The neurobiology of migraine
ANDREW CHARLES* AND K.C. BRENNAN
Headache Research and Treatment Program, Department of Neurology,
David Geffen School of Medicine at UCLA, Los Angeles, CA, USA
There is substantial clinical and imaging evidence for
changes in cortical activity associated with migraine.
Migraine is associated with a variety of symptoms that
can be attributed to changes in cortical function. The
most prominent among these are thevisualchanges asso-
in the occipital lobe. Migraine patients may also experi-
ence corticalsensory,motor, language,orothercognitive
dysfunction. These symptoms are typically described in
the context of migraine with aura, but they may also
for migraine with aura. Functional imaging studies con-
firm that migraine is associated with dramatic changes
in blood flow and metabolic activity in the cortex (Cutrer
and Black, 2006). These imaging phenomena have been
demonstrated primarily in patients with migraine with
aura, but may also occur in patients with migraine with-
out aura (Woods et al., 1994; Geraud et al., 2005).
Migraine-related changes in blood flow and functional
magnetic resonance imaging (fMRI) signal in the cortex
are propagated with temporal and spatial characteristics
that are remarkably similar to those of cortical spreading
depression (CSD), the spreading wave of depolarization
followed by suppression of electrocortical activity ori-
ginally described by Lea ˜o in 1944 (Woods et al., 1994;
Hadjikhani et al., 2001). The correlation between the
characteristics of the clinical symptoms of migraine
aura, CSD in animal models, and functional imaging
has provided support for the long-standing hypothesis
that CSD is a fundamental mechanism of migraine aura.
A much more controversial question continues to be
whether similar cortical phenomena may also occur in
migraine without aura. As discussed below, distinct pat-
terns of signaling in individual cellular compartments
could underlie cortical activity that does not necessarily
evoke classical aura symptoms.
Other evidence for fundamental changes in cortical
excitability in migraine comes from clinical electrophy-
siology studies. A significant number of studies find
an increased amplitude as well as a decreased habitua-
tion of cortical evoked potentials in migraine patients
compared with controls during the interictal period,
with normalization of these differences during the ictal
period (Schoenen, 2006). Other studies also show that
the threshold for generation of phosphenes by trans-
cranial magnetic stimulation is reduced in migraine
patients (Aurora et al., 1999, 2003; Gerwig et al.,
2005). Migraine patients have also been reported to
show reduced magnetic stimulation-induced suppression
of visual accuracy (Aurora et al., 2007). These findings
are consistent with an increased cortical excitability (or
decreased inhibition) in patients with migraine.
However, there are also a significant number of stu-
dies that show either no differences in clinical electro-
physiological parameters between migraine patients
and controls, or in fact changes in the opposite direc-
tion consistent with a reduced cortical excitability in
migraine patients (Ambrosini and Shoenen, 2006).
The discrepancies between these studies may arise in
part from methodological differences in the way the
studies were performed. But another key explanation
for these discrepancies is that the level of cortical
excitability in migraine patients may vary substantially
over time. Consistent with this idea, the thresholds for
phosphene generation evoked by consecutive trans-
cranial magnetic stimulation were found to be more
*Correspondence to: Andrew Charles MD, Headache Research and Treatment Program, Department of Neurology, David Geffen
School of Medicine at UCLA, 635 Charles Young Drive, Los Angeles, CA 90095, USA, Tel: 310-206-7226, Fax: 310-206-6906,
Handbook of Clinical Neurology, Vol. 97 (3rd series)
G. Nappi and M.A. Moskowitz, Editors
# 2011 Elsevier B.V. All rights reserved
variable in migraine patients than in controls (Antal
et al., 2006). This suggests that, rather than having cor-
tical excitability that can be characterized simply as
either increased or reduced, migraine patients have a
dysregulation of cortical excitability (Ambrosini and
Shoenen, 2006; Stankewitz and May, 2007). Abnor-
mally wide swings in cortical excitability are an appeal-
ing explanation for the complex variety of symptoms
that are experienced by migraine patients.
There is also strong evidence that brainstem mechan-
isms play a significant role in the pathophysiology of
migraine. Nausea, vertigo, and autonomic symptoms
are among the clinical features of migraine that may
arise from an alteration of signaling in the brainstem.
It has also been suggested that the pain of migraine
may arise primarily from the brainstem (Weiller
et al., 1995; Tajti et al., 2001). Functional imaging
studies of migraine patients consistently demonstrate
activation of the brainstem during migraine attacks,
particularly the region of the dorsolateral pons (Bahra
et al., 2001; Afridi et al., 2005; Denuelle et al., 2007).
Positron emission tomography (PET) and fMRI studies
suggest that metabolism and function in the brainstem
may also be chronically altered in patients with chronic
migraine (Welch et al., 2001; Aurora et al., 2007).
There have also been multiple reports of structural
lesions in the brainstem that appear to cause headache
in humans (Haas et al., 1993; Goadsby, 2002; Fragoso
and Brooks, 2007). Furthermore, electrical stimulation
in the region of the periaqueductal gray can evoke
headache (Raskin et al., 1987).
