Seth's Blog

Background. I took 4 T of flaxseed oil during the day (instead of just before bedtime) and measured its effect with a cursor test. The test was how accurately I move the cursor from one point to another with a single movement. The result was a sharp improvement — some of which lasted, some of which didn’t. (Just to be perfectly clear: what’s varied is not my daily amount of flaxseed oil. It’s the time of day I take it. I’m varying the time between a short-lived peak in omega-3 concentration, which happens shortly after ingestion, and doing the cursor test. Usually they are far apart. The interesting data are what happens when I move them close together.)

New data. I tried the same thing again. Here are the results.

The green line shows when I took 4 tablespoons of flaxseed oil. I took the oil at 8:30 am. The first test after that, at 9:30 am, showed the improvement. (In previous measurements of the short-term effects, it has taken closer to 2 hours to see the maximum effect.)

Here is a longer view, which emphasizes the constancy of the pre-test baseline.

For comparison, here are the earlier results.

Conclusions. When I take 4 T of flaxseed oil, it creates for a few hours a higher-than-usual concentration of flaxseed oil in my blood. I’m pretty sure the active ingredient is omega-3. This has two effects:

The Freakonomics blog posted a letter from reader Jeffrey Mindich about an interesting traffic experiment in Taiwan. Timers were installed alongside red and green traffic lights:

At 187 intersections which had the timers installed, those that counted down the remaining time on green lights saw a doubling in the number of reported accidents . . . while those that counted down until a red light turned green saw a halving in . . . the number of reported accidents.

Great research! Unexpected results. Simple, easy-to-understand design. Large effects — to change something we care about (such as traffic accidents) by a factor of two in a new way is a great accomplishment. This reveals something important — I don’t know what — about what causes accidents. I expect it can be used to reduce accidents in other situations.

It’s another example (in addition to obstetrics) of what I was talking about in my twisted skepticism post — the downside of “correctness”. There’s no control group, no randomization (apparently), yet the results are very convincing (that adding the timers caused the changes in accidents). The evidence-based medicine movement says treatment decisions should be guided by results from controlled randomized trials, nothing less. This evidence would fail their test. Following their rules, you would say: “This is low-quality evidence. Controlled experiment needed.” The Taiwan evidence is obviously very useful — it could lead a vast worldwide decrease in traffic accidents — so there must be something wrong with their rules, which would delay or prevent taking this evidence as seriously as it deserves.

Writing about advances in obstetrics, Atul Gawande, like me, suggests there is a serious downside to being methodologically “correct”:

Ask most research physicians how a profession can advance, and they will talk about the model of “evidence-based medicine”—the idea that nothing ought to be introduced into practice unless it has been properly tested and proved effective by research centers, preferably through a double-blind, randomized controlled trial. But, in a 1978 ranking of medical specialties according to their use of hard evidence from randomized clinical trials, obstetrics came in last. Obstetricians did few randomized trials, and when they did they ignored the results. . . . Doctors in other fields have always looked down their masked noses on their obstetrical colleagues. Obstetricians used to have trouble attracting the top medical students to their specialty, and there seemed little science or sophistication to what they did. Yet almost nothing else in medicine has saved lives on the scale that obstetrics has. In obstetrics . . . if a strategy seemed worth trying doctors did not wait for research trials to tell them if it was all right. They just went ahead and tried it, then looked to see if results improved. Obstetrics went about improving the same way Toyota and General Electric did: on the fly, but always paying attention to the results and trying to better them. And it worked.

Is there a biological metaphor for this? A perfectly good method (say, randomized trials) is introduced into the population of medical research methods. Unfortunately for those in poor health, the new method becomes the tool of a dogmatic tendency, which uses it to reduce medical progress.

Scientists are fond of placing great value on what they call skepticism: Not taking things on faith. Science versus religion, is the point. In practice this means wondering about the evidence behind this or that statement, rather than believing it because an authority figure said it. A better term for this attitude would be: Value data.

A vast number of scientists have managed to convince themselves that skepticism means, or at least includes, the opposite of value data. They tell themselves that they are being “skeptical” — properly, of course — when they ignore data. They ignore it in all sorts of familiar ways. They claim “correlation does not equal causation” — and act as if the correlation is meaningless. They claim that “the plural of anecdote is not data” — apparently believing that observations not collected as part of a study are worthless. Those are the low-rent expressions of this attitude. The high-rent version is when a high-level commission delegated to decide some question ignores data that does not come from a placebo-controlled double-blind study, or something similar.

These methodological beliefs — that data above a certain threshold of rigor are valuable but data below that threshold are worthless — are based on no evidence; and the complexities and diversity of research imply it is highly unlikely that such a binary weighting is optimal. Human nature is hard to avoid, huh? Organized religions exist because they express certain aspects of human nature, including certain things we want (such as certainty); and scientists, being human, have a hard time not expressing the same desires in other ways. The scientists who condemn and ignore this or that bit of data desire a methodological certainty, a black-and-whiteness, a right-and-wrongness, that doesn’t exist.

How to be wrong.

In this interesting video about losing weight, Paul McKenna, a British hypnotist, recreates a study in which people ate food blindfolded. In the study, they ate one-quarter less when blindfolded than when not blindfolded. This doesn’t impress me; nothing is stopping the blindfolded subjects from eating more at later meals. But it makes me wonder how not seeing your food affects flavor-calorie learning. It might make it stronger (you’re less distracted) or it might make weaker (the sight of food acts like glue to strengthen flavor-calorie associations — there is actually evidence for something like this).

