So I am reading a book on neuroscience and they mentioned in passing that the action potential is capable of travelling in either direction along the axon (orthodromic vs antidromic), The wikipedia antidromic article states that the effect is often used to confirm connections in laboratory experiments.

What I'm now wondering is if the phenomenon has been observed with normally behaving neurons in situ (as normal as a neuron under experimental observation can be anyway) and therefore possibly needs to be factored in when modelling neurons in software or is it safe/justified to treat the information as only flowing in one direction (orthodromic)?


Under physiological conditions, action potentials are generally assumed to travel one-way. Action potentials are generated in the dendritic region, and travel from the soma to the axon terminal. Because voltage-gated sodium channels are inactivated after having been active during action potential generation, the action potential cannot travel backwards because the tail-end of the action potential is basically temporarily shut down (Fig. 1). The duration of inactivation of sodium channels determines how fast a neuron can fire, i.e., it determines the refractory period of neurons (Purves et al., 2001).

Fig. 1. Action potential conduction and refractoriness. Source: Zoology.

When neural tissue is artificially electrically stimulated, however, action potentials can be generated anywhere along the neuron. When an axon is activated somehwere in the middle with an electrical stimulus, an action potential will travel both ways, i.e. normally to the axon terminal, but also antiodromically to the cell body.

However, it has been noted in vivo that some neurons do show antidromic action potentials under physiological conditions (Jansen et al., 1996). For modeling purposes I would not bother too much about this, though, because antidromic action potentials are generally only observed under artificial conditions.

- Jansen et al., J Neurophysiol; 76(6): 4206-9
- Purves et al., ed. Neuroscience. 2nd ed. Sunderland (MA): Sinauer Associates; 2001

  • $\begingroup$ Any chance you could expand on/summarize the Jansen et al. article (It's behind a $20 pay wall) $\endgroup$ – norlesh Sep 28 '15 at 14:17
  • $\begingroup$ @norlesh - what do you wish to know? The authors found antidromic responses in some neurons in an animal model, but they don't expand on how and where they were generated. If you can say what you are after I'm happy to elaborate. $\endgroup$ – AliceD Sep 28 '15 at 14:23
  • 1
    $\begingroup$ Guess I'm wondering in the article what effect/outcome the antidromic signal had in the animal model? did it just fizzle out once it reached the hillock or was there some observable function/benefit? $\endgroup$ – norlesh Sep 28 '15 at 14:44
  • $\begingroup$ @norlesh - It was merely observed. They just monitored this type of neuron and saw that some spikes were out of sync. By doing the math they figured they must have been antidromic. It's a short observatory report (4 pages only). Happy to email the pdf, grab me on chat $\endgroup$ – AliceD Sep 28 '15 at 14:59

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.