Cortical Spreading Depression and Migraine: Critical Review of a Neurophysiological Principle in the Third Millennium

Known since antiquity, migraine is a protean, common, primary, neurologically “self-cycling” (spontaneous onset and offset), lateralizing (unilateral or bilateral, sideshifting or side-fixed) headache with the term ‘megrim/hemicrania’ (Greek, halfskull) clearly indicating preponderance of half-sided / unilateral or bilateral discrete/episodic headache attacks involving self-limited fronto-temporal and / or nuchal pain lasting variably for 4-72 hours (Headache Classification Committee of the International Headache Society-International Classification of Headache Disorders-III (Pearce, 1986; ICHD-III, 2018). In 1597, the Shorter Oxford English Dictionary, even before The Plague and the Great Fire, described migraine as a far wider constellation of symptoms besides headache alone, such as the association with imbalance and motion sickness, including sea-sickness, with ‘emotional changes’ and ‘depression’. (Pearce, 1986) Migraine is one of the most prevalent disorders throughout the world, affecting approximately 1/5th of the adult population, but with widely varying intensity, frequency, social stigma, absenteeism and presenteeism, and largely incomprehensible genetic transmission, as well as varied prevalence in different age groups, including childhood or teenage, adults, trimesters of pregnancy, or advancing age/menopause (ICHD-III, 2018; Bigal and Lipton, 2009). Migraine headache is ranked among the top 10 debilitating diseases worldwide (Leonardi and Raggi, 2013). Prevalence of migraine in the population has remained relatively consistent for the past 30 years, ranging from 11.7% to 14.7% overall, 17.1% to 19.2% in women, and 5.6% to 7.2% in men in the studies reviewed.

(Cohen et al., 2024). Migraine predominantly affects females, and the F:M ratio is 3:1 or 4:1 (Buse et al., 2010). Migraine is far more prevalent among neurologists than in the general population. The 1-year and lifetime prevalences of migraine in a sizable cohort of neurologists is as follows: male neurologists, 34.7%, 46.6%; male headache specialists, 59.3%, 71.9%; female neurologists, 58.1%, 62.8%; and female headache specialists, 74.1%, 81.5% (Evans et al., 2003).

Besides neurological aura and headache, photophobia is another typical feature of migraine headache (Nir et al., 2018) and, along with phonophobia, is included as a diagnostic criterion (ICHD-III, 2018). Nausea and/or vomiting in migraine is a typical, diagnostic, and usually early feature, and can remarkably ameliorate or abort the headache – noted as early as by Hippocrates (c.460 B.C.-c.370 B.C.) (Pearce, 1986). The premonitory or prodromal phase of migraine involves tiredness, irritability, reduced concentration, and yawning, and can precede the aura or headache phase by up to 48 hours (Giffin et al., 2003; Kelman, 2007). Most migraine headache attacks are followed by hours or a day of feeling unwell, usually with tiredness, called the postdrome phase (Kelman 2006; Giffin et al., 2016). In one-third of the migraine cohort, headache attacks are associated with neurological phenomena, positive or negative, collectively termed migraine aura (MwA) (Rasmussen and Olesen, 1992). The scintillating scotomata (SS) are regarded as the only truly pathognomonic feature of the disorder (Hupp et al., 1989; Rose, 1995). Comprehension of migraine mechanisms is limited and controversial, despite an elaborate phenotype and sophisticated updated symptomatic classification (Pietrobon and Moskowitz, 2014; Goadsby et al., 2017; Lindquist and Shuttleworth, 2017; Schulte et al., 2017; Youssef et al., 2017; Cozzolino et al., 2018; ICHD-III, 2018). Migraine still does not have a specific diagnostic laboratory or radiologic test (Nalbant et al., 2017). Recurrent migraine headache is usually lateralized to the same side (right or left fronto-temporal region or both or nuchal) in subsequent attacks, or side-locked or side-shifting. Migraine headache is commonly retro-orbital and/or nuchal or, uncommonly, lower-half headache (facial pain) in distribution or location (Campbell 1988, 1990; Blau 1992a; Sjaastad et al., 1992; Constantinides et al., 2015). The protean nature of migraine results in a characteristically varied frequency, differential severity, duration and morbidity, and inconsistent autonomic features in successive headache attacks in the same patient as well as between different patients.             

Currently, the migraine attack is regarded as a paroxysmal “self-cycling” alteration in gain, or input-output modulation (Haider and McCormick, 2009) of multiple sensory systems involving multiple cortical, subcortical and brainstem regions with a strong genetic basis to account for the pain and the wide constellation of symptoms characterizing the attack (Ferrari et al., 2015; Goadsby et al., 2017; Brennan and Pietrobon, 2018). Understanding of migraine as disruption in sensory gain and plasticity is an oversimplified completely untested hypothesis (Brennan and Pietrobon, 2018). How a migraine attack starts or stops remains a complete mystery, even in familial hemiplegic migraine (FHM) (Hansen et al. 2011; Burstein et al. 2015). There is almost a complete lack of synaptic, let alone cellular/molecular understanding of “established” phenomena, such as CSD and trigeminovascular system (TVS) sensitization (Brennan and Pietrobon, 2018). Dural TVS nociception is a favoured hypothesis for the genesis of migraine headache involving the largely unmyelinated sensory nerve fibre plexus of the pan-trigeminal nerve (Moskowitz et al., 1979; Moskowitz,1990, 1991, 2007a, 2007b, 2008), even though the most typical features of migraine, i.e., lateralization of headache and V1 predominance of distribution of headache cannot be explained by diffuse dural activation (Gupta, 2009a, 2019, 2024) (discussed further). Does the dura “self-cycle” in migraine; if so, how? And how does diffuse dural “self-activation” produce lateralizing headache localized to V1?

Cortical spreading depression (CSD) is an electrophysiologic pathological state in which a wave of depolarization in the cerebral cortex is followed by the suppression of spontaneous neuronal activity (Hill et al., 2024). This transient spread of neuronal depolarization on the cortical surface is the hallmark of CSD. Numerous investigations have demonstrated that transmembrane ion transport, astrocytic ion clearing and fatigue, glucose metabolism, certain genetic markers, point mutations, and expression of the enzyme responsible for the production of various arachidonic acid derivatives that participate in the inflammatory response, namely, cyclooxygenase (COX), all influence CSD (Hill et al., 2024). The inflammatory response cannot explain key clinical features of migraine as well as of CSD (Gupta, 2009a, 2019). Since the seminal acknowledgment by Graham (1953) almost eight decades ago that the fundamental reason (“why” and “how”) of observed “immediate misbehavior” or “aberration” of cranial vessels in migraine remains a “mystery”, nothing substantive has changed. Graham (1953) also elaborated on a peculiar “affinity to stress” and “deficiency in the physiological adaptation to stress” of migraineurs. Graham (1953) additionally conceded that the nature of migraine was elusive, an analysis that seems to have held up to the early 21st century. The neuronal or vascular mechanisms of migraine have been debated vigorously over the last 150 years. In the last decade of the 20th century, a new beginning appeared essential (Edmeads 1991; Blau 1992a).

Migraine pathophysiology has remained impervious to technological advances, nosological “splitting”, statistical methods, animal experiments, randomized clinical trials (RCTs) and meta-analyses (Gupta, 2009a, 2010a; Borgdorff, 2018). End-points in migraine research are relatively soft and variable, even in the same patient on different occasions and are difficult to replicate; all therapies, including those linked to RCTs, remain empirical or experiential. Migraine – as well as its relevance to CSD -- stands today where ischemic heart disease stood before the discovery of the atheromatous plaque or diabetes mellitus stood before the discovery of insulin.

Limitations of current understanding of migraine pathogenesis

While the precise aetiology of migraine remains unknown and its genesis is the topic of much controversy and speculation, some researchers underscore seemingly significant progress along the vascular, neurogenic, neurotransmitter, neuroimaging and genetic / molecular biological paradigms (Moskowitz, 2008; Tfelt-Hansen and Koehler, 2011; May, 2013).

Dural TVS nociception – particularly involving the middle meningeal artery (MMA) – has been a compelling influence (Olesen et al., 2009; Akerman et al., 2013; Levy et al., 2019). In humans with inelastic, unyielding dura mater and the bony channels in the inner table of the calvarium, significant and pathogenetic dilatation of such vessels is quite unlikely (Shevel, 2009). In a cohort of 18 healthy volunteers, after correction for placebo-induced dilatation in the MMA, calcitonin-gene-relatedpeptide (CGRP) infusion (1.5 μg/min h-αCGRP over 20 minutes) induced only mild clinically insignificant dilatation in the MMA while there was no change in the middle cerebral artery (MCA) (Asghar et al., 2010). Similar changes were noted after glyceryltrinitrate (GTN)-induced migraine headache (Schoonman et al., 2008). With further studies, a shift from Wolff’s concept entrenched for over half a century was recommended (Amin et al., 2017). Two facts, however, ensure that the vascular component of migraine pathophysiology will forever remain relevant: i) GTN consistently causes an early atypical headache and a more delayed more typical headache in migraineurs but without SS or nausea/vomiting (Ashina et al., 2013; Maniyar et al., 2014; Nozaki et al., 2015) and ii) migraine attacks are remitted by triptans (Cameron et al., 2015) and ergotamine (Campbell, 1990). Another cranial vascular system appears to be involved besides the cerebrovascular system, including the MMA and the MCA (see below).

A central neuronal dysfunction of the central nervous system (CNS) is generally regarded as the underlying cause of migraine (Burstein et al., 2015; Ferrari et al., 2015; Goadsby et al., 2017; Hougaard et al., 2017; Schulte et al., 2017). Besides CSD (Ayata and Lauritzen, 2015), a “self-cycling” seemingly primary brain stem activation concept (Goadsby and Edvinsson, 1993; Akerman et al., 2011), and the theory of a primary “self-cycling” disorder of sensory processing (Goadsby et al. 2017), with a rhythmic cortical feedback to the thalamus has also been suggested to underlie amplification of thalamocortical oscillations and to influence neuronal excitability, early sensory processing, and multisensory integration (Hodkinson et al. 2016). All CNS theories of migraine, including CSD, suffer from the twin major disadvantages of a plethora of assumptions or arbitrary elements as well as from being unable to explain key phenotypic features of the disorder (Gupta, 2009a, 2019, 2024, 2025; Borgdorff, 2018).

Writing on Airy, Levene (1975) noted the vastness and unmanageability of the migraine literature. Almost fifty years later, the difficulty in assimilating the explosive accretion of data in the absence of an overarching theory is far more profound, a paradox that defies common sense and is very likely to worsen.

Cortical spreading depression (CSD)

In 1944, Leão serendipitously reported, while studying “experimental epilepsy”, the ability to produce temporary suppression of electrical activity in the brain of a rabbit under narcosis over areas such as the cerebral cortex, cerebellum, or hippocampus; such stimuli included weak faradic electrical stimulation of the brain surface, application of potassium chloride solution, and slight mechanical deformation (Engelhardt and Gomes, 2015). From these animal experiments, it was inferred that once induced, depression of spontaneous electroencephalographic (EEG) activity travelled slowly and contiguously in all directions into neighbouring regions of cortex – like ripples-in-a-pond. The frontal regions were stimulated more readily than regions near the occipital pole (Leão, 1944a). Leão also reported that CSD was associated with marked dilatation of the pial arteries and increased blood flow; a smaller vasoconstriction sometimes followed (Leão, 1944b). Later, Leão and Morison described a slow, negative-positive potential that accompanied the onset of EEG suppression (Leão and Morison, 1945). CSD, thus, is an electrophysiologic pathological state in which a wave of depolarization in the cerebral cortex is followed by the suppression of spontaneous neuronal activity. This transient spread of neuronal depolarization on the surface of the cortex is the hallmark of CSD. Numerous influences on the CSD have been detailed (Hill, 2024); the list is not exhaustive. (see above).

Leão’s observations have been repeatedly confirmed in a wide variety of animal species, both vertebrates and invertebrates, including mammals, birds, reptiles, locusts, cephalopods, and Drosophilia (Ayata and Lauritzen, 2015; Spong et al., 2016a, 2016b, 2016c). CSD is, thus, a transient wave of intense near-complete neuronal and glial depolarization observed in experimental animals and traumatic brain-injured (TBI) humans as well as in other acute cortical lesions such as malignant stroke, subarachnoid and intracranial hemorrhage (Lauritzen et al., 2011; Lauritzen and Strong, 2017; Hartings et al., 2017). A working hypothesis for translation to human disease, where complex patterns of depolarizations are observed in acute brain injury and appear to mediate and signal ongoing secondary damage, places CSD in a primary role in acute brain injury, a viewpoint inconsistent with animal experiments (Gupta, 2009a; Borgdorff, 2018). CSD slowly propagates (≈3-5 mm/min) in grey matter in a characteristic ‘creeping’ fashion by way of contiguity regardless of functional divisions or arterial territories, followed by hyperpolarization or long-lasting suppression or silence of neural bioelectrical activity, that, in turn, is associated with a dramatic failure of brain ion homeostasis as well as efflux of excitatory amino acids from nerve cells and increased energy metabolism (Pietrobon and Moskowitz, 2014; Somjen, 2001; Lauritzen, 1994, 2001; Lauritzen et al., 2011, see below). Depolarization of both astrocytes and neurons is associated with onset of CSD (Ghaemi et al., 2018). Following cardiac arrest cerebral ischemia, a protective influence of CSD has been demonstrated, a form of neuronal resistance to ischaemia by preceding spreading depression in Sprague-Dawley rats (Kawahara et al., 1995) (see below). Other features of a neuronal adaptive / protective role for CSD / SD have been discussed previously (Gupta, 2009a).

CSD in experimental animals is at least a hundred times slower than any other known brain neurological process in humans, vascular or neuronal (Blau 1992a; Kaufman et al., 2017). Since Leão’s discovery, CSD has been implicated in migraine due to the presumed similarity in electrophysiological characteristics, in particular the rate of propagation and SS (Milner, 1958; Lauritzen et al., 1985; Bolay et al., 2002). Nevertheless, lateralizing pain of migraine headache is the primary clinical event, although the mechanistic considerations of the SS (Pietrobon and Moskowitz, 2014) appear to have displaced it.                                                           

CSD has been maintained in migraine pathophysiology enthusiastically for several decades by a long series of critical omissions and assumptions (see Box 1, Gupta, 2009a). Several major issues, nevertheless, remain unresolved: (i) What triggers CSD itself, in the first instance, in the otherwise normal migraine patient is completely unknown, whereas migraine is triggered by a wide variety of well-known stimuli (Kelman, 2004a, 2004b, 2005, 2006, 2007; Sauro and Becker, 2009; Peroutka, 2014). (ii) CSD appears to arise only in damaged tissues but propagates through uninjured normal tissue (Leão, 1944a). Migraine is not associated with injured brain tissue (Gupta, 2009a). (iii) In contrast to the “wave-in-a-pond” propagation of CSD in rodents, in primates CSD may remain confined to a small local area contiguous with the stimulus, e.g., within one sulcus (Pearce, 1985). (iv) Without violating the principle of contiguity of spread of CSD, the full spectrum of frequently “skipping” migraine aura phenotype cannot be explained (Gupta, 2009a) (see below). (v) Psychogenic stress constitutes the commonest precipitant of migraine (Goadby, 2014; Borkum, 2016; Solotareff et al., 2017; Pellegrino et al., 2018). However, there is no known pathophysiologic link between psychogenic stress and CSD. (Gupta, 2024) (vi) Migraine headache is worsened by increase in heart rate or blood pressure and activities involving the Valsalva maneuver (VM) (Campbell, 1990; Gupta, 2010b), which have no link to CSD (Gupta, 2009a). (vii) In the majority of migraineurs i.e., migraine without aura (MwoA) cohort, the headache is not preceded by an aura – compelling the assumption of a “silent” aura (Moskowitz, 2008; Parisi, 2009; Lambert et al. 2011). (viii) There is a vast species and factual event gap between CSD-induced changes of behavior in experimental animals (monkeys or rats/mice) and migraine aura/headache in humans, as reviewed previously (Iadecola 2002; Gupta, 2009a). (ix) Pain of headache is believed, in the absence of definitive evidence, to be linked to CSD through behavioral changes in rats (Pietrobon and Moskowitz, 2014). Single unilateral waves of CSD in rats induce contralateral sensory neglect and motor impairment of the forepaw lasting for 15-30 minutes (Bureš et al., 1974, 1984), both features not unequivocally indicating pain. As reviewed, rats do not experience cortical or hippocampal CSD as aversive (Koroleva and Bures, 1993) (x) CSD is an all-or-none physiological phenomenon (Pearce, 1985). In striking contrast, the migraine phenotype is protean with markedly varied frequency, severity, and duration of headaches and strikingly varied accompanying clinical features, including ANS manifestations, disability and morbidity (Gupta 2009a; Borgdorff, 2018). Heterogeneity of CSD in the laboratory has been presumedly linked to clinical variations in the migraine phenotype, in particular the aura (Kaufman et al., 2017).

