Gua Sha for Migraine — Where Hands Often Fail

There is a kind of headache that responds beautifully to a thumb pressed firmly into the trapezius at the base of the skull. Most office workers know this point intuitively — the moment of relief when someone with strong hands works that exact knot, the sense that something has loosened that was about to ruin the afternoon. This works because the headache in question is a tension-type headache, and tension-type headaches respond to mechanical pressure on the muscles whose tightness is generating them.

But there is another kind of headache — and many people who think they have tension headaches actually have this one — that does not respond to thumb pressure at all. In fact, deep pressure on the upper neck can make it worse. This second kind of headache responds to gua sha, the scraping technique, in a way that has surprised many of my clients. It took me a long time to understand why; what follows is the picture I have come to, in plain language.

If you have ever had a headache that started at the base of the skull and crept forward over the top of the head, with a quality that was more “throbbing-pressure” than “tight-band,” and which got worse rather than better when someone tried to massage it — this piece is about why, and what to do instead. (For the same condition approached through needles rather than scraping, see the earlier migraine-relief acupuncture story.)

The two headaches that look the same

The tension-type headache and the cervicogenic migraine are often confused because they begin in similar places. Both start in the upper neck or base of the skull. Both can radiate forward. Both feel, to the person experiencing them, like “neck causing head.”

But they are different things, with different sources.

A tension-type headache comes from muscles in the neck and the head holding on for too long. The squeezing reduces the blood flow to the muscles themselves, the muscles get tired and irritated, and the irritation becomes pain. The feeling is steady, often “a tight band.” It responds to anything that relaxes the muscles: heat, gentle massage, careful stretching.

The other kind — what TCM teachers in my training called a “wind in the neck” headache, and what some Western doctors I have read call a “neck-origin” or “cervicogenic” migraine — comes from irritation of the small nerves in the upper neck. These nerves and the nerves of the face share part of the same pathway in the brain. When the neck nerves are angry, the brain often reads the signal as pain in the forehead and around the eyes. The feeling is more throbbing than tight, sometimes one-sided, sometimes with sensitivity to light or sound or with a little nausea — closer to a migraine, even though the source is in the neck.

The reason pressing the upper neck often makes this kind of headache worse is that the pressure adds irritation where the nerves are already irritated. The therapist is trying to release the muscle, but the muscle is not the real source. The muscle is holding on because the nerve beneath it is angry, and pressing harder makes the nerve angrier.

What gua sha does differently

Gua sha — the technique of scraping the skin with a smooth-edged tool after oiling — works on the body at a quite different level than direct pressure.

When applied to the upper neck and base of the skull, gua sha seems to do several useful things for this kind of headache.

It moves the fluid and blood in the surface layer of the skin and the soft tissue just under it, where things have often become quietly congested. The scraping clears this layer in a way nothing else really does.

It produces a strong but brief sensation — the warm scraping feeling — that the nervous system has to attend to. While it is attending to that, it cannot also amplify the background pain it has been amplifying. This is the same reason a hot shower briefly helps a headache: the nervous system has only so much attention to give.

It cools the tissue afterwards (the warmth of the scraping is followed, in the next minutes, by a slight cooling as the small capillaries settle), and the cooling seems to quieten the nerves further.

All of this happens without any pressure on the inflamed nerves themselves. The work is done on the layer above. That, I think, is why gua sha often helps where direct massage on the neck does not.

This is why gua sha often succeeds where massage fails on this kind of headache. The intervention is gentle in the right place rather than firm in the wrong place.

The protocol I use

For clients with suspected cervicogenic migraine — usually identified through a detailed history-taking that distinguishes their headache pattern from ordinary tension-type — the gua sha protocol is straightforward but specific.

The client lies on their stomach, with the neck and upper back exposed. A small amount of oil — usually a base of jojoba with a few drops of Mentha haplocalyx (Chinese field mint) — is applied to the area.

I use a porcelain spoon with a smoothed edge (a traditional gua sha tool that I have used for years and that fits the contours of my hand well). The strokes begin at the very base of the skull, moving downward, with moderate but not heavy pressure. The strokes follow the path of the bladder meridian along the spine, then move laterally to follow the trapezius along its upper border.

The treatment area extends from the occipital ridge down to the upper edge of the shoulder blades, and outward to about the level of the acromion (the bony point at the top of the shoulder). The whole region is treated systematically, with each strip of skin receiving five to ten strokes.

The marks that emerge — the sha — tell the story. In clients with cervicogenic patterns, the deepest sha is almost always in two specific zones: just below the occipital ridge on the affected side, and along the upper trapezius about two finger-widths from the base of the neck. These are the zones where the underlying nerve irritation is generating the surface congestion.

A session takes about twenty-five minutes. The relief, in clients for whom the technique is appropriate, usually begins to be felt within four to six hours after the session and is most pronounced the next day.

