Properties of Real Neurons
by David D. Olmsted (Copyright - 2000, 2006. Free to use for personal and
Last Revised November 4, 2006
Experimental Neural Signal Recording Methods
Experiments which seek to record the electrical response of individual neural cells
generally use hollow glass electrodes filled with some electically conducting solution.
These are pushed down into the brain region of interest until some signal is detected.
This signal is usually from a cell body since it is much larger than the fibers
but even then the signal will be noisy due to all the electrical events from nearby
fibers and cells. This usually requires that the recorded signals be averaged over
The type of signal which is detected greatly depends on the type of stimulus used
and the activation properties of the neural cell. Neurons having very specific activation
conditions will have a low probability of being detected while those with general
activation conditions will have a high probability of being detected. When neurons
were assumed be simple integrate and fire devices one could also assume that all
the cells related to interpreting a stimulus would be detected but that is no longer
Real neurons are not just simple integrate
and fire mechanisms but instead have many ways to modulate their temporal response.
Any information processing theory of the brain must account for this.
The Main Chemical Pathways in the Synapse (Shepherd 1994)
Typical Post-Synaptic Response to 10 Action Potentials (Wang and Stevens, published in Koch - 1999)
The principle biochemical mechanisms
involved in the chemical synapses are shown in figure 1. The neurotransmitters are
synthesized in the neural cell body and actively transported to the end of the dendrites
in little membrane spheres (vesicles). There they are stored until commanded to
merge with the cell body by an influx of calcium ions which enter the cell through
voltage gated ion channels. The released neurotransmitter and the cell membrane
are recycled back into the cell. (Many antidpressent drugs work by blocking the
re-uptake of certain neurotransmitters, especially that of serotonin.). The neurotransmitter
in the synaptic junction binds to ion channels selective for that type of neurotransmitter
located on the targeted post synaptic dendrite. These open thereby allowing
positively charged ions to enter. Since the cell membrane is normally charge polarized
to -70 millivolts this tends to depolarize the membrane towards 0 millivolts. The charge then
propagates towards the cell body following the path of least charge density since similar charges are repulsive.
synapse does not work perfectly or reliably. Its activation probabilities vary between
10 and 90% and even when it works the post-synaptic response is not always the same.
A typical synaptic response is shown in figure 2 from a Schaffer collateral axon
onto a CA1 pyramidal cell in a mammalian hippocampal slice. This synaptic unreliability
introduces uncertainty into the brain's information processing at a very low level indicating
that the brain must be able overcome this uncertainty
during its information processing.
Ionic Current Flow (downward direction) in a Single Channel (from Sigworth, in Corly = 1983)
The variability of the post-synaptic response seems to be due to the thermal noise affecting
the ion channels which tends to open and close at random after they have been
opened by a neurotransmitter. This is shown in figure 3. Numbered explanations
- Channel opens briefly with a 3 picoamp current.
- Channel opens for a longer amount
of time and shows a flicker of current interruption
- Channel closes
- Channel opens.
- Channel opens briefly and has a significant flicker
- A 20 millisecond channel opening
with several flickers (notice the time scale change).
- A long channel opening with little flicker.
Synaptic Response Governed by Type of Neurotransmitter Receptors
Ionotropic Receptor at Top, Metabotropic Receptor at Bottom
The type of neurotransmitter receptor affects the speed, time course, and modulation
potential of the post synaptic response. The most fundamental difference between
receptors is between the ionotropic and the metabotropic types as shown in figure
4. The ionotropic receptors begin and end quickly while the metabotropic receptors
take a longer time to respond and can remain open a longer time due to the cascade
of internal chemical reactions. The metabotropic receptors are also subject to modulation
by other cell processes.
The neurotransmitters themselves can be divided into three
classes depending on their chemical structure with the smaller, faster diffusing
neurotransmitters generally producing quick on and off post synaptic response
times. The type of receptor each neurotransmitter activates further subdivides their
function. The three major classes of neurotransmitter - receptor are shown below with some of their
more important sub-types:
1. Amino Acids (small sized, fast response and fast decay, found in the central nervous system)
A. Glutamate - facilatory
a.NMDA receptor - ionotropic, also voltage dependent
b. Non-NMDA receptor - ionotropic, very fast
B. Aspartate - facilatory
C. GABA- inhibitory
a. GABAA - ionotropic, fast inhibition
b. GABAB - metabotropic, slow inhibition
2. Biogenic Amines (medium sized, intermediate response and staying power)
a. Nicotinic, ionotropic, used for muscle activation
b. Muscarinic, metabotropic, decreases potassium (K) conductance
B. Dopamine - used in emotion and motivation circuits
C. Noradrenaline - used in emotion and motivation circuits
a. Alpha2, metabotropic, increases potassium K) conductance
b. Beta1, metabotropic, decreases potassium (K) conductance
D. Serotonin- used in emotion, motivation, and sleep circuits
3. Neuropeptides (large sized, long time scaled responses, also called neuromodulators, only a few of many listed)
A. Substance P
C. Proctolin Neurotensin
D. Luteinizing hormone releasing hormone (LHRH)
Two Main Types of Glutamate Receptors - One with an AND Response
The Two Main Types of Glutamate Receptors, Non-NMDA on Left, NMDA on Right. (Nicoli, et al - 1988, scanned from McCormick - 1988 )
Two main types
of glutamate receptors are responsible for the fast sensory information processing
in the brain. As shown in figure 5 the conventional non-NMDA receptor opens in response to a
glutamate molecule binding to its receptor site allowing sodium ions to pass in
and potassium ions to pass out.
