Brain Tracking: A Study on Soy

Posted March 17, 2104


Some reports have suggested that the consumption of soy products is harmful to your brain function and to your health in general.  I set out to test the hypothesis that my reaction time would be adversely affected by eating soy.

Experimental design

The study consisted of three week-long conditions (with a fourth, post hoc condition added later, as described further down):

  1. Baseline:  I made no effort to change my diet.  (My vegan diet does normally include a fair amount of soy-based products.)
  2. No soy:  I eliminated soy products from my diet.
  3. Heavy soy:  I made a conscious effort to eat more soy than I would usually eat.  I ate plain tofu, smoked tofu, fermented tofu, soy creamer, okara burgers, soy cheese, and soy-based “faux meats”.

On each day during this three-week period, I would take the reaction-time test twice: First in the morning-to-early-afternoon (I'm calling these the “early-day sessions”), and then again a few hours after dinner (“late-day sessions”).  Each testing session involved 35 reaction-time trials as described in more detail in my earlier write-up about my caffeine experiment.

A few days after the completion of these three phases, I showed the results to Seth Roberts.  He suggested adding a fourth phase: a return to baseline (in which I would again eat what I normally eat, without attempting to change my consumption of soy).  So I continued taking the reaction-time test daily, although I was not able to maintain the same regular pattern of early-day and late-day tests.  The third graph below shows all four phases plotted on a single graph.

Results and discussion

The results were not consistent with the hypothesis that eating soy is harmful to brain function.  Surprisingly, my scores became significantly faster during the study, possibly because I made a special effort to “psych myself up” to perform as well as I could on each test.  I think I tried harder than I would if I were just taking the test outside of a structured study.  The final graph below shows all of my data since I first started using the reaction-time application.  The improvements in performance do not occur in a regular pattern.

I don't know why my results are inconsistent with prior work.  Perhaps people differ in their sensitivity to soy.  Or perhaps I've been eating so much soy for so long that I've made myself resistant to any changes that might result from relatively short-term fluctuations in level of soy consumption.

While I was eating the high-soy diet, I did not notice any obvious changes in my health, in my ability to concentrate, or in the quality of my sleep.  I felt the way I usually do.  This fact is reassuring to me, as eliminating soy from my diet would be difficult for me to accomplish and is not something I intend to try.

In the three graphs below, the blue lines are linear regressions.  In the fourth (final) graph, the blue curve indicates loess smoothing.

The following graph shows all of my tests to date.  The vertical gray line marks the beginning of the soy study:

Readers are welcome to contact me at with any questions.