Breaking News
admin
Apr 21 2025
7How does the brain manage to learn new skills, remember a song, or navigate unfamiliar places? A groundbreaking study by researchers at the University of California San Diego has shed new light on how the brain encodes new information by revealing that learning involves diverse rules across individual neurons and synapses. These findings, published on April 17 in Science, challenge long-held beliefs about how learning occurs in the brain.
The research team, composed of neurobiologists William "Jake" Wright, Nathan Hedrick, and senior author Takaki Komiyama, utilized advanced brain imaging technology—including two-photon microscopy—to observe changes in the brains of mice as they learned new tasks. This high-resolution method allowed them to examine individual synapses and neural activity in real time during the learning process.
Traditionally, synaptic plasticity—the strengthening or weakening of synapses—has been considered a uniform mechanism across the brain. However, this new study found that neurons can follow multiple plasticity rules at the same time, depending on the synapse’s location. These site-specific modifications mean that rather than adhering to one overarching method of change, neurons process learning through varied, parallel strategies within their subcellular compartments.
“This fundamentally shifts our understanding of how learning is implemented in the brain,” said Komiyama, who holds joint appointments in UC San Diego’s Departments of Neurobiology and Neurosciences, as well as the Halıcıoğlu Data Science Institute and Kavli Institute for Brain and Mind.
The discovery also offers a new perspective on the so-called "credit assignment problem"—the question of how local synaptic changes contribute to global learning. The researchers liken this issue to how ants contribute to colony goals without knowing the bigger picture. Their study suggests that neurons, much like those ants, execute distinct computations in parallel, enabling complex learning behaviors through decentralized processes.
“These findings help explain how individual synapses, with access only to local information, contribute to broad learning outcomes,” said Wright, the study's first author and postdoctoral scholar at the School of Biological Sciences.
This deeper understanding of how synaptic changes occur not only informs neuroscience but also offers promising implications for artificial intelligence. Present-day AI systems typically operate on single-rule networks, but insights from this study could inspire more sophisticated neural networks that mirror the brain’s multi-rule processing.
The research also holds potential for advancing treatments for neurological and behavioral conditions such as Alzheimer’s disease, PTSD, addiction, and autism. Since many brain disorders involve disrupted synaptic plasticity, knowing how learning normally functions is key to understanding what goes awry.
“This work lays a foundation for understanding healthy brain function, which is crucial for diagnosing and treating various neurological diseases,” Wright added.
Backed by funding from the National Institutes of Health, the study marks the beginning of a deeper exploration into how neurons manage to apply multiple learning rules—and the advantages this flexibility offers in both biological and artificial systems.
Source:https://www.sciencedaily.com/releases/2025/04/250417144912.html
This is non-financial/medical advice and made using AI so could be wrong.
