What is the Source of the EEG?

Prévia do material em texto

CLINICAL EEG and NEUROSCIENCE 02009 VOL. 40 NO. 3 
What is the Source of the EEG? 
Tim0 Kirschstein and Rudiger Kohling 
Key Words 
Cortex 
Excitatory Postsynaptic Potential 
GABA 
Glutamate 
Inhibitory Postsynaptic Potential 
Synapse 
Thalamus 
ABSTRACT 
Neurons in the human cortex generally process their information by 
means of electrical signals and thus enable the electrical recording of 
their activity, the electroencephalogram (EEG). Due to their unique 
orientation with their long apical dendrites perpendicular to the cortical 
surface, large cortical pyramidal neurons in deep cortical layers play a 
major role in the generation of the EEG. Specific and non-specific 
thalamic nuclei, as well as distant cortical areas, terminate on these 
apical dendrites and form myriads of excitatory and inhibitory afferents. 
The release of excitatory and inhibitory neurotransmitters by these 
fibers activates specific postsynaptic receptors and generates 
excitatory and inhibitory postsynaptic potentials, respectively. By 
electrotonic spread of postsynaptic potentials along the apical dendrites 
and equivalent capacitive currents, they become electrical dipoles. 
Positive or negative deflections are generated by both excitatory and 
inhibitory afferents, depending on the location of these synapses on the 
apical dendrites. Negative (upward) deflections are due to superficial 
excitatory or deep inhibitory inputs, whereas positive (downward) 
deflections represent deep excitatory or superficial inhibitory inputs. 
ELECTRICAL POTENTIALS IN NEURONS: CHARGE 
TRANSFER ACROSS THE CELL MEMBRANE 
The EEG is the registration of neuronal activity within the human 
brain. Basically, this reflects changes in the resting membrane potential 
inside the cell which is measured negative against the extracellular 
space. On the one hand, the resting membrane potential is due to 
energy-consuming membrane pumps such as the sodium-potassium 
pump providing an imbalance of sodium (Na') and potassium ( K ) ions. 
While Na- ions are actively carried to the extracellular solution, K' ions 
are pumped into the cell. At rest, there is, on the other hand, a K 
conductance which allows a K' current from intracellular to 
extracellular according to the K' concentration gradient. This current is 
accompanied by an efflux of positive charge carriers resulting in a 
negative cytosolic charge (hyperpolarization), which, in turn, causes an 
electrostatic attraction on K' ions in the extracellular space back to the 
cytosol. Finally, a dynamic balance between K' influx and K' efflux is 
achieved. This dynamic equilibrium between concentration gradient 
and electrostatic forces of K' ions is called the K* equilibrium potential 
and essentially determines the resting membrane potential (approx. - 
90 mV). 
Changes of the membrane potential in neurons are commonly 
seen as action potentials. These are fast changes of the negative 
membrane potential due to the rapid opening (approx. 1 ms) of 
voltage-gated Na' channels that allows a Na' influx and reverses the 
polarity of the cytosol (depolarization). This depolarization ends by 
automatic closing of Na' channels (inactivation); in addition it opens 
voltage-gated K channels which help to restore the negative cytosolic 
potential by a K' efflux (repolarization). Measured intracellularly, action 
potentials exhibit the largest potential changes that occur in neurons, 
however, their extracellular amplitude is so small that a substantial 
summation of a large number of simultaneous action potentials would 
be required to produce an EEG signal. Lasting only about 1 ms, they 
are too short to sum up sufficiently to generate a potential that can be 
registered extracellularly on the scalp. 
So what is the source of the potential changes in the EEG? Neurons, 
in particular cortical pyramidal cells that are extremely important in terms 
of EEG production, form synapses at their dendrites to make contact 
with myriads of other neurons via corticocortical and thalamocortical 
nerve fibers (Figure l). ' At these synapses, neuronal activity is 
transferred chemically from one cell to another via a neurotransmitter 
released from the presynaptic cell and a specific receptor located on the 
postsynaptic cell. Based on the type of neurotransmitter and receptor, 
synapses can be excitatory or inhibitory. At the postsynaptic membrane, 
excitatory and inhibitory postsynaptic potentials are generated spreading 
electrotonically along the cell membrane of the dendrites to the soma. 