The trigeminal pathway
Although it is clear from functional imaging studies
that multiple brain regions involved in nociception are
activated during a migraine attack (Weiller et al.,
1995; Cao et al., 1999, 2002; Bahra et al., 2001; Afridi
et al., 2005; Denuelle et al., 2007), the site where the
initial activation of these nociceptive pathways occurs
has not been determined with certainty. The idea that
dilation of cerebral vessels is a primary cause of
migraine pain has been challenged by a variety of
evidence. However, the “trigeminovascular system”
continues to be widely accepted as an important com-
ponent of the headache. This concept developed based
on the early observations that stimulation of trigemin-
ally innervated vessels evoked pain in conscious
patients, whereas stimulation of the brain parenchyma
did not evoke discomfort (Penfield, 1935; Ray and
Wolff, 1940). A role for the trigeminovascular system
was supported by the observation that calcitonin
gene-related peptide (CGRP), a neuropeptide known
to be involved in cerebral vasoregulation, is increased
in jugular venous blood during a migraine attack
(Goadsby and Edvinsson, 1994). The idea of trigemi-
novascular activation has been reinforced by the
effects of triptans, and more recently CGRP antago-
nists, whose effects have been presumed to be in
large part mediated by inhibition of perivascular tri-
geminal nociceptive input. However, the findings that
activation of 5-HT1 receptors by triptans at multiple
other sites was involved in trigeminal nociception,
including the nucleus caudalis, periaqueductal gray,
and thalamus indicate that the perivascular trigeminal
nerve endings are not necessarily the only site of
action of these drugs (Goadsby and Hoskin, 1996;
Ellrich et al., 2001; Boers et al., 2004; Shields and
Goadsby, 2006). Other migraine therapies such as
ergotamines, that had initially been presumed to act
primarily at the peripheral trigeminal vascular com-
plex, have also been shown to have central mechan-
isms of action (Hoskin et al., 1996).
neurons are found through the central and peripheral
in the brainstem) (Tajti et al., 2001; Jenkins et al., 2004;
Offenhauser et al., 2005; D’Hanis, 2007), and these
neurons could represent an alternative source of the
CGRP that is released during a migraine attack. Thus,
although there is considerable indirect evidence for tri-
geminovascular activation in migraine, there is as yet
no direct evidence that confirms this hypothesis.
This may be because there is no reasonable way of
measuring such activation in humans. However, it is
interesting to note that trigeminal rhizotomy, even
when complete disruption of trigeminal sensory func-
tion has been achieved, is not consistently effective in
the prevention of migraine or cluster headache that is
experienced on the side ipsilateral to the trigeminal
deafferentation (Matharu and Goadsby, 2002). This
observation suggests that it is possible for nociceptive
pathways responsible for headache to be activated in
the absence of peripheral trigeminal input from the
same side on which the headache is experienced.
Further investigation regarding the sites of actions of
the triptans and CGRP receptor antagonists in humans
may shed important light on this issue.
Is there an anatomical sequence
The sequence of activation of different brain regions
in migraine remains uncertain. At this stage it is not
clear whether changes in cortical activity activate the
100 A. CHARLES AND K.C. BRENNAN
brainstem, or vice versa. Alternatively, changes in
these brain regions could occur in parallel, without an
orderly sequence from one to the next. The typical
occurrence of the migraine aura before migraine head-
ache supports the hypothesis that cortical activation
precedes brainstem activation. Consistent with this con-
cept, important studies in experimental models demon-
strate that it is possible for CSD to activate neurons in
the trigeminal nucleus caudalis via trigeminal afferents
(Bolay et al., 2002). Conversely, it has been shown that
brainstem activation can evoke changes in cortical
blood flow, raising the possibility that a process begin-
ning in the brainstem could secondarily evoke some of
the cortical phenomena of migraine (Adams et al.,
1989; Goadsby and Duckworth, 1989). But it is also
possible that both brain regions are activated simulta-
neously, a concept that is supported by the observa-
tions that, in some patients, clinical symptoms may
occur without any clearly defined sequence that indi-
cates alteration in the function of one region leading
to change in function of another. Regardless of the
order of their activation, however, it is clear that both
cortical and brainstem signaling mechanisms are
involved in migraine and are appealing targets for
new therapeutic approaches for migraine.
The concept of migraine as a primarily vascular head-
ache has given way to the understanding that it is an
episodic disorder of brain excitability that involves
coordinated changes in neuronal, glial, and vascular
function. Examination of the individual cellular com-
ponents of migraine identifies new potential pathophy-
siological mechanisms, and new targets for treatment.
It is generally assumed that migraine pain is initiated
by activation of meningeal sensory afferents that
synapse in the trigeminal nucleus caudalis, although,
as discussed above, this assumption has not been defi-
nitively confirmed. Trigeminal neurons that innervate
the meninges have properties that are in many ways
similar to other visceral sensory neurons (Strassman
and Levy, 2006). The only sensation that they appear
to transmit is pain. However, the specific stimuli that
trigger sensory neurons to produce migraine pain are
not certain. A variety have been proposed, including
mechanical pressure, changes in the ionic composition
of the extracellular space (e.g., increased Kþ, increased
osmolarity, and decreased pH), neuropeptides (e.g.,
bradykinin, substance P, endothelin, and CGRP), neuro-
transmitters (glutamate, serotonin, histamine, adenosine
triphosphate, adenosine), eicosanoids, and nitric oxide.
There may be subsets of trigeminal neurons that
respond specifically to different stimuli with different
thresholds. Individual neurons with conduction veloci-
ties consistent with A and C fibers have been identi-
fied, with different sensitivities to mechanical versus
neurochemical stimuli, and different responses to repe-
titive or sequential combinations of stimuli (Levy and
Peripheral trigeminal neurons, as with other neurons
in the trigeminal pathway, show the phenomenon of
sensitization, whereby exposure to an algesic stimulus
lowers the threshold for the response of the neuron to
the same stimulus, or a different stimulus (Strassman
et al., 1996). For example, exposure to inflammatory
mediators lowers the threshold of trigeminal neurons
for responding to a mechanical stimulus (Levy and
Strassman, 2002). This type of sensitization has been
suggested as a mechanism whereby otherwise painless
localized or diffuse changes in pressure (such as might
occur with vascular pulsations) could result in the pul-
sating pain of migraine. It has been proposed that both
peripheral and central sensitization are involved in the
generation and maintenance of migraine pain, as well
as the associated phenomenon of cutaneous allodynia.
