The Smarter Brain Burns Less Energy, Not More

Over half an hour, a group of people were injected with a small dose of radioactive glucose and asked to solve reasoning problems. The better they did, the less their brain lit up on the scan.

The higher-scoring people weren’t trying less hard in the lazy sense of the phrase: their brains seemed to need less fuel to reach the same place. That was the origin of the neural efficiency hypothesis, one of the most cited ideas about the biology of intelligence. This article tells the finding as Arthur Jensen presents it and then covers what later research has qualified, which is quite a lot.

The experiment: radioactive glucose and a matrix test

Psychologist Arthur Jensen covers this work in The g Factor (1998), his synthesis of the science behind the general factor of intelligence (g). The experiment in question uses a technique called PET (positron emission tomography):

• A small amount of a radioactive glucose isotope is injected into the bloodstream.
• The person spends about 30 minutes solving a nonverbal reasoning test, Raven’s Progressive Matrices, one of the most heavily g-loaded tests there is.
• The scanner detects which brain regions metabolized the most glucose, in other words, where the most neural energy was spent.

Researcher Richard Haier, a leading figure in this field, found something that at first glance runs against intuition: the amount of glucose consumed during the test correlated negatively with the score obtained, with values between -0.7 and -0.8. The higher the score, the less energy spent solving the same kind of problem.

One detail the book mentions in passing carries a lot of weight: the original study by Haier and his team (1988) was done with eight people. With samples that small, a correlation can come out very high by chance, which is why the finding was taken as a starting point, not a conclusion.

Why “more horsepower” is the wrong metaphor

It’s tempting to picture a smart brain as a bigger engine that burns more fuel and therefore performs better. The early PET data pointed in exactly the opposite direction:

• A more efficient brain seems to need less metabolic activity to reach the same correct result.
• That inverse relationship shows up across most of the cerebral cortex, though it’s more pronounced in the temporal regions.
• Jensen sums it up with a simple idea: g appears to be related to the efficiency of neural processing, not its volume.

In Jensen’s version, a more capable brain doesn’t work harder: it gets more done for every unit of energy it spends.

The experiment that set it in motion: learning to play Tetris

The most interesting part isn’t a single snapshot, but what happens when someone gets better at something. Haier and his team repeated the PET scan with a group of people while they learned to play Tetris, a video game that demands fast information processing, hand-eye coordination and on-the-fly planning.

• In the first rounds, playing Tetris burned a relatively high amount of brain glucose.
• After practice sessions of 30 to 45 minutes, up to five a week, over 4 to 8 weeks, skill improved noticeably.
• At the same time, brain glucose consumption during the game dropped, and the correlation between that metabolic activity and the g factor weakened.

The size of that saving was in turn correlated with scores on the Raven’s test: those who started with higher scores were the ones who gained the most energy efficiency with practice.

A maturing brain also gets cheaper to run

The same pattern, according to Jensen, also appears in anyone’s normal development. PET studies across different ages show that brain glucose consumption decreases progressively from childhood to adulthood, precisely while mental capacity keeps growing.

One plausible explanation the book discusses is so-called neural pruning: the spontaneous reduction, especially in the first years of life, in the number of synaptic connections that aren’t useful.

The idea is that fewer redundant connections allow cleaner processing with less wasted energy. Jensen adds that insufficient pruning during early development is associated with certain types of intellectual disability.

Having fewer active neurons, at the right point in development, isn’t always a loss. Sometimes it’s the necessary condition for the rest of the system to work better.

What came next: efficiency has conditions

Over the following two decades, dozens of studies using PET, functional MRI and EEG put the hypothesis to the test. Aljoscha Neubauer and Andreas Fink, of the University of Graz, reviewed that literature in 2009 and reached a more nuanced conclusion than the original headline:

• Most of the early studies confirmed the hypothesis, but later ones found contradictory results and several variables that modulate it: sex, the type of task, its complexity and the brain region.
• Efficiency shows up mainly when the task is of low or moderate difficulty for the person doing it, and is observed mostly in frontal areas.
• On very complex tasks, the pattern reverses: more capable people seem to invest more cortical resources, and the correlation between activation and ability turns positive.
• Practice also matters: efficiency depends on the amount and quality of learning, as the Tetris study already suggested.

In parallel, Rex Jung and Haier himself (2007) reviewed 37 neuroimaging studies and proposed the Parieto-Frontal Integration Theory (P-FIT): intelligence would depend less on how much energy the brain spends overall than on how well a network of frontal and parietal regions communicates. It’s a shift in focus, from total expenditure to how the network is organized.

That’s why the title of this article is true only with the fine print: the more capable brain spends less energy on tasks it finds manageable, not always and not on every problem.

What this finding changes (and what it doesn’t)

None of this turns brain efficiency into something you can measure at home, or Tetris practice into a shortcut to a higher IQ.

• The energy cost of solving a problem and the capacity to solve it are different things: here they’re related, not confused.
• Practice reduces the energy cost of any task that becomes automatic, in anyone, though the starting point varies.
• Neural efficiency is one of the most studied biological correlates of the g factor, but it isn’t a general law: it depends on task difficulty, brain region and practice.

What this finding does leave is a more interesting picture of what “a powerful brain” means: not the one that always works hardest, but the one that knows when to spend little and when to spend a lot.

If you’re interested in the same idea seen from another angle of the brain, what your brain decides in a fifth of a second explains how even the simplest reaction speed, measured in milliseconds, relates to this same variable.

Sources

• Jensen, A. R. (1998). The g Factor: The Science of Mental Ability. Praeger.
• Haier, R. J. (1993). Cerebral glucose metabolism and intelligence. In P. A. Vernon (Ed.), Biological Approaches to the Study of Human Intelligence (pp. 317-373). Ablex.
• Haier, R. J., Siegel, B. V., Nuechterlein, K. H., Hazlett, E., Wu, J. C., Paek, J., Browning, H. L. and Buchsbaum, M. S. (1988). Cortical glucose metabolic rate correlates of abstract reasoning and attention studied with positron emission tomography. Intelligence, 12(2), 199-217. doi:10.1016/0160-2896(88)90016-5
• Haier, R. J., Siegel, B., Tang, C., Abel, L., & Buchsbaum, M. S. (1992). Intelligence and changes in regional cerebral glucose metabolic rate following learning. Intelligence, 16(3-4), 415-426. doi:10.1016/0160-2896(92)90018-M
• Jung, R. E. and Haier, R. J. (2007). The Parieto-Frontal Integration Theory (P-FIT) of intelligence: Converging neuroimaging evidence. Behavioral and Brain Sciences, 30(2), 135-154. doi:10.1017/S0140525X07001185
• Neubauer, A. C. and Fink, A. (2009). Intelligence and neural efficiency. Neuroscience & Biobehavioral Reviews, 33(7), 1004-1023. doi:10.1016/j.neubiorev.2009.04.001