In contrast the non-conventional NMDA channel has greater restrictions
on its opening conditions. It only opens when glutamate binds to its receptor site AND the receptor region is substantially depolarized, that is, the internal region
is less negative allowing the release of the blocking magnesium ion. When the channel
is open potassium ions escape and both sodium and calcium ions enter the cell.
Membrane Voltage Effects on the Different Glutimate Receptors (Stern, et al - 1992)
The two types of channels are often found together as shown in figure 6. The left
column (A) shows the voltage response of both together. The middle column (B) shows
only the NMDA voltage response. The right column shows only the non-NMDA voltage
response. The rows show their responses at different membrane depolarizations.
The resting voltage polarization of a cell is approximately -70 millivolts and at
that membrane potential the NMDA channel does not operate. It begins to operate
in the stellate cell around -50 millivolts reaching it maximum around -20 millivolts.
It is much longer lasting than the non-NMDA channel. The brief spike seen in the
non-NMDA response is so brief that it very likely has no effect on the neuron decision
on whether to fire off an action potential yet is would seem to be enough to act
as an enabler for the NMDA receptor by producing the initial membrane depolarization.
The response of the NMDA receptor is proportional to the depolarization voltage.
This is what would be expected if it was implementing a multivalued logic AND operation
which passes the least value among its imputs.
This is an another important piece of evidence that the brain impliments logic like
operations. This AND like effect of the NMDA receptor has been noted by several
neural researchers (especially Koch - 1999) but no specific link to multivalued
(fuzzy) logic has ever been made.
Two Types of Microcircuits (from Shepherd and Koch - 1998)
of neural microcircuits add even more time modulation potential to neurons. Significantly,
they are only found in non-cortical regions. (Koch - 1999)
The most typical microcircuits
are shown in figure 7. Illustration “A” is called the spine triad by Koch and it
consists of two excitatory synapses from the same axon branch (a) projecting onto
two different dendrites with the inhibitory (dark) dendrite inhibiting the common
target (b). This type of microcircuit is common in the retina and the thalamus.
Illustration “B” shows a synaptic junction having bidirectional synapses with one
being inhibitory and the other excitatory. This dendro-dendritic reciprocal inhibition
occurs between the mitral and granual neurons in the olfactory bulb. One assumes
that all these different synapses have different time responses and/or have different
Signal Spread in the Neuron
Rate of Charge Dissapation in a Pyrimidal Neuron (Koch, et al - 1996)
Charge Amplification in Pyramidal Neurons (Stuart and Sakmann - 1994)
The long pyramid cell shown in figure
8 is the typical cell of the cerebral cortex. Experimental injections of a finite
charge propagate in both directions along the dendrites. The voltage drops off rapidly
and significantly by a factor of 33 from the base of the apical (top)
tuft of dendrites to the cell body (Koch - 1999, p68).
The total charge is mostly conserved with only a little being lost to leakage.
Figure 8 shows how charge injections decay over time in different regions of the pyramidal cell. The graphs showing the charge are paired. The bottom graph
shows the decay in the electrode while the top graph shows the decay in the neuron.
The number give the time to the centroid of the curve. In accordance with the high
resistance shown by the trunk dendrite linking the top dendritic bush with the cell
body the decay time of the trunk is large being 14.6 milliseconds. In contrast the
top and bottom dendrite bushes have fast having decay times of 6.5 and 2 milliseconds
This suggests that the top bush is for inputs that need to produce long lasting
context signals for the quicker reacting signals terminating at the bottom bush.
The charge injected
into the cell body has the longest decay time of all being 18.7
milliseconds which lengthens the neural signals even more.
In order to keep the amplitude of the charge signal up high eventhough it is being
spread out while it travels down the trunk the pyramidal cells have trunk signal
amplification as shown in figure 9. The top illustration shows the result after a 100 picoamp current
injection at the cell body (soma) with the recording electrode 443 micrometers up the dendrite (almost to the top bush). The dendritic charge pulse
takes about 3 milliseconds (room temperature) to
reach the electrode and is smaller and broader than the axon action potential. The
bottom illustration shows the same
neuron 100 seconds after the injection of QX-314,
a chemical which inactivates the sodium channels. The passive pulse is now quite
small being about 15 millivolts high. Notice the length of both is about 30 milliseconds.
Cory, D.P. (1983). Patch clamp: current excitement in membrane physiology.
Neuroscience Commentary 1:99-110
Koch, C., Rapp, M., and Segev, I. (1996). A short history of time (constants).
Cerebral Cortex 6:93-101
Koch, C (1999). Biophysics of Computation, Information
Processing in Single Neurons. Oxford University Press
McCormick, D.A. (1998). Membrane
properties and neurotransmitter actions. In Gordon Shepherd (ed.)The Synaptic Organization
of the Brain - fourth edition. Oxford University Press
Nicoll, R.A. (1988). The
coupling of neurotransmitter receptors to ion channels in the brain. Science 241:545-551
Shepherd, G. M. (1994). Neurobiology. Oxford University Press
Shepherd, Gordon, M. (1998) editor The Synaptic Organization of the Brain - fourth edition,
Oxford University Press
Stuart, G.J. & Sakmann, B. (1994). Active propagation
of somatic action potentials inot neocortical pyramidal cell dendrites. Nature 367:69-72
Stern, P., Edwards, F.A. and Sakmann, B. (1992). Fast and slow components of unitary
EPSC’s on stellate cells elicited by focal stimulation in slices of rat visual cortex.
Journal of Physiology 449:247-278