The brain's dendrites, often overlooked in favor of the cell body, are taking center stage in a groundbreaking study. This research reveals that dendrites can independently store information and make predictions, challenging the traditional view of neurons as uniform units. The study, published in Science, demonstrates that dendrites can dissociate from cell body activity, depending on the animal's goals, providing the first in-vivo evidence of this long-standing theoretical prediction. This finding is particularly intriguing, as it suggests that dendrites play a more active role in cognitive processes than previously thought.
The research team, led by Attila Losonczy, used ultrafast voltage imaging to record electrical activity in the dendrites of pyramidal place cells in the hippocampus of mice. They found that dendrites retained information about the original reward locations even when the virtual environment changed, while the cell body caught up later. This decoupling of dendritic and somatic activity may enable cells to rapidly adapt to new environments without losing prior spatial learning.
The study also highlights the technological advancement of single-cell voltage imaging over long periods in awake animals. When the reward location changed, dendrites closer to the cell body remapped faster, but when the entire environment changed, distal dendrites learned the new spatial code first. This suggests that the way synapses in distal and proximal dendrites store information may differ, with distal synapses achieving long-term potentiation without an action potential if enough of them become active simultaneously.
The findings have significant implications for our understanding of learning in neurons. Eilif Muller, an associate professor of neurosciences, suggests that learning in the dendrites starts out unsupervised, providing an opportunity to uncover the fundamental learning rules in neurons. However, the mechanism behind the split between dendritic and somatic activity remains unclear, as the dense recurrent connections in the CA3 region may make some independent dendritic computations appear inherited from the surrounding network.
Despite these challenges, the study opens up exciting possibilities for further research. Losonczy and his colleagues are working to decipher the mechanism, emphasizing the need for a flexible gating mechanism that can couple and uncouple local clusters of synapses to the soma. This research not only advances our understanding of the brain's inner workings but also has the potential to inspire new approaches in artificial intelligence, where dendrites are often underappreciated.