A complex interplay between neurons, glial cells, and
vascular cells may be particularly critical in cortical
mechanisms of migraine. As discussed above, the ana-
tomical spread of migraine aura symptoms as well as
the propagated changes in blood flow and metabolism
observed in migraine patients has implicated the phe-
nomenon of CSD in migraine. CSD was originally
reported by Lea ˜o in 1944 as a suppression of electrical
activity that spread slowly across large areas of the
cortex (Lea ˜o, 1944). Subsequent recordings demon-
strated that the suppression of activity was preceded
by a dramatic depolarization, resulting in a “DC [direct
current] shift” of the electrocortical signal, indicating
that the CSD wave consists of a propagated wave of
profound cortical activation followed by sustained
inhibition of activity (Lea ˜o, 1947).
The clinical symptoms of migraine aura, as well as
the clinical electrophysiological and transcranial mag-
netic stimulation responses of migraine patients, indi-
cate a fundamental role for changes in neuronal
excitability as a basis for increased cortical excitability
in migraine. Other evidence comes from genetic and
pharmacological studies. Mutations in neuron-specific
genes have been identified as the cause of familial
hemiplegic migraine in some families. These include
mutations in the P/Q-type calcium channel (FHM1)
THE NEUROBIOLOGY OF MIGRAINE101
(Ophoff et al., 1996; van den Maagdenberg et al., 2004),
and SCN1A voltage-gated sodium channel (Dichgans
of the neuron.
Pharmacological studies indicate that inhibition of
neuron-specific receptors or channels can inhibit CSD.
Inhibition of N-methyl-d-aspartate (NMDA) receptors,
which are expressed primarily on neurons, can inhibit
the occurrence of CSD in vitro and in vivo preparations
(Hernandez-Caceres et al., 1987; Peeters et al., 2007).
CSD in vitro can also be blocked by inhibitors of P/Q-
type calcium channels, which are specific to neurons
(Kunkler and Kraig, 2004). Conversely, studies with a
knock-in mouse model expressing a mutation of the
P/Q calcium channel gene associated with familial hemi-
plegic migraine show that alteration in the function of
this neuronal channel lowers the threshold for evoking
CSD, and increases its rate of propagation (van den
Maagdenberg et al., 2004). These studies indicate that
neuronal hyperexcitability is involved in the triggering
of CSD and its propagation.
Glial cells may also play a key role in the changes in
cortical activity associated with migraine. Direct evi-
dence for such a role comes from the discovery that
a mutation in an Naþ/KþATPase that is expressed pri-
marily in astrocytes is responsible for familial hemiple-
gic migraine type 2 (De Fusco et al., 2003; Vanmolkot
et al., 2006). In vitro studies indicate that this mutation
reduces the function of the enzyme (Segall et al., 2005;
Vanmolkot et al., 2006), an effect that would be
expected to increase excitability by increasing extracel-
lular Kþ. Astrocytes are abundant cells in the central
nervous system that have been traditionally viewed as
playing only a passive and supportive role in nervous
system function. But recent studies demonstrate that
astrocytes are capable of extensive intercellular sig-
naling that can modulate both neuronal and vascular
activity. Astrocytes express a variety of neurotrans-
mitter receptors that allow them to respond to neuro-
nal activity (Fellin et al., 2006). Conversely, they are
capable of active release of transmitters, including
glutamate and adenosine triphosphate (ATP), that can
modulate neuronal function (Haydon and Carmignoto,
2006). Astrocytes are also in close contact with vascu-
lar cells via their endfeet, that enwrap blood vessels.
Astrocyte signaling has been shown to modulate
vascular tone directly, resulting in either vasoconstric-
tion or vasodilation via release of eicosanoids, Kþ,
and ATP (Zonta et al., 2003; Mulligan and MacVicar,
2004; Filosa et al., 2006; Takano et al., 2006).
Astrocytes are capable of extensive intercellular
communication via increases in intracellular calcium
concentration that are propagated from cell to cell in
a wave-like pattern (Charles et al., 1991). The primary
mechanism for these intercellular calcium waves
appears to be release of ATP into the extracellular
space and activation of purinergic receptors on adja-
cent cells (Guthrie et al., 1999). These intercellular cal-
cium waves in astrocytes can be triggered by chemical,
electrical, or mechanical stimuli, and spread with tem-
poral and spatial characteristics that are remarkably
similar to those of CSD (Charles, 1998). In fact, recent
microscopic imaging studies have shown that calcium
waves in astrocytes consistently occur in association
with spreading depression, both in vitro and in vivo
(Peters et al., 2003; Chuquet et al., 2007). However,
each phenomenon can occur independently of the
other. Inhibition of astrocyte calcium waves does not
block spreading depression, and astrocyte calcium
waves occur in the absence of spreading depression
(Peters et al., 2003; Chuquet et al., 2007). Thus, they
appear to be related phenomena that occur in parallel
but without a requisite interdependence.
