Table 1 Main features of ASL studies applied to migraine

Cadiot et al., 2018 [16]

17 MwA pts (median time of 6 h after the symptoms onset)

MRI (including TOF-MRA and 2D pulsed ASL)

- ASL: 16 children presented hypoperfusion in one or more cerebral lobes

- TOF-MRA: 12 children presented vasospasm within the intracranial vasculature

- 100% of the abnormal TOF-MRA images were associated with homolateral hypoperfusion

The association of perfusion abnormalities on ASL in brain regions, not corresponding to vascular territories, and vascular vasospasm in TOF-MRA, can aid diagnosis of MwA attack

Kim et al., 2016 [17]

1 pt with hemiplegic migraine during attack (R sided)

3 T MRI (including 3D pseudo-continuous ASL and DSC)

- DSC showed hypoperfusion with posterior predominance in the L cerebral hemisphere

- ASL perfusion correlated well with DSC findings

In pts with hemiplegic migraine, during an acute migraine attack, hypoperfusion in the affected cerebral hemisphere may be visualized with ASL technique

Boulouis et al., 2016 [18]

10 MwA pts during attack (median onset-to-MRI delay was 12 h)

1.5 T MRI (including 3D pseudo-continuous ASL and DWI)

CBF was:

- decreased in a brain region consistent with symptoms when MRI was performed less than 14 h after onset

- increased if the MRI was performed 17 h or more after

- DWI was normal

ASL provides insight in understanding of pathophysiology of MwA and helps ruling out acute ischaemic stroke

Pollock et al., 2008 [19]

3 migraine pts (1 hemiplegic migraine, 2 with visual aura) between 6 and 24 h from neurological symptoms onset

MRI (including ASL perfusion)

Pt 1:

- marked regional hyperperfusion in the R frontal and parietal cortex

Pt 2 and 3:

- ASL showed hyperperfusion in the medial occipital lobes bilaterally

ASL perfusion analysis may help identify pts with vascular or hemiplegic migraines by showing associated hypoperfusion during the aura phase and hyperperfusion during the headache phase

Burns et al., 2017 [20]

1 pt with CM and MwA during and after aura (R upper limb paresthesia and aphasia)

MRI (including ASL)

- Normal DWI

- ASL showed:

◦ hypoperfusion extending over the entire L hemisphere during aura

◦ hyperperfusion over the L parieto-occipital cortex 28 h later (headache still present, aphasia resolved)

ASL allows to detect hypoperfusion during the initial phase of aura extending over multiple vascular territories, and delayed hyperperfusion several hours after the beginning of migraine attack and potentially persisting after resolution of the aura during the headache phase of migraine cycle

Law-Ye et al., 2017 [21]

1 MwA pt (sudden aphasia fellowed by headache) during aura

MRI (including ASL)

- Normal axial T2-FLAIR and DWI

- ASL showed an extensive hypoperfusion of the L hemisphere, involving anterior and posterior vascular territories

ASL is especially useful in acute neurological deficits helping to diagnose mimics of stroke such as MwA

Wolf et al., 2017 [22]

4 MwA pts during aura, headache, and asymptomatic phases

MRI (including 3D pseudo-continuous ASL)

- Aura: adjacent hypoperfused and hyperperfused areas not respecting the boundaries of major cerebral vascular territories

- During headache: regional hyperperfusion

- Asymptomatic phase: normal perfusion

ASL perfusion imaging is a contrast agent-free

method suitable for assessment of reversible perfusion changes during or immediately after aura

Cobb-Pitstick et al., 2018 [23]

12 hemiplegic migraine pts between 3 and 11 h from neurologic symptoms onset

MRI (including 3D pseudo-continuous ASL)

All had normal DWI and abnormal ASL findings with differences in cerebral perfusion compared with the unaffected hemisphere:

◦ 9 hypoperfusion

◦ 3 hyperperfusion (indicating the transition to hyperperfusion by 11, 5.5, and 7.5 h)

