Brain Training Study

[Indy3/4]

Without analyzing this study, that impressive '29%' figure sorta sounds to me as if it is a relative risk measurement.

In regards to relative risk, below is a very simplified example of how researchers and pharma companies could portray the relative risk benefits of statin use. Let's say this pretend study includes 5,000 participants in the statin testing segment and 5,000 participants in the placebo segment.

At the end of this pretend study, only 1.0 statin participant (0.00020%) had a cardiovascular event versus the 1.4 (0.00028%) placebo participants had a cardiovascular event. That is an absolute risk change of only 0.00008%. Yet, the pharma company claims a 29% reduction of risk benefit.

How so? It's simple when misleading relative risk arithmetic is at work: 29% equals (0.00008/0.00028) times 100. Voila, a minor absolute improvement turns magically into a "OMG, I think I need that statin-stuff, Doc!" moment.

Another relative risk example is from my own Prometheus analysis of genes/snps: 18x risk of Type-1 Diabetes...this variation appears to significantly increase risk of Type-1 diabetes...your relative risk approximately 18x, which changes your lifetime risk from a very small 0.04% to a still very small 0.75% chance. Your odds are not 75%, they are 0.75%, which is less than 1%.

[floramaria]

Unfortunately, a closer look at the original paper reveals a different picture.

Supplemental Table 2 shows that the speed processing training can be relied on (with 95% statistical confidence) to have an effect of only 0.2 percent. Hazard ratio 0.71 (0.5 – 0.998, 95%CI).

In other words, the claimed 29% reduction in risk is actually 29% +/- 28.8% (95% confidence interval).

What does that mean? Play Double Decision if you enjoy it and have time and money to spare. But don't rely on it to reduce your risk of dementia.

Thank you for the clarification, Searcher. I am a ;little sad at the thought that the Brain HQ Training has no value. I have been dedicating about 90 minutes a week to it. Time I could be using to learn Chinese!

[NF52]

Like Searcher, I also followed the link to the study, While I wish I had Searcher's expertise with the statistical analysis, I also noted the following methodological issues with this study:

It appears that fully 50% of the original participants dropped out, some after one "training" session. It is impossible to know how the cognitive status of that group would have affected the results. [This is the reason why high schools in the U.S. now have to count graduation rates based on how many students were in the 9th grade, not how many were in 12th grade who graduated. I once visited a failing NYC high school that had a 75% graduation rate if you looked at 12th graders, and ignored the 50% of the 9th grade class who dropped out before ever getting to 12th grade, They were closed by the state.)
Only people who finished the first round of 8 training sessions were "offered" the chance to have two separate booster sessions. Again, that's like offering AP classes only to kids who got all A's in 9th grade and then claiming that AP classes make kids more likely to do well on AP tests than those kids who never were offered the chance to take the classes. People with good cognitive reserve, good health and strong interest in brain games, continued through all 18 possible training sessions. Those people would have been less likely to develop dementia than those who elected not to participate.
They allowed people or family to self-identify who had been given a diagnosis of dementia!

I think anytime someone claims that a specific produce reduces dementia by 29% when used 18 times, you should be very cautious. Doesn't mean it might not be one arrow in your quiver--but don't count on it working.

Here is what the authors acknowledged in their Discussion section: (emphasis added)
Limitations are also noted including the absence of a clinical diagnosis, attrition during follow-up, and the method of booster training assignment. ACTIVE did not have dementia as a primary outcome, so results are from secondary analyses. We acknowledge that the association between number of training sessions and the risk for dementia could be due to reverse causality. As such, we have appropriately moderated the interpretation of the exposure to training results toward association with risk. Our dementia criteria were defined a priori [8]. There are of course limitations to these criteria, for example, self- and proxy-reports of dementia diagnosis are not infallible, MDS-IADL function was self-reported and thus biased, low MMSE is not a sensitive dementia marker, and overlap among the dementia criteria was low. A definitive study of the efficacy of cognitive training on dementia requires a clinical diagnosis as the primary outcome.[/quote]
http://www.trci.alzdem.com/article/S235 ... 8/fulltext

[floramaria]

NiHao~
I have written to Brain HQ to see if they care to comment on Searcher's assessment of the results of the study. Will post any reply "One arrow in my quiver" is all I was ever hoping for from the Brain HQ Training. I don't count on anything working in isolation.
Zai jian

[circular]

Searcher I nominate you, if you are ever so inclined and have the time, to write us a Wiki page about interpreting research papers. You explain it all so clearly.