It has been assumed for decades that spreading
depression as it is observed in animal models is the
physiological basis for migraine aura. But the classical
electroencephalogram (EEG) changes of CSD have not
been observed in migraine patients. This could be
because surface EEG recordings have not been sensi-
tive enough to detect it. However, CSD in animal mod-
els is a profound neurophysiological event. It would
be expected to produce not only EEG changes, but
also a greater degree of neurological impairment than
is observed in most migraine patients. A speculative
alternative to classical CSD as a mechanism of the pro-
pagated cortical changes in migraine is a phenomenon
that involves astrocyte waves. Astrocyte calcium waves
are associated with release of significant amounts of
ATP, glutamate, and vasoactive mediators (Haydon
and Carmignoto, 2006). Based on this function, and
based on their close spatial relationship with both
neurons and vascular cells, astrocytes are ideally posi-
tioned to modulate widely propagated changes in both
vascular activity and neuronal activity with the pattern
that is observed in migraine patients. A primary role
for astrocytes could explain both the propagated
changes in blood flow and metabolism that are
observed with functional imaging studies, as well as
the cortical symptoms that in most patients are not as
profound as might be expected to result from classical
CSD. Even if astrocytes do not play such a primary
role, it is likely that their activity represents a sub-
stantial component of the cellular pathophysiology of
102 A. CHARLES AND K.C. BRENNAN
In addition to the changes in neuronal and glial cell
function, migraine clearly involves significant changes
in the activity of vascular cells. Mutations in genes
encoding proteins expressed predominantly by vascu-
lar cells have been found in families with vasculo-
pathies that have migraine as part of the clinical
phenotype. Mutations in the gene encoding Notch3,
which in adult humans is expressed predominantly in
vascular smooth-muscle cells, cause cerebral autoso-
mal-dominant arteriopathy with subcortical infarcts
and leukoecephelopathy (CADASIL). This condition
begins with migraine with aura in approximately
one-third of patients (Dichgans et al., 1998). Muta-
tions in the gene encoding the widely expressed exo-
nuclease TREX1 also cause a vasculopathy that may
have migraine as part of the phenotype (Richards
et al., 2007).
As with astrocytes, it has long been assumed that
vascular cells simply respond passively to changes in
neuronal function. But it is important to keep in mind
that vascular cells are also capable of active intercellu-
lar signaling, and can release messengers like nitric
oxide that could potentially modulate the function of
surrounding neurons and glia. As mentioned above,
the longstanding hypothesis that migraine pain is a
direct consequence of vasodilation has been increas-
ingly challenged by functional imaging studies in
migraine patients. The classic blood flow studies by
Olesen and colleagues (1990) clearly demonstrated an
increase in blood flow in patients with migraine. How-
ever, they found that the onset of pain preceded the
onset of increased blood flow, and in fact began dur-
ing the period of hypoperfusion that was associated
with migraine aura. Conversely, the phase of increased
blood flow often persisted for a significant duration
after the pain of migraine had stopped. These studies
suggest that vasodilation and the pain of migraine are
not necessarily temporally correlated.
A variety of other functional imaging studies with
PET and fMRI techniques also show cortical hypoper-
fusion during the onset of migraine pain (Woods et al.,
1994; Hadjikhani et al., 2001; Geraud et al., 2005).
Still other more recent studies with pharmacologically
induced migraine also demonstrate that the onset of
pain is not correlated with vasodilation. Olesen and col-
leagues found that, with migraine induced by sildena-
fil, the pain of migraine did not begin until after
recovery from the drug-induced cerebral vasodilation,
as indicated by transcranial Doppler (Kruuse et al.,
2003). Similarly, Schoonman and colleagues (2008)
used magnetic resonance angiography techniques to
show that migraine pain evoked by nitroglycerin did
not occur until after recovery from nitroglycerin-
induced cerebral vasodilation.
The complex vascular changes that are observed
with CSD may provide some insight into the complex
vascular phenomena seen with functional imaging stu-
dies in migraine patients. Spreading depression in mice
is associated with a multiphasic vascular response.
There is an initial dilation of cortical surface vessels
that may be actively propagated along the vessel via
an intrinsic vascular mechanism (Brennan et al.,
2007b). This is followed by a profound constriction
of the vessels, after which there is subsequent vasodila-
tion and then eventual recovery to normal caliber
(Ayata et al., 2004; Brennan et al., 2007b). There may
be significant species-specific and methodology-related
differences in the presence and extent of the vasocon-
striction component of the surface arteriolar response
to CSD. In humans, a significant vasoconstriction is
supported by functional imaging studies that consis-
tently show a hypoperfusion associated with migraine
aura, and even in migraine without aura (Woods
et al., 1994; Cao et al., 1999; Geraud et al., 2005).
In vivo imaging studies in rodent models provide evi-
dence that astrocyte calcium waves mediate the propa-
gated vasoconstriction associated with CSD, indicating
a primary role for glial cells in this process (Chuquet
et al., 2007) (Figure 7.1).
Vasoconstriction could be involved in the generation
of migraine pain via a variety of mechanisms. First,
constriction is associated with the release of messen-
gers and peptides that are believed to be involved in
the generation of migraine. CGRP, widely accepted as
a potentially important mediator of migraine pain, is
released by perivascular neurons in response to vaso-
constriction as part of a reflex arc whose purpose is
to maintain vascular caliber. Nitric oxide may also be
released by vascular cells in response to constriction.
Vasoconstriction associated with CSD and/or astrocyte
calcium waves could also result in an uncoupling of
blood flow and metabolic activity. Reduced parenchy-
mal blood flow in the face of intense neuronal and
glial activity could cause the release of cellular metabo-
lites and lower extracellular pH to trigger a nociceptive
response. For example, transient receptor potential
channels or acid-sensing ion channels on nociceptive
trigeminal neurons could be activated under these con-
ditions (McCleskey and Gold, 1999).