(see below). (xi) Episodic migraine attacks stop by themselves or remain self-limited (4-72 h) but remain unlinked to CSD, the neurological basis of which cessation is unknown (Houle et al., 2012; Villanueva, 2012) (xii) There is no residual functional neurological impairment after a wave of CSD in experimental animals, maximally observed at one hour (Leão, 1944a). Conversely, persistent neurological deficits and/or prolonged headache over several hours or days, however, are reported in migraine (Hupp et al., 1989; Campbell, 1990; Maxner and Moeller, 2005; ICHD, 2018) (xiii) Migraine-related critical cerebral ischemia, systemic vasculopathy, endothelial dysfunction and a hypercoagulable state are not found with any frequency or consistency in the uncomplicated adult migraine patient (Waters et al., 1983; Kurth et al., 2007; Rambarat et al., 2017). Critical (>75%) ischemia and cerebral infarct have also has not been documented in experimental animals after induction of CSD, as reviewed (Gupta 2009a). Stroke in migraineurs, nevertheless, is believed to be consequent to CSD (Eikermann-Haerter, 2014; Guidetti et al., 2014; Malik et al., 2015). Although migraine attacks can epidemiologically be considered as a risk factor for ischemic or hemorrhagic stroke, the overall clinical impact is limited (Gupta, 2010a, 2017d), and the functional outcome after ischemic strokes is good (Sacco and Kurth, 2014). The key limitation is the plethora of other known vascular risk factors (Eikermann-Haerter, 2014). (xiv) The pathophysiological basis of brain white matter lesions (WMLs) in migraine patients or their link to CSD has not yet been established. (Hamedani et al., 2013; Tana et al., 2013; Montieth et al., 2014; Erdélyi-Bótor et al., 2015). WMLs are always distributed bilaterally; lateralizing WMLs have never been reported in migraine (Gupta, 2009b). (xv) Goadsby et al. (1990) reported non-lateralizing release of calcitonin gene-related peptide (CGRP) in jugular venous blood during migraine attacks. It has been generally accepted that CGRP, expressed throughout the CNS, plays an important role in migraine pathophysiology (Edvinsson, 2008). The source of episodic release of CGRP in humans in synchronization with precipitation of migraine headaches and its link to CSD in experimental animals remains speculative (Gupta, 2017c; Gupta, 2019). No systemic influence (via diffuse release of neuropeptides, including 5-HT or noradrenaline or CGRP or adenosine or oxidants / reactive oxygen species (ROS) or nitric oxide whether released intrinsically in the brain or at the level of the heart – with or without patent foramen ovale (PFO) – or in the circulation can rationally explain the typically recurrent lateralizing headache of migraine (Gupta, 1996a). Neuropeptides do not work in splendid isolation in the human body; immediately after secretion or administration of any agent, drug, pharmaceutical, or biological substance into the body or into the brain, another set of counter-regulatory adaptive chemicals are released to maintain homeostasis across the blood-brain barrier (BBB) (Gupta, 2017e, 2019, 2024). (xvi) BBB issues cloud CSD and pharmacologic absolutes (Borgdorff, 2018; Gupta, 2009a, 2019). One of the key assumptions about BBB disruption during MwA attacks and / or MwoA attacks has been disproved (DosSantos et al., 2014; Amin et al., 2017; Hougaard et al., 2017) (discussed further). (xvii) Tonabersat, an experimental novel benzopyran derivative that serves as a gap-junction modulator and blocks CSD, has not emerged as a useful migraine preventive agent (Hauge et al., 2009; Dodick, 2009; Silberstein et al., 2009; Cao and Zheng, 2014). (xviii) During CSD, influx of Na+, Ca++, and Cl- ions into neuronal cells along with water, causes swelling, which consequently reduces the volume of the extracellular compartment to ~50% of the control value (Hansen and Olsen, 1980; Gorji, 2001). Conversely, significant change in water diffusion during or after spontaneous migraine aura was not demonstrated in a human fMRI study. (Cutrer et al., 1998). (xix) In CSD experiments in rats, cerebral metabolism is distinctly enhanced (Piilgaard and Lauritzen, 2009; Mayevsky and Weiss, 1991) but remained unchanged in MwA patients (Friberg et al., 1994). (xx) EEG activity is depressed to flattened in CSD (Leão, 1944a; Duckrow, 1991, 1993), but EEG depression is not found in migraine aura (Borgdorff, 2018). (xxi) Regular physical activity might reduce migraine attack frequency and pain intensity (Darling, 1991; Linde, 2006; Busch and Gaul, 2008; Irby et al., 2016) but regular physical activity has no known link to CSD. Conversely, intentional avoidance of moderate- and vigorous-intensity physical activity in women is a common migraine management strategy (Farris et al., 2018). (xxii) Perhaps the most significant component of CSD that is also least well understood is the fact that CSD is not an ordinary physiologic action potential but a severe, profound, hyperpolarization following long-lasting suppression or silencing of neuronal bioelectrical activity associated with dramatic failure of brain ion homeostasis, with total neuronal inexcitability resulting from a massive efflux of K+ with extracellular potential shift of several millivolts up to 20 mV (Leão 1944a; Pearce, 1985; Borgdorff, 2018). In the face of such profound neuronal silencing, generation of the well-patterned positive SS or of lateralizing headache of migraine is highly unlikely.

CSD-Magnesium-Migraine nexus

Magnesium (Mg²⁺) lies central to the proposed schema for the genesis of CSD and/or vasoconstriction to effect migraine (Welch et al., 1990, see Figure 4). Spreading depression (SD) / CSD can be more readily evoked in Mg²⁺ deficient brain cortex in rat (Mody et al., 1987) and in human (Avoli et al., 1991) cortical slices. The range of pathologies associated with Mg²⁺ deficiency/depletion, acute or chronic, is very wide (Johnson, 2001), but is unlinked to CSD / SD (Gupta, 2024).    

Mg²⁺ has differential modulatory effects on vascular tone and reactivity. Lowering Mg²⁺ concentration enhances in vitro vascular tone and contractile responses to various agonists (Altura and Altura, 1978; Altura et al., 1987; Laurant et al., 1997). Exogenously administered Mg²⁺ is a systemic vasodilator and a competitive calcium antagonist (Pfefferanferth et al., 1996; Wang et al., 2003). Vasodilators, generally, precipitate migraine attacks. Additionally, why migraine patients should primarily or spontaneously become brain Mg²⁺-depleted or repleted is a far more important issue (Gupta, 2009a, 2017b, 2024).

Mg²⁺ concentration in CSF is significantly greater than in plasma and tends to remain stable or rise mildly and slowly in the face of changing plasma concentration (Watchorn and McCance, 1932; Oppelt et al., 1963). Transport of Mg²⁺ from blood to CSF across the BBB is limited in normal humans and has little effect on post-operative pain scores (Ko et al., 2001; Harris, 2013; Albrecht et al., 2013).

Additionally, in principle (Gupta, 2005c, 2009a) as well as in practice (Mees et al. 2012; Golan et al., 2013; Leijenaar et al., 2015; Dorhout Mees et al., 2015), supplementation or prophylactic use of magnesium sulphate (MGSO4) in subarachnoid hemorrhage to counter delayed cerebral ischemia does not appear useful. In stroke management also, i.v. MGSO4 did not offer significant neuroprotection when administered within two hours of symptom onset (Saver et al., 2015). Intriguingly, exogenous MGSO4 administration itself attenuates enhanced BBB permeability in different experimental models (Esen et al., 2003; Kaya et al., 2004; Esen et al., 2005). The belief that exogenous Mg²⁺ offers neuroprotection by suppressing development of CSD (Welch et al., 1990; Van Der Hel et al., 1998) does not appear to be tenable any longer in human patients, including in traumatic brain injury (TBI).

A low serum Mg²⁺, as is found in migraine patients, during and between headache attacks, tends to limit or prevent vasodilatation, systemic as well as cranial, thereby restricting the cephalalgogenic effect of various headache triggers. Hypomagnesemia, thus, appears to be a generalizable, logically-defensible, physiological adaptive evolutionary mechanism across the interdisciplinary neurosciences (Gupta, 2004d, 2009a, 2009b) and the cardiovascular sciences (Gupta, 1996b) that enables optimal operation of a host of life-preserving calcium-related cellular-level functions during acute stress of life threatening conditions (such as acute myocardial infarction) or chronic stress of migraine-like conditions. Like all other adaptive functions, e.g., fluid retention, hypomagnesemia is simplistic in nature. Severe hypomagnesemia is linked with hypocalcaemia, refractory arrhythmia, and a poor prognosis. A therapeutic role for i.v. MGSO4 supplementation in refractory atrial fibrillation in combination with other atrio-ventricular nodal blockers is proposed (Bouida et al., 2019).

For almost 100 years, widespread use of MGSO4 has been in practice for pre-eclampsia and eclampsia, based on the premise that it has a powerful central anti-epileptic action (Alton and Lincoln, 1925; Lazard, 1925; Sadeh, 1989). No study of toxemia of pregnancy has found that Mg²⁺ freely, rapidly, or significantly crosses the BBB to prevent convulsions. Conversely, occurrence of neuromuscular blockade and the resultant potentiation of muscle relaxants during MgSO4 administration has been well-documented (Giesecke et al., 1968; Skaredoff et al., 1982; Sinatra et al., 1985; Ramanathan et al., 1988). The obstetric anesthesiologist is most concerned about the action of Mg²⁺ at the myoneural junction rather than at the CNS. In addition, during i.v. MGSO4 administration, there is a marked decrease in the serum calcium levels in patients with preeclampsia (Cruikshank et al., 1979). High serum levels of MGSO4 in the range of 12 to 15 mEq/L can cause paralysis and respiratory depression (Sibai, 2004), giving further credence to neuromuscular junction blockade as an alternative or apparent anticonvulsant mechanism in pre-eclampsia/toxemia of pregnancy (Gupta, 2023). The so-called central anti-convulsant action of Mg²⁺ does not support a CNS / brain site of action that might be critical in the genesis of CSD.

A review of the literature from 1990 and of four intervention trials assessing the independent efficacy of oral Mg²⁺ supplementation in the prevention of migraine concluded that the evidence supporting oral Mg²⁺ supplementation is limited (Teigen and Boes, 2015). Coupled with the in-depth evaluation of Mg²⁺ induced cellular, cardiovascular, myoneural, BBB-related kinetics, and basic brain functional aberrations, including CSD, presented in this article, the review of Teigen and Boes (2015) indicates that the concept of a primary aberration of Mg²⁺ in migraine and its correction by exogenous administration of MGSO4 has proved erroneous in migraine therapeutics (Gupta, 2004d, 2009a, 2009b).

CSD and Photophobia

A diagnostic feature of both MwA and MwoA (Drummond, 1986; Vanagaite et al., 1997; ICHD - III, 2018), photophobia is defined as abnormal sensitivity or painful intolerance to light (Miller, 1985). Approximately 80-90% of migraine patients with normal eyesight experience photophobia (Selby and Lance, 1960; Miller, 1985; Drummond, 1986) (see below).

Defining photophobia nearly eight decades ago, Lebensohn and Bellows, (1934) underscored: “exposure” of the eye to light definitely induces or exacerbates pain. The neurobiology of photophobia is important to understand the fundamental basis of migraine (Maniyar et al., 2014a). Exacerbation of headache by light (ambient) or more correctly photoallodynia or photo-oculodynia (Fine and Digre, 1995) is commonly associated with primary intraocular disorders such as corneal abrasion /ulcer, iridocyclitis/uveitis, cone dystrophies, chiasmal compression, psychiatric disorders such as agoraphobia, as well as intracranial pathologies such as essential blepharospasm, almost all forms of migraine, meningitis, subarachnoid haemorrhage, and traumatic brain injury (Kawasaki and Purvin, 2002; Digre and Brennan, 2012; Friedman, 2013; Rossi and Recobar, 2015). Vanagaite et al. (1997) reported that photophobia is reported in 66-88% of migraine patients, and, that 3057% of migraine attacks can be precipitated by different light and glare sources such as sunlight, flicker of moving pictures, or television. Photophobia is more marked on the side of the headache (Drummond, 1986; Lance, 1993). Some regard photophobia as merely a functional symptom (Trobe, 2002). CSD has been suggested to underlie photophobia, through a rather convoluted process that involves activation of peripheral and central trigeminovascular neurons (Noseda and Burstein, 2013).

Painful photophobia depends on the functional integrity of both the optic and the trigeminal nerves (Lebensohn and Bellows, 1934; Maleki et al., 2012).

Photophobia is not associated with “positive” or “negative” migraine aura, and begins with and subsides with episodic migraine headache. Diplopia, palinopsia (illusory rather than hallucinatory), basilar-type migraine, retinal migraine, migraine with unilateral mydriasis, persistent visual aura, visual snow, phosphenes, halos, and migraine with binocular blindness are not associated with photophobia (Hupp et al., 1989; Schwartz and Robbins, 2012; Gersztenkorn and Lee, 2015; Kalita et al., 2016). Similarly, migraine-related vertigo and migraine-associated motion sickness are not associated with photophobia (Crevits and Bosman, 2005).

Usually very uncomfortable and disabling, the exact nature and causative mechanisms of headache-related photophobia in migraine are still uncertain (Chronicle and Mulleners, 1996; Rossi and Recober, 2015). Clinical and preclinical research has identified several potential pathways involved in enhanced light sensitivity. Some of these structures include trigeminal afferents in the eye, second order neurons in the trigeminal nucleus caudalis, third-order neurons in the posterior thalamus, modulatory neurons in the hypothalamus, and fourth order neurons in the visual and somatosensory cortices (Rossi and Recobar, 2015). The comfort of the dark and that of dark sunglasses afforded to the migraine patient during and inbetween headache attacks suggests a definitive role for miosis in the genesis of dominantly lateralizing photophobia.