[slacker]

Floriamarie - I suspect learning Chinese would be an equally helpful way of building brain neurons...and if you enjoy it, perhaps superior!

[Searcher]

Searcher I nominate you, if you are ever so inclined and have the time, to write us a Wiki page about interpreting research papers. You explain it all so clearly.

Thanks for your kind words, circular. I try to understand things.

I'd recommend any short textbook on statistical inference.

See also Karl Popper's little book (I think it's called Conjectures and Refutations).

[edit: The book I was thinking about is called [url = http://strangebeautiful.com/other-texts ... covery.pdf ]"The Logic of Scientific Discovery"[/url]. I realize it's not as little as I remembered.]

That should get most people quite far.

Family and friends keep pestering me for advice. Because they'd have to pay other people for it. I keep assembling practices that are based on sound scientific research, for my own benefit and wellbeing. It's good to have my views tested and refined in the crucible of this enlightened and thoughtful group. I do write regularly to my mailing list of family and friends. Some of them are adventurous (foolish?) enough to join me in healthy living practices.

For those who are really keen to take independent decisions based on primary research, I'd recommend further steps in addition to the two books mentioned above.

Study medicine.

Study psychology.

Study epidemiology.

Study economics.

Study business.

Pick a narrow topic and explore it every which way. Write a convincing overview that would convert a sceptic.

Enjoy the process of separating the wheat from the chaff.

All that should keep you off the streets and out of mischief for a while.

Incidentally, I've switched off notifications (as I tend to do everywhere). So if I miss some comments in this group, please excuse me.

[Bettylacy]

Thanks for the clarification searcher and Indy 3/4. I was horrible at statistics, but you make it easier. I’m still confused about Relative Risk verses Absolute risk. Joseph Jebelli, a British neuroscientist, has published at book called “In Pursuit of Memory...” and was recently featured on NPR https://www.npr.org/templates/transcrip ... =575055148 . He spoke about relative risk of the APOE4 gene saying we have relative risk but no absolute risk. Can you clarify this statement? I agree with circ that adding to knowledge base on stats would be really helpful. Just thinking about a brief talk at our next meet-up on stats might prove interesting.

For me I’m in line with Stavia doing functional acitivites to boost our BDNF rather than the brain training. Like her I started taking music lessons with an instrument I had never learned before- the ukulele. THat was 5 years ago and now I’m having a ball. I started a local group that is flourishing. We now are requested to play at local gigs like senior centers, winery’s and art galleries.

Wish list for me would be a future study using N-Quant, which measures hypocampal volume and compares to age appropriate normals, after doing some of these activities to see if volume is changed.

[chrissyr]

I have been enjoying exercise while learning dance moves and socializing--- I know it's good for me! And working, writing reports, going to the women's march, chillin with the dog. Cooking and enjoying good food.
I have a hard time sticking with Brain HQ. However I do not discount it totally. There is very little other training for me where I would be asked to work on processing speed, visually scanning the periphery while keeping my eye on another target, quickly shifting focus (these are difficult). And I feel they could translate to seeing a car in the periphery and quickly slamming on the brake or turning to avoid an accident. Or in general, working the eyes and brain a way that they don't normally do.

[Searcher]

we have relative risk but no absolute risk. Can you clarify this statement?

Bettylacy, I doubt he said exactly that.

A made-up example might help.

The absolute risk of a man getting pregnant might be estimated at 1 in a trillion. (10^-12)

Administering female sex hormones to the man might multiply that risk by 1,000 times.

The relative risk of pregnancy attributable to the female sex hormones is 1000.

Despite the relative risk being 1000, the treated man still has an absolute risk of only 1 in a billion. (10^-9)

Forgive the outlandish example, but I hope it helps clarify the terms used.


Congrats on your music-making activities, which seem to bring you a lot of enjoyment and a rich social life. At the same time, they're intrinsically beneficial to your brain. Here's an example of how:

http://www.jneurosci.org/content/37/24/5948

Structural changes in the brain are useful for monitoring outcomes, but functional changes are even more impactful. There are often surprises when anatomy or pathology are compared to function.