THE BLOOD^BRAIN BARRIER
In addition to changes in vascular caliber, cortical
waves may be associated with changes in the perme-
ability of the vasculature. The blood–brain barrier is a
structural and functional barrier comprised of both
THE NEUROBIOLOGY OF MIGRAINE103
vascular and astrocytic components. Tight junctions
between endothelial cells, and astrocyte endfeet that
are closely apposed to the abluminal surface of the
vessel, create a specialized structure with extremely
limited permeability (Persidsky et al., 2006). This struc-
ture, in combination with a variety of membrane
pumps and transporters, results in a barrier that allows
only highly selective entry of glucose, amino acids, and
other specific molecules from the vasculature into the
brain. It has long been speculated that migraine is asso-
ciated with a breakdown of the blood–brain barrier,
partly because of temporal differences in the efficacy
of abortive medications based on the phase of the
migraine attack in which they are delivered (Harper
et al., 1977). There are a few case reports of cortical
gadolinium enhancement on MRI in migraineurs dur-
ing attacks, indicating increased permeability of the
blood–brain barrier (Smith et al., 2002; Dreier et al.,
2005) (although this is a rare exception to the typically
normal MRI studies of most migraine patients). CSD
in rats has been associated with increased permeability
of the blood–brain barrier as indicated by activa-
tion and upregulation of matrix metalloproteinase-9
(Gursoy-Ozdemir et al., 2004). This study provides evi-
dence that one of the consequences of the propagated
cortical activity in migraine may be an increase in
blood–brain barrier permeability, and raises the possi-
bility that leakage across the blood–brain barrier could
be a mechanism for migraine pain. While inhibitors
of plasma protein extravasation have not been effec-
tive in treating migraine, these agents were developed
with models where the trigger for plasma extravasation
was neurogenic inflammation, rather than spreading
depression. Thus, they do not necessarily exclude a
role for blood–brain barrier breakdown that is due to
GENETIC MODULATION OF MIGRAINE
A wide variety of factors may influence the individual
cellular components of migraine. As discussed above,
genetic variations in the function of neurons, astro-
cytes, and vascular cells have each been implicated in
different forms of migraine. It is important to note
that mutations in genes expressed in multiple cell types
and responsible for a variety of cellular functions all
result in a similar clinical phenotype. This indicates that
there are multiple points of entry to a final common
pathway leading to migraine. For the FHM mutations,
5 s10 s
10 s 30 s
Fig. 7.1. Cortical spreading depression (CSD) and astrocyte calcium waves. (A) Optical imaging of the cortical surface in a
mouse. Far left panel shows image of mouse cortex visualized through the thinned skull of an anesthetized mouse (a record-
ing electrode is seen in the upper right of the field; image scale ¼ 1200 ? 1200 mm). Subsequent images show change in
reflectance of this area of cortex over time associated with a CSD wave. A change in optical signal of the parenchyma
spreads slowly across the cortex. Dilation (darkening) of surface arteries propagates ahead of the CSD wavefront, followed
by constriction of these vessels accompanying the CSD wave. (B) Astrocyte calcium wave. Images show fluorescence of the
calcium indicator fluo-4 with an inverse gray scale (darker gray indicates higher intracellular calcium; image scale ¼ 400 ?
400 mm). Mechanical stimulation of a single cell with a micropipette evokes a wave of increased calcium concentration that
spreads from cell to cell over as many as hundreds of cells. The temporal and spatial characteristics of this wave are remark-
ably similar to those of CSD.
104A. CHARLES AND K.C. BRENNAN
the resulting alterations in function are all likely to lead
to an increase in cortical excitability. A knockin mouse
expressing one identified mutation in the P/Q-type
calcium channel associated with FHM1 provides com-
pelling evidence for increased cortical excitability
related to migraine (van den Maagdenberg et al.,
2004). The threshold for evoking CSD in this mouse
is significantly reduced compared with wild-type con-
trols, and the rate of propagation of CSD is increased.
For the FHM2 mutation in the astrocyte Naþ/Kþ
ATPase, there is as yet no mouse model expressing this
mutation. However, a reduced function of this pump,
as has been shown in vitro in cells expressing the
FHM2 mutations, would be expected to increase corti-
cal excitability. Consistent with this concept, the Naþ/
Kþinhibitor ouabain reliably evokes CSD in rodent
models (Basarsky et al., 1998). The increased excitabil-
ity of neurons associated with the voltage-gated Naþ
channel mutations responsible for FHM3 would also
be expected to increase cortical excitability. Thus, at
least for FHM genes, an increased cortical excitability
potentially predisposing to the occurrence of cortical
waves may represent a common mechanism leading
to the similar clinical phenotype. It will be interesting
to determine if other migraine genes share this
SEX MODULATION OF MIGRAINE
Sex is another important modulator of migraine patho-
physiology. The mechanisms underlying the dramati-
cally greater prevalence of migraine in adult females
are poorly understood. But as with genetic factors,
sex may influence migraine through modulation of
cortical excitability. We have found that the threshold
for induction of CSD in mice is significantly reduced
in female mice as compared with males (Brennan
et al., 2007a). This result was obtained in randomly
sampled mice without monitoring of the estrous cycle,
such that the results cannot be easily explained by a
reduced threshold for CSD occurring only at a single
specific phase of the hormonal cycle. This increased
susceptibility to CSD could involve changes in cortical
excitability that are mediated by sustained exposure
to gonadal hormones over days to months, develop-
mental effects of hormones, chromosomal effects
that are independent of hormones, or any combi-
nation of the above. Another potential mechanism
for modulation of migraine by sex is alteration of
sensitization of trigeminal nociceptive pathways.
