A number of years ago, I met a Manhattan psychologist who  had the idea of adapting stroke rehab software for average folks who were experiencing memory problems.  These were not people who were sick but, rather, people in middle age who were not as sharp as they once were.  They would come to his office and sit in front of a computer screen for an hour, negotiating mazes and manipulating images.  The psychologist convinced me that this was the wave of the future, and that some day, in addition to working out at the health club, we'd be working out at the brain gym.  Of course now that we know that physical exercise improves memory and may even slow dementia, there's reason to question  if brain gyms will become obsolete before they become mainstream. 

      While the Manhattan doctor's approach--take something that was made for people with brain injuries and tailor it for people whose pathology was simply that they were getting older--attempted to be based in science since the software had been tested on and was known to work for stroke victims, the doctor wasn't trying to figure out if, in fact, his approach really worked.  In other words, he wasn't doing a placebo controlled double blind study to assess whether the exercises he was using actually improved anyone's memory.  

       This is true, it turns out, for most brain exercise promoters.  This is worrisome, because people with memory problems can be desperate, and reach for any rope that's tossed their way, even if it's expensive as well as unproven.  If you find yourself tempted to plunk down a couple of hundred dollars for a hand-held brain game machine, don't settle for some user testimonials or the manufacturers hype:  ask the company to direct you to the research on which it make its claims.  Make sure the research wasn't done by someone affiliated with the company, that it was done in a reputable place by reputable researchers, with a super-long digit span of participants.  Chances are the company--though it will tout its board of scientific advisors--will come up empty-handed. Caveat emptor.

      One online set of "brain" exercises that attracted me right off the bat was one that promised to help me improve my "digit span."  A digit span is a string of random numbers, like 23938393835575, that one can hold in one's memory.  Typically people can remember 7 numbers going forward and six in reverse.  Testing digit span--which happens, for instance, when you take an IQ test--is a way of getting at how good your working memory is.  It stands to reason that the longer the span you can recall, the better your memory.  It also stands to reason that you can both boost your memory and increase your IQ if you can increase the number of numbers you can recall.

      It was with this in mind that I spent about half an hour a day for a month, sitting in front of my computer, learning to remember longer and longer strings.  Was I able, at the end of the month, to increase my base-line digit span?  Absolutely, just as the company and its august board of advisors claimed, and for this reason the exercises could be considered a success.  But did this translate into remembering to buy avocados at the grocery store?  Well, no.  Did it make me smarter--not obviously, though if I were to take an IQ test before and after my score might have have gone up considerably.  (That should suggest something about the fallacy of quantifying intelligence.)

     But here is another thing that happened:  armed with my newfound ability to recall twenty numbers at a time, I showed up one day to participate in a memory study at New York University.  One part of the study required participants to go through a battery of neuropsychological tests, and one of those tests was of digit span.  I was pretty psyched, since I knew I'd do well, which would boost my overall score, which could be the difference between being told I had some kind of cognitive impairment and being told I was normal or even better than that. 

     The tester said a bunch of numbers and instead of zipping them back to her, I was lost.  I could remember a couple at the beginning and a couple at the end and that was that.  This is because the brain gym I was using not only spoke the numbers outloud, it flashed the string on the screen.  I heard them, I saw them, and then, when I said them outloud to myself, I heard them again. No wonder my digit span expanded--I was processing the numbers three times, using two different senses.  But in the testing room there was only one modality and it was not enough.  In the end, despite my efforts, I pretty much remembered the same number of numbers as everyone else, which made me, well, average.