Is Dyslexia a superpower?

Everybody kind of wondered why Ali had this miraculous ability to
dodge a punch, and some doctors I talked to said it might've actually
been connected to his dyslexia, that when you're dyslexic your brain
works differently. When you learn to read, your brain gets re-wired so
that you focus really carefully on one thing, you can concentrate
really hard on those letters on the page.

But when you never learn to read, your brain remains more accessible
to outside forces. You can maybe hear and understand two conversations
at one time, because your brain hasn't been re-wired by the process of
learning to read.

Ali, because he learned to read very late and never really very well,
may have been better at picking up visual clues than most people. He
may have been able to see little signs in his opponent's body that
suggested when and where the punch was going to come. It's a
fascinating theory, I think.

This is an excerpt from an interview on Fresh Air with Jonathan Eig, the author of a biography of Muhammad Ali ¹. I was intrigued but skeptical of this idea, so I decided to do some research.

I quickly learned that boxing isn't the only domain in which dyslexics have a purported advantage. Malcom Gladwell popularized the idea that dyslexia is a "desirable difficulty" in his 2013 book David and Goliath, because dyslexics are forced to develop alternative learning strategies.

There might be something to this. Evidently, there are a surprising number of talented dyslexic artists ². Also, apparently a lot of famous, successful people are dyslexic. Like, a lot.

Here are just a few lists of famous, successful people with dyslexia. Combine all of those names and you basically get a Who's Who list of the 19th and 20th century: Albert Einstein, Thomas Edison, Nikola Tesla, Steve Jobs, Bill Gates…

Thumb through an American history book and pick a name at random: George Washington? Dyslexic. Abraham Lincoln? Yup. Ben Franklin? Check. Nelson Rockefeller, Henry Ford, the Wright Brothers, Walt Disney, General Patton, Dwight Eisenhower? You guessed it, all dyslexic.

At this point, I decided something weird was going on. Either:

1. A shadowy cabal with the power to manipulate Google search rankings is running a sophisticated, pro-dyslexia, propaganda campaign
2. A lot of well-meaning people for some reason mistakenly believe that dyslexia is a superpower, or…
3. Dyslexia is legitimately a superpower

My prior for #1 is pretty low, but I wasn't sure what to make of #2 vs. #3. It's well-established that the brain can compensate for the loss of one sense by reallocating the unused gray matter to another, as when a blind person's visual cortex is appropriated for speech-processing ³. So the idea that decreased verbal ability could be offset by increased abilities in another domain sounds plausible, at least.

Fortunately, psychologists have spent a lot of time and grant money studying this.

Most of the papers I found focused on the purported advantage in visuo-spatial abilities. I started with this study ⁴ by Winner, von Károlyi, et. al., which was funded by the International Dyslexia Association. This study was cited by almost every other paper that I read, so apparently it was pretty important.

The authors ran three studies in which they administered various tests of spatial abilities. They even threw in a drawing exercise, to test for artistic ability. Contrary to their predictions, not only did the dyslexic group perform no better than the control group, they actually performed worse on almost every test.

von Károlyi was busy in 2001 - in addition to the previous paper, she published a study that focused on a possible link between dyslexia and enhanced "global" visuo-spatial processing ⁵. The premise of the study is fascinating and goes something like this:

Studies of brain-damaged artists suggest that the left-brain handles local, detail-oriented, visual-spatial processing, whereas the right-brain does global, holistic spatial processing. Language processing, which dyslexics struggle with, is mediated by the left-brain. Therefore, if dyslexics did have an advantage in visual-spatial processing, it may be confined to right-brain dominant tasks that involve global (as opposed to local) processing.

It's a fascinating hypothesis, but the the experimental results were mixed. In the test of local/left-brain visual-spatial ability, there was no significant difference between the two groups. In the test of global visual-spatial ability, there was no difference in accuracy, but the dyslexic group did respond faster (N=62, p=.03). The author took this as evidence that global visual-spatial processing might be slightly more efficient in the dyslexic brain.

This was promising enough for the International Dyslexia Association to break out the checkbook for another study, so von Károlyi and Winner reunited for this 2003 paper ⁶, in which they successfully replicated von Károlyi's results (N=64, p=.0003): the dyslexic group recognized "Impossible Figures" (a test of global spatial-processing ability) more quickly than the control group.

