Reaction Time as a Measure of Health

Six years ago I started using a reaction-time (RT) test (a test where you press a key in response to something as fast as possible) to track my brain function. I took the test daily. It must use only a small part of the brain but I assumed that something that made me faster would probably improve overall brain function. Behind this belief, which I call better RT, better brain, were countless studies of brain anatomy and physiology, which had shown that neurons and glial cells all over the brain share many features. Cells in different parts of the brain are much more alike than different. More support for this assumption was that certain doses of flaxseed oil improved both RT and other measures of brain function, such as balance.

I also assumed that changes that improved RT would probably improve overall health — what I call the better RT, better body assumption. It was less plausible than the better RT, better brain assumption because the cells in different organs of the body differ so much. They have many similarities but also many differences. I believed it for two reasons. (a) Flaxseed oil not only improved several measures of brain function, it improved my gums, no doubt because it reduced inflammation. It had been far from obvious that improving gums was so easy or that flaxseed oil (in the right dosage) would do so. The assumption better RT, better body had made a surprising prediction, you could say, that turned out to be true. (b) The brain gets much the same blood as the rest of the body. (Not exactly the same, because of the blood-brain barrier.) In the same way, all plug-in electrical appliances use the same house current. Just as all appliances have been designed to work well with that current, all our organs should have been shaped by evolution to work well with same mix of nutrients. You can’t feed your brain differently than your heart.

When I discovered that butter improved RT, the better RT, better body assumption made a second even more surprising prediction: Eating more butter improved my health. This contradicted the claims of all mainstream health experts, who say saturated fats cause heart disease. I stuck with my assumption — I still eat a lot of butter. The data I’ve seen since then has supported my conclusion. For example, my Agatston score got better, not worse, after a year of eating lots of butter. The Agatston score is currently the best predictor of heart disease.

I recently found more support for the better RT, better body assumption. Several studies have found that RT is a good predictor of health (better RT, better health). Even more impressive, it is a better predictor than many of the predictors we already know of. The RT test used in these studies is close to the test I now use, which I developed independently. The RT test in these studies involves showing a digit (0-4), after which the subject presses one of five keys (labelled 0-4) as fast as possible. My current RT test is very similar but uses 7 digits instead of 5.

A 2005 study looked at the oft-reported correlation between higher IQ and lower mortality. The IQs and RTs of about 900 persons were measured in 1988. Deaths until 2002 were noted. RT was associated with lower mortality, even after taking out associations with smoking, education and social class. RT and IQ are correlated (better RT, higher IQ). When the RT-death association was removed, IQ no longer predicted death. So RT does a good job of capturing whatever it is about IQ that predicts mortality.

A 2009 study compared RT to more conventional health predictors (“risk factors”). About 7,000 subjects were followed from 1984 to 2005. RT in 1984 was a good predictor of all-cause mortality compared to classic risk factors. Smoking was by far the best predictor, followed by RT. RT was a better predictor than physical activity, blood pressure, a questionnaire measuring “psychological distress”, resting heart rate, waist/hip ratio, alcohol intake, and body mass index.

A third study, based on the same subjects as the 2009 study, found that amount of decline (slowing) in RT (from one test to a second test seven years later) predicted death. People with more decline were more likely to die.

All this supports studying how your RT is controlled by your environment, especially what you eat. You have to choose wisely what to study. The point is not to be as fast as possible regardless of everything else. Lots of drugs (stimulants, such as caffeine) decrease RT for short periods of time. I doubt they improve health. (If they harm sleep, they probably worsen health.) What makes sense is to look for two things: 1. Poisons. Things that slow you down. I discovered that tofu did so. I gave several reasons for thinking that tofu affects many people this way, not just me. Billions of people eat tofu, unaware of this possibility. 2. Deficiencies. Study things that are missing from your life now but were likely to be present when we evolved. It is quite plausible that our ancient ancestors ate more omega-3 (in fish, but also in flaxseed) and more animal fat (from big animals, but also in butter) than we do now. My data suggest omega-3 and animal fat are nutrients necessary for health whose importance mainstream nutrition researchers have not fully appreciated.

My RT data have shown me there’s a lot I didn’t/don’t know about how my food affects me. Maybe everyone can say that. Unlike almost anyone else, however, I can reduce my ignorance myself. I don’t need to rely on experts.

17 Responses to “Reaction Time as a Measure of Health”

  1. Dirk Says:

    Hi Seth, which tool do you use to measure RT?