The alteration of ASL sequences with normal DWI help to distinguishing hemiplegic migraine from stroke

Uetani et al., 2018 [24]

49 migraine pts (33 only migraine; 16 with other neurological symptoms: 8 aura, 3 visual symptoms, 9 motor disabilities, 7 confusion) during attack

3 T MRI (including 3D pulsed ASL)

- 11 pts exhibited perfusion abnormality, especially in the occipital lobe (73%)

- no abnormality on non-ASL MRI

- Higher prevalence of aura, motor disabilities, confusion, hospitalization in pts with perfusion abnormalities

In paediatric and adolescent pts with migraine, ASL shows a high prevalence of perfusion abnormality, especially in the occipital lobe

Kato et al., 2010 [25]

One MwoA scanned during 2 episodes of migraine:

- during attack (within 1 h post-onset)

- 30 min after oral administration of rizatriptan 10 mg attack-free period

3 T MRI (including pulsed ASL)

During attack (compared to migraine-free period):

- significant relative hypoperfusion in bilateral median thalamic areas and hypothalamus

- significant relative hyperperfusion in the frontal cortex

30 min after treatment (compared to images during attack):

- relative improvement of perfusion in the bilateral median thalamic areas and hypothalamus

Hypothalamus and its surrounding areas may participate in the pathogenesis of migraine attack

Younis et al., 2021 [26]

26 MwoA pts before and 6 h after administration of CGRP and sildenafil to induce MwoA attacks

3 T MRI (including pseudo-continuous ASL) and proton MR spectroscopy

- CBF increased in dorsolateral pons, ipsilateral to pain side during attacks, compared to outside attacks

- Glutamate levels in the same area remained unchanged during attacks, total creatine levels increased

Dorsolateral pontine activation during migraine was not associated with higher glutamate levels, while the concurrently increased total creatine levels may suggest an altered energy metabolism

Gil-Gouveia et al., 2017 [27]

13 MwoA pts during attack and headache-free period

3 T MRI (including 3D pseudo-continuous ASL)

Average total CBF values were similar within and outside the attack

Significant perfusion changes are not expected to occur during MwoA

Michels et al., 2019 [28]

17 migraine pts (13 MwA, 4 MwoA) free from migraine attacks at least 48 h before and after MRI, 19 HC

3 T MRI (including 2D pseudo-continuous ASL)

Hyperperfusion in the extrastriate cortex (area V5) in migraineurs in comparison to HC, more pronounced in MwA during the interictal state

In the interictal state, hyperperfusion was found in the supposed region of CSD onset, located occipito-temporally

Cucchiara et al., 2013 [29]

117 migraine pts (56 MwA, 61 MwoA) 53 HC

3 T MRI (including 3D pseudo-continuous ASL)

- Incomplete circle of Willis was significantly more common in the MwA compared to HC and MwoA

- MwA had a higher burden of variants than HC

- Pts with an incomplete circle had greater asymmetry in hemispheric CBF

- Specific posterior cerebral artery variants were associated with greater asymmetries of blood flow in the posterior cerebral artery territory

An incomplete circle of Willis is more common in MwA pts than HC, and is associated with CBF abnormalities

Puledda et al., 2021 [30]

24 VSS pts (15 with concomitant episodic migraine diagnosis) and 24 HC

scanned at rest and during “snow-like” visual stimulus

3 T MRI (including 3D pseudo-continuous ASL)

- Higher rCBF in VSS pts vs HC at rest and during “snow-like” visual stimulus in the cuneus, precuneus, inferior and superior parietal lobule superior parietal lobule, supplementary motor area, frontal eye fields, premotor cortex, posterior cingulate cortex, middle frontal gyrus, angular gyrus, post central gyrus, middle and superior occipital lobule