So the theory that Ali's dyslexia contributed to his success in boxing is sounding more plausible. After all, "holistic visual-spatial processing" sounds like psychology-speak for what boxers are doing when they anticipate their opponent's next punch. In a sport where you can literally die ⁷ by failing to move your chin quickly enough, you want that right-hemisphere firing on all cylinders.

At this point you're probably thinking, "If only we had high-resolution brain scans of dyslexics performing tasks that involve global visual-spatial processing, then we'd have indisputable evidence of a dyslexic advantage!"

Which do you want first: the good news or the bad news?

The good news is that Diehl et. al. stuck a bunch of people with dyslexia in an fMRI machine and quizzed them with von Károlyi's Impossible Figures ⁸, and they did see marked differences between dyslexic and typical brains.

The bad news is that the results still weren't 100% conclusive.

Their findings were consistent with von Károlyi's results in that the dyslexic group was slightly faster on some tests of visual-spatial ability, namely the Impossible Figures test and a test of mental rotation abilities. But, because of course they were, the dyslexic group was only faster when tested outside of the fMRI machine.

Also, the evidence with regards to von Károlyi's theory re: global vs. local processing was mixed: the dyslexics were faster to recognize Impossible Figures, but no faster on the other test of global spatial ability. However, it's difficult to draw any conclusions from this second test because it required the subject to recognize a target letter which is itself composed of smaller letters, like this letter "H" which is composed of the letter "S":

SSS      SSS
SSS      SSS
SSS      SSS
SSS      SSS
SSSSSSSSSSSS
SSSSSSSSSSSS
SSS      SSS
SSS      SSS
SSS      SSS
SSS      SSS

I don't mean to go full, arm-chair, experimental psychologist, but did no one that reviewed the design of this experiment notice a potential issue? Like the fact that their test of visual-spatial ability involves the alphabet? For a group whose membership criteria is essentially, "Has trouble with the alphabet?"

To their credit, the authors do acknowledge this shortcoming in their discussion of the results. And despite these frustrations, the fMRI data was revealing.

In neuro-typical brains, the Impossible Figures caused the brain to light up with activity, whereas printed text barely provoked a response at all. Relative to the neuro-typical group, the dyslexic brains were less stimulated by the Impossible Figures but significantly more stimulated by the text. In other words, the dyslexic group had to work harder than the neurotypical group to process text, but less hard to process the Impossible Figures.

Not only was the overall level of brain activity different between the two groups - the location of the activity differed, too. In neuro-typical subjects, text caused the left-hemisphere to light up, whereas the Impossible Figures predominantly activated the right-hemisphere. But there was little differentiation in the dyslexic brains: both the left and right hemispheres stayed active, regardless of the task.

The authors concluded that there is probably some trade-off between language processing and spatial processing, and that faster response times for a given task (i.e., reading vs. visual-spatial processing) are correlated with more efficient neural circuitry.

If there is a trade-off between language and spatial processing, then this raises an interesting question: are dyslexics born better at spatial processing? Or does the process of learning to read reorganize the brain at the expense of spatial processing? That is, does becoming a better reader make you worse at spatial visualization?

Unsurprisingly, the authors conclude that answering this question will require further study. If I have learned one thing from reading all of these psych papers, it's that the only thing that psychologists can ever be sure of is that, while their results are promising, nothing is settled and everything warrants further study.

As for me, I've had enough studies: I'm ready to answer the question that started this whole investigation: yes, Muhammad Ali probably had some atypical neuro-circuitry in his right-hemisphere that made him slightly better-than-average at dodging punches.

I'm still confused by something, though: what about Einstein and all of the dyslexic super-geniuses? Surely a small advantage in global visual-spatial processing can't account for what seems to be a massive overrepresentation of dyslexics in the Who's Who list of modern, Western history?

Ok, I lied. We need more studies.

First, let's get one thing out of the way: the case for dyslexics having generally superior spatial abilities is pretty weak. If people with dyslexia do have enhanced visual-spatial abilities, then they hide them very well when subjected to formal testing.

In "Reading disability and enhanced dynamic spatial reasoning: A review of the literature" ⁹, the authors compiled and analyzed the results of 21 studies that examined spatial-processing abilities in people with reading disabilities ("RD", in their terminology). The conclusion:

RD samples do not perform better on dynamic spatial reasoning tasks
than nRD samples. In fact, subjects with RD most often perform worse
than, or equal to, controls... The few studies that found a
significant RD advantage tended to deal with virtual or Impossible
Figures types of tests, and the RD advantage was most often limited to
measures of [reaction time].

