The neurone shortcut

The traditional model of neuronal computation is that a neurone receives inputs on it’s dendrites, all these inputs are integrated at the soma, and if the sum of these inputs is large enough to reach a threshold value, the soma sends an output along the axon. This paper shows that some CA1 neurones in the hippocampus don’t follow this simple model – these neurones have a dendrite directly coupled to the axon.

AcD Schematic
Schematic showing two neuronal morphologies. Tradtional neuronal morphology with the axon (green) attached to the soma. Novel AcD neurones with an axon (red) attached to a privelleged dendrite where input (red arrow) has a stronger effect on output compared with input on an neighbouring dendrite (grey arrow).

Anatomy and physiology go hand in hand and in CA1 neurones the axon was coupled to a basal dendrite much more often than expected – over 60% in the middle of CA1 and ~20% at the boarders. This in turn functionally privileged the dendrite coupled to the axon for 2 reasons. Firstly, input onto this dendrite was more capable of generating dendritic spikes.  Secondly, input onto this dendrite has a shortcut to the axon, and less input is needed to trigger an output.

This is an intriguing finding for a number of reasons. An axon connected to a dendrite in CA1 is a novel morphology. Connecting a dendrite and axon together has a strong functional effects. And it occurs so frequently. The immediate implications of this data  is that it changes how we think neurones compute input/output relationships. How connecting dendrites and axons together affect the network and behaviour are the next most relevant questions for me. What are the inputs onto this privileged dendrite and how are they different from the neighbouring dendrites? Where does the output of these cells go? What is the function of these cells? Are they place cells? These are the questions I hope to be tackling next.