Human brain cells are far more powerful than scientists thought

A recent study published in the Proceedings of the National Academy of Sciences (PNAS) has revealed that individual human neurons exhibit far more complex computational capabilities than neurons found in other mammals. This discovery sheds new light on what might account for the exceptional cognitive abilities of humans, such as language, imagination, mathematics, and invention.

For years, scientists have largely attributed human brain capabilities to its size and connectivity, with neurons connected through an extensive network. However, a new study suggests that the key may lie at the cellular level, in the remarkable computational prowess of individual brain cells.

Researchers from Hebrew University and the Free University of Amsterdam found that neurons in the human cortex function as more advanced information processing units than those in other mammals. Each human cortical neuron, with its intricately branched dendritic tree and unique electrical properties, can carry out sophisticated computations on incoming data, including visual input (like distinguishing between cats and dogs). This indicates that these neurons are more than simple "on-off" components.

The research team developed a new method to measure the computational complexity of individual brain cells, combining computer modeling with artificial intelligence. By comparing the difficulty of learning and reproducing the relationship between input and response in an artificial neural network (ANN) to that of a biological neuron, the team discovered that neurons in the human cortex possess significantly greater computational power.

This suggests that individual neurons may have played a crucial role in the evolution of human cognition, challenging the notion that intelligence primarily stems from the sheer number of neurons and their connections. The findings also have implications for artificial intelligence, hinting at the potential for brain-inspired AI that features deeper, more powerful artificial units, more closely mirroring the capabilities of biological neurons.