Myelin & g
So, myelin made brains faster, more compact, and less pound-for-pound energy intensive. Now let's jump ahead about 380 million years or so; you couldn't have, say, hummingbirds without myelin, and you couldn't have primates either. Something so crucial to information processing capacity has to be important for intelligence, right?
The earliest research into the biological correlates of the g factor (which is highly heritable, recall) found that it correlates moderately with brain volume and with inspection & response time on elementary cognitive tasks (extremely simple tasks expressly designed for clocking simple brain processes). Digging a little deeper, they found that the inspection & response time correlation was mediated by nerve conduction velocity, as you might expect. An inverse correlation was also found between g and energy expenditures per unit volume in the brain (measured by glucose metabolism rates)-smarter people's brains are more energy-efficient. Just as a point of reference, nervous transmission and the cellular mechanisms that support it take up about half the energy expended in the brain.
So thicker myelination could account both for the improvement in energy expenditure (remember that the increased resistance means less energy is required to push current through axons) and the increased conduction velocity (due to increased capacitance), and also incidentally some of the difference in brain volume. And there's more circumstantial evidence backing this up: Ashkenazi Jews, who have an average IQ about 12 points or so higher than the average European population, are unusually succeptible to several lysosomal storage diseases such as Gaucher's, Canavan & Neimann-Pick, which follow a straightforward Mendelian inheritance pattern and are known to affect myelin. (The idea being that having one copy of the mutant gene enhances myelin's effectiveness while having two copies results in too much of a good thing.) Also, people with multiple sclerosis show progressive deterioration in both reaction time and intelligence that tracks the progressive deterioration of their white matter.
Another interesting facet of this is that less intelligent people show greater *variability* in their reaction times on elementary cognitive tasks than more intelligent people. This could well be caused by noisier transmission in the brain, and the likely culprit for that would be thinner myelination. If you got occasional random crosstalk between excitory and inhibitory channels, this would create greater ambiguity which would be subsequently resolved by further stimulus as the milliseconds ticked by, slowing reaction time. (This is because neural firing is based on thresholds, and amount of simultaneous stimulus and duration of stimulus are partial substitutes in the short run. One dendrite being hit with a packet of glutamate twice in rapid succession can have nearly the same effect as two dendrites in the same neuron each being hit once simultaneously.)
Moreover, as the cognitive tasks get gradually more complex, reaction time increases for all intelligence levels (slightly more rapidly for the less intelligent, as you might expect), and *variance* in reaction time for each trial also increases-and this effect is also more pronounced for the less intelligent, which you might not expect if you weren't thinking in these terms already. This seems consistent with the crosstalk hypothesis, since the more hops a signal has to go through to complete its journey, the more important error-propagation becomes and the longer it would tend to take for the resolution of ambiguity. And since this means it takes more firing to achieve the same result, this would also tend to increase the amount of glucose being burned per unit volume of brain.
Of course, myelin is important but can't be the whole story behind g-at most it explains, apparently, about 1/4th of it. Posthuma et al. (2002) reported a correlation of 0.24 between g and white matter volume, which is the same as they found between g and grey matter volume (0.25). But this leads me to some interesting conjectures which I'll save for later.
(No references for all these claims because I'm feeling lazy, but you can find most of this discussed in Arthur Jensen's The g Factor if so inclined.)
The earliest research into the biological correlates of the g factor (which is highly heritable, recall) found that it correlates moderately with brain volume and with inspection & response time on elementary cognitive tasks (extremely simple tasks expressly designed for clocking simple brain processes). Digging a little deeper, they found that the inspection & response time correlation was mediated by nerve conduction velocity, as you might expect. An inverse correlation was also found between g and energy expenditures per unit volume in the brain (measured by glucose metabolism rates)-smarter people's brains are more energy-efficient. Just as a point of reference, nervous transmission and the cellular mechanisms that support it take up about half the energy expended in the brain.
So thicker myelination could account both for the improvement in energy expenditure (remember that the increased resistance means less energy is required to push current through axons) and the increased conduction velocity (due to increased capacitance), and also incidentally some of the difference in brain volume. And there's more circumstantial evidence backing this up: Ashkenazi Jews, who have an average IQ about 12 points or so higher than the average European population, are unusually succeptible to several lysosomal storage diseases such as Gaucher's, Canavan & Neimann-Pick, which follow a straightforward Mendelian inheritance pattern and are known to affect myelin. (The idea being that having one copy of the mutant gene enhances myelin's effectiveness while having two copies results in too much of a good thing.) Also, people with multiple sclerosis show progressive deterioration in both reaction time and intelligence that tracks the progressive deterioration of their white matter.
Another interesting facet of this is that less intelligent people show greater *variability* in their reaction times on elementary cognitive tasks than more intelligent people. This could well be caused by noisier transmission in the brain, and the likely culprit for that would be thinner myelination. If you got occasional random crosstalk between excitory and inhibitory channels, this would create greater ambiguity which would be subsequently resolved by further stimulus as the milliseconds ticked by, slowing reaction time. (This is because neural firing is based on thresholds, and amount of simultaneous stimulus and duration of stimulus are partial substitutes in the short run. One dendrite being hit with a packet of glutamate twice in rapid succession can have nearly the same effect as two dendrites in the same neuron each being hit once simultaneously.)
Moreover, as the cognitive tasks get gradually more complex, reaction time increases for all intelligence levels (slightly more rapidly for the less intelligent, as you might expect), and *variance* in reaction time for each trial also increases-and this effect is also more pronounced for the less intelligent, which you might not expect if you weren't thinking in these terms already. This seems consistent with the crosstalk hypothesis, since the more hops a signal has to go through to complete its journey, the more important error-propagation becomes and the longer it would tend to take for the resolution of ambiguity. And since this means it takes more firing to achieve the same result, this would also tend to increase the amount of glucose being burned per unit volume of brain.
Of course, myelin is important but can't be the whole story behind g-at most it explains, apparently, about 1/4th of it. Posthuma et al. (2002) reported a correlation of 0.24 between g and white matter volume, which is the same as they found between g and grey matter volume (0.25). But this leads me to some interesting conjectures which I'll save for later.
(No references for all these claims because I'm feeling lazy, but you can find most of this discussed in Arthur Jensen's The g Factor if so inclined.)