the study, which could be reflective of healthcare access, education, and just general knowledge about dementia risk reduction. The booster sessions, why they are important is this idea of reinforcement. And it's the same sort of thing,  when you go and visit a physio or you try and increase your capacity at the gym, it's that reinforcement that's really important for giving you a benefit. If you just did a task, a lot of times in an intensive period, and then just gave up on it, we might not see a benefit in that task unless you continue to reinforce that behavior and continue to reinforce those new kind of cognitive pathways in the months and the years that follow down the track.

Speaker: 15:00
Neil Broomfield:

Okay. Let's now look at what they did find. The key result behind the headline seems to be that people in the speed training group who completed booster sessions were about 25% less likely to be diagnosed with dementia over the following 20 years. That sounds like quite a lot to me.

Speaker 1: 15:21
Eddy Roccati:

It does sound like a lot, but what we really need to be cognizant of here is that there's a difference between relative and absolute risk. Let's look at that. I want you to imagine you have two towns, you have town A and town B. In one year in town A, two people get an illness and only one person in town B gets that illness. If we were to compare town A and town B, you could accurately say that town B has 50% lower risk than town A. That is relative risk. That is the difference, that's what we're looking at here. We're looking at the relative risk between two towns. But really, in absolute terms, only one person extra got the disease. In this study, we are looking at relative risk. The big headline has come from the fact that the relative risk between the groups, there was a 25% difference in them. But when we look at the actual, prevalence of the diagnoses in the group, it's 49% versus 40%. You really have to look at the absolute reduction, which is only nine percentage points. Basically, only nine fewer people out of every 100 being diagnosed over 20 years. And you might say that that is not really a big difference. The broadcasting of this study has lent in on a big difference, 25% risk reduction. But in reality, it's probably a little bit less than that.

Speaker: 16:51
Neil Broomfield:

Do you know what this reminds me of, Eddy, is these adverts that maybe cosmetic companies put out. 99% of people say that this works. You've opened the can here for us. I thank you for that.

Speaker 1: 17:06
Eddy Roccati:

I've seen those ones. There's always an asterisk after that. If you untangle that asterisk and see where that rabbit hole goes, I'm sure you'll find that there's similar things going on in how they massage group differences and and numbers, to get the result they want.

Speaker: 17:22
Neil Broomfield:

Quite so. A really important point here is that other training didn't show the same effect. Memory training didn't reduce dementia risk. Reasoning training didn't either. And even speed training without booster sessions didn't show a benefit. Therefore, we can see that the finding is really quite specific. What can we surmise from that, Eddy?

Speaker 1: 17:45
Eddy Roccati:

This finding is very specific. Although the headline implies it's quite broad, the finding with this particular study, is really specific. This study wasn't cognitive training reduces dementia risk broadly. It was a narrow specific combination of speed of processing training plus completing booster sessions. When you compare two groups, there was a difference. Speed training alone, without any boosters, that had no benefit. Memory and reasoning had no benefit with or without boosters. Really, only one of six possible combinations tested out significant in terms of difference in dementia diagnoses. We can't really extend this study to brain training works. This was one very specific, very structive, adaptive computer task that showed a benefit over time.

Speaker: 18:41
Neil Broomfield:

Well, naturally, people will wonder why might this type of training make a difference?

Speaker 1: 18:46
Eddy Roccati:

That's the natural next question. The paper offers a few explanations as to why might be the case. This is classic when you read a scientific article. The discussion section is really just a way of researchers trying to justify their conclusions, trying to link it back and synthesize it with previous research. The first explanation that these researchers are offered, is that it targets a different kind of learning. This was different to memory or reasoning learning, where you might learn a strategy and once you have that strategy, that unlocks completing a test. Speed training doesn't have a strategy. It's all about performance and it's about doing things as quick as possible. This study, in that it's targeting a different learning system, that might provide more resilience. That's one explanation. Another explanation is that the task was adaptive. It never really got easy. It got continually harder, and that continually pushed people to the edge. So prior research in meta-analysis has shown how adaptive speed training tasks outperform non-adaptive ones. This concept might be supporting neuroplasticity, which is the brain's ability to form and strengthen neural connections. We might be seeing a benefit to just challenging cognitive behaviors and making the brain work and do more than it has done. The act of doing something difficult might actually be why we're seeing positive results two decades later.

Speaker: 20:22
Neil Broomfield:

Excellent. Okay. Now let's spend a moment on limitations, because this is where a lot of the learning sits. Are there any major limitations people should think about here?