In migraine patients with normal eyesight, green light appears to exacerbate migraine headache significantly less than white, blue, amber, or red lights. (Noseda et al., 2016) Migraineurs experience with colour, and, colour-selective migraine photophobia could originate in cone-driven retinal pathways, fine-tuned in relay thalamic neurons outside the main visual pathway, and preserved by the cortex (Zervas and Smith, 1987; Noseda et al., 2016). Photophobia of migraine appears to be an “irritability/hyperexcitability” of non-cortical (retinal) origin rather than cortical (Brennan, 2011). Importantly, no difference in light sensitivity or photophobia has been found between MwA patients and MwoA patients during attacks as well as between attacks (Vanagaite et al., 1997). Pain thresholds are generally lower in light sensitive than in non-light-sensitive migraine patients. (Vanagaite et al., 1997)

Remarkably, photophobia, reported by visually impaired adult migraine patients (n=6) in whom the headache preceded the blindness in time, disappeared after the development of the visual impairment (Silva et al., 2014) (see below).

Clearly, the anterior visual pathway rather than the visual cortex-related CSD is predominantly involved in the genesis of photophobia in migraineurs. Pupillary miosis and normal retinal function appear essential in the development of photophobia of migraine.

CSD and meninges

Migraine pain is thought to originate from chemical irritation of the meninges leading to transmission of nociceptive signals from the dura mater to the brain via the TVS (Blau, 1992a; Noseda et al., 2010; Noseda and Burstein, 2013). Such a long-term (frequently over decades), recurrent, episodic, self-limited, non-infective, uncomplicated meningeal activation/irritation is a unique or idiosyncratic, purely arbitrary mechanism or phenomenon that cannot be generalized (Table 1). Given that a far less severe but arbitrary “variant” of meningitis (? meningeal irritation) is prominently proposed to underlie episodic migraine, the concept is presumptive, raises more questions than answers, and encourages further research endeavors in this speculative vein without elucidating the basic neurobiology of CSD or allowing a comprehensive synthesis to emerge. Table 1 presents a comparison between migraine and the long-sought meningitis-like or meningeal “irritation/activation” state to position the discipline in order to confront a mythical neurological concept.

There is, apparently, no link between CSD and meningeal ‘irritation’, and hence, by extrapolation, between CSD and migraine (Table 1).

CSD and Ophthalmic division of the Trigeminal Nerve (V1)

Migraine is not a pan-trigeminal disorder but primarily affects the ophthalmic division (V 1) of the trigeminal nerve (Gupta, 2009a; Akerman et al., 2013). The pain of migraine headache is not, as has been generally assumed (Moskowitz et al., 1979; Andreou et al., 2010; Edvinsson, 2011), a diffusely distributed involvement throughout the three divisions of the trigeminal nerve. Unilateral or bilateral frontotemporal, parietal, vertex, retro-orbital or nuchal segmental pain is common in many migraine patients (Obermann et al. 2007a; Campbell 1990), and indicates predominant or selective and partial involvement of V 1 (Gupta, 2006b, 2009a, 2019, 2024). Migraine pain uncommonly involves the second (V 2) or maxillary division and generally spares the third (V 3) or mandibular division (Campbell, 1988). However, the pain of migraine may extend to C2 and C3, and neck stiffness is common; also, the pain wanders during individual attacks and in different episodes (Campbell, 1990; Blau and MacGregor, 1994; Bartsch, 2005). In another study with location frequencies, the highest were as follows: eyes (67.1%), temporal (58%), and frontal (55.9%) while the lowest were: diffusely (17.5%) and vertex (24.1%); intermediate frequencies were reported in the occipital (39.8%) and neck areas (39.7%) (Kelman, 2005). Lower-half headache / facial syndrome involving orbital pain spreading into the antral region, lower jaw and upper neck is an uncommon variant, with some patients responding to ergotamine (Campbell 1988; Peñarrocha et al., 2004). There is another variant with atypical pain localization but with typical deep seated unilateral pain, bizarre distribution, and other features of migraine including response to triptan but without V 1 involvement (Obermann et al., 2007a, 2007b).

Spread of pain to the back of the head as well as to the nuchal region has been, again, attributed to a diffuse undifferentiated involvement of all three divisions of the trigeminal nerve and upper cervical spinal segments – “trigeminovascular system” (Bartsch, 2005) or to the hypothetical phenomenon of “convergence” of sensory input (Kelman, 2005). The primary afferent fibres of the three divisions of the trigeminal nerve are arranged in a complex manner at the spinal nucleus with only pain and temperature fibers from the ophthalmic area descending densely to the lower limit of the first cervical spinal segment (Bannister et al., 1995; Bogduk, 1995). It is not possible to conceive that an appropriate V 1 segment or a matching segment of the somatosensory cortex could be selected or activated in isolation, cyclically or periodically or episodically, by any brain neuronal / neuropeptide process, including CSD, to generate neuroanatomically corresponding lateralized/localized cephalic or nuchal pain or both episodically for over decades in any one migraine patient (Gupta, 2009a, 2017e, 2019).

Peripheral sensitization of the trigeminal primary nociceptors via neurogenic inflammation is another favoured pathogenetic hypothesis (Strassman et al., 1996; Olesen et al., 2009; Noseda and Burstein, 2013). While CSD is the prototypic putative central sensitization process, activation of peripheral nociceptors is also believed to underlie neuronal hyperexcitability in migraine. Nociceptive input from perivascular sensory, particularly extracranial, dural, and pial arteries rather than veins, may be responsible for the headache phase of migraine. Such evidences are, however, limited and controversial (Olesen et al., 2009). It is likely, as suggested previously (Olesen et al., 2009), that afferent contributions from all threeterritories of the trigeminal nerve vary between migraine patients, with a predominant role for V 1 (Gupta, 2009a, 2019, 2024). Nevertheless, diffuse peripheral afferent trigeminal nociceptor activation, in isolation or in combination with central sensitization, cannot explain persistently or frequently lateralizing headache. In an exhaustive review of abnormal peripheral sensitization and a discussion of the origin of this input, (Olesen et al., 2009) acknowledge that there is little or no evidence with certitude regarding which pain fibers are activated, or how they are activated, or what activates them.

CSD has no conceivable link with peripheral sensitization of the trigeminal primary nociceptors in migraine whereas there are key limitations for central sensitization of trigeminal primary nociceptors through CSD in the disorder, as further discussed below.

CSD and Migraine Therapies – Major

Pharmacologic-Pathophysiologic Disconnects  

A wide spectrum of treatment options is currently available to manage migraine including medical, surgical (Mathew, 2014), and botulinum toxin.

(Ramachandran and Yaksh, 2014; Jackson et al., 2012; Gupta, 2005h, 2006a), closure of patent / persistent foramen ovale (Wilmhurst, 2018; Gupta, 2009c, 2010a), anesthesiologic blockade (Inan et al., 2016; Dilli et al., 2015; Bogduk, 2004), manual (Chaibi et al., 2011), cognitive (Harris et al., 2015), transcranial magnetic stimulation (Kumar et al., 2018; Conforto et al., 2014), acupuncture (Gu et al., 2018), external electrical trigeminal neurostimulation (Schoenen and Coppola, 2018), and noninvasive electrical vagal nerve stimulation (Tassorelli et al., 2018). This list is not exhaustive but underscores the profound confusion in the principles of theory and therapy of migraine. All extant pathogenetic therapies are empirical or experiential, including those supported by RCTs (Gupta, 2009a).

The sustained thrust for CSD is augmented by the assumed potential for development of new therapeutic drugs that could directly act on or modulate CSD to prevent migraine attacks (Kaufman et al., 2017; Galletti et al., 2009; Sanchez-del-Rio et al., 2006; Hadjikhani et al., 2001). It is widely believed that all migraine preventive agents suppress CSD as a common mechanism of action that will eventually lead to targeted and specific therapy (Ayata and Lauritzen, 2015; Ayata, 2010, 2009; Ayata et al., 2006). Widely disparate pharmacologic classes of drugs are used for migraine prevention. (Silberstein, 2015; Bogdanov et al., 2011; Jackson et al., 2015; 2012; 2010; Galletti et al., 2009). To significantly influence the occipital cortex or CSD, the pharmaceutical agent must freely cross the BBB by lipophilicity, carrier mediated influx transport, or receptor- or absorptive-mediated transcytosis (Edvinsson and Tfelt-Hansen, 2008; Gupta, 2019, 2009a). Lipophilicity, significant CSF levels, and prominent CNS side-effects of the pharmaceutical indicate significant BBB and CNS penetration and underscore its potential to influence CSD. While disruption of BBB has been excluded during both MwoA and MwA attacks, respectively, (Amin et al., 2017; Hougaard et al., 2017), penetration of the BBB by the pharmacologic agent inbetween attacks is, intriguingly, not considered important in migraine therapeutics. An assumed combination of peripheral and central sensitization (Goadsby, 2005, 2001; Malick and Burstein, 2000; Burstein et al., 2005; 2000; Knight and Goadsby, 2001) may have theoretically supplanted this pharmacologic absolute (Gupta, 2024, 2019, 2009a). The causal role of CSD appears to be more probable in MwA patients (Bogdanov et al., 2011) but significant limitations persist (see below) (Table 2).

No significant modification of CSD was observed after acute systemic administration of metoprolol, dihydroergotamine, acetylsalicylic acid, lignocaine, clonazepam, and valproate in cat brain (Kaube and Goadsby, 1994), and after acute systemic administration of metoprolol in rat brain (Alemdar et al., 2007). Conversely, chronic daily administration of topiramate, valproate, propranolol, amitriptyline, and methysergide resulted in a dose- and time-dependent suppression of CSD frequency by 40-80% (Ayata et al., 2006). Topical application of propranolol – as well as norepinephrine or clonidine or yohimbine – inhibited migration of a CSD wave in the adult rat grey matter (Richter et al., 2005).

No experimental data have determined definitely the mechanism(s) through which propranolol reduces migraine attack frequency in both MwA and MwoA patients (Weerasuriya et al., 1982; Andersson and Vinge, 1990; Limmroth and Michel, 2001; Gobel and Heinze, 2003; Diener, 2003). The efficacy of propranolol (80-240 mg /day), metoprolol (200 mg/day), and timolol (20-30 mg/day) is supported by clinical studies (Tfelt-Hansen et al., 1984; Tfelt-Hansen, 1986; Linde and Rossnagel, 2004).

While propranolol, metoprolol, and timolol are lipophilic and hence penetrate well into the CNS, atenolol and nadolol are only slightly or not at all lipophilic (Weerasuriya et al., 1982; Cruickshank and Prichard, 1994). Since atenolol (Stensrud and Sjaastad, 1980; Forssman et al., 1983; Johannsson et al., 1987) and nadolol (Ryan et al., 1983; Ryan, 1984; Freitag and Diamond, 1984; Olerud et al., 1986; Sudlikovski et al., 1987) have demonstrated prophylactic efficacy, BBB penetration does not appear to be required for prevention of MWA as well as MwoA (Gupta 2009a; Hougaard et al, 2017; Amin et al, 2017). Atenolol is currently established as a drug of first choice for episodic migraine prophylaxis (Andersson and Vinge, 1990; Emilien and Maloteaux, 1998; Limmroth and Michel, 2001; Edvardsson, 2013; Silberstein, 2015). Atenolol and nadolol are not known to critically alter any brain function possibly related to CSD. Nadolol has shown migraine preventive activity comparable to propranolol (Olerud et al., 1986; Sudlikovski et al., 1987; Andersson and Vinge, 1990). The concentration in the CSF of propranolol, oxprenolol, and metoprolol approximates that of the free drug concentration in the plasma but appears in brain tissue at concentrations 10-20 times greater than that of the hydrophilic atenolol (Neil-Dwyer et al. 1981).Remarkably, oxprenolol does not have any migraine preventive activity despite being lipophilic with CSF concentration approximating the plasma concentration and the brain/plasma ratio being higher than that for propranolol (Neil-Dwyer et al., 1981). This indicates that the pathophysiologic system involved in migraine lies outside the CNS, as is also indicated by other migraine attack-preventive drugs that do not cross the BBB, such as atenolol, nadolol, and verapamil, with no link to CSD (Gupta, 2004a). Barring insomnia, a side effect that is less often seen than with propranolol users -- as a class effect for ßblockers -- atenolol has no central or brain neuronal side effect (McAinsh and Cruickshank, 1990; Scheer et al., 2012; Chang et al., 2013). All ß-1-selective blockers including hydrophilic atenolol and moderately lipophilic metoprolol and bisoprolol, suppress nighttime melatonin (Stoschitzky et al., 1999; Nathan et al., 1997; Van Den Heuvel et al., 1997; Brismar et al., 1988; Cowen et al., 1985; Arendt et al., 1985). In one polysomnographic study, 100 mg atenolol in healthy volunteers led to decreases in urinary 6-sulphatoxymelatonin and total sleep time, whereas 5 mg of melatonin supplementation while on atenolol restored sleep to baseline levels (Van Den Heuvel et al., 1997). Because the pineal gland lies outside of the BBB, both lipophilic and hydrophilic beta-blockers can selectively suppress melatonin production (Scheer et al., 2012).

Verapamil, a widely used migraine prophylactic agent but with limited evidence of benefit (Solomon, 1989; Andersson and Vinge, 1990; Goadsby et al., 2002), also does not cross the BBB (Agnoli, 1988; Gupta, 2009a). Some therapists prefer to use verapamil prophylactically in MwA patients, with or without headache, particularly when the auras are frequent or associated with hemiparesis, and in sporadic hemiplegic migraine (Silberstein and Young, 1995; Yu and Horowitz, 2003), although verapamil is not known to affect CSD. Nifedipine, like verapamil, does not freely cross the intact BBB, but both drugs can abort migraine aura (Silberstein and Young, 1995; Hupp et al., 1989). Amyl nitrite (Schumacher and Wolff, 1941), isoproterenol (Kupersmith et al., 1979), and glyceryltrinitrate (GTN) are regarded as the best vasodilator drugs to abort migraine aura (Silberstein and Young, 1995) but are not known to influence experimental CSD. GTN does not cross the BBB but releases NO radicals that freely cross the BBB. While GTN can precipitate an early and a delayed atypical migraine headache without vomiting (Maniyar et al., 2014), it cannot provoke the SS in MwA patients (Christiansen et al., 1999; Afridi et al., 2004), and is unlikely to affect CSD.              

The exclusive focus on brain theories of migraine genesis, including CSD, is not supported by the repeatedly confirmed migraine-preventing action of atenolol, nadolol, and verapamil discovered almost 3-4 decades ago, a feature that remains an enigma. Reduction of intraocular pressure (IOP) likely underlies the anti-migraine effect of beta-blockers and calcium-channel antagonists, a feature that cannot be discussed in detail in this article (Gupta, 2009a, 2019, 2023a, 2024).    

Evidence-based guidelines include antidepressants for prevention of migraine.

