4
$\begingroup$

Consider the following synaptic connections (from here):

axodendritic - A term pertaining to an excitatory or inhibitory synaptic connection between the presynaptic axon of a transmitting neuron and the postsynaptic dendrite(s) of a receiving neuron in a nerve impulse pathway; such connections can influence whether or not a graded potential will be generated on a postsynaptic dendrite.

axosomatic - A term pertaining to an excitatory or inhibitory synaptic connection between the presynaptic axon of a transmitting neuron and the postsynaptic cell body/soma of a receiving neuron in a nerve impulse pathway; such connections can influence whether or not an action potential will be generated in the postsynaptic axon trigger zone at the axon hillock.

axoaxonic - A term pertaining to an excitatory or inhibitory synaptic connection between the presynaptic axon of a transmitting neuron and the postsynaptic axon hillock or axon of a receiving neuron in a nerve impulse pathway; such connections can influence whether or not an action potential will be generated in the postsynaptic axon trigger zone at the axon hillock.

What exactly are the differences between the connections in terms of their influence on the neuron?

To my understanding:

  • axodendritic: influences a dendrite $\rightarrow$ hence influences the graded potential of the neuron
  • axosomatic: influences the graded potential of the neuron directly
  • axoaxonic: influences the axon only (independently of the neuron?)

Hence, to my understanding the axodendritic and axosomatic connections yield the same result: they influence the graded potential of the neuron (e.g. whether the neuron "fires"), where the axoaxonic connection only influences whether the axon fires. Is this correct?

If so, then what is the purpose of dendrites? If axons can connect to the cell body directly, then why would they need dendrites (apart from yielding a larger connection surface to the neuron)?

Note: I have read this answer, which partially explains the use of dendrites. However, it also states that the signal decay due to the synaptic location is compensated for by the local potential change. Therefore, IMO, the answer states that the main advantage of having dendrites is that inhibitory axosomatic synapses can capitalize on the location-dependence of the axodendritic synapses. Is that the true and only purpose of axosomatic synapses? (Please, correct me if my inference is wrong!)

$\endgroup$
  • $\begingroup$ I wrote that linked answer. I did not mean to imply that synaptic location is always compensated by local potential change. $\endgroup$ – yamad Nov 1 '15 at 19:30
  • $\begingroup$ @yamad Thank you for clarifying. Do you happen to know the differences (in functionality) between axosomatic and axodendritic synaptic connections? $\endgroup$ – Jean-Paul Nov 1 '15 at 19:32
4
$\begingroup$

Neurons are all about specialized structures having specialized roles. You've given a good justification for dendrites---an efficient way to fit lots of connections in a small space. A pyramidal neuron in cortex, for instance, will have tens of thousands of synapses. If you only had axosomatic connections, the soma would have to be enormous to fit all of them (and this would cause lots of other issues).

Dendritic structure also allows the cell more efficient electrical properties (see cable theory) and to process inputs independently. In fact, different parts of the dendrite are now thought of as independent computational compartments. For instance, the cell might want to selectively strengthen a single synapse or all the synapses on one dendritic branch. By physically isolating those synapses in small dendritic structures, it is easier to make those targeted changes without unwanted "crosstalk" on other synapses or cell activities. What's more, incoming connections from different areas often segregate into regions of the dendrite (e.g. in CA1 hippocampal pyramidal neurons, cortical connections to the most distant part of the dendrite, internal hippocampal connections to the part of the dendrite closer to the soma).

When a cell has a large dendrite (some neurons have no dendrites), its useful to think of axodendritic synapses as the normal inputs that contribute to fine-grained computation, and axosomatic and axoaxonic synapses as more global gates. These latter types take advantage of their location to short circuit the "normal" computation in the dendrites. An inhibitory axosomatic synapse, for example, will reduce the effect of all dendritic inputs when activated.

$\endgroup$
  • $\begingroup$ Extremely informative. Thank you. If I may, I would like to know your opinion about the theory that the "firing" of neurons is somehow connected to the specific combination of excitatory dendrites. E.g. if we neglect differences in potential change and assume 3 dendrites to excitate with the same potential effect on the membrane potential, do you think that different combinations of these 3 dendrites have a different effect on whether the neuron fires? Or are the potential changes from the three dendrites always simply summed up? So, given equal potential change, does location matter? $\endgroup$ – Jean-Paul Nov 1 '15 at 20:34
  • 2
    $\begingroup$ @Jean-Paul Yes, location matters. In general, the spatial and temporal pattern of synaptic activation matters a lot. Two synapses activated at the same time on different dendritic branches will have a different effect from two synapses on the same branch. Two synapses on the same branch will have a different effect depending on the order of activation. Good rule of thumb for thinking about neural computation: if something can vary, it probably matters (and if it doesn't matter, there is probably a very interesting story about how the neuron ignores the variation). $\endgroup$ – yamad Nov 1 '15 at 22:48
  • $\begingroup$ Thank you very much for your time and effort! This is very helpful. I'm just curious now: does location matter in terms of signal strength or does it matter for the pattern in which the neuron fires? $\endgroup$ – Jean-Paul Nov 2 '15 at 7:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.