- rCBF increased in R insula in VSS pts vs HC during visual stimulation

Pts with VSS show an increased rCBF in a wide network of intrinsic brain areas involved in the processing of complex sensory and cognitive states independently of the presence of an external visual stimulus, suggesting that these abnormalities could be a causal factor of this disorder

Hodkinson et al., 2015 [31]

17 MwoA pts free from migraine attacks at least 72 h before and 24 h after MRI, 17 HC

3 T MRI (including pseudo-continuous ASL)

- Migraineurs had significantly higher rCBF in both the R and L post-central gyrus compared to HC

- CBF values in post-central gyrus were positively correlated with attack frequency, but not the duration of illness or the age of the pts

These results demonstrate the presence of a disease-specific functional deficit in a known region of the trigemino-cortical pathway, which may be driven by adaptive or maladaptive functional plasticity

Youssef et al., 2017 [32]

22 MwoA pts free from migraine attacks at least 72 h before and 24 h after MRI, 22 HC

3 T MRI (including pseudo-continuous ASL)

- Significant resting rCBF increases within bilateral post-central gyrus compared with HC

- Within the R post-central gyrus positive correlation between blood flow value with migraine attack frequency and cutaneous allodynia

These alterations within post-central gyrus may be consequence of repeated migraine attacks

Chen et al., 2018 [33]

15 MwoA pts free from migraine attacks at least 72 h before MRI, 15 HC

3 T MRI (including 3D pseudo-continuous ASL)

- Increased CBF in L BA38

- No significantly decreased CBF

- HAMD scores significantly positive correlation with the CBF value of the L BA38

Pattern of cerebral hyperperfusion may elucidate the neurogenic mechanism in the MwoA genesis

Fu et al., 2022 [34]

88 migraine pts (32 MwA, 56 MwoA) during interictal period

44 HC

3 T MRI (including 3D pseudo-continuous ASL)

Pts with MwA vs MwoA and HC present:

- higher CBF levels in superior frontal gyrus, postcentral gyrus and cerebellum,

- lower CBF levels in the middle frontal gyrus, thalamus and medioventral occipital cortex

- these variations significantly correlated with clinical rating scales

The alternated CBF in brain regions overlaid important component of neuro-networks and

circuits modulating migraine pain processing and showed consistency with previous studies in differentiating MwA from MwoA and HC

Zhang et al., 2021 [35]

40 MwoA pts free from migraine attacks at least 48 h before and after MRI, 42 HC

3 T MRI (including 3D pseudo-continuous ASL)

- MwoA patients had higher CBF in R middle frontal orbital gyrus and R middle frontal gyrus, while that in Vermis declined

- Increased CBF positively correlated with VLSQ-8 and monthly attack frequency score

- Decreased CBF connectivity between R middle frontal orbital gyrus and L superior frontal gyrus, R putamen, R caudate, R angular gyrus

- Increased CBF connectivity between L calcarine cortex and R middle frontal orbital gyrus

MwoA pts exhibit abnormalities in regional CBF and feature CBF connection defects at the RS in areas involve information perception, information integration, and emotional, pain and visual processing

Liu et al., 2022 [36]

13 CM pts (during interictal period)

15 HC

3 T MRI (including 3D pseudo-continuous ASL)

- Lower CBF value in CM pts vs HC in L nucleus accumbens

- Negative correlation between CBF value of the L nucleus accumbens and VAS score

- No significant difference for the volume of bilateral nucleus accumbens in CM vs HC

Hypoperfusion of the L nucleus accumbens could be considered as a potential diagnostic imaging biomarker of CM pts

Bai et al., 2022 [37]

18 CM during interictal period, 15 NDPH, 15 HC

3 T MRI (including 3D pseudo-continuous ASL)

NDPH vs HC:

- decreased CBF and aCBV values in R hemisphere (posterior and middle orbital gyrus, ventral anterior nucleus of thalamus)

CM vs HC:

- increased CBF and aCBV values in the ventral lateral nucleus of thalamus bilaterally

Being the thalamus a relay centre for ascending nociceptive information, recurring attacks of migraine in CM pts, could lead to increase neuronal thalamic activation and regional CBF

Meylakh et al., 2020 [38]

34 migraine pts (25 MwoA, 9 MwA)

- 22 scanned during interictal period (free from migraine at least 72 h before and 24 h after MRI)

- 13 scanned immediately following the end of attacks (within 72 h)

- 7 scanned prior to an attack (within 24 h)

26 HC

3 T MRI (including pseudo-continuous ASL)

- Prior to an attack, RS-CBF decreased in the lateral hypothalamus

- Resting functional connectivity strength decreased between the lateral hypothalamus and regions of the pain processing pathway (midbrain periaqueductal gray, dorsal pons, rostral ventromedial medulla, and cingulate cortex) only before an attack

Altered hypothalamic function and connectivity in the period immediately prior to

a migraine headache supports the hypothesis that hypothalamus is involved in migraine attack ignition

Datta et al., 2013 [39]

50 migraine pts (25 MwA, 25 MwoA) during interictal period; 25 HC

3 T MRI (including pulsed ASL and BOLD-fMRI)

BOLD fMRI response to visual stimulation within primary visual cortex and in the lateral geniculate nuclei was:

- greater in MwA compared to MwoA and HC

No difference between MwoA and HC

Increased activation in MWA was confined to the occipital pole

- Regional resting cerebral blood flow did not differ between groups

MwA subjects have cortical hyperresponsiveness to visual stimulus, suggesting a direct connection between cortical hyperresponsiveness and aura itself

Zhang et al., 2017 [40]

116 migraine pts (56 MwA, 60 MwoA) during interictal period, 54 HC

3 T MRI (including 3D pseudo-continuous ASL)

- MwA pts with high white matter hyperintensities load had lower CBF

- No association between white matter hyperintensities load and CBF was seen in HC and MwoA pts

White matter hyperintensities in MwA may be related to alterations in resting CBF (or viceversa)

Hu et al., 2019 [41]

24 CM pts; 24 HC

3 T MRI (including 3D pseudo-continuous ASL-fMRI and BOLD-fMRI images)

NVC biomarkers were:

- lower in L inferior parietal gyrus, L superior marginal gyrus, and L angular gyrus

- higher in R superior occipital gyrus, R superior parietal gyrus, and precuneus

These areas were related to visual or sensory information processing

ALFF-CBF:

- in R superior parietal gyrus positively correlated with disease history while R precuneus negatively correlated with migraine persisting time

fALFF-CBF:

- in L superior marginal gyrus and in L angular gyrus were negatively related to headache frequency and positively related to health condition

Multi-modal MRI could be used to detect NVC dysfunction in CM pts to assess the impact of chronic pain to the brain

Park et al., 2022 [42]

58 migraine pts (11 MwA, 47 MwoA) during interictal period

3 T MRI (including 3D pseudo-continuous ASL)

Pts with MwA vs MwoA present:

- no differences in the network measures of global structural connectivity

- differences in global functional connectivity (lower assortative coefficient)

Low network global assortativity may result in more vulnerable, easily disrupted, highly synchronized networks and, therefore more excitable brain connectome in MwA

Stankewitz et al., 2021 [43]

- 12 MwoA pts during attack (within 6 h after the beginning) rescanned every 1–4 days (final recording within 48 h before the subsequent attack)

3 T MRI (including pseudo-continuous ASL)

- Cyclic changes of brain perfusion in the limbic circuit, with the highest perfusion during attack

- Increase of hypothalamic connectivity to the limbic system over the interictal interval towards the attack, then collapsing during the headache phase

The hypothalamus seems to act as a metronome of internal processes, able to control limbic pathways, and migraine attacks may represent the result of the lost hypothalamic inhibition on the limbic circuit