(Hershey et al., 1999; Colombo et al., 2004; Silberstein et al., 2012; Hershey et al., 2013; Kalita et al., 2013; Hickman et al., 2015). Amitriptyline (AMT) is known to consistently and cumulatively prevent migraine attacks in low doses in placebocontrolled trials (Gomersall and Stuart, 1973; Couch and Hassanein, 1979; Ziegler et al., 1993; Jackson et al., 2015; Silberstein, 2006, 2015). AMT enhances brain intrinsic serotonergic and noradrenergic neuronal activity and can rarely lead to seizure (Andermann et al., 1987, 1987a; Schacter, 1998; Mullins et al., 1999; Reichert et al., 2014). AMT, like cocaine and amphetamine, freely crosses the BBB and stimulates brain noradrenergic and serotonergic function by synaptic reuptake inhibition, as reviewed (Gupta, 2006b; 2006d). In higher doses and in combination with monoamine-oxidase inhibitors (MAOI), like clomipramine, AMT has the potential to cause the serotonin syndrome (Sternbach, 1991; Sporer, 1995; Gillman, 1999, 2010; Boyer and Shannon, 2005). Development of the serotonin syndrome clearly establishes that AMT is a brain serotonergic agonist. For several decades, AMT has erroneously been regarded as a CNS serotonin antagonist in the migraine literature. Theparadoxical combination of propranolol and AMT is also effective for the prevention of migraine attacks (Ziegler et al., 1987; Ziegler et al., 1993). Moreover, AMT releases arginine-vasopressin (AVP), the latter, in conjunction with increased migraine attack-related bioavailability of serotonin and noradrenaline within and outside the CNS, possibly maintains ocular vascular integrity, antinociception, and behavior control during early “pre-prodromal” or “pre-premonitory” and prodromal phases of vasodilatory antidromic V 1 nerve discharges (Gupta, 1997, 2023a, 2024, 2025). This “adaptive” or “protective” system likely keeps the migraine attack at bay for a limited but variable period of time (post-stress headache) in migraineurs, and, while “stress” is ubiquitous, it possibly offers unlimited neuroendocrine “protection” to the approximately 80% of the general population that does not suffer recurrent migraine attacks – one of the greatest mysteries of migraine (Gupta, 2024; 2025). There is no link between this human intrinsic and/or extrinsic eye/brain neuroendocrine adaptive “system” and experimental CSD in lower animals.

While topiramate at 30 mg kg_1 was able to inhibit both regional CBF changes and cortical depolarization in all rats and in 8 of 13 cats (Akerman and Goadsby, 2005), the drug suppresses migraine aura variably (Lampl et al., 2004; Silberstein et al., 2006; Karadas et al., 2014) or completely (Wolthausen et al., 2009).

Nevertheless, in one study, topiramate was at least as effective as AMT in terms of reducing the rate of mean monthly migraine headache episodes and all prespecified secondary efficacy endpoints (Dodick et al., 2009). In yet another study, topiramate appeared to reduce migraine auras in parallel with headache reductions (Reuter et al., 2010). A recent meta-analysis of placebo-controlled studies also demonstrated that topiramate in a 100 mg/day dosage is effective in reducing headache frequency and is reasonably well-tolerated in adult patients with episodic migraine (Linde et al., 2013a). Meta-analysis, however, risks drawing the clinician away from clinical reality, clinical judgment, and clinical commonsense (Horton and Kendall, 1991; Feinstein, 1994). A dose-response association has not been found with use of topiramate in migraine; a dose of 200 mg does not appear to be more effective than 100 mg (Linde et al., 2013a). Topiramate is a classic example of the difficulty of replicating results in migraineurs and of the variable therapeutic dose paradigm even in the same patient from time to time.

Lamotrigine, which is selectively effective on the migraine aura but not on the headache, had a suppressive effect on CSD of 37% and 60% at posterior (parieto-occipital) and anterior (frontal) electrodes, respectively, as well as on frontal Fos expression (Bogdanov et al., 2011). Valproate, which has a non-selective effect on both migraine aura and headache, had no effect on posterior CSDs but reduced anterior ones by 32%, slowed propagation velocity, and decreased frontal Fos expression (Bogdanov et al., 2011). Riboflavin also has a non-selective, unpredictable, non-dose-responsive effect in migraine prophylaxis (Schoenen et al., 1998; Colombo et al., 2014; Shaik and Gan, 2015; Namazi et al., 2015) but did not suppress CSD or its propagation (Bogdanov et al., 2011). Valproate and riboflavin have no known effect on the triggering of CSD, although they are seemingly effective in preventing MwoA (see above, Table 2). These studies highlight the unexpected, unpatterned, differential, inconsistent, irreplicable, and irreconcilable differences between animal experimental CSD results and human pharmacologic evidence.

Pregabalin is effective in increasing the threshold for CSD in wild-type mice, in slowing the propagation velocity of CSD in familial hemiplegic migraine (FHM)-1 mutant mice, and in preventing CSD subcortical propagation associated with the milder R192Q FHM-1 mutation (Cain et al., 2017). Early reports on the preventive use of pregabalin in migraine are encouraging (Calandre et al., 2010; Pizzolato et al., 2011; Bakhshandeh Bali et al., 2015). In routine clinically efficacious oral doses, chronic administration of pregabalin inhibits measures of central sensitization in the electrical hyperalgesia model in human volunteers (Chizh et al., 2007). Central sensitization is a key in the development and maintenance of chronic pain, particularly neuropathic pain, such as postherpetic neuralgia (Dworkin et al., 2003) and painful diabetic peripheral neuropathy (Didangelos et al., 2014). The pain of migraine, however, is not neuropathic in nature. The precise mechanism(s) of the anti-migraine effect of pregabalin remain unknown.

Acetylsalicylic acid (ASA) prevents migraine attacks, the effects being small, statistically insignificant, not dose-dependent, and variable (Benseñor et al., 2001). In other studies, ASA reduced the attack frequency of migraine to below 50% in relatively large doses (500 mg t.i.d.), the effect being comparable to metoprolol (200 mg daily) (Grotemeyer et al., 1990) or propranolol (O’Neill and Mann, 1978; Baldrati et al., 1983; Savi et al., 2013). ASA (1000 mg) is an effective treatment for acute migraine headaches (Kirthi et al., 2013). In a retrospective study, Anoaica et al., (2014) found ASA to be particularly useful for acute attacks of headache in MwA patients. Both cyclooxygenase-1 and -2 isoenzymes are expressed in peripheral tissues and the CNS. Although a solely peripheral mechanism of action was suggested for several years, various studies have suggested an additional central effect of ASA on the CNS, including pain processing structures (Ferreira et al., 1978; Maihöfner et al., 2007; Kröger and May, 2014).

Flunarizine, a Ca++ antagonist (see Table 2), appears to parallel topiramate.

(Kim et al., 2013) and propranolol (Verspeelt et al., 1996) in the prevention of migraine attacks. Following oral administration, flunarizine shows a rise in plasma concentration, reaching the steady state in five weeks. Drowsiness, weight gain, depression (Veerspeelt et al., 1996), and a Parkinsonian syndrome are the most important side effects (Martinez-Lage, 1988; Centonze et al., 1990), none of which can be related to CSD in lower experimental animals.

In summary, BBB penetration and modulation of CSD by migraine preventive drugs is non-uniform, essential, non-dose dependent, and controversial. Drugs that do not cross the BBB significantly have as good migraine preventive action as do drugs that readily cross the BBB. Modulation or abolition of presumed CSD is not an essential prerequisite to prevent MwA or MwoA. The primary site of pathophysiologic aberration in migraine appears to lie outside of the brain / CNS (Gupta, 2023a, 2024, 2025).

CSD, brain cortical excitability, epilepsy, transcranial magnetic stimulation, glyceryl dinitrate, depression, sleep, and anxiety

Central neuronal hyperexcitability (CNH) is regarded as an essential association or precursor / prelude to the presumed spontaneously-generating idiopathic or de novo CSD-migraine algorithm (Welch et al. 1990). How or why such CNH should be triggered periodically and result in lateralizing headache as well as non-homonymous SS (see below) in the first place remains unknown.               

The link between migraine and epilepsy began with the descriptions of Hughlings Jackson (1834-1911) and later Gowers (Pearce, 1986). The “self-cycling” nature of epilepsy / CNH appeared as the most appropriate model for the CSD-migraine nexus. Childhood epilepsy and migraine are two of the most common conditions seen individually in pediatric neurology and separating the two is not always easy (Oakley and Kossoff, 2014). Despite major phenotypic differences between migraine and epilepsy, the epidemiological bidirectional link in children is overemphasized, especially in those with benign partial epilepsies with occipital paroxysms (Bauer et al., 2013; Oakley and Kossoff, 2014). When either migraine or epilepsy is present in an individual, the other is more likely to be present (Haut et al., 2006; Kelley et al., 2012; Winawer et al., 2013), the reasons for which might vary from a shared genetic susceptibility to “stress” to a shared high commonality in the community. Several clinical observations circumstantially relate migraine to epilepsy, including the following: (i) electroencephalographic (EEG) abnormalities occur in migraineurs who have not had seizures; (ii) patients can have a seizure following onset of migraine; (iii) both conditions may be preceded by an aura, and (iv) anticonvulsants can be effective in both conditions (Andermann, 1987, 1987a; Ottman and Lipton, 1994; Oakley and Kossoff, 2014). Many patients may have both epilepsy and migraine simultaneously (Oakley and Kossoff, 2014). Triggers common to both migraine and epilepsy include sleep deprivation and stress (Neut et al., 2012; Balamurugan et al., 2013). Differentiation between migraine and epilepsy becomes more difficult in individuals whose seizures present with headache or when there is an atypical clinical presentation suggestive of both disorders (Peatfield and Rose, 1981; Gelfand et al., 2013; Bauer et al., 2013; Oakley and Kossoff, 2014). A detailed consideration of the complex and controversial relation between migraine and epilepsy is beyond the scope of this review article. Focal aberrant brain neuronal activation with or without partial or diffuse spread underlies seizure disorder. Migraine headache attacks are, remarkably, generated over a much-slower time framework than epilepsy (hours versus seconds and minutes, respectively). Even the rate of propagation of migrainous SS (discussed below) is at least a hundred times slower, in general, as compared with other neurogenic vascular/neuronal phenomena (Blau, 1992a). The ICHD-II (International Classification of Headache Disorders – II edition--2004) defines “migralepsy”, “hemicrania epileptic”, and “postictal headache” (Olesen, 2004), while the International League against Epilepsy (ILAE) does not include headache at all (Berg et al., 2010). Key phenomenological and therapeutic differences, nevertheless, exist between epilepsy and migraine (Table 3).

An idiopathic, self-generating CNH / hyperresponsiveness of trigeminal neurons is regarded as a pivotal physiological possibility underlying pathogenesis of migraine (Welch et al., 1990, 2005; Aurora et al., 1999; Aurora and Wilkinson 2007; De Simone et al., 2007; Akerman and Goadsby, 2005). Instantaneous but variable responses to therapeutic transcranial magnetic stimulation (tTMS) applied during the aura or headache phase may reflect abnormal impulse activity in cortical pyramidal neurons (outer layers II and III of the cerebral cortex) of migraine patients in the form of a spreading fast oscillatory neuronal activity rather than CSD (Brighina et al. 2004; Fisher, 2004; O’Reardon et al., 2007; McComas and Upton, 2015). It is essential to establish the physiological process(es) that initiate as well as terminate(s) the putative or presumed CNH / trigeminal V1 neuronal hyperresponsiveness. Moreover, the striking and lateralizing confinement of pain of migraine headache to the trigeminal V1 and/or nuchal region cannot be explained by diffuse CNH or holotrigeminal neuronal hyperresponsiveness (Gupta, 2024).Finally, while the site of therapeutic transcranial magnetic stimulation (tTMS) is distant from the proposed site of origin of headache (somatosensory area), in a severely affected patient the optimal site for tTMS may move and the “therapy” may eventually lose its efficacy (McComas and Upton, 2015).

Glycerol trinitrate (GTN) precipitates delayed-onset (1.5-6 hours after i.v. administration) migraine-like attacks (Thomsen et al., 1994; Schoonman et al., 2008). GTN is a nitric oxide (NO) donor, which easily crosses the BBB. GTN (or NO) remarkably, does not generate SS in MwA patients. Conversely, GTN (0.5 mg sublingually) is the best known treatment to manage SS (Silberstein and Young, 1995).

Additionally, peripheral fronto-temporal scalp application of GTN ointment precipitates fronto-temporal lateralizing migraine attacks, a peripheral feature inexplicable with CNH or CSD or diffuse trigeminal neuronal hyperresponsiveness. (Bonuso et al., 1989) This experiment underscores the importance of the extracranial fronto-temporal region in genesis of migraine (Bonuso et al., 1989), an importance that is as yet undeciphered but is clearly unlinked to CSD. While NO does not itself generate CSD, brain tissue NO is increased by CSD; NO is essential for recovery of ionic homeostasis following CSD (Read et al., 1997; Obrenovitch et al., 2002).

Migraine is comorbid with depression, anxiety, somatic amplification, perceived stress and sleep disorders, and deficiency in problem solving, which observational relationships could be bidirectional (Yavuz et al., 2013; Eskin et al., 2013; Dikmen et al., 2015; Baldacci et al., 2015). A brain cortical hyperactivity has been suggested to underlie the complex neurolimbic dysfunction seen in these comorbid disorders (Baldacci et al., 2015). There is no known parallel or link between CSD in experimental animals and depression, anxiety, or stress generated in humans. The neurocircuitry between migraine and its comorbid neuropsychiatric associations remains unknown (Minen et al., 2016). Sleep can trigger as well as alleviate migraine headache, lending no further clarity to the underlying mechanisms in the current theoretical framework (Bigal and Hargreaves, 2013; Kelman and Rains, 2005). Regardless of the stage of sleep, AVP is released and can abort migraine headache (Gupta, 2024). A fuller synthesis of these factors is beyond the scope of this review.

CSD / SD or CNH or trigeminal neuronal hyperresponsiveness have not been linked to the often-lateralizing autonomic nervous system (ANS) dysfunction known to prevail in migraine since approximately 150 years (Hsu et al., 1977; Havanka-Kanniainen et al., 1988; Peroutka, 2004) (discussed further below).

In summary, none of the elaborations of this section clarify the link between migraine and CSD. The extracranial fronto-temporal scalp region appears to have a definitive pathophysiologic significance that needs further elucidation.

CSD, Migraine, Stress, ANS, Adaptation, Stroke

Sudden Death, and Evolution: “Stress” is a ubiquitous clinical influence. While far from fully comprehending the subject and its link to “stress”, we have been enabled to visualize its general landscape (Chrousos and Gold, 1995). What was intuitive and remote for many generations of physicians and thinkers, including Hippocrates, Galen, Aristotle,

Sydenham, Bernard, Cannon, and Selye, can now begun to be rationally understood (Goldstein and Kopin, 2007). The ANS is central to adaptation. Different stressors elicit different patterns of activations of adaptive systems, the outcome of one of such incredibly complex systems concept of ANS adaptation being migraine.

“Stress is life and life is stress” – Hans Selye’s words could easily be paraphrased for migraine sufferers – “Stress is migraine and migraine is stress”. “Stress” fully pervades our life and influences us as individuals, communities, and humanity. Presently, “stress” serves in migraine, as it does in other medical fields, to confuse and confound rational thought as a “learned” euphemism for scientific ignorance (Charlton, 1992; Chrousos and Gold, 1995; Gupta 2004e, 2004f; 2019; 2024). “Stress” or “tension” and its defervescence (post-stress period) is one of the most commonly cited causes of migraine, and up to two-thirds of migraineurs experience “stress” and its aftermath (Spierings et al., 2014; Kurlander et al., 2014; Lipton et al., 2014; Peroutka, 2014; Goadsby, 2014; Pellegrino et al., 2018). Whatever cannot be understood in medicine, is attributed to “stress” or “stressors”, as if the underlying pathophysiological mechanism thus becomes apparent or that a fundamentally biological or a major therapeutic advance has been attained (Gupta, 2019; 2024). “Stress” has, however, never been linked to CSD, but migraineurs do indicate a failure to maintain the “milieu intérieur” (Claude Bernard), “homeostasis” or failure of the “fight or flight” response (Cannon) (Goldstein and Kopin, 2007). The need for near-constant adjustments to maintain physiological, behavioral, cognitive and emotional homeostasis in migraineurs is important. (Noseda et al., 2014).