This will be integrated with postsynaptic potentials coming from other 
dendrites. If this integration results in a supra-threshold depolarization at 
the initial axon segment, an action potential is elicited which propagates 
to the synaptic terminal to cause transmitter release. 
Besides direct afferents from the thalamus and other cortical areas, 
cortical pyramidal cells also receive synaptic input by interneurons being 
themselves activated via corticocortical or thalamocortical fibers. 
Thalamocortical afferents are not evenly distributed. Rather, non-specific 
thalamocortical afferents ascend toward superficial cortical layers to 
terminate on pyramidal cells. In contrast, specific thalamocortical 
afferents related to the diverse sensory organs make synaptic contacts 
with pyramidal cells preferably in deeper cortical layers. Corticocortical 
afferents form synapses with pyramidal cells in virtually all cortical layers. 
POSTSYNAPTIC POTENTIALS OF CORTICAL 
PYRAMIDAL CELLS: THE SOURCE OF EEG 
Whereas action potentials are too short to sufficiently sum up, 
postsynaptic potentials having durations of up to several 10 ms are able 
to produce potential changes to be recorded extracellularly on the scalp 
From the Institute of Physiology, University of Rostock, Germany 
Address requests for reprints to Timo Kirschstein. MD, University of Rostock. Institute of 
Physiology. Gertrudenstrasse 9, 18055 Rostock, Germany. 
Email: timo. kirschstein@uni-rostock.de 
Invited. November 10, 2008. accepted March 25, 2009. 
146 
CLINICAL EEG and NEUROSCIENCE 02009 VOL. 40 NO. 3 
I 
II 
111 
IV 
V 
VI 
1 I 
Figure 1. Cortical pyramidal neurons 
Large pyramidal neurons have a long apical dendrite ascending up to the 
surface as well as several basal dendrites, and are innervated by certain brain 
areas in distinct cortical layers (Laminae I to VI, left margin). Non-specific 
thalamocortical afferents form synapses with pyramidal neurons in superficial 
layers, whereas specific thalamocortical afferents predominantly terminate in 
deep layers. Corticocortical afferents make synaptic contacts in all layers of 
the cortex. lnterneurons are also found in the cortex and are interconnected. 
Modified from ref.' 
Figure 3. Inhibitory synapse 
The prototypic inhibitory synapse in the cortex uses GABA as the 
neurotransmitter. Here, the presynaptic action potential opens voltage-gated 
Ca" channels, and the presynaptic Ca" influx causes the GABA release 
stored in presynaptic vesicles. GABA binds to specific postsynaptic receptors 
and leads to a C f influx and K' efflux producing the negative inhibitory 
postsynaptic potential (IPSP). Modified from ref.' 
(EEG). How are postsynaptic potentials generated? The scheme in 
Figure 2 illustrates an excitatory glutamatergic synapse. The presynaptic 
excitation presents as an action potential (opening of voltagegated Na' 
channels) and causes the opening of voltagegated Ca2+ channels at the 
presynaptic terminal. The Ca2+ influx is the trigger for the release of 
neurotransmitter (glutamate) from presynaptic vesicles which diffuses 
across the synaptic cleft and binds to specific receptors at the 
postsynaptic membrane. These AMPA and NMDA receptors (named 
after synthetic agonists) are ion channels with Na' or Na'lCa'' 
conductances, respectively resulting in an inward current of positive 
Figure 2. Excitatory synapse 
The prototypic excitatory synapse in the cortex uses glutamateas the 
neurotransmitter. Here, the presynaptic action potential opens voltage-gated 
Ca'' channels, and the presynaptic Ca" influx causes the glutamate release 
stored in presynaptic vesicles. Glutamate binds to specific postsynaptic 
receptors and leads to a Na' and CaZ' influx producing the positive excitatory 
postsynaptic potential (EPSP). Modified from ref.' 
charge carriers (Na', Ca''). This positive potential at the postsynaptic 
membrane (depolarization) is called the excitatory postsynaptic potential 
(EPSP). In the extracellular space, however, a negativity is recorded due 
to the preponderance of negative charge carriers. 
