It is estimated that the in the human brain there are about 100 billion neurons.

With thousands of synapses for each of them, the total number of synapses is over 100 trillion.

Yet each neuron has one axon with which it can output signal. Obviously each axon forms more than one synapse (else the number wouldn't add up), but how many? It should be thousands per axon, with several different neurons. Is that so?


You're right - from the simple math in the numbers you've given, you can deduce that each average neuron makes thousands of connections (sending and receiving).

Of course, this varies a bit by cell type:

  • There is one special synapse called the Calyx of Held which is a nearly one-to-one connection between certain types of cells in the auditory brainstem: one globular bushy cell makes only 1 or 2 calyx synapses with MNTB neurons (though they also have collaterals to other targets) which themselves receive 1 or at most 2 calyx synapses (though they also receive a small fraction of input from other sources) (Smith et al 1991, Smith et al 1998, Rodríguez‐Contreras et al 2006)

  • Napper and Harvey 1988 estimate about 175,000 synapses arrive at each Purkinje cell in the cerebellum.

  • The ballpark number of "thousands of synapses per neuron" is reasonable for most cells, however.

Neurons can and do make multiple synaptic connections with the same post-synaptic cell - that's one of the ways to regulate the strength of connection between two neurons - but most synapses will be from different neurons.

Synapses on different compartments of a cell can have different effects, because they tend to sum more with other local connections and less from far ones, and because synapses closer to the cell body are more likely to directly control firing than synapses far away.

Axons from many cell types in neocortex have a complex branching pattern, where axons often branch both locally as well as sending various collaterals to other areas of neocortex and even to subcortical structures (see some examples in Ojima et al. 1991 & 1992). They make synapses all along these branches except for the longest projection stretches.

Hawkins, J., & Ahmad, S. (2016). Why neurons have thousands of synapses, a theory of sequence memory in neocortex. Frontiers in neural circuits, 10, 23.

Napper, R. M. A., & Harvey, R. J. (1988). Number of parallel fiber synapses on an individual Purkinje cell in the cerebellum of the rat. Journal of Comparative Neurology, 274(2), 168-177.

Ojima, H., Honda, C. N., & Jones, E. (1991). Patterns of axon collateralization of identified supragranular pyramidal neurons in the cat auditory cortex. Cerebral Cortex, 1(1), 80-94.

Ojima, H., Honda, C. N., & Jones, E. G. (1992). Characteristics of intracellularly injected infragranular pyramidal neurons in cat primary auditory cortex. Cerebral Cortex, 2(3), 197-216.

Rodríguez‐Contreras, A., de Lange, R. P., Lucassen, P. J., & Borst, J. G. G. (2006). Branching of calyceal afferents during postnatal development in the rat auditory brainstem. Journal of Comparative Neurology, 496(2), 214-228.

Smith, P. H., Joris, P. X., Carney, L. H., & Yin, T. C. (1991). Projections of physiologically characterized globular bushy cell axons from the cochlear nucleus of the cat. Journal of comparative neurology, 304(3), 387-407.

Smith, P. H., Joris, P. X., & Yin, T. C. (1998). Anatomy and physiology of principal cells of the medial nucleus of the trapezoid body (MNTB) of the cat. Journal of neurophysiology, 79(6), 3127-3142.

  • 1
    $\begingroup$ It has been years but my recollection is that the end bulb and calyx of Held are really one to many in terms of connections (i.e., one GBC connects to many principal cells) and the one-to-one relationship refers to neural firing. The principal cell fires if, and only if, the GBC fires. $\endgroup$
    – StrongBad
    Jan 8 '20 at 2:37
  • $\begingroup$ @StrongBad Thanks, I'll look in to it. My recollection is 1:1 although the cell on each end itself has many inputs/outputs in the rest of the pathway. It may be that there are collaterals and modulatory inputs besides the specialized calix. $\endgroup$
    – Bryan Krause
    Jan 8 '20 at 4:10
  • $\begingroup$ @StrongBad I went back to some old papers from my colleagues and indeed, at least in the cat, there can be more than one calyx from each GBC and possibly more than one calyx to each MNTB cell. However, it seems like the ratios are 1:2 or 2:1 rather than 1:many. I also clarified that the GBCs have other collaterals and that the MNTB cells receive other (at most minor) inputs. I did a quick look at some other papers and this looks consistent in other species, with most evidence suggesting the majority MNTB cells receive only 1 calyx while 80%+ of GBCs make only 1. $\endgroup$
    – Bryan Krause
    Jan 8 '20 at 5:52
  • $\begingroup$ I think I was most remiss in forgetting about the GBC axon collaterals as these are of course crucial to the circuit function. Indeed the calyx itself would have no purpose otherwise. $\endgroup$
    – Bryan Krause
    Jan 8 '20 at 5:56

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.