Graham (1953) underscored that migraine patients manifest an “undue tendency to seek “stress” and at the same time a deficiency in physiological adaptation to “stress”. Graham (1953) also alluded to a mechanistic basis for the self-limited nature of migraine attacks (4-72 hours); he proposed that “they (the migraine attacks) are terminated…when the stress of the attack itself finally stimulates the patient’s own adaptation machinery into proper action…”. Graham spoke presciently of “bodily economy”, a phrase that can today be interpreted in terms of exhaustion of the entrained physiologic processes, in particular, the proposed vasopressin-serotonin-noradrenaline “system / nexus” and / or ANS as a “limited, variably exhaustible biologic commodity in a large segment of the general population afflicted with migraine attacks” (Gupta, 2009a, 2019, 2024). Most importantly, Graham (1953) found that migraine attacks were triggered in 75% of patients after the stress / emotion had resolved, a “precipitation-by-exhaustion” (Gupta, 2009a, 2009b, 2017a, 2024). A slow, steady draining of physical and psychical energy that could better be termed today as “ANS exhaustion or depletion”, appears central to the genesis of episodic post-stress migraine attacks. A “protective” or “evolutionary defense” or “warning” advantageous role for stress/stressors for a psychosocially-critical but limited-period in migraineurs could thus be postulated (Blau, 1992a, 1992b, 1992c; Loder, 2002; Devineni, 2005; Buzzi et al., 2005), through “physico-psychical collapse” or “enforced rest”, including “sleep” or a periodic “upgrading” of the proposed “brain intrinsic vasopressin-serotonergicnoradrenergic” adaptive “system” (Gupta, 2009a, 2017a, 2019, 2024).

Importantly, while fear- or anxiety-related emotional stress may actively suppress AVP secretion, sleep stimulates AVP release, as reviewed (Gupta,1997). Sleep enhances AVP secretion regardless of EEG sleep stage, remarkably at a time when redistribution of the blood volume from the extremities to the central vessels would otherwise suppress AVP secretion (Valiquette, 1987). AVP effects dose-related moderate to substantial reductions in rapid-eye-movement (REM) sleep, which effects were primary and not mediated by changes in cortisol or blood pressure(Born et al., 1992). The relationship between sleep and headache has been known for over a hundred years. Migraine seems to occur in association with REM sleep, and in association with excessive Stage III, IV and REM sleep stages, an association seemingly directly related to the sleep state rather than to circadian rhythmicity (Sahota and Dexter, 1990). Also, preponderance of migraine in women could reflect their need for greater biological fitness (Friedman and De ver Dye, 2009). Early and frequent exposure to “visceral” pain of migraine in childhood and adolescence could also prepare females evolutionarily for the most severe “visceral” pain of unassisted cervical/vulval dilatation during parturition, the most vital futuristic adaptive function to the preservation of humans since antiquity.

Migraine could also promote survival (Blau, 1991a, 1992b, 1992c). Normotensive migraineurs maintain a relatively low blood pressure, particularly the diastolic value, before or very early, during, and after migraine attacks (Seçil et al., 2010). Aortic distensibility and aortic stiffness (beta index) were significantly impaired in patients with migraine compared to control subjects (Akturk et al., 2014), indicating a central component to peripheral BP measurements in migraine that limits antigravity cranial blood flow. Epidemiological evidence of an inverse relationship between systolic and diastolic blood pressure as well as pulse pressure in a headache-free population at baseline and subsequent development of both migraine and tension-type headache 11 years later has been presented (Fagernæs et al. 2015), in effect confirming that migraine is associated with low blood pressure and low pulse pressure. Migraine patients have fewer than expected cerebrovascular and cardiovascular disease events and fewer electrocardiographic abnormalities (Featherstone, 1986; Waters et al., 1983); an inconsistent statistically heterogeneous outlook expectedly prevails in this comorbidity (Mahmoud et al., 2018). Sudden unexpected death (SUD) in adult migraineurs is rare -- except with arrhythmias (Monroe et al., 2015) -- to practically unknown (Cozzolino et al., 2018) (discussed above). Comorbid depression, fibromyalgia, and severe pain are associated with a high suicide risk. (Breslau et al. 2012; Kim and Park, 2014; Liu et al., 2015) Migraine-associated suicide is not reported in populations where medical needs are comprehensively covered by the Government, such as Taiwan (Harnod et al., 2018). Additionally, the risk of all-cause mortality was not significantly higher in migraineurs (Mahmoud et al., 2018). Finally, the link to ischemic or hemorrhagic stroke is devoid of any defensible mechanistic link to CSD. (Gupta, 2017d, 2024).

The migraineur with frequent attacks remains in a near-constant state of unstable homeostasis or overt/subliminal perception of stress, while the migraine attack reflects an intermittent total failure of homeostasis or adaptive mechanisms. The biology of any disease, including migraine, is the integrated composite of pathogenetic, adaptive, and epiphenomenal features (Gupta 2009a, 2009b, 2017d,2019; 2024). The hyper-focus on biologic “markers” detected in the laboratory over the last several decades has distorted the meaning of the term “biological” in migraine research. The term “biological” is not synonymous with “physiological”, “organic”, “laboratory” or “non-environmental” (van Pragg et al., 1979). “Good medicine is indivisible…(has) always (seen) the patient balanced between opposing forces pushing him towards health or disease…” (Wall and Melzack, 1989). The concatenation of physiological forces that envelops the migraine patient is not a nihilistic, hopelessly complex obfuscation but a well-orchestrated, intricate combination of opposing influences pushing or pulling the patient towards or away from migraine attacks (Gupta, 2019, 2024).

Adaptive forces are always generated by the organism itself (Carrel, 1959). Migraine may reflect an evolutionary human ability or “readiness” to always identify and to adapt to a changing environment, albeit with varying success, frequently negative consequences, and comorbidity. Migraine may also be a dynamic psychophysical fulcrum or equilibrium to allow subliminal stress to surface -- an evolutionary “alarm”. Migraine, like life itself, is a high-order dynamic equilibrium that preserves and prepares the patient against many (but not all) odds and offers alarm “systems”, enforced rest, and now medications to restore equilibrium. The neuroendocrine adaptive system evolved in migraineurs likely involves a parallel activation of the CNS / extra-CNS “vasopressinergic-serotonergic-noradrenergic” nexus (Gupta 1997, 2009a). Like all homeostatic systems, however, the adaptive mechanism evolved in migraine has both limitations as well as a distinct biological cost (Gupta, 2024). There appears to be no link between a pathogenetic role for CSD in migraine and this putative adaptive system.

A significant delay (running upto several hours or few days) in onset of migraine attacks, occurs in the following clinical circumstances: (i) stress, exercise or consumption of alcohol (Blau, 1992a); ingestion of tyramine-containing foods, involving a delay of upto 22 hours (Poole, 1986); cocaine abuse or abstinence (Satel and Gawin 1989; Dhuna et al., 1991); caffeine withdrawal over the weekend (Couturier et al., 1992); GTN ingestion or infusion (Thomsen et al., 1994); and, exposure to m-chlorophenylpiperazine (Brewerton et al., 1988). Primary headache associated with catastrophic psychological (non-physical) trauma (Shulman, 1991) or physical trauma (Haas et al., 2004) is similarly associated with a significant delay in onset of migraine (Gupta 2009a, 2009b). In the last 24 hours before headache onset, hypothalamic activity, as seen by MRI as a response to trigeminal nociceptive activity, is altered and increases towards the next migraine attack (Schulte and May, 2016b). Intriguingly, such hypothalamic activity (Maniyar et al., 2014; Schulte et al., 2017) is neither primary (Gupta, 2006c) nor lateralizing (Gupta, 2004a, 2009a, 2019).

Significant delay in onset of migraine headache encompasses the proposed subliminal “pre-prodromal / pre-premonitory” phase as well as the covert/overt prodromal/premonitory phase of migraine and is entirely analogous to the period after an “acute injury” -- as outlined by Wall and Melzack (1989) -- during which the migraine patient remains largely or completely unaware of the threat of impending painful cephalic disturbance and continues to function appropriately (Gupta 2009a, 2017a, 2019, 2024, 2025). During this “pre-attack” period, a host of linked adaptive physiological features are activated and operate efficiently to keep the pre-attack (“pre-prodromal” / “pre-premonitory”) phase, or prodrome, or aura, or headache, or any combination thereof at bay (Gupta, 2019; 2024; 2025). The physiological system(s) primarily affected in migraine, therefore, must necessarily be afforded a considerable (but limited and exhaustible) degree of protection by homeostatic defense mechanisms, thereby allowing the subject to continue to function for some considerable time despite exposure to the headache-provoking stimulus or clinical situation. This protection is apparently unlimited or inexhaustible in the majority of the general population who are not susceptible to migraine headaches but are practically exposed to similar stressful situations. Such “adaptive” or “protective” function is liable to fatigue variablybothin different migraine patients as well as in the same migraine patient on different occasions. Onset of primary migraine headache attacks, therefore, likely represents periodic “overwhelming” or “fatigue” of this adaptive “system” (Gupta, 2024; 2025).

The adaptive role of CSD as a protective biological strategy in the form of vascular preconditioning and consequent delayed intracellular neuronal protection is emerging (Gupta 2005a, 2017a, 2024; Kobayashi et al., 1995; Matsushima et al., 1996; Shen et al., 2016). CSD effects the expression of several potentially neuroprotective genes and induces subsequent or secondary tolerance to permanent focal ischemia in rat brain (Kawahara et al., 1995; Matsushima and Hakim, 1995; Yanamoto et al., 2000; Yanamoto et al., 2004; Yanamoto et al., 2005; Otori et al., 2003). Uncoupling protein 5 (UP-5) is triggered by CSD in rats after 24 hours in the ipsilateral cortex in all cortical strata, suggesting that CSD-stimulated UC-5 can have a neuroprotective effect at this time (Viggiano et al., 2016).

Under conditions of anoxia or high-temperature stress many insects enter a reversible coma, which is protective, a temporary shutdown of neuromuscular systems that prevents irreversible consequences of complete energy loss (Rodgers et al., 2010). SD can be induced in the CNS of the invertebrate, African Locusta migratoria, to evoke stress-induced coma as a protective strategy in response to anoxic and other extreme environmental conditions (high ambient temperature, flash floods, and monsoons in their natural habitat) with a silencing of the activity of the nervous and muscular systems until the stress is removed (Rodgers et al., 2010). During stress-induced coma, SD-like events occur in the locust metathoracic ganglion (MTG) that closely resemble CSD in many respects, including mechanisms of induction, extracellular potassium ion changes, and propagation in mammalian grey matter. These locust SD-like events conserve energy and prevent cellular damage (Rodgers et al., 2010). Such coma is associated with dramatic surges of extracellular K+ and temporary electric silence in the CNS, remarkably bearing all the hallmarks of SD in the vertebrate CNS grey matter. In the locust thoracic nervous system, the SD process can be studied in preparations that have an intact neuronal architecture allowing the simultaneous investigation of cellular and subcellular neuronal responses without anesthetic. After return to normoxia, locusts completely recover from up to at least 6 hours of anoxia-induced coma (Wu et al., 2002; Armstrong et al., 2009). SD was first proposed to occur in insects in the cockroach approximately 50 years ago. (Rounds, 1967).

Neuroimaging “correlates” of Cortical Spreading

Depression in humans: Due to the protean nature of migraine as well as the short duration of the migraine aura, neuroimaging during spontaneous or triggered attacks has proved difficult (Denuelle et al., 2008). Some scientists regard neuroimaging as an exciting development that has provided unprecedented insights and significantly enhanced understanding of migraine as a CNS disorder (Schwedt and Dodick 2009; May, 2013; Lakhan et al., 2013; Schulte and May, 2016a). Descriptions of brain regions and networks that are atypical between migraine attacks are believed to have helped identify the neuroanatomical basis for migraine symptomatology and to bridge the wide gaps between neurophysiological studies and clinical findings (May, 2013; Schwedt and Chong, 2015). It is also firmly believed that functional imaging has confirmed the importance of CSD as the underlying pathophysiological mechanism of migraine aura in human beings (Bogdanov et al., 2011; Lauritzen and Strong, 2017). Mechanisms involved in the transformation of migraine from an episodic disorder to one with near-continuous symptomatology are also believed to have been unveiled through advanced neuroimaging (Maniyar and Goadsby, 2013). Additionally, brain neuroimaging has stoked the hope that novel diagnostic and therapeutic interventions would emerge through further pursuit of CSD (Ayata et al. 2006; Ayata, 2009, 2010; Tedeschi et al., 2012; Costa et al., 2013).

A profound confusion, however, still prevails between CSD-related CBF patterns in experimental animals and altered CBF patterns in humans (Gupta 2009a, 2009b; Charles and Baca, 2013; Borgdorff, 2018). Issues of the basis of “self-cycling”, onset and offset of CBF alterations, and other key limitations (see above) critically limit the pathophysiologic validity of the CSD-linked neurovascular theory of migraine (Edmeads, 1991; Blau, 1992a; Gupta, 2009a). Next, the true beginning of a migraine attack remains preclinical and inaccessible to investigators (Friedman, 1985; May, 2013). Moreover, the symptoms and signs of migraine develop at least a hundred times slower than those seen in brain attacks of vascular or epileptic origin (Blau, 1992a) (discussed above). Finally, the notion of clinically “silent aura” for MwoA patients (Moskowitz, 2008) is simply an assumption (Charles and Baca, 2013; Gupta, 2009a).

Localized decrease in cerebral blood flow (CBF) that appeared to spread contiguously along the cerebral cortex was observed in MwA patients (Olesen et al., 1981). This pattern of “spreading oligemia” was confirmed in subsequent studies (Olesen et al., 1982; Lauritzen et al., 1983; Olesen et al., 1990). The “spreading hypoperfusion” in CBF was also seen in MwoA (Gelmers, 1982; Woods et al., 1994).

Whereas an aberrant hyperemia-oligemia CBF pattern has been linked to CSD (Leão, 1944b; Lauritzen, 1994), the converse oligemia-hyperemia pattern prevails during migraine attacks (Lauritzen and Olesen, 1984; Olsen et al., 1987; Kobari et al., 1989; Olesen et al., 1990). Artifacts related to Compton’s scatter (Olsen et al., 1993) and CBF changes induced by carotid artery puncture and catheterization complicated interpretation of such studies. Other studies also demonstrated a noncritical (<25%) oligemia-hyperemia pattern in MwA patients (Andersen et al., 1988; La Spina et al., 1997; Soriani et al., 1997; Sanchez del Rio et al., 1999) and in MwoA patients (Deneulle et al., 2008). In a human model of sildenafil-induced migraine, blood flow velocity in the middle cerebral artery (MCA) as recorded by transcranial Doppler ultrasonography and regional CBF in the territory of the MCA measured using SPECT and xenon-133 inhalation remained unchanged (Kruuse et al., 2003). In these studies in humans, the degree of cerebral oligemia was relatively small and not critical (i.e., not >75%) to affect occipital neuronal function negatively (Gupta, 2009a, 2019). Painful stimulation can itself, on the other hand, induce a robust (22.8%) statistically reliable decrease in global CBF, possibly involving sympathetic nervous system activation (Coghill et al., 1998). It is, therefore, unclear whether the oligemia seen in migraine patients was associated with pain or was a manifestation of CSD (Schwedt and Dodick, 2009).