In contrast to this glutamatergic synapse, an inhibitory postsynaptic 
potential (IPSP) is observed at a GABAergic synapse (Figure 3). The 
most important inhibitory neurotransmitter, y-amino butyric acid (GABA), 
binds to GABAA and G A B h receptors. lonotropic GABAA receptors are 
chloride (Cl-) channels and mediate a CI- influx, whereas metabotropic 
G proteincoupled GAB& receptors induce a K outward current. Both 
ionic currents, however, produce a stronger negative membrane 
potential (hyperpolarization), the inhibitory postsynaptic potential. Similar 
to the EPSP described above, the extracellularly recorded IPSP is 
positive due to the preponderance of positive charge carriers. 
CORTICAL PYRAMIDAL NEURONS: DIPOLES IN THE EEG 
Cortical pyramidal neurons are excellent dipoles due to their unique 
anatomical structure with a long apical dendrite perpendicular to the 
cortical surface. The direction of such a dipole is determined by the 
superficial or deep location of synaptic input. Excitatory synapses are 
surrounded by negativity due to the cationic inward current (Na', Ca"). 
As the evolving EPSP spreads electrotonically along the dendritic cell 
membrane, an equivalent capacitive efflux of positive charge occurs on 
the apical dendrite. Distant to the synapse, this produces a positive 
extracellular potential making the pyramidal neuron an extracellular 
dipole. A superficial excitation causes a negative scalp potential (Figure 
4A), while a positive scalp potential (Figure 5A) is due to a deep 
excitation. By convention, a negative potential is recorded as an upward 
deflection in the EEG. Thus, superficial excitatory inputs are followed by 
upward deflections, whereas deep excitatory inputs are registered as 
downward deflections that are commonly somewhat smaller due to the 
distance to the scalp. 
The situation at inhibitory synapses is reversed. The local positivity 
brought about by the inward current of negative charge (CI- influx or K' 
efflux) is contrasted by the distant negativity due to the electrotonic 
147 
CLINICAL EEG and NEUROSCIENCE 02009 VOL. 40 NO. 3 
A 
+ 
/ 
B &I- 
+ 
/ 
Figure 4. Negative (upward) deflections 
Negative (upward) deflections are due to superficial excitatory inputs (A) or 
deep inhibitory inputs (6) to the pyramidal neurons. A superficial excitatory 
input (A) causes an influx of positive charge carriers (EPSP) leading to a 
negative polarity in the extracellular space. Positive charges spread within the 
apical dendrite, and by capacitive efflux along the dendrite they evoke an 
extracellular positivity distant to the synapse (i.e., at the soma). This produces 
a dipole with negative polarity at the scalp which by convention is displayed 
with an upward deflection. The same dipole can also be observed when a 
deep inhibitory synapse is active (B). Here, an inward current of negative 
charge carriers (IPSP) leads to an extracellular positivity while the capacitive 
outward current along the apical dendrite causes a negative extracellular 
polarity at the scalp. However, when compared to the superficial excitatory 
input, the EEG wave following a deep inhibitory input is smaller because the 
source is more distant to the scalp and the IPSP charge carriers have a 
smaller electrochemical gradient. Modified from ref.’ 
I I S 
Figure 6. Synchronization and desynchronization 
Synchronization (A) and desynchronization (B) is shown for three pyramidal 
neurons each of them receiving one superficial excitatory input (EPSP). For 
clarity’s sake, all afferents have the same firing rate (3 Hz). In the 
synchronized EEG (A) all three afferents are fired synchronously so that three 
EPSPs can sum up to produce an EEG wave of 3 Hz at the scalp. A 
desynchronized EEG (B) with lower amplitude and higher frequency occurs, 
when these three afferents fire in an alternating manner. As the example 
demonstrates, the EEG amplitude is no longer summed up, but its frequency 
reaches 9 Hz. Modified from ref.’ 
