Scientists Challenge the 70-Year-Old 'Lizard Brain' Myth
Georgia Tech researchers say brain evolution is better explained by competing wiring strategies than by a primitive 'lizard brain' layered beneath a newer, rational one.
Step by step
- 1
1950s theory proposes layered brain evolution
- 2
Georgia Tech compares brains and AI networks
- 3
182 species show coordinated size trade-off
- 4
AI tests confirm competing wiring strategies
Georgia Tech researchers are challenging the idea of a primitive "lizard brain" wired for instinct beneath a newer, rational human brain -- a 1950s theory describing evolution as layers stacked atop each other. "This is not how an evolutionary biologist would think about the problem," said Nabil Imam, an assistant professor in Georgia Tech's School of Computational Science and Engineering. The new study, published in Science Advances, suggests brain evolution is better explained by wiring than by newer regions being added on top of older ones.
Imam and colleagues compared biological brains with artificial neural networks and found that evolution may divide limited brain space between two competing wiring strategies. The , involved in vision, perception and reasoning, is organised as a spatial map: regions processing nearby body parts, such as the thumb and index finger, sit close together. The , often called the "reptilian brain," works more like a barcode instead, using distributed patterns of activity to represent smells or complex memories.
Comparing 182 species, the researchers found that when one part of the limbic system grew larger, other limbic regions tended to grow too, while the neocortex became smaller -- evidence of a coordinated trade-off rather than independent evolution. In artificial-network tests, localised connections suited vision, sound and touch, while distributed "barcode-style" networks were needed for smell and memory. When a simulated environment rewarded smell, the distributed system expanded and the neocortex shrank -- and the pattern reversed for vision. The smell-reliant nine-banded armadillo has a very large limbic system, while the vision-reliant squirrel monkey has a neocortex-dominated brain.
The findings could also point toward AI systems that need far less training data and energy, the researchers say. "Today's artificial neural networks are trained by vast amounts of data -- it's about nurture," Imam said. "But the brain is not a blank slate that gets trained by experience. It is a mix of nature and nurture."
Terms explained
The story so far
- Scientists Identify a Brain 'Brake' That Can Shut Down Chronic Nerve Pain
- Open ASR Leaderboard Adds Its First Indian-Language Speech Recognition Test
- Scientists Find a Molecular 'Brake' That Keeps Damaged Nerves From Healing
- Your Sleep May Hide an Early Clue to Alzheimer's Risk, Study Finds
- Nanoparticle Injection Reverses Cognitive Decline in Mice With Alzheimer's Model
- AI Screens 150 Million Compositions to Find New Lead-Free Electronics Materials
- Scientists Challenge the 70-Year-Old 'Lizard Brain' Myth