Cutrer and colleagues studied cerebral hypoperfusion in patients with migraine aura using perfusion- and diffusion-weighted MRI (Cutrer et al., 1998). Decreases in CBF (16-53%) and cerebral blood volume (6-33%) and an increase in tissue mean transit time (10-54%) were seen in the grey matter of the occipital lobe contralateral to the affected visual hemifield (Cutrer et al., 1998). No changes in the apparent diffusion coefficient were observed either while the patients were symptomatic or after resolution of the visual symptoms but before the onset of headache (Cutrer et al., 1998). With a red-green checkerboard stimulus in patients with migraine, the functional MRI-BOLD (blood-oxygenation-level-dependent) technique was used to record the SD of prior activation of BOLD signal before the onset of visual change and headache in migraine patients (Cao et al., 1999). Hadjikhani and colleagues also studied spontaneous visual auras in four migraine patients using BOLD-functional MRI (fMRI) (Hadjikhani et al., 2001). Following a radial flickering checkerboard stimulus, certain CSD-linked CBF characteristics were recorded in migraine patients. Despite direct evidence for the neuronal origin of BOLD-source localized magnetoencephalography (MEG) in resting humans of frequency-specific electrophysiological correlation structure (Hipp and Siegel, 2015), the CSD issue in migraine remains uncertain (Chong et al., 2016).

Structural changes in the brain, using high-resolution cortical measurement and diffusion tensor imaging (DTI), in the visual motion processing network involving bilateral non-lateralizing abnormal thickness in areas MT+ and V3A (both cortical and in the underlying white matter), and, significantly lower fractional anisotropy in subcortical areas (superior colliculus and lateral geniculate nucleus) compared to healthy controls have been detected in some MwA and MwoA patients with longstanding disease (~20 years) that could account for, or be consequent to CNH (Granziera et al., 2006). These and other vascular, morphological, functional, chemical composition, metabolic, and low-frequency oscillatory brain changes in both MwA and MwoA patients suggest a basis for dysfunction between attacks as well as structural and functional plasticity in the “pain matrix”, showing reduced cortical thickness and surface area (P < .01). These two metrics were increased in regions involved in executive functions and visual motion processing (Messina et al., 2013; Schwedt et al., 2015; Hodkinson et al., 2016; Chong et al., 2016). Questions arise regarding the origin, specificity, selectivity, and duration, as well as the nature of such changes in discrete brain areas such as the hippocampus, anterior insula, and primary somatosensory cortex in different migraine patients with varying (low versus high) frequency of attacks over decades. Furthermore, fMRI detected modular involvement of the salience network in the interictal period, which includes the anterior cingulate cortex, and, the frontoinsular cortex in another cohort of MwoA patients (Yu et al., 2017), and higher amplitude of the resting state fMRI fluctuation in MwA than in MwoA patients in the fronto-parietal and anterior cingulate cortices and the cerebellum (Faragó et al., 2017), again underscoring the variability of the results of neuroimaging between different cohorts and the absence of any functional link between the neuroimaging changes and the neuro-anatomical lateralization of neurologic aura or headache or both (Schmitz et al., 2008). Conversely, other migraine cohorts have an age-related advancing non-headache lateralizing negative and compounding thinning of regions that do not thin in healthy controls, including bilateral postcentral, right fusiform (occipitotemporal gyrus), and right temporal pole areas (Schmitz et al., 2008a; Chong et al., 2014).

The specificity of such non-lateralizing and conflicting functional and structural brain neuronal plasticity changes to migraine attack frequency/duration (DaSilva et al., 2007; Maleki et al., 2011; Messina et al., 2013; Rocca et al., 2014) that may also be conspicuous by their absence (Matharu et al., 2003), the influence of psychological and emotional comorbid factors such as pain, fear, stress, stresssusceptibility, stress-resilience, and helplessness besides anxiety and depression, and their possible link to CSD are other key factors that remain indeterminate. Migraine cohorts differ significantly, as a recent study using highly sensitive surface-based morphometry found no difference in cortical thickness between controls and patients, both in MwA and MwoA (Datta et al., 2011). In still another cohort of MwoA patients, the mean cortical thickness of many brain regions in the frontal lobe, temporal lobe, occipital lobe, parietal lobe, and insula in the migraine patient group was significantly decreased compared with that in the healthy control group (Yu et al., 2016). The mean cortical thickness of the insula anterior was positively correlated with the duration of the disease course, while the mean cortical thickness of insula superior and insula inferior was negatively correlated with the duration of the disease course (Yu et al., 2016). The complex interactive reaction involving many brain regions and many brain network systems in MwoA patients does not clarify the cause-and-effect enigma (Hadjikhani, 2008; Yu et al., 2016). In yet another cohort of MwA patients, sensory migraine aura was not associated with altered grey matter structure, and patients with visual aura had normal cortical structure of areas involved in visual processing; the previously reported decrease in grey matter volume of the cingulate gyrus in patients compared to controls in MwA and MwoA most likely reflects a general clinical bias between patients and healthy controls (Hougaard et al., 2017). Therefore, besides the cause or consequence dilemma that remains unresolved (Hadjikhani, 2008; Yu et al., 2016), a more fundamental limitation of physical measurement and interpretation of cerebral cortex thickness in migraine patients per se, besides technical constraints, appears to be involved.

Experience-dependent secondary reactive or adaptive cortical plasticity – thickening or thinning -- is neither unique nor specific to MwA or MwoA patients only, and is quite unlikely to emerge as the long-elusive brain biomarker for the disorder.

Many factors including age (Sowell et al., 2003; Salat et al., 2004; Westlye et al., 2010; Chong et al., 2014; Wierenga et al., 2014), gender (Savic and Arver, 2014), genetics (Fjell et al., 2015), stress-induced alteration of hippocampal neurogenesis and its potential contribution to stress resilience / susceptibility (Levone et al., 2014), perception of pain (Schwedt and Chong, 2014), stress- and hibernation-linked bidirectional neuroplasm fluxes (Sotnikov et al., 2015), attention (Shaw et al., 2012, 2013), state of hydration (Biller et al., 2015), local topological constraints through childhood, adolescence, and young adulthood with white matter organization and expansion driving the thinning of sulcal cortex and synaptogenesis or increases in neuropil causing thickening in gyral cortex (Vandekar et al., 2015), apolipoprotein E genetic variations (Espeseth et al., 2012), influence of medications including opioid dependency in pain-free patients decreasing regional brain volume and connectivity in the amygdala (Upadhyay et al., 2010), cardiovascular fitness improvement (Scheewe et al., 2013), diabetes mellitus (Brundel et al., 2010; Chen et al., 2015; Moran et al., 2015), systemic hypertension (Alosco et al., 2014), and neuropathology and psychopathology influence the thickness of the brain cortex non-specifically (Rosas et al., 2002; Salat et al., 2004; Kuperberg et al., 2003; Rauch et al., 2004; Papagni et al., 2011).

The prevalence of migraine dramatically declines with advancing age (Blau, 1992a), but not necessarily in menopause (Ripa et al., 2015; Carturan et al., 2016); however, the relevance of age-related structural changes in the cerebral cortex to the pathogenesis of migraine appears incredibly complex (see above). Depression appears to be a prominent comorbidity with migraine and affects cortical thickness, with reduced right parahippocampal and fusiform gyrus thickness appearing to serve as a familial trait marker for vulnerability to mood disorders, and with progressive cortical thinning in the left inferior frontal and precentral gyri in subjects who remain well but with an increased risk for mood disorders (Papmeyer et al., 2015; Carturan et al., 2016). Conversely, onset of depression is associated with increasing left inferior frontal and precentral thickness (Papmeyer et al., 2015). With advancing age, increasing local arealization and gyrification occur with decreasing cortical thickness; this pattern of regional “cortical stretching” reflects the well-established phylogenetic principle of maximizing surface area and gyrification rather than increasing thickness to facilitate brain connectivity and functional development (Hogstrom et al., 2013).

Cessation of migraine attack-related pain of headache, i.e., 4-72 hours after onset, does not necessarily indicate a return in the postdrome phase of associated pathophysiologic anomalies to baseline (Gupta, 2005a, 2019). Also, the state of stress-associated arousal accompanying perceptions of challenge, threat, harm/loss, and coping options - primary and secondary appraisals (Schwarzer, 1998) – as well as subliminal stress (Passchier, 1994), besides innate stress-susceptibility or stress-resilience, suggest the operation of a continuum of stress-associated clinical, laboratory, and tissue responses (Gupta, 1997; 2004a; 2009a). Additionally, the prodrome can last several hours to a few days (Kelman, 2004b; Peatfield, 1988), while the postdrome lasts for an average duration of 25.2 hours (Kelman, 2006; Bose et al., 2018; Dodick, 2018). Most neuroimaging studies of migraine patients place a cut-off point of no migraine attack 72 hours before the study session, as well as no migraine attacks during or after scans, as a primary selection criterion (Hodkinson et al., 2015; Youssef et al., 2017). Besides variability between subjects in a cohort (Hodkinson et al., 2015) as well as of the test-retest reliability of CBF measurements in a clinical model of on-going post-surgical pain (Hodkinson et al., 2013), given the characteristically protean nature of migraine, no two attacks in any two migraine patients – and even in the same migraine patient on any two occasions -- would be strictly comparable in terms of length and nature of prodrome, aura manifestation and its attendant morbidity, severity and duration of headache, associated pre-attack and post-attack morbidity, associated symptoms like nausea/vomiting/dizziness, change in systemic BP, postdrome symptoms intensity and duration. (Gupta, 2010a, 2019, 2025) Meta-analyses add to the extant confusion in migraine pathophysiology (Gupta, 2010a; 2018a).

In a typically headache-lateralizing disorder such as migraine (Campbell, 1990), besides non-lateralizing bilateral rCBF increases in the primary somatosensory cortex (Kobari et al., 1989), a chance positive lateralizing correlation was found only within the right primary sensory cortex between rCBF value, migraine attack frequency, and cutaneous allodynia profile (Youssef et al., 2017). Such neuroanatomically inexplicable and conflicting findings across different investigations deepen the mystery of migraine. Since no two cohorts of migraine can be identical (Gupta, 2010a), replication of rCBF patterns is rather unlikely, and inexplicable results in neuroimaging are the rule rather than the exception, as the time from onset of migraine attacks to scanning and the quantum of stress-pain neuro-endocrinologic and neuro-psychologic responses will never be constant parameters.

In summary, in such a clinical scenario in migraine, statistical significance more often than not does not correlate with biological significance and is clinically misleading (Gupta, 2010a; 2018a; 2019).

Cortical spreading depression and migrainous

Scintillating scotoma: The cardinal attraction of CSD for maintaining the “neuronal” theory of migraine is the presumed ability of the CNS phenomenon (in rabbits and rodents) to explain positive visual aura in humans, in particular the onset and offset of SS and its propagation in MwA patients (Lauritzen, 1985; Vincent and Hadjikhani, 2007; Moskowitz, 2008; Pietrobon and Moskowitz, 2014; DeLange and Cutrer, 2014). The pathophysiology of the trigger(s) of CSD in the migraine population, nevertheless, remains unknown. (Charles and Baca, 2013).

Myriad neurological symptoms precede the headache phase in MwA patients (Hupp et al., 1989; Campbell, 1990; Vincent and Hadjikhani, 2007). Visual, sensory, motor, or psychic disturbances in migraine can occur that characteristically last for 10 to 25 minutes, but may be as brief as a few minutes or as long as an hour (Alvarez, 1960; Campbell, 1990; Friedman, 2004, 2008). Positive, negative, autonomic, or efferent ocular symptoms and signs are associated with MwA (Hupp et al., 1989; Friedman, 2004, 2008).

Migraine aura can be stereotyped or can vary greatly in severity on different occasions even in the same patient (Lauritzen, 1994; Schott, 2007; Hansen et al., 2013). Diverse visual (Queiroz et al., 1997), sensory (Jensen et al., 1986), motor (Thomsen and Olesen, 2004), language (Almeida et al., 1999), psychic, bizarre or brain stem disturbances can occur (Hupp et al., 1989; Silberstein and Young, 1995;

Foroozan and Cutrer, 2009), indicating that other areas of the cortex distinct from the occipital lobe must be affected (Vincent and Hadjikhani, 2007). The ability of CSD / SD to originate at discrete “skip” areas of the brain cortical region has been proposed (Vincent and Hadjikhani, 2007) but not established and remains inexplicable as the “dark unknown side of CSD / SD”. (Cozzolino et al., 2018) Besides SS, complex visual phenomenon (micropsia, macropsia, altered colour perception), lacrimation, monocular vision loss, diplopia, pupillary abnormalities, ophthalmoplegic migraine, and recurrent facial nerve palsy are also described in migraine (Hupp et al., 1989; Chronicle and Mulleners, 1996; Friedman, 2004; Maxner and Moeller, 2005). Importantly, photophobia, is common to both MwA and MwoA (Chronicle and Mulleners, 1996). The visual aura of migraine is purely symptomatic and is largely inaccessible to others, including investigators (Schott, 2007).
C.H. Parry in 1778 described his own scotomas precisely with the phenomenon never being followed by headache (Alvarez, 1960). Fothergill’s personal description of the visual aura emphasized movement, brightness, and the angles he likened to a fortification figure (Fox, 1919). Airy, unaware of Fothergill’s earlier description (Liveing, 1873), emphasized expansion of his own visual field defect from a “germ” in the paracentral area to a luminous crescent with a “boiling and trembling” motion and coloured angles (Airy, 1870). Individuals, including artists, have been drawing and describing their own migraine visual auras from centuries (Hare, 1966; Wilkinson and Robinson, 1985; Fuller and Gale, 1988; Rose, 2004; Podoll and Robinson, 2008; Schott, 2007; Hansen et al., 2013; Zehentbauer, 2015). Jolly (1902) studied the aura in one eye, the other having been destroyed at the age
of nine (Alvarez, 1960).

A positive central scotoma enlarging over 20-30 minutes and then disappearing is probably the only pathognomonic feature of migraine (Fisher, 1980; Rose, 1986). Migraineurs may not be aware of negative scotomata, but the positive and vivid SS cannot fail to be noticed (Fuller and Gale, 1988). The jagged effect of water, silhouette of a man with a spike edge, and a black sun motif intruding into an interior scene are some examples of illustrated migraine attacks by de Chirico (Fuller and Gale, 1988). In a series of paintings, MwA sufferers have shown stars or flashes of light (70%), fortification spectra (48%), and visual loss (16%) illustrated in different styles (Singer, 1928; Wilkinson and Robinson, 1985; Hupp et al., 1989; DeLange and Cutrer, 2014).

Lashley (1941) provided the first quantitative recording of the temporal spread of the migrainous SS completely without headache attacks. Another individual made detailed drawing of his visual percept of migraine aura in real time in over 1000 attacks of migraine aura without headache over 18 years (Hansen et al., 2013). The predominant visual percept was a curvilinear band of flickering (10-15 Hz) colourless brilliant-and-black gridlines, followed by an advancing area of impaired vision (Hansen et al., 2013). Consistent patterns of aura initiation, propagation, and termination were observed in both right and left visual fields. Most aura attacks originated centrally (within 10° eccentricity) but there were also other distinct sites of initiation in the visual fields (Hansen et al., 2013). Auras beginning centrally dominantly propagated first through lower nasal field (69-77% of all auras) before travelling to upper and outer temporal fields. The speed of the visual event was
estimated as 2-3 mm/min. Negative visual auras range from tunnel vision and scotomas to homonymous hemianopia and transient cortical blindness (Hupp et al., 1989; Campbell, 1990; Chronicle and Mulleners, 1996; Maxner and Moeller, 2005).