I 
Figure 5. Positive (downward) deflections 
Positive (downward) deflections are due to deep excitatory inputs (A) or 
superficial inhibitory inputs (B) to the pyramidal neurons. A deep excitatory 
input (A) causes an influx of positive charge carriers (EPSP) leading to a 
negative polarity in the extracellular space. Positive charges spread within the 
apical dendrite, and by capacitive efflux along the dendrite they evoke an 
extracellular positivity distant to the synapse. This produces a dipole with 
positive polarity at the scalp which by convention is displayed with a 
downward deflection. The same dipole can also be observed when a 
superficial inhibitory synapse is active (B). Here, an inward current of negative 
charge carriers (IPSP) leads to an extracellular positivity while the capacitive 
outward current along the apical dendrite causes a negative extracellular 
polarity distant to the synapse (i.e.. at the soma). Generally, the amplitude of 
the EEG wave depends on the distance between the source and the scalp 
and the electrochemical gradient of the charge carriers. Thus, it is difficult to 
predict whether deep excitatory or superficial inhibitory inputs produce larger 
EEG signals. Modified from ref.’ 
, 
spread of the IPSP and equivalent capacitive efflux of negative charge. 
Again, the cortical pyramidal neuron becomes a dipole with a negative 
scalp polarity when inhibited in deeper layers (Figure 4B), or a positive 
scalp polarity when inhibited superficially (Figure 58). The deep 
inhibition is also smaller compared to the superficial inhibition due to 
the distance to the scalp. 
Thus, the EEG is the total of all excitatory and inhibitory 
postsynaptic potentials of cortical pyramidal neurons that produce a 
vertical dipole perpendicular to the scalp. Since CI- and K as main 
charge carriers of the IPSP have a smaller electrochemical gradient 
than Na’ and CaZ’ (the charge carriers of the EPSP), excitatory 
postsynaptic potentials predominate as generators of the EEG waves. 
Further, since the cortex is structured in gyri and sub, pyramidal cells 
in the gyral walls form horizontal dipoles which are seen as “neutral 
charges” by the scalp electrode and thus are hardly registered in the 
EEG. Rather, the EEG preferably detects pyramidal dipoles on the top 
of a gyrus or - with less amplitude due to the distance to the scalp - 
pyramidal dipoles in the depth of a sulcus. However, certain EEG 
patterns contain horizontal dipoles such as centrotemporal spikes in 
Rolandic epilepsy or small sharp spikes during sleep. 
SYNCHRONIZATION AND DESYNCHRONIZATION: FREQUENCY 
AND AMPLITUDE OF THE EEG 
How do postsynaptic potentials of cortical pyramidal neurons 
become EEG waves? As already mentioned, single EPSPs and IPSPs 
are too small, so that a sufficiently large number of postsynaptic 
148 
Douglas Canone Garcia
Highlight
Douglas Canone Garcia
Highlight
CLINICAL EEG and NEUROSCIENCE 02009 VOL. 40 NO 3 
potentials being simultaneous and in the same direction are needed 
to sum up in order to generate EEG waves on the scalp. Due to the 
local vicinity to the scalp, superficial postsynaptic potentials from non- 
specific thalarnocortical afferents are the main source of the EEG. The 
synaptic activity of these afferents is differently synchronized and 
depends on the level of vigilance. The moresynchronously 
postsynaptic potentials are evoked in a pyramidal neuron by its 
afferents, the larger the amplitude of the summed EEG wave gets - 
a process which is called synchronization (Figure 6). But this also 
implies that the frequency of these waves decreases. If all 
thalarnocortical afferents were active synchronously, the frequency of 
the EEG waves would be identical to the synaptic activity of these 
afferents. Such an extreme synchronization takes place during deep 
sleep (theta waves [4-7.5 Hz] and delta waves [0.5-3.5 Hz]). But also 
during awake resting (relaxed, closed eyes) sensory inputs are 
reduced (especially the visual system), and a "relatively synchronous 
EEG (alpha waves [&I3 Hz]) is obtained. By opening the eyes the 
visual information desynchronizes the thalamocortical afferents and 
produces a "relatively" non-synchronous EEG (beta waves [13-30 
Hz]). The alpha blockade following eye opening, the so-called Berger 
effect, is the clinically most important desynchronization, but 
experimentally there is also an acoustic, tactile or mental (e.g., 
arithmetic problem) desynchronization. 
DISCLOSURE AND CONFLICT OF INTEREST 
in relation to this article. 
Tim0 Kirschstein and Rudiger Kohling have no conflicts of interest 
REFERENCE 
1 Kirschstein T Wie entsteht das EEG7 Neurophysiol Lab 2008, 30 29-37 
149 
Douglas Canone Garcia
Highlight