Science in Action: Mysterious Mental Improvement

For a few years, I’ve been making daily measurements of how well my brain works. I got the idea after I found that omega-3 (from flaxseed oil) improves my balance. It improved other mental functions as well. Tim Lundeen, using an arithmetic test, found similar results. These results suggested to me there might be a lot we don’t know about how our environment affects our brain.

If so, tracking myself might turn up interesting anomalies — clues to big environmental effects. The first one I found involved flaxseed oil. There turned out to be a short burst of improvement after I took it. The second anomaly I found also involved flaxseed oil. When I switched from Chinese flaxseed oil to American flaxseed oil (Spectrum Organic), a few days later my arithmetic scores suddenly improved. Something was wrong with the Chinese flaxseed oil.
The third revealing anomaly — which doesn’t involve flaxseed oil — happened yesterday (see below). Each point on the graph is one testing session.  Each session consists of 32 simple arithmetic problems (e.g., 3+5, 7-6) and takes about 3 minutes. I use R on my laptop to collect the data. I type the answer or the last digit of the answer (e.g., if the answer is 13 I type “3”) as fast as possible. Here are the results from almost a year of this task:
2010-03-09 arithmetic time vs time of test

The Y axis is the time it took to do one problem. Yesterday, the graph shows, I suddenly got much faster. My score dropped about 50 msec — far more than normal variation.

What caused the drop? I can think of four possibilities:

1. The test was standing. Usually I test myself sitting.

2. The test happened after I’d been walking on my treadmill for 10 minutes. That too was very rare.

3. I’d had about 30 g of butter 2 hours earlier.

4. I’d stood on my cobblestone mat 2 hours earlier.

My guess is that it’s #2 (10 min walking). The previous record low score, in January, might have come after I did Dance Dance Revolution for 30 minutes or so.

17 Responses to “Science in Action: Mysterious Mental Improvement”

  1. q Says:

    laptop clocks are not always good for calculating time intervals. they sometimes ‘jump’. (the reason for this is that it’s potentially very bad for computer clocks to ever go backwards — software almost never covers that possibility) so manufacturers calibrate clocks to go slowly and then jump forward discontinuously. one laptop i worked with jumped 7.2 seconds once or twice an hour. if you get more than one such outlier in a row you can probably rule out a clock anomaly though.

  2. Timothy Beneke Says:

    What strikes me about the flaxseed oil is how emotionally and physically soothed I feel for about half an hour after I take it. It has a calming effect; I’m tempted to say that it induces a state of serenity, or near-serenity. It certainly assuages hunger far more than an equivalent number of carbohydrate of protein calories does.

    I’d be very curious to see ANS physiology measures or other measures of things associated with calmness and soothing — oxytocin or vagus nerve activation.

    Also, does anyone else find flaxseed oil soothing?

  3. seth Says:

    q, the faster-than-usual performance I saw on the latest day came from many individual trials, not one. And of course the problem you describe would make performance slower than usual, not faster.

  4. tetsuwanatom Says:

    that’s interesting!

    i suppose the average of the scores is probably more to do with repetition of the task => being able to complete the task faster.

    it would be interesting to see if the average for a month would change is you only take the tests after an exercise.

  5. Michael Says:

    You should consider varying the math problems so you don’t get used to them. To help my arithmetic skills when I was a trader, I set up an excel worksheet which made random two digit numbers and added or subtracted them (also randomized) to a second two digit number. I hide the solutions column when I printed off the worksheet.

  6. Scot Says:

    Arousal facilitating a dominant/well-learned response?

  7. Alex Chernavsky Says:

    Seth, do you have a suggested list of computer-based tests (memory, arithmetic, reaction time, etc.) that self-experimenters can download to assess mental acuity?

  8. Nancy Lebovitz Says:

    Have you tried Dual N Back— it’s supposed to improve fluid intelligence. I haven’t been keeping records, but I think it’s been doing my mood and mental focus some good.

  9. seth Says:

    No, Alex, sorry, this is in the home-brew phase.

    Nancy, I’ve heard of the task but haven’t tried it. Thanks for the link.

  10. vimal Says:

    What brand of flax seed oil do you use?
    I use

    Also is the math test here comparable to your homebrewed test?

  11. seth Says:

    I’ve used Spectrum Organic and Barlean’s flaxseed oil. I never found a difference.

    The math test I do is simpler than the math test on that page. I only use addition, subtraction, and multiplication — no division. And the two numbers being combined are always single digits (0-9). My test doesn’t have the awkward interface of the web test–you keep your hands on the number keys and type a single digit. And each trial is separated from previous trials by a few seconds at least. I go at my own pace.

  12. q Says:

    you’re right about the problem i described making it slower.
    guess i failed the math test.

  13. Dan Says:


    Yes, very soothing. I started taking flaxseed oil a month ago, and noticed the calming effect pretty much immediately. I’m not a particularly anxious person, so I’m curious to see if people with anxiety problems respond the same way.

  14. ben lipkowitz Says:

    it seems like you are measuring response time more than anything, with times in the sub-second range. perhaps you should do less trials of larger digit additions, so that you’d have to do multiple operations and juggle numbers. I think long pauses for “duh” moments ought to be included in the data.

    could you post your R code?

  15. seth Says:

    ben, the reason for using very simple arithmetic problems is so that little learning will be involved — I already know really well that 3 + 1 = 4. Use of larger numbers will allow more learning. I don’t know really well that 17+39 = 56.

    I have posted most of the R code — see the R code category.

  16. peter Says:

    interesting results. now to turn it into a real self-experiment, you might need a few more controls to turn it into a significant trial. placebo, dose-dependent use, time of day (stress v. relaxed times), some positive control (e.g. ritalin), etc…

  17. Physiological Computing : Better living through affective computing Says:

    […] In my opinion, both arguments are sound.  I would add another strand to the same debate – it seems to me that interest in self-tracking systems, such as those lifelogging experiments described regularly on the excellent Quantified Self blog, is on the rise.  The technology to self-monitor is used principally for sports training, see this range of apparatus described recently in Wired.  Ambulatory systems to monitor physiology are generally used for fitness training but their mere existence opens the door for self-experimentation, outside the laboratory (to borrow from Prof. Picard’s title), for the people by the people.  As one example, see this blog posting by Seth Roberts on the effects on omega-3 on physical and mental ability: if you’re really interested in self-tracking and self-experimentation, I’d recommend his paper published recently in Medical Hypotheses – full text is linked from this QS post. […]