Scintillating scotoma (SS) is characterized most frequently by a wide spectrum of flashes of light or bright zig-zag, positive hemianopic (one sector of the visual field in 79.3% up to of 99% cases at least some of the time), horseshoe-shaped expanding visual perception with the open end of the horse-shoe towards the area of fixation that spreads from the paracentral region towards the peripheral temporal visual field usually in 10-20 minutes, and other less common but equally fascinating symptoms in quality, duration, and distribution i.e., unilateral (uniocular) or bilateral (binocular) (Alvarez, 1960; Manzoni et al., 1985; Hupp et al., 1989; Friedman, 2004). SS and other positive phenomenon possess a brightness that can be seen with eyes either opened or closed (Campbell,1990; Friedman, 2004). The bright positive image of the SS has a zig-zag quality with several short bands of bright colours or white light running parallel to one another (Alvarez, 1960; Hupp et al., 1989; Friedman, 2004). Spreading across the visual field, this positive phenomenon is generally followed by a negative scotoma that creates an area of obscured vision, wherein words are often read incompletely or only half a face might be seen (Hupp et al., 1989; Campbell, 1990; Friedman, 2004). Besides SS, positive visual aura symptoms include phosphenes or photopsiae, heat waves, kaleidoscope effects, and fragmented ‘cracked glass’ vision (Friedman, 2004). SS and other positive phenomena of the aura are uncertainly declared to be either monocular or binocular in distribution. Many patients will initially maintain that the observed SS was a monocular phenomenon (Hupp et al., 1989; Campbell, 1990) but by
suggestion acknowledge a binocular even homonymous SS defect (Campbell, 1990; Friedman, 2004). Bilateral visual field involvement without a definitive homonymous pattern is described in a retrospective study wherein a purely symptomatic phenomenon is complicated unsurprisingly also by passage of time involving memory and introspection (Queiroz, et al., 2011). Central vision disturbances, as is typical of SS, are commonly retinal in origin (Hupp et al., 1989).

Besides important characteristics of SS (Hupp et al., 1989), to further localize its source as being either related either to the anterior or to the posterior visual pathway, the following self-examination maneuvers are essential.

BOX 1. Clinical self-examination by migraine patients in SS / Negative Scotoma

1. The patient should be instructed to occlude each eye separately and alternatively during an attack of SS or negative scotoma, and, to observe the field of vision of the uncovered eye carefully. Occurrence of a monocular non-homonymous SS is an absolute sign of the retinal origin of the phenomenon.
2. Displaceability of the visual phenomenon horizontally or vertically or both offers an absolute clue, with only retinal visual phenomena moving with digital displacement of the eyeball (Hupp et al., 1989). Digital displacement of SS is another absolute proof of its origin in the anterior visual pathway at the level of the retina.
3. Movement of the SS synchronously (horizontally or vertically or both) with the movements of the eye is the third absolute proof of the retina being the site of origin of the phenomenon of SS.
4. Central vision disturbances, as are common in SS, are retinal in origin.
5. The MwA patient should be encouraged to draw the SS as a self-perceived visual illustration.

SS is not consistently associated with neuroanatomically lateralizing headache (Pearce, 1985; Blau, 1992a). Macular retinal lesions are well known to be associated with visual distortions (Arndt et al., 2000). Micropsia and macropsia have been described most frequently in the literature (in 58.6% and 45.0% of all patients, respectively), which might indicate that they are the most prevalent types of distortion but also that they are the best known and therefore studied most frequently, including the Alice in Wonderland syndrome (Blom, 2016). A combination of cultural factors and visual distortions at the macula may underlie the Alice in Wonderland phenomenon as part of positive scotomata in migraine.

The high percentage of neurologists and headache-specialists suffering migraine (Evans et al., 2003) or other knowledgeable medical practitioners / scientists are best placed to provide such critical information about migrainous SS. Fisher (1999), a physician, reported his own personal experience with SS in the form of an achromatic zigzag line over 27 years but did not describe a homonymous phenomenon or a simultaneous involvement of either a single or both nasal visual fields. Another neuro-physician with many migraine attacks over seven decades never described a homonymous scintillating / positive visual field defect. (Aring, 1991) Oliver Sachs, a neurologist by profession, described in his foreword for the book by Podoll and Robinson (2008), that he himself suffered left-sided non-homonymous visual phenomena and SS followed by blindness for a few minutes. No artist has as yet drawn their own positive migrainous visual aura to portray a homonymous SS i.e., involvement of both the nasal half of one visual field and the temporal half of the other visual field with lateralizing headache to support the principle of CSD or CSD-like cerebral phenomenon (Chronicle and Mulleners, 1996;

Gupta 2006b; Vongvaivanich et al., 2015). Additionally, the SS of migraine has never been represented diagrammatically by an expanding positive horse-shoe shaped scotoma beginning from the center and occupying the major portion of the nasal half of the visual field, right or left. Even Lashley (1941) did not describe spread of his own SS in the nasal halves of his own visual fields. The patient reported by Hansen et al. (2013) never drew a dominantly nasal expansion of the horse-shoe shaped SS in either right or left visual field. Conversely, migraineurs may experience almost any type of negative scotoma unaccompanied by positive scintillations (Hupp et al. 1989). In order of frequency, these negative scotomata may take the character of homonymous, central, tunnel, or altitudinal visual field defect, and mimics of positive migraine visual aura (see above) (Hupp et al., 1989 (see Table 2, p. 226); Alvarez, 1960; Fisher, 1980; O’Connor, 1981). It is, thus, critical to distinguish between positive and negative scotomata to understand the pathognomonic phenomenon of SS of migraine and to determine its source (retina v/s visual cortex).

Several unresolved issues persist and none of the extant theories offers a defensible hypothesis (Chronicle and Mulleners, 1996). While CSD and its propagation is the favoured theoretical basis for SS (Vincent and Hadjikhani, 2007; Noseda and Burstein, 2013), just like for migraine headache, there is no clue for the nature or basis of “self-cycling” genesis as well as for subsidence of SS. Moreover, there has been a profound conceptual mix-up between the clinical properties of positive and negative visual scotomata in migraine as well as in their neurobiology. Negative homonymous hemianopia of migraine (of cortical origin) must not be confused with positive non-homonymous  (of retinal origin). Finally, the profound or strong neurosilencing of CSD is not expected to be associated with florid positive
symptoms such as SS, flashing lights, castellations or pins and needles (Pearce, 1985; Borgdorff, 2018). In other migraine patients, visual aura symptoms are superimposed on a normal vision or numbness and tingling exist without loss of touch; such preservation of function seems incompatible with CSD that causes a reduction in spontaneous synaptic activity and action potential firing lasting over an hour (Sawant-Pokam et al., 2017). Typical paracentral horse-shoe-shaped expanding positive SS and typical migraine headache and are not reported after enucleation of one or both eyes (Peatfield and Rose, 1981; Nicolodi et al., 1997; Sörös et al, 2003; Noseda et al., 2010). Sightlessness markedly attenuates migrainous SS as well as photophobia (Silva et al., 2014).

A migraine-like aura with atypical features in perception and duration is a frequent concomitant or sequela of seizure (Shams and Plant, 2011). Such migraine-like aura is often associated with recurrent non-postular loss of consciousness as in the case-report of Peatfield and Rose (1981), in a woman who underwent bilateral enucleation of the eyes but also developed several episodic unrelated neurological features. Syncopal loss of consciousness in migraine is, conversely, generally orthostatic in origin. (Thijs et al., 2006; Curfman et al., 2012) The migraine aura and / or SS, in general, is much sharper in onset and shorter in duration than the headache. Yet, despite its relatively exceptional clinical pace, the rate of spread of the SS of migraine is remarkably slower than other brain phenomena involving vascular or epileptic / neuronal aberration (Blau, 1992a). Migraine-like visual aura otherwise fulfilling the ICHD (2018) criteria, but not the typical migrainous SS in time, pattern, appearance, or manner of spread -- have been reported to occur with a variety of focal cerebral lesions, i.e., an occipital seizure focus including tumour,
arteriovenous malformation or unknown (Wittkowsky, 1952; Battistella et al., 1987; Hupp et al,. 1989; Bruera et al., 1999; Shams and Plant, 2011). Moreover, in six migraine patients with bilateral enucleation of the eyes or damage to the optic nerve due to other illnesses, photophobia was absent (Noseda et al., 2010b), indicating that the postulated image formation input at the level of the central neurons that processes nociceptive signals from meninges is not obligatory for migraine-induced photophobia. Conversely, blind migraineurs with retained eyeballs (i.e., without enucleation), retained light perception, and retained pupillary responses, and with retained circadian photoentrainment reported sensitivity to light (Noseda et al., 2010b), underscoring the link between pupillary responsiveness and photophobia or light sensitivity. In male Sprague-Dawley rats (270-350 g, Harlan, Indianapolis, IN), intense light activates nociceptive neurons in superficial laminae of trigeminal subnucleus caudalis through an intraocular mechanism driven by a luminance-responsive circuit and increased parasympathetic outflow to the eye (Okamoto et al., 2010). Whereas visual pattern perception is not essential to experience photophobia (Amini et al., 2006), a minor or subsidiary central component through optic nerve input is required since intolerance to light due to eye injury is eliminated by enucleation of the eye (Custer and Reistad, 2000).

In summary, the oculo-brainstem-ANS-pupillary nexus in migraine does not appear to have any link to CSD.

Cortical spreading depression, nausea / vomiting of migraine, motion sickness, and cyclic vomiting syndrome

Early nausea and / or vomiting is a cardinal and diagnostic feature of migraine, both in episodic MwA and in MwoA; other forms of primary headache are generally not associated with early or prominent nausea and / or vomiting (ICHD, 2018). In MwA, nausea and / or vomiting, headache, and photophobia generally follow neurological aura symptoms either directly or within an hour (Campbell, 1990). One of the key limitations of the CSD theory appears to be its inability to explain the biology and the mechanism of occurrence of nausea and / or vomiting early in the migraine attack.

Conventionally, early nausea and / or vomiting in migraine attacks is assumed to reflect activation of the same putative pathogenetic cascade in migraine as is for SS or headache or both, with presumed SD-related orbitofrontal cortical, hypothalamic, and brain stem activation along with 5-HT release forming the final common pathway (Willoughby 1981; Welch, 1987; Welch et al. 1990; Schulte and May, 2016a). Nausea and vomiting are among the most frequently occurring symptoms observed by clinicians, a feature that involves vagal sensory inputs to the tractus solitarius that has, in turn, direct or indirect connections with several other hindbrain, midbrain and forebrain structures responsible for coordination of multiple organ systems (Babic and Browning, 2014). Several brain stem nuclei--including the
nucleus tractus solitarius, dorsal vagus and phrenic nuclei, medullary nuclei that regulate respiration, and nuclei that control pharyngeal, facial, and tongue movements-coordinate the initiation of emesis. Neurotransmitters postulated to be involved in this coordination include neurokinin-1, serotonin 5-HT3, and vasopressin (AVP) pathways (Hasler and Chey, 2003). Some believe that nausea and headache, two key symptoms of migraine, interact pathogenetically by reinforcing each other in a vicious circle during attacks (Drummond and Granston, 2004).

Somatic and visceral muscles exhibit stereotypic responses during emesis. Vomiting, coordinated by the brain stem, is effected by neuromuscular responses in the gut, pharynx, and thoraco-abdominal wall across the diaphragm (Hasler and Chey, 2003; Babic and Browning, 2014). The medulla oblongata is phylogenetically the oldest part of the nervous system. The emetic response was selected early in humans for protective purposes and serves as a safeguard against poisoning of the body (Chai et al., 2013). Biologically, nausea and vomiting are important components of an organism’s defences against ingested toxins (Andrews and Horn, 2006; Chai et al. 2013). This response gets toxins and drugs out of the body by summoning control over motor neurons either spontaneously or in an induced fashion by reflux gagging (Chai et al., 2013).

Emesis and migraine provoked by unpleasant thoughts or smells is believed to originate in the cerebral cortex (Stankewitz and May, 2001; Friedman and De Ver Dye, 2009; Sjostrand et al., 2010; de Lima et al., 2011). The area postrema, a medullary nucleus, however, responds to bloodborne emetic stimuli and is termed
the chemoreceptor trigger zone (CTZ). The area postrema is richly served by nerve fibers acting on 5-HT3 , muscarinic (M1), histaminergic  (H1), and dopamine  (D2) receptor subtypes. With the area postrema lying outside the BBB, many drugs and other agents act on the CTZ such as bacterial toxins and metabolic factors produced during uremia, hypoxia, and ketoacidosis. This means that large polar molecules, such as emetic toxins, can diffuse through to and reach the CTZ quite easily (Babic and Browning, 2014). Early adaptive release of AVP that stimulates the CTZ is the likely cause of early nausea and vomiting in migraine (see below) (Gupta, 2004j, 2005b, 2019).

Motion sickness represents a condition often associated with migraine that can reduce quality of life (Furman and Marcus, 2012). Motion sickness associated nausea and vomiting is believed to be due to a primary stimulation of the labyrinthine apparatus. Additionally, stress and perceived stress in a wide variety of clinical circumstances e.g., cyclic vomiting syndrome (CVS) (Li and Balint, 2000; Gupta, 2004j, 2005b; McRonald and Fleischer, 2005) also stimulate nausea and / or vomiting in an anticipatory manner to mitigate, alleviate, draw or divert attention from the stressful event. “Anticipatory stress” is a possible mechanism for the occurrence of CVS with migraine (Gupta, 2004g; 2004j). Self-induced vomiting (“gagging”) can also be a manipulative event. The pathophysiologic cause-effect mechanisms of “anticipatory stress” remain unknown but is likely linked to AVP secretion early in the course of migraine attack-related stress. The discussion of other clinical circumstances of nausea and / or vomiting are beyond the scope of this review.

As discussed in this review, the term “stress”, howsoever generated, does not translate itself into any pathophysiological mechanism. The central processing of signals for nausea and vomiting is particularly problematic in the light of observations that vomiting is more readily amenable to pharmacological treatment than is nausea, despite the assumption that nausea represents “low” intensity activation of pathways that can evoke vomiting when stimulated more intensely (Andrews and Horn, 2006). A central role for AVP in migraine pathophysiology has been suggested (Gupta 1997, 2009a, 2024). Spontaneous migraine-related stress (Hampton et al., 1991; Hasselblatt et al., 1999) and apomorphine-induced nausea (Sørensen and Hammer, 1985) both are powerful stimuli for AVP secretion. Continuous subcutaneous infusion of apomorphine (a classical dopamine receptor agonist with a rapid and powerful emetogenic effect) for three weeks was effective in reducing the frequency and the severity of migraine attacks (Lai et al., 1997).

AVP can itself generate vomiting, while vomiting, in turn, significantly increases AVP secretion resulting in a vicious circle. AVP administered intravenously results in vomiting and retching in humans (Rowe et al., 1979; Kim et al., 1997) and dogs (Carpenter et al., 1984). A low dose of exogenous AVP, which has no effect when given peripherally, causes an emetic response given centrally (Ikegaya and Matsuki, 2002). This suggests that the effects of AVP on nausea and / or vomiting are mediated via the CNS. Plasma AVP concentration is rapidly elevated in response to diverse emetic stimuli such as motion stimuli (Kim et al., 1997), and, apomorphine (Nussey et al., 1988) or cisplatin administration (Edwards, et al., 1989; Cubeddu et al., 1990).

