The brain runs on about 15 to 20 watts, less than most light bulbs, but has still managed to evolve a voracious appetite for energy. In humans, it accounts for only about one-fiftieth of weight but consumes about 20% of our daily energy. “That is an off-the-charts fitness cost,” says SFI Professor David Wolpert. “Nothing else even comes close.” Intelligence must confer some survival advantage to sustain such a formidable energy share, says Wolpert, but until now, researchers haven’t rigorously examined that trade-off.
In February, Wolpert and SFI External Professor Van Savage, a mathematical biologist at UCLA, hosted a working group, “Evolutionary Consequences of Energetic Cost of Intelligence,” at SFI. The meeting brought together 15 neuroscientists, evolutionary biologists, physicists, and other experts to develop multidisciplinary strategies aimed at modeling the evolutionary tradeoffs of intelligence. Wolpert says SFI-affiliated researchers have long been interested in extending the mathematical tools used by evolutionary biologists to neuroscience, and the meeting offered a chance to introduce experts in those communities to each other’s ideas.
The gathering opened with a challenge: To clearly define concepts like energetics, evolution, and intelligence. The first two weren’t so difficult, says Savage, as physicists could reasonably describe what’s known about how the brain uses energy, and evolutionary biologists could describe methods for identifying evolutionary advantages of phenotypes. “But intelligence was very difficult to define,” he says.
“We started with a little bit of locking horns,” agrees Wolpert. They debated whether intelligence was unique to humans or something broader, like the ability to do nonrandom things in the environment that would increase their ability to acquire energy. They found common ground in the core concept, however: Whatever model they used, it had to incorporate how living systems use energy while reflecting the real-world limitations of time and size encountered by an intelligent system.
“The meeting was really useful to figure these issues out,” says Savage. “You can do a lot of intelligent things in an energy-efficient way if you have unlimited time. But in reality, you don’t have unlimited time.”
By the end of the meeting, says Wolpert, the researchers had established both a clear articulation of the importance of understanding this evolutionary tradeoff and begun to chart a road forward. They’re drafting a perspective paper on the subject, he says, and researchers are now collaborating on projects aimed at laying down a mathematical foundation. The participants are eager to keep the conversation advancing, he says. “I’m getting emails from people who are saying things like, unfortunately, I couldn’t sleep all week because there are so many ideas percolating in my head.”
This working group was partially funded by a grant from the Robert Wood Johnson Foundation.