Three of the four "studies that found a significant RD advantage" are ones that we've already seen: two are from Winner and/or von Károlyi, and the other is Diehl's fMRI study.

The fourth study is the triumphantly-titled, "A Virtual Reality Test (FINALLY!!!!) Identifies the Visuospatial Strengths of Adolescents with Dyslexia, Thereby Justifying The International Dyslexia Foundation's Continued Funding Of This Type of Research and Vindicating The Unsubstantiated Diagnoses of Dyslexia in Many Famous, Dead Scientists."

Just kidding, the actual title is just, "A Virtual Reality Test Identifies the Visuospatial Strengths of Adolescents with Dyslexia" ¹⁰. Also, my lawyers ¹¹ have asked me to emphasize that the previous paragraph constitutes a joke and that the authors of this paper include the Disclosure Statement: "The authors have no conflict of interest."

This study tested participant's spatial memory in "virtual reality," which apparently in 2009 meant, "a rendering of a 3D-environment on a laptop screen." The study's enthusiastic-sounding title belies the authors' circumspection about the experimental results, which suggested only a modest increase in spatial memory among the dyslexic group. A subsequent study failed to replicate their results ¹².

So to summarize our findings, so far: dyslexics may be slightly faster at certain types of spatial processing (e.g., the Impossible Figures task), but in general they have slightly lower spatial abilities than the rest of the population.

One possible criticism of this meta-analysis is that most of the experimental data that it aggregates is based on paper-and-pencil tests. And though the Virtual Reality study failed to replicate, there was at least one other study that found a speed-advantage among the dyslexic group when the tests were administered via a laptop ¹³.

From a purely intuitive perspective, it wouldn't be surprising if paper-and-pencil tests failed to accurately measure visual-spatial ability, especially in dyslexics. In fact, it would be more surprising if "sitting quietly at a desk answering multiple-choice questions about abstract shapes" was an effective way to gauge spatial ability, especially among a group of kids with a poor track-record in a classroom setting.

Also, the fMRI data is a pretty strong signal that there is something different about the way that the dyslexic brain performs spatial processing. It would be surprising if that neurological difference didn't manifest as a difference in ability.

So we're no closer to an explanation of Einstein. Perhaps another meta-analysis holds the key to this mystery?

Here's where things get interesting. The meta-analytic findings of "Meta-analytic findings reveal lower means but higher variances in visuo-spatial ability in dyslexia" ¹⁴ reveal lower means but higher variances in visuo-spatial ability in dyslexia (sorry).

This study collated the experimental results of 28 other studies. Presumably their dataset has a lot of overlap with the 21 studies from the previous meta-analysis, so all of the same caveats re: the validity of paper-and-pencil testing still apply. But whereas the conclusion of the other study focused on the lower average scores for the dyslexic population, the authors of this study noticed that the variance was higher. In their words:

An alternative way to account for the inconclusive results regarding
visuospatial processing in dyslexia would be to posit that there exist
subsets of people with dyslexia who have extremely enhanced or
impoverished visuospatial processing abilities, arising from greater
variance in dyslexic samples.

A subset of people with dyslexia with "extremely enhanced" visuospatial abilities? Tell me more…

The notion of increased variance in one subpopulation has been most
notably employed in the study of sex differences in intelligence and
mathematical ability (Arden & Plomin, 2006; Halpern et al., 2007; Hyde
& Mertz, 2009; Irwing & Lynn, 2005), leading some authors to claim
that, ‘a small difference in variance [between males and females] may
have consequences at the extremes of ability resulting in visibly
unequal numbers of one sex among the less able or among the elite’
(Arden & Plomin, 2006, p.46).

Uhhhhh, OK. Not going to comment on this one, other than to note that "a slight increase in variance leading to overrepresentation of one group at the extreme ends of ability" would fit perfectly with the observation that a shocking number of famous people are supposedly dyslexic.

So maybe dyslexics tend towards the extremes of the spatial abilities bell-curve, and Einstein et. al. were the lucky ones who landed way out on the right tail. This seems plausible, and there could be something to it.

But after taking a closer look at the diagnoses on a case-by-case basis, I have a simpler explanation: most of these people weren't actually dyslexic.