Treatment with antagonists of the AVP receptor efficiently prevents motion sickness (Kohl and MacDonald, 1991; Cheung et al., 1994). This suggests that endogenous AVP is involved in naturally or spontaneously occurring emesis, including the characteristic nausea and/or vomiting of migraine. self-limited ressin- One of the most fascinating and vexing issues is how vomiting stops a migraine attack (Chai et al., 2013) and the linked question about how the migraine attacks stop (Villanueva, 2012; Houle et al., 2012). The vasop serotonergic-noradrenergic adaptive “system” apparently is a powerful central brain influence that can have a substantial role in both clinical circumstances( Gupta 19972009a). Intracerebroventricular and intravenous (i.v.) injections of AVP cause opposite effects on intraocular pressure (IOP), with i.v. administration causing a fall in IOP(Gondi M, et al., 2001). (Gondi M, et al., 2001). The central effect of AVP on IOP and the peripheral effects of AVP on IOP and pupil size (miosis) are due to stimulation of the V1 receptors Negativ e fluid balance, peripheral increase in AVP secretion, and fall in IOP are likely physiological influences that are possibly linked to the mechanisms involved in stopping of migraine attacks with or without nausea and / or vomiting (Gupta, 2023a). Moreover, in migraine patients (MwA or MwoA), with or without nausea and / or vomiting, the tamponade effect of the corneoscleral envelope also likely limits ocular distensibility and consequent trigeminal V 1 afferent nociceptive neural traffic, thereby limiting the duration of migraine attacks (Gupta, 2023a). in these key(Gupta, 2019). In summary, CSD has not been postulated to play a role phenotypic events involving AVP and tamponade effect of the eye(s) ;

Box 2. Clinical migraine phenomena that cannot be explained by CSD

Spontaneous (“self-cycling”) genesis/onset of migraine attacks (both for SS and headache).

Pain/headache of migraine.

Gender asymmetry – female preponderance. 

Protean nature of migraine, including predictable (menstrual) as well as unpredictable occurrences.

Prodrome and aura of migraine with multi-faceted “skip” neurological symptoms.

Delayed onset of headache for several hours or even a few days after exposure to triggers, including “stress”, significant alcohol consumption, and ingestion of tyramine-containing foods.

Spontaneous defervescence/cessation of migraine attacks (both for SS and headache).

Self-limited duration of migraine headache attacks (4-72 hours).                                                  

Variable frequency, severity, and comorbidity of migraine headache in discrete attacks dispersed in time, whereas CSD is an all-or-none physiological phenomenon.

Link to a large number of heterogeneous clinical conditions labelled as migraine triggering factors, particularly “stress” with no commonality of pathophysiologic mechanism(s).

“Post-stress” occurrence of migraine attacks.

Link to migraine remitting factors, particularly sleep, later trimesters of pregnancy. menopause and advancing age.

Precipitation of delayed atypical migraine-like headache attacks by GTN (human model) in the absence of aura / SS/nausea and/or vomiting.

Relief of SS with GTN / isoproterenol without any definitive effect on CSD.

Larger cohort ~66% of MwoA completely unconnected to CSD.

Typical lateralization of migraine headache.

Migraine-linked photophobia.

Similar light-sensitivity between MwA and MwoA.

Greater photophobia and light sensitivity on the side of the headache.

Postdrome of migraine.

Sparing of nasal visual field by SS.

Non-homonymous limited spread by SS in temporal half of visual field.

Markedly slow pattern of development and propagation of SS as well as headache in terms of or in relation to other neurological vascular or epileptic / neuronal disorders.

Loss of migrainous SS in sightless eyes in patients with previous migraine and non-enucleated eyes.

Loss or attenuation of photophobia in sightless eyes in patients with previous migraine and non-enucleated eyes.

Loss of SS, typical migraine headache, and photophobia in enucleated eyes in patients with previous migraine.

Autonomic manifestations of migraine—miosis, mydriasis, ptosis, lacrimation without salivation.

Early nausea / vomiting in migraine attacks.

Frequent abortion / cessation of headache following vomiting.

Link to motion sickness.

Link to cyclical vomiting of migraine.

Maintenance of systemic hypotension, vasovagal attacks with postular syncope, relative tachycardia, beat-to-beat heart rate and respiratory rate variability in between attacks.

Prevention of MwA and MwoA attacks by hydrophilic atenolol, nadolol, verapamil that do not cross the BBB readily.

Relief of migraine attacks by amitriptyline in both MwA and MwoA cohorts, a potentially seizure-inducing drug.

Relief of migraine attacks by a paradoxical combination of propranolol and amitriptyline in both MwA and MwoA cohorts.

The chance discovery of CSD in experimental animals by the alert scientist, Leão (1944a; 1944b), the presumed cortical origin of personal experience of SS without headache attacks earlier by another meticulous investigator (Lashley, 1941), and the assumed linkage between these two sets of cross-species events by yet another researcher (Milner, 1958), set the stage for a prolonged philosophic commitment by several migraine researchers to CSD well into the 21st century (Hartings et al., 2017). Another momentous but serendipitous event in migraine research six decades ago was the discovery of the preventive role of beta-blocker drugs (Rabkin et al., 1966), placing the otherwise well-known adaptive function of the ANS into pathophysiologic algorithms (Gupta, 2019; 2024). The bio-behavioral model of migraine emerged in the late 20th century, placing emphasis on aberration of the intrinsic brain noradrenergic function (Welch, 1987). Since the science of migraine has evolved by chance alone, all therapeutic options, including triptans, anti-convulsants, PFO-closure, botulinum toxin, and CGRP-antagonists, remain empirical or experiential (Gupta, 2009a, 2010a; Borgdorff, 2018). Most leading researchers acknowledge large gaps in the “theory” of CSD (and in other braincentric hypotheses including the assumed in vogue but unsustainable “self-cycling” concept), but simultaneously point to incredible and futuristic research progress over the last 100 years.

Do we face a scientific illusion through the lens of CSD or has indeed remarkable progress in the science of migraine been made? Do we know how or why episodes of CSD-linked migraine should start or stop de novo in almost one-fifth of humankind? Why is the remaining four-fifths of the human population spared from migraine, although stress is almost ubiquitous? (Gupta, 2019) What tissue or organ determines the typical duration (4-72 hours) of migraine attacks? Or why should vomiting commonly abort migraine attacks? Have we truly defined the biology of migraine? Does the diverse and increasing number of available therapies generate confidence in our comprehension of this discipline? Has the need for an overarching and robust theoretical matrix for migraine, outside of the putative vascular/neuronal hypotheses, become any lesser? (Gupta, 2009a, 2010a). Fundamentally, what is that which makes the CNS in migraine patients intermittently and unpredictably either more or less susceptible to the presumed attack-provoking influence of CSD (or other pathogenetic influences, including stress / stressors)?

(Maagdenberg et al., 2004; Teive et al., 2005; Cozzolino et al., 2018; Borgdorff, 2018).

Genetic mutations and amino acid substitutions in high-voltage sensing calcium channel proteins (Cav.2.1) have generated much excitement and are believed to increase CSD and aura susceptibility (Moskowitz, 2008), but in the context of a comprehensive evolution of the migraine entity, the contribution of these findings is essentially debatable. Several medical illnesses have a geneticpredisposition, such as diabetes mellitus or coronary artery disease, but it is the discovery of insulin and the atheromatous plaque, respectively, that dominates their pathophysiology, management, and research. In the absence of such a pathogenetic equivalent, migraine pathophysiology will remain rudimentary even with a rapidly increasing data bank, application of futuristic technology, and statistics. Assumptions and myths sustain CSD as a viable concept for migraine pathophysiology (see Box 2) (Gupta, 2009a).               

The need for a major change, and to de-link migraine from both the “vascular” and the “neuronal” theories was perceived almost three decades ago (Edmeads, 1991; Blau, 1992a). The key to understanding migraine mechanism(s) lies in the unlocking of the physiological processes underlying post-stress headache, an evolutionary protective mechanism (Graham, 1953; Blau, 1992a). A psychophysiologic and neuroendocrine mechanism underlying an early phase (preprodromal and prodromal) “protective” role for stress or emotion was proposed over two decades ago (Gupta, 1997), and further elucidated more recently (Gupta, 2004a, 2009a, 2010a, 2019, 2023a, 2024, 2025). Enhanced migraine attack-related bioavailability of AVP and 5-HT coupled with noradrenergic hyperfunction likely constitutes the adaptive “system” that maintains vascular integrity, antinociception, and behavior control during the early preictal “pre-prodromal / pre-premonitory” phase of stress of migrainous vasodilatory antidromic V1 trigeminal nerve nociceptive discharges (Gupta, 1997, 2009a, 2023a, 2024, 2025). Such an adaptive “system” helps create a conceptual divide between primary and secondary clinical phenomena (Gupta, 2004g, 2009a, 2019, 2024, 2025). The quest for a defensible theoretical matrix to place migraine research data in a logical, increasingly strengthening, robust, and generalizable synthesis appears feasible now.

Whereas CSD / SD has been detected in a few human brain diseases (Strong et al., 2002; Lauritzen et al., 2011; Lauritzen and Strong, 2017) with elements of acute tissue disruption, rheological changes, and pressure alterations such as is found in malignant stroke, subarachnoid and intracranial hemorrhage, and traumatic brain injury, CSD / SD has neither been detected in migraine patients without brain injury nor has any definite pathogenetic role been established for this electrical phenomenon (Gupta 2005a, 2009a; Yanamoto et al., 2000, 2004, 2005; Cozzolino et al., 2018). BBB disruption in migraine has also been practically eliminated in acute attacks of migraine (Hougaard et al., 2017; Amin et al., 2017, see above).

Conversely, the neuroprotective effect of CSD is known for the last three decades. Why might CSD / SD not be viewed as being protective in both children and adult migraineurs, as it is in experimental animals? (Gupta, 2009a; Siniatchkin et al., 2009). The rare clinical presentation of combined seizure-disorder and migraine (migralepsy) must not be allowed to confuse our perception (see above). Recording of primary or causative changes in migraine in the laboratory or by neuroimaging is difficult at present (Danuelle and Fabre, 2013). Since the migraine attack begins well before onset of the prodrome/premonitory phase, as well as the SS/aura, headache, or nausea/vomiting, and an extended postdrome is frequently manifest, researchers are invariably recording secondary changes in the laboratory or by neuroimaging and extrapolating them back to the tenets of experimental CSD / SD or to other central brain theories (Gupta 2009a, 2019, 2024, 2025). The focus on one or two factors recorded in the laboratory or through neuroimaging is generally diversionary as it ignores the whole and limits the perspective (Blau, 1992a).

A large spectrum of factors affects brain function and structure in migraine, and these features cannot be dissociated without insight into the relationship between primary and secondary factors (Gupta, 2019, 2024, 2025). As migraine is a protean disorder, these changes will be reflected variably rather than consistently -- even in the same patient on different occasions -- in advanced neuroimaging, including fMRI and BOLD-related studies, and thus the primary versus secondary or cause-effect debate will persist indefinitely (Gupta 2004a; 2024; 2025). Paradoxically, with the accumulation of more data in migraine research, pathophysiological confusion will worsen. SD-like phenomenon (like the CSD of Leão) has also been recorded in the retina, the only other avascular neural tissue (Martins-Ferreira and de Castro, 1966; Wiedemann et al., 1996; Duarte et al., 2003; Dahlem and Müller, 1997; Brand et al., 1998; Weimer and Hanke, 2005). Primary retinal stimulation in migraine patients to generate retinotopic localization of cortical changes cannot be excluded at this stage of our mechanistic comprehension. As discussed in this article, vision as well as the eye appears essential for the generation of typical photophobia, SS, and migraine headache (Gupta, 2019, 2023a, 2024, 2025). Association of low-tension glaucoma and normal-tension glaucoma in migraine (Bar et al., 2025) also has no conceivable link with CSD.

In a broader, more comprehensive, less confused, and more realistic canvas, both CSD in experimental animals and presumed CSD “surrogates” or “homologues” in humans in headache as well as in SS can be discussed afresh in this framework presented to evolve insight into migraine mechanisms. It is vital to discern the elusive pathophysiological commonality between widely disparate but well-established triggering and remitting factors—one single neurophysiological mechanism. Such an approach will provide a sound and logical matrix for the extensive and expanding phenotype of migraine. Conceptually, the adaptive “system” of migraine marks the first major defensible distinction between pathogenetic and non-pathogenetic phenomena in migraine (Gupta 2009a, 2023a 2024; 2025).

Remarkably little has been written about nausea and/or vomiting that rapidly follows the migraine attack and can frequently abort the headache. Over 2000 years ago, Hippocrates described “…vomiting, when it became possible, was able to divert the pain and render it more moderate” (Allory, 1859). In 1790, Tissot wrote: “…vomiting might herald the termination of the headache” (Pearce, 1986). The final common pathway for vomiting at the level of the brainstem or CTZ remains uncertain (Babic and Browning, 2014). AVP, as a neurotransmitter released in the brain, can itself trigger vomiting (see above) (Mathias and Bannister, 2013). Instantaneous increases in plasma AVP from 100 to 1000 times basal level are not unusual even when the nausea is transient and unaccompanied by vomiting or changes in blood pressure (Valiquette, 1987). Mild nausea can distort physiological AVP responses. because nausea is a more potent and predominating stimulus to AVP release. The onset of the migraine attack possibly represents overwhelming fatigue of the AVP- noradrenergic-serotonin linked adaptive “system” (Gupta 1997, 2009a, 2019, 2024, 2025). Once the migraine attack is triggered, the stress-stimulated release of AVP can trigger nausea or vomiting or both, while migraine attack-related nausea/vomiting may further augment the attack-abortive AVP release associated “system” in an adaptive circle. The tamponade effect of the eye, in the midst of such ANS and neuroendocrinal chaos, probably aids in aborting migraine headache(Gupta, 2009a, 2023, 2025).

While physiological and pharmacological concepts of CSD continue to advance (Spong et al., 2016a, 2016b, 2016c; Wainzstein and Lucci, 2017; Cozzolino et al. 2018; Klass et al. 2018; Close et al. 2019), the connect between CSD and migraine pathophysiology, based as it is on a presumptive link to the SS with many assumptions or arbitrary elements (Gupta 2023a, 2025) has made little progress. “The specific philosophic approach of the investigator to the problem at hand may itself be(come) one of the critical variables in the design and conduct of a given research strategy” (Watson, 1968).

In summary, the time to de-link migraine from CSD, a long-sustained mythical scientific connection, has arrived. Given the enormous homeostatic potential of the human body, everything that is recorded in the laboratory or through neuroimaging cannot be simply or simplistically presumed to be pathogenetic.

The article is a critical review and not a systematic one. Articles for the critical review have been collected from PubMED; Cochrane database; and, references from selected articles over the last 100 years using the terms “cortical spreading depression; spreading depression; migraine; aura; vomiting; vasopressin; serotonin; catecholamines, traumatic brain injury, photophobia”. A synthetic review is presented towards a gestaltic understanding of the disorder.

The role of CSD in migraine pathophysiology is fraught with assumptions and well-entrenched myths, including deeply ingrained link(s) to SS. The neuroanatomic, neurophysiologic, and biologic basis of scotomata / SS cannot be understood without definitive data coming from MwA in a systematic manner, as detailed in this article. Migraine pathophysiology is becoming increasingly blurred when viewed through the lens of CSD with additional data and statistics. This article provides a defensible, robust, generalizable, and progressive concept of the adaptive role of CSD / SD as well as of CNS / extra-CNS pathophysiology surrounding migraine and its main components. It is hoped that this matrix will further evolve the understanding of migraine, and, ultimately, its management and prognosis. The key event in deciphering migraine and creating a primary and secondary (cause / effect) divide is an unravelling of the “vasopressin-serotonin-norepinephrine” nexus and an understanding the fundamental laws that govern migraine pathophysiology and biology, its triggers, its link to stress, and its spontaneous onset, offset, and prevalence.