Your Brain Decides Far Earlier Than Scientists Thought and It Could Reshape How We Build AI

Scientists have found that the human brain begins making decisions far earlier in its processing chain than researchers long believed, a discovery that challenges a foundational assumption behind modern artificial intelligence and could point the way to computers that think more like biological brains.

In a study published in the Proceedings of the National Academy of Sciences, researchers at the University of Illinois Urbana-Champaign detected decision-making signals appearing as early as the primary somatosensory cortex—the brain's first cortical relay for touch and movement—rather than emerging only later in the frontal and premotor regions that neuroscientists have traditionally treated as the brain's command centers.

The finding overturns a tidy but tidy-perhaps-too-tidy picture of how the brain works: that raw sensory information flows upward through a hierarchy, is analyzed at the top, and only then produces a decision. Instead, the team led by Yurii Vlasov found that even primary sensory regions are shaped by higher brain areas through rapid feedback loops, meaning the brain is effectively weighing choices from the very first moments it processes a stimulus.

That more dynamic, back-and-forth architecture matters well beyond neuroscience. Most modern AI systems—including the large models powering today's chatbots—are built on the assumption that decisions are made at the top of a hierarchy, after information has flowed all the way up from the inputs. The new results suggest that biological brains distribute the work far more broadly, blending sensing and deciding in ways that artificial networks do not.

Engineers may be able to exploit that insight. The researchers argue that AI systems designed to mimic the brain's feedback-rich, distributed style of computation could make decisions faster and, crucially, using far less energy than the power-hungry data centers that current models require. As the AI industry confronts soaring electricity demands, a blueprint drawn from the efficiency of the human brain—which runs on roughly the power of a dim light bulb—could prove as valuable to technology as it is to our understanding of the mind.

To reach their conclusions, the team recorded neural activity as subjects performed simple sensory tasks, then traced precisely when and where in the cortex the first signatures of a choice appeared. The early emergence of those signals in a region long regarded as a mere relay station suggests that the brain does not wait to gather all the evidence before beginning to commit to an outcome—a finding that fits a growing body of research portraying perception and action as deeply intertwined. The authors caution that translating the discovery into working technology will take years, but they say it adds to mounting evidence that the next leap in artificial intelligence may come not from ever-larger models but from a closer study of how living brains actually compute.