recruitment of the NMDA receptors [65, 66]. The neck of the spine is, then, a
powerful modulator of the synaptic information transfer depending on the excitation (depolarization) level of the membrane potential in the dendrites [65, 66].
By considering the huge number of inputs (from 3 Â 10
3
to 3 Â 10
4
) received,
the dendritic arborization is not an electrical isopotential compartment. Differences
of potential among the different branches can be due to different input activities
and by spike backpropagation [71] in those areas where it can occur. The spike back
propagation depends on the presence of Na
+ and/or Ca
2+ voltage-gated channels in
the dendrites [12, 63, 70, 72]. The presence and density of these channels differ both
among neurons and, in the same neuron, among different dendritic regions [12, 63,
70, 72], and consequently differences of potential can produce complex potential
waves transiting the dendrites, and this wave can reach the PSD trough the neck
resistance influencing the single synaptic event. In our recent papers, we have
studied a possible effect of potential waves produced by excitatory synaptic activity
on the single synaptic response independently of the spike backpropagation
[65, 66]. We have found that, depending on the number of active synapses and on
their mean firing frequency, the amplitude, peak level, and time to peak of the
response vary in a complex nonlinear fashion (see Figure 3) [65, 66]. The number
of active synapses in some way simulates the input, for example, received from one
area of the brain where a group of active neurons fire in a more or less synchronous
way (in response to a stimulus) on the same neuron in a restricted dendritic area.
For the case already mentioned, for example, a neuron of a hippocampal subfield
can receive synchronous inputs from a large area (many neurons) of the dentate
gyrus but also from areas of the Entorhinal cortex in separate regions of the dendritic branches (see, e.g., [7]). The neurons from one of these two areas fire with a
mean frequency and a standard deviation which depends on the degree of their
synchronization. Such a condition produces waves into the dendritic area interested
to the stimulation which directly influence any single synapse which is active in the
same time window [65, 66]. The membrane potential of the receiving neuron
oscillates between two levels forming a voltage “band.” The amplitude of this
voltage “band” depends on the number of active synapses and on their mean firing
frequency [65, 66]. The EPSP of a given synapse can occur at any level inside the
“band.” According to Eqs. (1) and (2), depending on the level at which the EPSP of
a given synapse starts, its properties (amplitude, peak level, NMDA contribution,
etc.) will change [65, 66]. In this band it is possible to identify a mean value which
can be considered as the maximal likelihood level of V m at which the EPSP can
occur. This mean level increases (more depolarized) by increasing the number of
active synapses and/or their firing frequency [65, 66]. The existence of this “band”
of voltage furnish a large gamma of possible levels of V m at which EPSP can occur
and consequently it represents a very powerful regulator of the single EPSP
depending on the time of occurrence (phase of the oscillation inside the band)
[65, 66]. Said in a different way, the coincidence of the EPSP with the particular
level of determines the type and amount of information the synapse transfers.
NMDA receptors, being dependent on the membrane voltage for their activity, are
especially sensitive to this kind of regulation, and in fact, the “coincidence” of the
EPSP with the activity of other synapses is considered crucial for phenomena like
LTP and memory which are NMDA dependent. These are the basic mechanisms
who suggest that neurons, mostly in producing LTP and memory phenomena, act as
coincidence detectors (among many others, see, e.g., [73]). The dendritic activity
modulatory effect on the transfer of a single bit of synaptic information depends
essentially on the variation of potential in the membrane and can be summarized as
due to:
100
Advances in Neural Signal Processing
powerful modulator of the synaptic information transfer depending on the excitation (depolarization) level of the membrane potential in the dendrites [65, 66].
By considering the huge number of inputs (from 3 Â 10
3
to 3 Â 10
4
) received,
the dendritic arborization is not an electrical isopotential compartment. Differences
of potential among the different branches can be due to different input activities
and by spike backpropagation [71] in those areas where it can occur. The spike back
propagation depends on the presence of Na
+ and/or Ca
2+ voltage-gated channels in
the dendrites [12, 63, 70, 72]. The presence and density of these channels differ both
among neurons and, in the same neuron, among different dendritic regions [12, 63,
70, 72], and consequently differences of potential can produce complex potential
waves transiting the dendrites, and this wave can reach the PSD trough the neck
resistance influencing the single synaptic event. In our recent papers, we have
studied a possible effect of potential waves produced by excitatory synaptic activity
on the single synaptic response independently of the spike backpropagation
[65, 66]. We have found that, depending on the number of active synapses and on
their mean firing frequency, the amplitude, peak level, and time to peak of the
response vary in a complex nonlinear fashion (see Figure 3) [65, 66]. The number
of active synapses in some way simulates the input, for example, received from one
area of the brain where a group of active neurons fire in a more or less synchronous
way (in response to a stimulus) on the same neuron in a restricted dendritic area.
For the case already mentioned, for example, a neuron of a hippocampal subfield
can receive synchronous inputs from a large area (many neurons) of the dentate
gyrus but also from areas of the Entorhinal cortex in separate regions of the dendritic branches (see, e.g., [7]). The neurons from one of these two areas fire with a
mean frequency and a standard deviation which depends on the degree of their
synchronization. Such a condition produces waves into the dendritic area interested
to the stimulation which directly influence any single synapse which is active in the
same time window [65, 66]. The membrane potential of the receiving neuron
oscillates between two levels forming a voltage “band.” The amplitude of this
voltage “band” depends on the number of active synapses and on their mean firing
frequency [65, 66]. The EPSP of a given synapse can occur at any level inside the
“band.” According to Eqs. (1) and (2), depending on the level at which the EPSP of
a given synapse starts, its properties (amplitude, peak level, NMDA contribution,
etc.) will change [65, 66]. In this band it is possible to identify a mean value which
can be considered as the maximal likelihood level of V m at which the EPSP can
occur. This mean level increases (more depolarized) by increasing the number of
active synapses and/or their firing frequency [65, 66]. The existence of this “band”
of voltage furnish a large gamma of possible levels of V m at which EPSP can occur
and consequently it represents a very powerful regulator of the single EPSP
depending on the time of occurrence (phase of the oscillation inside the band)
[65, 66]. Said in a different way, the coincidence of the EPSP with the particular
level of determines the type and amount of information the synapse transfers.
NMDA receptors, being dependent on the membrane voltage for their activity, are
especially sensitive to this kind of regulation, and in fact, the “coincidence” of the
EPSP with the activity of other synapses is considered crucial for phenomena like
LTP and memory which are NMDA dependent. These are the basic mechanisms
who suggest that neurons, mostly in producing LTP and memory phenomena, act as
coincidence detectors (among many others, see, e.g., [73]). The dendritic activity
modulatory effect on the transfer of a single bit of synaptic information depends
essentially on the variation of potential in the membrane and can be summarized as
due to:
100
Advances in Neural Signal Processing