Speaker 1: 20:33
Eddy Roccati:

Yes there are. Every study has strengths and limitations. Some of the strengths of this study, if we want to do the positive things first, it was very long-term. We had 20 years of follow-up. It had a very strong, rigorous research design of randomizing and controlled trials. It was a very large and diverse sample, almost 3,000 people. And the outcomes were very objective, real-world diagnoses of dementia. We know that for certain because it was linked with Medicare records. There's always limitations. There's always going to be strengths, there's always going to be limitations. There were some reasons that we've spoken about, where some people didn't make it to the analysis, and that happened at various stages throughout the study. Some people didn't have healthcare access, might not have had education or a family situation to actually get to the diagnoses. We've spoken about how booster sessions were only offered to people who completed eight out of the 10 initial sessions. We can see that although they've adopted a really strong research design, there's always going to be some bias inherent in selecting people and doing anything over a long term. Overall, these findings are really specific and they're not really generalizable to global or overall brain health. But it is a really promising finding and one that shows how brain training does have potential to have positive impacts longer term. We might just be starting to scratch the surface over the longer-term impacts of these studies. So this is really positive to see one study showing an objective long-term outcome.

Speaker: 22:15
Neil Broomfield:

You know what, Eddy, this leaves me pondering how we best translate all this to our real lives. It seems that this is a good moment to bring this back to our island participants. Therefore, if someone hears this and thinks, should I start doing brain training, what would you say?

Speaker 1: 22:34
Eddy Roccati:

I would always say, with any activity advertise caution and look a little bit deeper from the findings. I would say, yes, always engage in something that is cognitively challenging. What we know is that anything that is cognitively demanding or cognitively stimulating is going to be protective for the brain. So doing things like crosswords, sudokus and wordles each day, that's a good thing, if it's challenging you. But not all of those training regimes have a strong evidence base. We're only starting to scratch the surface over these brain training regimes, which were a bit of a buzz in the 2000s. We're now two decades after that. We're only really now beginning to see the potential long-term impacts. So I would say the main takeaway from this study was that because it was adaptive and because it was really the act of the task getting harder and harder that we saw a benefit. I would say take that into your own life and look for things that are challenging, look for things that are cognitively demanding and try and engage with those because those are the things that are going to enhance your neuroplasticity, increase your cognitive load and really just buffer your brain to the longer-term effects of aging.

Speaker: 23:51
Neil Broomfield:

That is very helpful indeed. And of course, through the ISLAND project, we know that dementia risk isn't just about one thing. It's shaped by many factors: lifestyle, health, environment, engagement and so on. And just as we found in other episodes of the Brain Matters podcasts, this study is just one piece of that puzzle. As a summary, is there anything further you'd like to add to this very interesting piece of information?

Speaker 1: 24:18
Eddy Roccati:

Yes, thank you. We know there's lots of types of brain training. We know the brain is very diverse. Some levels of brain training require a different  cognitive capacity. Some are more enjoyable than others. What I try and say for for listeners, is try and look for challenges, look for cognitively demanding activities, but also see if you can find ones that you enjoy. Look for opportunities to combine a cognitively stimulating activity with another one, like physical activity, for example, going for a walk with a friend, picking up a new musical instrument, learning a language, joining a choir. There's so many things that can improve our brain health and be cognitively stimulating at the same time. Looking to multiply the benefits of that activity is only going to be a good thing. We know dementia risk can be modified throughout the life course. We know cognitive stimulation is important throughout life. Education shouldn't just be an early life endeavor. We should be constantly looking for things that are educating us and improving our brain health over time.

Speaker: 25:26
Neil Broomfield:

Well, Dr. Eddy Roccati, thank you so much for helping us understand all of this today.

Eddy Roccati:
Yeah, thank you for having me, Neil. It's always a pleasure to chat.

Speaker: 25:33
Neil Broomfield: 

The takeaway is cautiously encouraging. This study suggests that a very specific kind of cognitive training, done in a structured way and reinforced over time, may be linked to a lower risk of dementia diagnosis. But, it's not all brain training, it's not a guaranteed prevention, and it doesn't replace the bigger picture of brain health. Before we wrap up, here are a few questions to reflect on. Did hearing more about the study change how convincing the brain training reduces dementia risk headline felt to you? And why? What kinds of mentally stimulating activities do you already enjoy? And Eddy has covered some of this just a moment ago. And how do they fit into your wider brain health habits? Do you think it would help you to consider taking on a new challenge, like perhaps studying a new language, and so on? And now it's time for me to thank not only Dr. Eddy Roccati for sharing his expertise and knowledge with us today, but also to thank Drew Stansbury for his behind-the-scenes technical skill in enabling this recording, and of course, Flo Sward for her production of this whole program. Most of all, thank you, dear listeners, for listening, and we'll see you next time. This has been Brain Matters.