[I striked through some passages that might be misleading, resp. where I have been misled.]
Is the following - very simplified - model biologically (or physically) plausible, explaining how spatiotemporal processing on dendrites works, especially how dendritic spikes and short-term
potentiation modification of synaptic strength might work together? (I've read the wikipedia article on NMDA receptors but found it quite hard to digest. It's not a gentle introduction into the topic.)
Let the blue dots be two ligand- and voltage-dependent ion channels
(eg. NMDA-gated channels) sitting on two synapses.
Let the green dot be a voltage-dependent ion channel that generates dendritic spikes.
Let the red dots be voltage-dependent ion channels that generate action potentials.
Without the green channel (generating dendritic spikes), two simultaneously synaptic inputs with depolarization V0 which sum up at the branching point of the dendrite would not suffice to generate an action potential:
But with it, two subsequent synaptic inputs (e.g. in two synchronous spike trains) would, because the postsynaptic response is stronger due to an additional voltage signal from the dendritic spikes.
At time t2 another synaptic input and the dendritic spikes arrive at the blue channels. The resulting depolarization V1 > V0 suffices to evoke an action potential.
Is this picture correct in principle? Do such processes actually occur? Or are the real processes quite different? (Of course, they will be much more complicated and superimposed with other processes.) At least some NMDA receptors (which are ligand- and voltage-gated) sit on synapses.