Martin et al. (2007) found that there was increased
sensitization of nociceptive neuronal responses in
the trigeminal nucleus caudalis in association with
specific phases of the estrous cycle in rat, suggesting
that alteration in the frequency and pattern of
migraine associated with the menstrual cycle could
involve changes in the sensitization of the trigeminal
The following is a hypothesized model for the
sequence of events leading to migraine. In the cortex,
a variety of different factors (genetic, neurochemical,
ionic, and/or hormonal) lead to a dysregulation of
excitability. This altered excitability triggers propa-
gated waves of neuronal and glial activation, including
astrocyte calcium waves, that are similar to, but not
necessarily identical to, classical CSD. These cortical
waves are associated with propagated changes in vas-
cular caliber, and also breakdown of the blood–brain
barrier. Cortical waves are also associated with release
of a wide variety of neurotransmitters and neuromodu-
lators, as well as changes in the ionic composition of
the extracellular space that can activate nociceptive sig-
naling pathways. These signaling pathways may also be
activated by vascular constriction associated with corti-
cal waves, through the release of messengers such as
CGRP and nitric oxide that also evoke vasodilation,
as well as by changes in the extracellular chemical
and ionic condition resulting from vascular-metabolic
uncoupling. Events occurring primarily in the cortex
are then transmitted via peripheral trigeminal pathways
to the brainstem, where second-order neurons are acti-
vated and eventually become sensitized. Alternatively,
second-order neurons in the brainstem are activated
in parallel to, or even in the absence of, cortical phe-
nomena by changes in cellular excitability that may
be similar to those described in the cortex. Third-order
neurons in the thalamus and cortex are then activated
Specific neuronal, glial, and vascular signaling path-
ways may represent distinct targets for acute and pre-
ventive migraine therapies (Figure 7.2). An increased
understanding of these pathways is now more accessi-
ble with advanced imaging and physiological recording
techniques in combination with novel genetic models
and molecular and pharmacological approaches. It is
likely that there is extensive variation in the specific
pathways that lead to migraine in different individuals.
There may be opportunities to tailor new treatments to
these specific molecular and cellular pathways in order
to maximize efficacy and tolerability of therapy for
this complex neurobiological disorder.
THE NEUROBIOLOGY OF MIGRAINE105
Adams RW, Lambert GA, Lance JW (1989). Stimulation of
brainstem nuclei in the cat: effect on neuronal activity
in the primary visual cortex of relevance to cerebral blood
flow and migraine. Cephalalgia 9: 107–118.
Afridi SK, Matharu MS, Lee L et al. (2005). A PET study
exploring the laterality of brainstem activation in
migraine using glyceryl trinitrate. Brain 128: 932–939.
Ambrosini A, Shoenen J (2006). Electrophysiological res-
ponse patterns of primary sensory cortices in migraine.
J Headache Pain 7: 377–388.
Antal A, Arlt S, Nitsche MA et al. (2006). Higher variability
of phosphene thresholds in migraineurs than in controls: a
consecutive transcranial magnetic stimulation study.
Cephalalgia 26: 865–870.
Aurora SK, Cao Y, Bowyer SM et al. (1999). The occipital
cortex is hyperexcitable in migraine: experimental evi-
dence. Headache 39: 469–476.
Aurora SK, Welch KMA, Al-Sayed F (2003). The threshold
for phosphenes is lower in migraine. Cephalalgia 23:
Aurora SK, Barrodale PM, Tipton RL et al. (2007). Brain-
stem dysfunction in chronic migraine as evidenced by
neurophysiological and positron emission tomography
studies. Headache 47: 996–1003; discussion 1004–1007.
Ayata C, Shin HK, Salomone S et al. (2004). Pronounced
hypoperfusion during spreading depression in mouse
cortex. J Cereb Blood Flow Metab 24: 1172–1182.
Bahra A, Matharu MS, Buchel C et al. (2001). Brainstem
activation specific to migraine headache. Lancet 357:
Basarsky TA, Duffy SN, Andrew RD et al. (1998). Imaging
spreading depression and associated intracellular calcium
waves in brain slices. J Neurosci 18: 7189–7199.
Boers P, Donaldson C, Zagami A et al. (2004). Naratriptan
has a selective inhibitory effect on trigeminovascular
neurones at central 5-HT1A and 5-HT1B/1D receptors
in the cat: implications for migraine therapy. Cephalalgia
Bolay H, Reuter U, Dunn AK et al. (2002). Intrinsic brain
activity triggers trigeminal meningeal afferents in a
migraine model. Nat Med 8: 136–142.
Brennan KC, Romero-Reyes MR, Lopez-Valdes HE et al.
(2007a). Reduced threshold for cortical spreading depres-
sion in female mice. Ann Neurol 61: 603–606.
Brennan KC, Beltran-Parrazal L, Lopez-Valdes HE et al.
(2007b). Distinct vascular conduction with cortical
spreading depression. J Neurophysiol 97: 4143–4151.
Cao Y, Welch KM, Aurora S et al. (1999). Functional MRI-
BOLD of visually triggered headache in patients with
migraine. Arch Neurol 56: 548–554.
Cao Y, Aurora SK, Nagesh V et al. (2002). Functional MRI-
BOLD of brainstem structures during visually triggered
migraine. Neurology 59: 72–78.
Charles A (1998). Intercellular calcium waves in glia. Glia
Charles AC, Merrill JE, Dirksen ER et al. (1991). Intercellu-
lar signaling in glial cells: calcium waves and oscillations
in response to mechanical stimulation and glutamate.
Neuron 6: 983–992.
Chuquet J, Hollender L, Nimchinsky EA (2007). High-
resolution in vivo imaging of the neurovascular unit
during spreading depression. J Neurosci 27: 4036–4044.
Cutrer FM, Black DF (2006). Imaging findings of migraine.
Headache: J Head Face Pain 46: 1095–1107.
D’Hanis W, Linke R, Yilmazer-Hanke DM (2007) Topogra-
phy of thalamic and parabrachial calcitonin gene-related
peptide (CGRP) immunoreactive neurons projecting to
subnuclei of the amygdala and extended amygdala.
J Comp Neurol 505: 268–291.