A small parallel from my grandmother

My grandmother in Liaoning, when one of us children had a headache that would not go away, used to apply a small piece of cloth dipped in slightly warm sesame oil to the back of the neck, and rub the area firmly in slow downward strokes. She would do this for ten or fifteen minutes. We thought it was a kind of comforting gesture; she was, in fact, performing a household version of gua sha. The cloth was rougher than a porcelain tool but the principle was the same: slow downward stroking of the upper neck and base of the skull.

The technique turns up under many names — the Vietnamese cạo gió, the Indonesian kerokan, and others besides. It seems that almost any culture with access to a little oil and a smooth-edged object eventually discovers that slow scraping of the neck and back eases certain kinds of pain. That so many places arrived at the same gesture independently suggests the underlying mechanism is a robust one.

When to use it, when not to

Gua sha for headache is particularly useful when:

The headache has a cervicogenic quality — it begins in the upper neck, has throbbing rather than tight-band character, and either does not respond to manual massage or worsens with deep neck pressure.

The pattern is recurrent rather than acute. A single severe headache should be evaluated by a doctor to rule out serious causes (sudden onset severe headaches, headaches with neurological symptoms, headaches after head injury, headaches with fever).

The client is otherwise healthy and has no contraindication to the technique. The contraindications are similar to cupping: bleeding disorders, anticoagulant medication, very fragile skin, recent radiotherapy to the area.

It is not the right intervention for: classic migraine without obvious cervical involvement (where the headache is more clearly a central neurological event); cluster headaches (a different condition entirely); medication-overuse headache (where the problem is rebound from frequent painkiller use and gua sha will not address the underlying issue); and acute severe headaches of unknown cause.

For chronic cervicogenic migraine, a typical treatment course is six to eight sessions across two to three months, with progressive reduction in headache frequency over the course of treatment. Most clients see meaningful improvement by the fourth session.

What it feels like the first time

The first gua sha session is, for many clients, an unusual experience. The scraping sensation is firm but not painful. The aromatic oil produces a pleasant cooling effect on the skin. There is often a moment, around the third or fourth stroke in each zone, where the client feels a sudden release — a small ease somewhere in the head or neck that they have not felt in some time.

The post-session feeling is usually one of considerable relaxation, sometimes mild fatigue, occasionally a brief sense of emotional release. Clients often sleep deeply that night.

The marks — the sha — are visible the next morning. They fade over three to seven days. As with cupping, they are not bruises and are not painful to the touch. They are the record of work done.

If your headaches have a cervicogenic quality and you have been frustrated by approaches that have not worked, this is one of the older techniques worth trying. Sometimes the right help is gentle in the right place rather than firm in the wrong one — which, now that I think of it, is also the only sensible way to drive through the roundabout at Zogu i Zi.


Yang Wang practises Chinese medicine at Chinese Massage – Tai Chi Tirana.

When Migraines Stop Coming on Tuesday

Erjona’s migraines used to arrive on Tuesday afternoons, between three and five o’clock, with a regularity that had made her stop scheduling client meetings for that window. She had tracked them for nearly a year before she came to see me — a small notebook with dates, weather, sleep hours, what she had eaten, what she had drunk, where she had been in her cycle. The notebook was the kind of document a person keeps when they have stopped expecting doctors to solve their problem and have decided to solve it themselves.

She was thirty-four. She worked in marketing for a logistics company with an office near Bllok. The migraines had started in 2024, six months after she moved back from Milan to take care of her mother through a long illness. They had outlasted her mother’s recovery by two years.

The neurologist she had seen — a kind, competent specialist at a private practice in Tirana — had given her sumatriptan for the acute episodes and propranolol as a preventive. The sumatriptan worked. The propranolol made her tired without reducing the frequency. She had tried magnesium, riboflavin, coenzyme Q10, the elimination of nightshades, the elimination of dairy, the elimination of red wine. The migraines kept arriving on Tuesday afternoons.

When she sat down in my office and showed me the notebook, I asked her one question.

“What happens on Monday evenings?”

She paused. She had not asked herself that question.

The pattern under the pattern

Migraines have a reputation for being unpredictable, but a careful look at the data — one’s own data, kept in a notebook — almost always reveals a pattern. The body does not produce a fifteen-hour vascular and neurological event at random. Something precedes it, sometimes by hours, sometimes by a full day. The trick is to find the upstream signal.

In Erjona’s case, the answer turned out to be small but consistent. Monday evenings, after work, she met three friends at a café in Bllok for what they called the catch-up. Two glasses of wine, often three, conversation that ran late. The wine she had eliminated once and then re-introduced after the migraines kept coming anyway. She had concluded the wine was not the trigger. She was wrong, but in a more interesting way than she had assumed.

