Individual human neurons could also be way more highly effective computing models than beforehand understood.
Inside the human cortex, a single neuron could also be doing excess of merely deciding whether or not to ship a sign. New analysis means that one mind cell can perform computations advanced sufficient to rival the work of a deep synthetic neural community.
That risk adjustments the place scientists would possibly search for the origins of human skills corresponding to language, arithmetic, creativeness, and invention. Intelligence could rely not solely on the big scale of the mind, but in addition on what every of its particular person cells can accomplish.
For a long time, researchers largely defined the mind’s energy by means of its dimension and connectivity. The human mind accommodates near 100 billion neurons, linked by means of an unlimited community. Research printed within the Proceedings of the National Academy of Sciences (PNAS), nevertheless, signifies that the distinctive skills of the human mind may additionally come up from the unusually subtle processing carried out inside particular person neurons.
Intelligence could start inside single neurons
Neurons from the human cortex, the mind’s outer layer concerned in superior thought, seem to operate as unusually advanced info processing models (“microchips”). Rather than merely gathering indicators and producing a easy response, they’ll mix incoming info by means of intricate inside processes.
This risk may assist clarify how the human cortex helps cognitive skills that exceed these of different mammals. If every cell performs extra computation, the mind positive aspects further processing energy earlier than info even strikes by means of its wider community.
The analysis was led by Hebrew University Profs. Idan Segev and Mickey London, together with PhD college students Ido Aizenbud and Daniela Yoeli on the Edmond and Lily Safra Center for Brain Sciences (ELSC). Prof. Chris de Kock of the Free University, Amsterdam, additionally collaborated on the work.
“People often think of a neuron as a simple switch that either turns on or off,” mentioned Segev. “What we show is that a single human neuron is itself an extraordinarily sophisticated computing device.”
AI reveals every neuron’s computing energy
The researchers first wanted a constant solution to evaluate the computational skills of neurons from completely different mammals. Simply inspecting a cell’s dimension or form wouldn’t reveal how a lot info it may course of.
They approached the issue by constructing a digital imitator for every neuron. Using pc modeling and artificial intelligence, they tested how difficult it was for an artificial neural network (ANN) to learn the relationship between the signals entering a biological neuron and the response coming out.
A relatively simple neuron could be copied by a small artificial model. A more capable biological cell required a deeper and more elaborate network before the artificial version could reproduce its behavior accurately.
This imitation test gave the researchers a practical measure of neuronal complexity. The more difficult the neuron was to reproduce, the greater its apparent computational power.
Human neurons outperform other mammals
Human cortical neurons consistently required more complex artificial networks to imitate their behavior than neurons from other mammals. Their advantage appears to arise partly from their richly branched dendritic trees, the structures that receive signals from neighboring cells, and from their distinctive electrical characteristics.
Those features allow a neuron to analyze combinations of incoming signals instead of simply adding them together. In principle, this could support demanding distinctions within sensory information (e.g., distinguishing between images of cats versus dogs).
The results portray a human cortical neuron as much more than an “on-off” component. One cell can operate as a layered computing system with capabilities comparable to those of a deep neural network.
That conclusion challenges the long-standing assumption that human intelligence depends mainly on neuron count and the number of connections among cells. The sophistication built into individual neurons may also have contributed to the evolution of human cognition.
Smarter artificial neurons could reshape AI
The researchers also introduced a general framework for connecting the physical features of a neuron with the computations it can perform. That approach could help scientists investigate how cellular structure contributes to learning, thought, and other forms of cognition.
The findings may also influence the design of brain-inspired AI. Most current systems are assembled from highly simplified artificial units, even when the resulting networks contain many layers.
Future models could instead use artificial components with more processing ability inside each unit. Such systems would more closely resemble biological neurons and could offer a different path for developing state-of-the-art machine-learning technology.
Reference: “Dendritic morphology and synaptic nonlinearities enhance functional complexity in human cortical neurons” by Ido Aizenbud, Daniela Yoeli, David Beniaguev, Christiaan P. J. de Kock, Michael London and Idan Segev, 7 July 2026, Proceedings of the National Academy of Sciences.
DOI: 10.1073/pnas.2533168123
This work was supported by the Office of Naval Research Grant Award No. N00014-24-1-2055 and Grant Award No. N00014-23-1-2051.
Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.