De Fusco M, Marconi R, Silvestri L et al. (2003). Haploinsuf-
ficiency of ATP1A2 encoding the Naþ/Kþpump alpha2
subunit associated with familial hemiplegic migraine type
2. Nat Genet 33: 192–196.
Fig. 7.2. Schematic of hypothesized sequence of events in migraine. Each step in the sequence may have discrete thresholds,
and involve distinct molecular, cellular, and neurochemical pathways. This complexity may underlie the heterogeneous and
variable clinical presentation of migraine. BBB, blood–brain barrier.
106 A. CHARLES AND K.C. BRENNAN
Denuelle M, Fabre N, Payoux P et al. (2007). Hypothalamic
activation in spontaneous migraine attacks. Headache 47:
Dichgans M, Mayer M, Uttner I et al. (1998). The phenoty-
pic spectrum of CADASIL: Clinical findings in 102
cases. Ann Neurol 44: 731–739.
Dichgans M, Freilinger T, Eckstein G et al. (2005). Mutation
in the neuronal voltage-gated sodium channel SCN1A in
familial hemiplegic migraine. Lancet 366: 371–377.
Dreier JP, Jurkat-Rott K, Petzold GC et al. (2005). Opening
of the blood–brain barrier preceding cortical edema in a
severe attack of FHM type II. Neurology 64: 2145–2147.
Ellrich J, Messlinger K, Chiang CY et al. (2001). Modulation
of neuronal activity in the nucleus raphe magnus by the 5-
HT(1)-receptor agonist naratriptan in rat. Pain 90: 227–231.
Fellin T, Sul JY, D’Ascenzo M et al. (2006). Bidirectional
astrocyte–neuron communication: the many roles of
glutamate and ATP. Novartis Found Symp 276: 208–217;
discussion 217–221, 233–207, 275–281.
Filosa JA, Bonev AD, Straub SV et al. (2006). Local potas-
sium signaling couples neuronal activity to vasodilation
in the brain. Nat Neurosci 9: 1397–1403.
Fragoso YD, Brooks JB (2007). Two cases of lesions in
brainstem in multiple sclerosis and refractory migraine.
Headache 47: 852–854.
Geraud G, Denuelle M, Fabre N et al. (2005). [Positron
emission tomographic studies of migraine.] Rev Neurol
(Paris) 161: 666–670.
Gerwig M, Niehaus L, Kastrup O et al. (2005). Visual cortex
excitability in migraine evaluated by single and paired
magnetic stimuli. Headache: J Head Face Pain 45:
Goadsby PJ (2002). Neurovascular headache and a midbrain
vascular malformation: evidence for a role of the brain-
stem in chronic migraine. Cephalalgia 22: 107–111.
Goadsby PJ, Duckworth JW (1989). Low frequency stimula-
tion of the locus coeruleus reduces regional cerebral
blood flow in the spinalized cat. Brain Res 476: 71–77.
Goadsby PJ, Edvinsson L (1994). Human in vivo evidence
for trigeminovascular activation in cluster headache. Neu-
ropeptide changes and effects of acute attacks therapies.
Brain 117: 427–434.
Goadsby PJ, Hoskin KL (1996). Inhibition of trigeminal neu-
rons by intravenous administration of the serotonin (5HT)
1B/D receptor agonist zolmitriptan (311C90): are brain
stem sites therapeutic target in migraine? Pain 67: 355–359.
Gursoy-Ozdemir Y, Qiu J, Matsuoka N et al. (2004). Cortical
spreading depression activates and upregulates MMP-9. J
Clin Invest 113: 1447–1455.
Guthrie PB, Knappenberger J, Segal M et al. (1999). ATP
released from astrocytes mediates glial calcium waves.
J Neurosci 19: 520–528.
Haas DC, Kent PF, Friedman DI (1993). Headache caused
by a single lesion of multiple sclerosis in the periaque-
ductal gray area. Headache 33: 452–455.
isms ofmigraine aura revealed by functional MRI inhuman
visual cortex. Proc Natl Acad Sci 98: 4687–4692.
Harper AM, MacKenzie ET, McCulloch J et al. (1977).
Migraine and the blood–brain barrier. Lancet 1: 1034–1036.
Haydon PG, Carmignoto G (2006). Astrocyte control of
synaptic transmission and neurovascular coupling. Phy-
siol Rev 86: 1009–1031.
Hernandez-Caceres J, Macias-Gonzalez R, Brozek G et al.
(1987). Systemic ketamine blocks cortical spreading
depression but does not delay the onset of terminal anoxic
depolarization in rats. Brain Res 437: 360–364.
Hoskin KL, Kaube H, Goadsby PJ (1996). Central activation
of the trigeminovascular pathway in the cat is inhibited
by dihydroergotamine. A c-Fos and electrophysiological
study. Brain 119: 249–256.
Jenkins DW, Langmead CJ, Parsons AA et al. (2004). Regu-
lation of calcitonin gene-related peptide release from rat
trigeminal nucleus caudalis slices in vitro. Neurosci Lett
Kruuse C, Thomsen LL, Birk S et al. (2003). Migraine can
be induced by sildenafil without changes in middle cere-
bral artery diameter. Brain 126: 241–247.
Kunkler PE, Kraig RP (2004). P/Q Ca2þchannel blockade stops
spreading depression and related pyramidal neuronal Ca2þ
Lea ˜o AAP (1944). Spreading depression of activity in the
cerebral cortex. J Neurophysiol 7: 359–390.
Lea ˜o AAP (1947). Further observations on the spreading
depression of activity in the cerebral cortex. J Neurophy-
siol 10: 409–414.
Levy D, Strassman AM (2002). Mechanical response proper-
ties of A and C primary afferent neurons innervating the
rat intracranial dura. J Neurophysiol 88: 3021–3031.