The wine was not triggering the migraine. The wine was triggering a small, predictable disruption of her sleep architecture, which combined with elevated cortisol on Tuesday morning meetings, which combined with the late lunch she usually skipped on Tuesdays because of those meetings, which combined with — the actual immediate trigger — a small drop in blood sugar around three in the afternoon.

It was a stack of four small factors, none of them sufficient on its own. The pattern was visible only if you looked at all four together.

What acupuncture seems to do for migraine

I have read, over the years, that acupuncture is one of the better-tested traditional treatments for migraine, and that some European countries now suggest it for migraine sufferers when the usual medicines have not worked. The mechanism is still being worked out, as far as I understand it, but the broad picture goes something like this — and here I am simplifying what doctors much more qualified than me have written.

Migraine is, very roughly, a kind of over-reactive event in the brain. Something — stress, a hormonal shift, a missed meal, a glass of wine the night before — pushes the system over a threshold, and the migraine arrives. The brain of a migraine sufferer sits closer to that threshold than the average brain.

Acupuncture, done well and consistently, seems to raise the threshold by a small but useful amount. The needles work, as I understand it, partly on the nerves of the face and head, partly by quietening down inflammation, and partly through the nervous system as a whole — bringing the over-alert “fight or flight” mode back to a calmer baseline.

It is not a cure. It does not eliminate migraines for most patients. What it reliably does, in clients who respond, is reduce the frequency by between thirty and sixty percent and reduce the severity of the episodes that still occur. For someone having eight migraines a month, this means three or four migraines a month. The change is significant. It is also rarely dramatic enough to convince a sceptic in a single session.

What we did

I told Erjona the truth: her notebook had already done eighty percent of the work. The single most useful intervention she could make, before any needle touched her, was to address the Monday evening pattern. Not eliminate the friends. Not eliminate the wine. Reduce the wine to one glass, eat a real lunch on Tuesday before any meeting, carry a small protein snack for the three-o’clock window. She agreed to try this for six weeks.

In parallel, we began acupuncture. Twice a week for the first three weeks, then once a week. Points: Taiyang (the extra point at the temples), Fengchi (the gallbladder-20 point at the base of the skull), Hegu (large-intestine-4, in the hand), Taichong (liver-3, between the toes), and a rotating selection of secondary points based on what her pulse and tongue showed at each visit. A familiar pattern for migraine prevention, the kind of protocol many TCM practitioners have used for many years. Nothing experimental.

The first two weeks: no change. She had two migraines, both on Tuesday.

The third week: one migraine, on Wednesday rather than Tuesday. We discussed whether this was meaningful or coincidence. I told her honestly that I did not know yet.

The fourth week: no migraine.

The fifth week: no migraine.

The sixth week: no migraine.

She came in on the Tuesday of the seventh week, smiling for the first time since I had met her.

“I think I have not had a Tuesday afternoon in three years.”

What I told her, and what I tell anyone with a similar story

Erjona’s migraines have not vanished. She had one in the eighth week, mild, lifted with rest and a single dose of sumatriptan. She has had three more in the eleven months since. Three migraines in eleven months, compared with the previous frequency of two to three per month, is a meaningful change. It is not perfection. It is enough that she has her Tuesdays back.

The lesson I take from her case — and I have seen variations of it many times — is that acupuncture rarely fixes migraine on its own. It works best as part of an honest investigation into what the body is reacting to. The needles raise the threshold. The lifestyle changes lower the load. Together, the gap between threshold and load opens up enough that the migraines do not arrive. There is a short earlier piece on a different migraine-relief acupuncture story, if you want a second example.

If you are reading this and considering acupuncture for your own migraines, the suggestion I would make is this: keep a notebook for at least three months before your first session. Write down everything that seems irrelevant. Hours of sleep, what you drank the day before, where you were in your cycle if you have one, the weather, whether you ate breakfast. The patterns will surface. The acupuncturist’s job is then to help you address what the patterns are showing.

The needles are useful. They are not the whole story.

A small footnote about the diaspora

I notice in my practice that the migraine clients who improve the most are often those who have moved between countries — diaspora returnees, expatriates, people whose nervous systems have adapted twice or three times to new climates, new schedules, new languages. The migrating nervous system seems to be more sensitive to small ongoing disruptions. It also seems to respond particularly well to interventions that respect its complexity rather than trying to override it with a single drug.

I do not have data to back this up, only an impression from many client conversations. But it matches something I notice in myself, having moved from Liaoning to Tirana a few years ago: a body that has crossed borders carries a different kind of attention.

The work is to give it the right kind of quiet.


Yang Wang practises acupuncture at Chinese Massage – Tai Chi Tirana. The parlour is in central Tirana, near Bulevardi Gjergj Fishta. Names in client stories have been changed.