Dynamic compartmentalization in neurons enables branch specific learning.

The dendritic trees of neurons play an important role in the information processing in the brain. While it is tacitly assumed that dendrites require independent compartments to perform most of their computational functions, it is still not understood how they compartmentalize into functional subunits. Here we show how these subunits can be deduced from the structural and electrical properties of dendrites. We devised a mathematical formalism that links the dendritic arborization to an impedance-based tree-graph and show how the topology of this tree-graph reveals independent dendritic compartments. This analysis reveals that coopera-tivity between synapses decreases less than depolarization with increasing electrical separation, and thus that surprisingly few independent subunits coexist on dendritic trees. We nevertheless find that balanced inputs or shunting inhibition can modify this topology and increase the number and size of compartments in a context-dependent, temporal manner. We also find that this dynamic recompartmentalization can enable branch-specific learning of stimulus features.