Martin VT, Lee J, Behbehani MM (2007). Sensitization of
the trigeminal sensory system during different stages of
the rat estrous cycle: implications for menstrual migraine.
Headache: J Head Face Pain 47: 552–563.
Matharu MS, Goadsby PJ (2002). Persistence of attacks of
cluster headache after trigeminal nerve root section. Brain
McCleskey EW, Gold MS (1999). Ion channels of nocicep-
tion. Annu Rev Physiol 61: 835–856.
Mulligan SJ, MacVicar BA (2004). Calcium transients in
astrocyte endfeet cause cerebrovascular constrictions.
Nature 431: 195–199.
Offenhauser N, Zinck T, Hoffmann J et al. (2005). CGRP
release and c-fos expression within trigeminal nucleus
caudalis of the rat following glyceryltrinitrate infusion.
Cephalalgia 25: 225–236.
Olesen J, Friberg L, Olsen TS et al. (1990). Timing and topo-
graphy of cerebral blood flow, aura, and headache during
migraine attacks. Ann Neurol 28: 791–798.
Ophoff RA, Terwindt GM, Vergouwe MN et al. (1996).
Familial hemiplegic migraine and episodic ataxia type-2
are caused by mutations in the Ca2þchannel gene
CACNL1A4. Cell 87: 543–552.
Peeters M, Gunthorpe MJ, Strijbos PJ et al. (2007). Effects of
pan- and subtype-selective NMDA receptor antagonists on
cortical spreading depression in the rat: therapeutic poten-
tial for migraine. J Pharmacol Exp Ther 321: 564–572.
THE NEUROBIOLOGY OF MIGRAINE107
Penfield W (1935). A contribution to the mechanism of Download full-text
intracranial pain. Assoc Res Nerv Ment Dis 15: 399–416.
Persidsky Y, Ramirez S, Haorah J et al. (2006). Blood–brain
Barrier: structural components and function under physio-
logic and pathologic conditions. J Neuroimmune Pharma-
col 1: 223–236.
Peters O, Schipke CG, Hashimoto Y et al. (2003). Different
mechanisms promote astrocyte Ca2þwaves and spread-
ing depression in the mouse neocortex. J Neurosci 23:
Raskin NH, Hosobuchi Y, Lamb S (1987). Headache
may arise from perturbation of brain. Headache 27:
Ray BS, Wolff HG (1940). Experimental studies in head-
ache: Pain-sensitive structures of the head and their struc-
tural significance. Arch Surg 41: 813–856.
Richards A, van den Maagdenberg AMJM, Jen JC et al.
(2007). C-terminal truncations in human 3[prime]-5
[prime] DNA exonuclease TREX1 cause autosomal
dominant retinal vasculopathy with cerebral leukodystro-
phy. Nat Genet 39: 1068–1070.
Schoenen J (2006). Neurophysiological features of the
migrainous brain. Neurol Sci 27(Suppl. 2): S77–S81.
Schoonman GG, van der Grond J, Kortman C et al.
(2008). Migraine headache is not associated with cerebral
or meningeal vasodilation – a 3T magnetic resonance
angiography study. Brain 131: 2192–2200.
Segall L, Mezzetti A, Scanzano R et al. (2005). Alterations
in the alpha2 isoform of Na,K-ATPase associated with
familial hemiplegic migraine type 2. Proc Natl Acad Sci
U S A 102: 11106–11111.
Shields KG, Goadsby PJ (2006). Serotonin receptors modu-
late trigeminovascular responses in ventroposteromedial
nucleus of thalamus: A migraine target? Neurobiol Dis
Smith M, Cros D, Sheen V (2002). Hyperperfusion with
vasogenic leakage by fMRI in migraine with prolonged
aura. Neurology 58: 1308–1310.
Stankewitz A, May A (2007). Cortical excitability and
migraine. Cephalalgia 27: 1454–1456.
ceptors in relation to headache.J Neurophysiol 95: 1298–1306.
Strassman AM, Raymond SA, Burstein R (1996). Sensitiza-
tion of meningeal sensory neurons and the origin of head-
aches. Nature 384: 560–564.
Tajti J, Uddman R, Edvinsson L (2001). Neuropeptide loca-
lization in the ‘migraine generator’ region of the human
brainstem. Cephalalgia 21: 96–101.
Takano T, Tian GF, Peng W et al. (2006). Astrocyte-mediated
control of cerebral blood flow. Nat Neurosci 9: 260–267.
van den Maagdenberg AMJM, Pietrobon D, Pizzorusso
T et al. (2004). A Cacna1a knockin migraine mouse
model with increased susceptibility to cortical spreading
depression. Neuron 41: 701–710.
Vanmolkot KRJ, Kors EE, Turk U et al. (2006). Two de
novo mutations in the Na,K-ATPase gene ATP1A2 asso-
ciated with pure familial hemiplegic migraine. Eur J.
Human Genetics 14: 555–560.
Weiller C, May A, Limmroth V et al. (1995). Brain stem
activation in spontaneous human migraine attacks. Nat
Med 1: 658–660.
Welch KM, Nagesh V, Aurora SK et al. (2001). Periaqueduc-
tal gray matter dysfunction in migraine: cause or the bur-
den of illness? Headache 41: 629–637.
Woods RP, Iacoboni M, Mazziotta JC (1994). Bilateral
spreading cerebral hypoperfusion during spontaneous
migraine headache. N Engl J Med 331: 1689–1692.
Zonta M, Angulo MC, Gobbo S et al. (2003). Neuron-to-
astrocyte signaling is central to the dynamic control of
brain microcirculation. Nat Neurosci 6: 43–50.
108 A. CHARLES AND K.C. BRENNAN