The exact site of action potential generation is debated, as is the underlying physiological mechanism (Colbert & Johnsten, 1996). It is generally agreed upon that the initiation site is beyond the soma, at the axon hillock, or even further in the initial segment of the axon.
The reason for the distal axonal initiation site of action potentials is, according to the classical view, that the density of Na+ channels is not high enough in the soma, while the density in the axon initial segment is substantially higher and sufficient. However, it has also been proposed that the threshold of action potential generation is lower at the initial segment (Kole et al., 2008).
Regarding your questions in the comments:
So can't action potentials be initiated at dendrites and soma with the help of VGCCs?
Answer: No. Channel density is too low;
What are the channels required for AP?
Answer: Basically Na+ and K+ channels;
How is it that action potentials are initiated at axon hillock, although the input is at dendrites? Why can't enough depolarization happen at dendrites and soma?
Answer: Either due to a low density of Na+ channels, and/or a higher action potential threshold (e.g., due to the presence of K+ channels in the soma counteraction depolarization);
As a corollary,if enough depolarization cannot occur at dendrites and soma,how then is it propagated to the axon, which is further down anatomically , in the neuron?
Answer: If the EPSP reaches a region with sufficient Na+ channels, depolarization becomes sufficient to activate a critical amount of Na+ channels that causes a sufficient increase in the membrane potential to sustain a full-blown action potential. Along the way, leaky K+ channels decrease the depolarization so a critical amount of channels is necessary, otherwise the EPSP dies out.
What about backpropagating APs from axon hillock to dendrites ? How is that possible?
Answer: the soma and dendritic region are believed to represent too big an electrical load to ensure action potential propagation in both the orthodromic and the antidromic direction. This impedance mismatch can be overcome for an action potential propagating from the axon back into the somato-dendritic region when the initial segment and axon hillock provide additional current. This can be achieved with a high density of sodium channels in that region (Lüscher & Larkum, 1998).
- Colbert & Johnsten, J Neurosci (1996), 16(21): 6676–86
- Kole et al., Nature Neurosci (2008); 11: 178-86
- Lüscher & Larkum, J Neurophysiol (1998); 80(2): 715-29