Book Review: The Neuroscience of Intelligence

A popular question on the question site Quora is “how do I increase my IQ/intelligence?”  There are a number of schemes advertised to do just that: exposure to Mozart, memory practice, video games, early childhood interventions, plus various pills, supplements, and nostrums. 

Unfortunately, says Richard Haier, writing in his book, The Neuroscience of Intelligence, none of them appear to work.

I have long been an IQ skeptic, with the core of my skepticism being based on the lack of identifiable neurobiological correlates of IQ.  Such correlates, based mainly on brain imaging studies that are of relatively recent availability, are the major theme of the book.

So what are those neural correlates?  Many of them seem to be connected to communication and connection between the frontal and parietal regions of the brain.  Functional magnetic resonance imaging (fMRI) of the brain reveals the most active regions of the brain during task performance.  The brains of high IQ persons have both better communication and much sparser activation during problem solving, as if they are just a lot more efficient at the task.  There is also some evidence that these patterns persist from childhood to adulthood.

One of the largest physical correlations is brain size (30%, significant but hardly gigantic).  That may explain why Ed Witten is so much smarter than me.

The author also claims that the evidence for a large component of heritability of IQ is ironclad, and I found his evidence persuasive.  What is not known is what genes are responsible for IQ differences and how they work their effects.  Large scale genome wide association studies have revealed only genes with very modest effects.  The assumption is that a large number of genes each contribute a small effect.  This is not terribly surprising, since a similar pattern holds for adult height, another highly heritable trait.

Animal studies have shown similar patterns.  Mice with enhanced learning capabilities have been bred (they are known as “dougie mice” after the Neil Patrick Harris character Dougie Howser, MD.)

There have been attempts to identify very gifted individuals at young, including a math talent program that gave the SAT math exam to seventh graders.  A number of them scored over 600.  These students have been followed for a number of years and have done very well.  Three notable students were Sergei Brin (cofounder of Google), Mark Zuckerberg, and Lady Gaga.

The book is short, only about 200 pages if you don’t count reference, glossary, and index, and it is written for a sophisticated reader with no special expertise in neurobiology.  One somewhat annoying feature is the plethora of acronyms.

tACS was given for 15 minutes prior to completing two tests of fluid intelligence. The tests were a modified version of the RAPM and the paper folding and cutting test (PF&C) of spatial ability from the Stanford–Binet IQ test battery. EEG were also obtained during both tests. The authors concluded that, “Left parietal tACS increased performance on the difficult test items of both tests (RAPM and PF&C) whereas left frontal tACS increased performance only on the easy test items of one test (RAPM).

Haier, Richard J.. The Neuroscience of Intelligence (Cambridge Fundamentals of Neuroscience in Psychology) (p. 161). Cambridge University Press. Kindle Edition.

UPDATE: One point the author emphasizes is that IQ scores are statistical.  An IQ of 150 is not 1.5 times as high as one of 100 in any meaningful sense.  Imagine, for example, that running speed were judged on a similar scale.  In that case Usain Bolt would have an RQ of 200+, but in fact he would be perhaps 1% faster than a dozen or two other guys, and maybe 10% faster than a fair high school sprinter with an RQ of 130 or so.  The author laments the lack of an absolute scale for Intelligence.  Of course that neglects the obvious - how many questions did the test takers get right.  So how many more right answers would an IQ 160 individual get right more than an 1Q 135 scorer?  Maybe two or three.