Dendritic Computations in a Rall Model with Strong Distal Stimulation

Abstract

Rall’s work is the basis for investigating dendritic computations, but only recently the technology became available to study their properties experimentally. Empirical evidence supports the idea that synaptic inputs at distal dendritic locations set the context for recognizing synaptic activation patterns of synapses proximal to the soma. Such a context-dependence is fundamental for action selection and decision making. It is usually assumed that active channels in dendrites are necessary. Here we investigate under which conditions of synaptic drive, a passive dendrite model can realize such a context-dependence, and we find that stronger distal than proximal activation, paired with delayed inhibition, is sufficient to produce so-called up states. Testing the model on a different protocol (selectivity to synaptic activation sequences: distal to proximal vs. proximal to distal) shows that it is more similar to recent experimental findings than Rall’s original parameterization, and similar to a model with active dendrites. Our results show that, given stronger distal activation, context-dependent pattern recognition can be implemented in passive dendrites. As a consequence, future experimental studies need to determine on a case-by-case basis the contribution of active channels in dendrites (a single neuron property) vs. synaptic drive (a network property) in context-dependent pattern recognition.