While wearing noseclips while eating with others is too weird, wearing dark sunglasses might not be. And what about listening to music (for distraction) while you eat? My calorie learning experiments are continuing; eventually I should be able to test these possibilities.

Thanks to Gary Skaleski.

Scientific American Online has started a eight-part series about self-experimentation. Part 2 is about Morgan Spurlock of Supersize Me.

The results I described in the previous post surprised me because (a) my performance suddenly got better after being stable for many tests and (b) after the improvement, further practice appeared to make my performance worse. I’d never before seen either result in a motor learning situation. If you can think of an explanation of the result that practice makes performance worse, and animal learning isn’t your research area, please let me know.

Learning researchers used to think of associative learning as a kind of stamping-in process. The more you experience A and B together, the stronger the association between them. Simple as that. In the 1960s, however, several results called this idea into question. Situations that should have caused learning did not. The feature that united the various results was that in each case, learning didn’t happen when the animal already expected the second event. If A and B occur together, and you already expect B, there is no learning. Theories that explained these findings — the Rescorla-Wagner model is the best known, but the Pearce-Hall model is the one that appears to be correct — took the discrepancy between expected and observed — an event’s “surprise factor” — rather than simply the event itself, to be what causes learning. We are constantly trying to predict the future; only when we fail do we learn.

In my motor-learning task, imagine that the brain “expects” a certain accuracy. When actual accuracy is less, performance improves. Performance stops improving when actual accuracy equals expected accuracy. The effect of more omega-3 in the blood, and therefore the brain, was to increase expected accuracy. (One of the main things the brain does is learn. If we do something that improves brain performance in other ways, it is plausible that it will also improve learning ability.) Thus the sudden improvement. The decrement in accuracy with further practice came about because, when the omega-3 concentration went down, actual accuracy was better than expected accuracy. Accuracy was “over-predicted,” a learning theorist might say. So the observed change in performance was in the opposite-from-usual direction. Accuracy got worse, not better.

Related happiness research. “Christensen’s study was called “Why Danes Are Smug,” and essentially his answer was it’s because they’re so glum and get happy when things turn out not quite as badly as they expected.”

The more I played racquetball, the more accurate my shots became — the more control I had. It was a kind of learning: learning to place the ball. I was fascinated by how little we knew about how that learning took place. I studied associative learning in my own research. The motor learning during racquetball resembled associative learning in the sense that my actions (hitting the ball with the racket) were shaped by what happened next (accuracy of placement). Yet I knew nothing non-obvious about motor learning.

This background of ignorance is why I find my latest flaxseed oil results so interesting. As I’ve posted, I’ve started using a new test in which I use the touchpad to “toss” the cursor from one spot to another (that is, move the cursor with a single finger movement), and measure how close it “lands” to the target. The function relating cursor position to finger position on the touchpad isn’t simple.

Of course I wanted to see how flaxseed oil affected performance on this task. I doubted that it would. This task is untimed. No time pressure. It is like shooting free throws. Most of the previous tasks I’ve used that have shown a flaxseed-oil effect have been tasks where you respond as fast as possible. My balance test was go at your own pace, but it involved a huge amount of computation. Balancing my body on one foot for several seconds seemed to involve a lot more computation than moving a finger about an inch.

Usually I take 4 tablespoons of flaxseed oil just before bedtime. One recent day I took it much earlier and did the toss test at 30-minute intervals before and for several hours afterward.

Here are the results plotted as a function of test session number.

Here are the same results plotted versus the time of the test:

Here is a close-up of the crucial data:

About two hours after I drank the flaxseed oil, my accuracy got worse. Then it slowly got much better. The amazing thing about the improvement is that it reached a maximum long after you would think that the effects of the flaxseed oil had worn off. My overall level of omega-3 is high because I take 4 T flaxseed oil per day. The effect of shifting when I drink the 4 T is just to change the timing of a short-lived peak. Usually that peak happens when I’m asleep and my omega-3 levels are reasonably constant while I’m doing the test. In this case the peak happened while I was doing the test.

I’ll discuss what this might mean in a later post.

You’ve heard a million times that there is “good” cholesterol (HDL) and “bad” cholesterol (LDL). Recently I got my cholesterol measured. My LDL was 151 mg/dl. The test results were written on a form that said your LDL should be “Below 100 mg/dl. Below 70 mg/dl if High Risk.” The person who handed the results to me said, “These are not the best results . . . ”

How concerned should I be? A 2005 study in the Journal of the American Geriatric Society surveyed several thousand “elderly people [who] were recruited from a general Italian population, and mortality was monitored from 1983 to 1995.” The emphasis of the study was on whether LDL was good or bad.

People with more LDL lived longer. You read that correctly. For women, mortality was lowest at the highest level of LDL. For men, mortality was higher at the highest level of LDL (60 deaths/1000 patient years) than at the next highest (50), but still lower than at the lowest level of HDL (90). Going from the lowest to the highest levels of LDL is associated with a one-third decrease in mortality, in other words.

What should I make of my 151 mg/dl? To convert to the units of the paper (mmol/L), I needed to divide by 39. 151/39 = 3.9. Looking at the graph relating mortality to LDL, an LDL concentration of 3.9 mmol/L is where the mortality vs LDL function reaches a minimum — the lowest mortality. According to this study, my LDL is optimal.

Thanks to Joel Kauffman.

A recent EW Popwatch post compared several YouTube versions of Leonard Cohen’s Hallelujah, much like I did here. It’s like Leno and Letterman telling the same joke. We have an Instinct of Connoisseurship, Veblen would say.