Recently, two properties of action potential generation in
cortical
neurons were highlighted:
1) their rapid rate of rise at
onset;
2)
their variable threshold (Naundorf et al., Nature 440:1060).
Neither of these properties are predicted by
Hodgkin-Huxley style
single compartment models of action potential generation.
Could they come from the
properties of spike initiation in the axon?
Figure 1. Spikes recorded
simultaneously from the soma (a) and axon initial segment (d) of a
cortical pyramidal neuron, as per the method of Shu et al. Nature
441:
761 and J.
Neurophysiol, in press. Note the "kink" in the action
potential in the soma and the
smooth nature of the action potential in the axon initial
segment. B, E. Phase plots of dV/dt versus membrane
potential for the somatic and axonal action potentials. Insets
illustrate the initial rate of rise of the action potentials.
Note the biphasic nature of spike generation in the soma (parts labeled
1, 2). D,
F. Injection of noise into the cell results in a more variable
spike onset voltage in the soma than in the axon initial segment.
What then are the mechanisms for these properties of spike generation?
Here we demonstrate that these properties arise from
the
initiation of action potentials in the axon followed by
back-propagation of the action potentials into the soma. They are
predicted by Hodgkin and Huxley style models, if the spatial geometry
of the cell is taken into account. We
have
published a Brief Communication Arising in Nature on this topic.
And we
have
published a full length paper in Journal of Neuroscience on this topic.