  2. sam Says:

    I’ve been thinking about trying RT tests for a Long time now. As far as I remember you show calculations and have to enter the last digit of the result?

    Or do you have a plain reaction test now?

    Seth: The test I use is what psychologists call a “choice reaction time” task. See a digit, type that digit.

  3. Mehmet Says:

    I did not know there was such a strong correlation between overall health and reaction time. That was very interesting. Measuring brain function is certainly easier than measuring your liver function :)

    Regarding the RT test, I have a similar question with sam. Have you changed the test from a simple arithmetic test to a digit retyping test?

    Previously, I coded another version of your simple arithmetic test. I started experimenting with it but I saw that trying to be fast reduced the accuracy of my results. I accepted a maximum 4 errors in 30 questions for the experiment to be valid. However, there is a significant difference in trying to make zero errors and allowing a margin of four errors.

    In a digit retyping test, I think it maybe possible to be more accurate and reduce the margin of error. In the end, if you spend enough time, it seems to me that simple arithmetic test boils down to a digit retyping test.

  4. Jeff Winkler Says:

    That’s really interesting. Remember in Eat, Fast where they did the balancing and ruler drop tests? Backs up your thesis (though Mosley did not re-test those measures after the intervention)

    What I find fascinating about your RT tests is that they’re sensitive cheap and quick, with lots of longitudinal data. So you use them like a tremor detector. Signal detection. That turns up things like soybeans.

    Seth: I remember those two tests only vaguely. I will watch it again. I agree, my RT tests are sensitive cheap and quick. And easy. Room for improvement: Less sensitive during learning period, comparisons across people a bit unclear (timing may depend on your computer), not especially accessible, speed/accuracy tradeoff a possible problem.

  5. Ripken Holt Says:

    Is there a application online that could be used to test one’s reaction time? I would like to test mine but do not have any knowledge of coding?

    Seth: I am distributing the reaction time test I use. If you are interested in using it, please contact me. It is not online, it runs under R.

  6. as Says:

    This is off topic, but for flaxseeds, isn’t phytic acid an issue? Or do you assume that if you have you have good results with regards to reaction time and dental health, the flaxseeds can’t be harming you too much?

  7. Nancy Lebovitz Says:

    Any thoughts about psychological(?) factors like ADD or hypervigilence?

  8. Alex Says:

    Mehmet brings up a good question: Seth, in your testing, what do you do with wrong answers? Ignore, track separately, allow some number of them, just never make mistakes?

  9. Alex Says:

    Nancy Lebovitz, the distractability and hyperfocus of ADHD are two sides of the same coin: impaired ability to *direct* attention. ADHD is badly named. It’s not really attention deficit; it’s attention-control deficit. It would not surprise me if attention control rose and fell with general brain function.

    I wish I had a dime for every friend who’s told me, “My kid doesn’t have ADHD; s/he can focus on a video game for hours on end.”

  10. Nancy Lebovitz Says:

    Hypervigilance isn’t the same thing as hyperfocus. Hyperviligance is constant awareness of possible threats, and it’s associated with PTSD and/or anxiety.

    http://en.wikipedia.org/wiki/Hypervigilance

  11. Seth Roberts Says:

    to measure RT I use an R program I wrote myself.

  12. Adam Says:

    Koanic Soul turned me on to this online arithmetic game: http://arithmetic.zetamac.com/game?key=d10589f4

    You can File > Save Page As to save it to your HD if you like

  13. Mike Garcia Says:

    I would like for you to send me the test if you would? Thanks, Mike

  14. GTR Says:

    Apparently in 19th centaury people had lower reaction times than today. More sleep perhaps?

    http://phys.org/news/2013-05-victorian-era-people-intelligent-modern-day-counterparts.html

  15. Ian Says:

    An online reaction test can be found here.

    http://www.humanbenchmark.com/

  16. Steve R Says:

    I’d think testing RT and tracking results would be a great app for a smartphone, but all I see are games, nothing that really keeps track of the times and could be used for research. Have you heard of anything for iPhone or Android?

  17. Steve R Says:

    Just discovered this after leaving the above comment, there’s some interesting stuff here, both some possible apps and some discussion about whether these devices are really usable for testing RT

    http://cogsci.stackexchange.com/questions/506/are-there-any-cognitive-test-or-test-suites-available-on-the-ipad