The first squeeze of my left thumb is gentle, almost reassuring. I rate it as 0 out of 100 on the pain scale.

But as a technician ramps up pressure on the custom-made thumb-squeezing device, it becomes less pleasant. I give ratings of 2, 6 … then 36. A few squeezes later, I’m at 79.

At 84, I’m glad the test is over as I put my tender thumb to my lips.

I’ve offered myself up for a pain study at the University of Michigan, in a long, low-slung building northeast of the university’s main campus in Ann Arbor. As the day wears on, I’ll undergo needle pokes, leg squeezes and an MRI scan — all part of a grand bid to better understand the root cause of an individual’s pain, and point to the best solutions.

It’s an understanding that’s sorely needed. Lucky for me, I’m just a control in this experiment, and I can cry for mercy whenever I want. That’s not the case for the multitudes of people — 50 million in the US alone — who have ongoing, chronic pain, for whom the medical pause buttons are far from adequate.

A person’s thumb is wedged in a hole in a hand-held device with a wire coming off it. Next to it is a screen showing a pain scale from 0 to 100.

The thumb pressure test, in which participants rate their pain level on a scale from 0 to 100 as their thumbs are subjected to increasing pressure, is one of several ways that clinicians and researchers can evaluate a person’s pain responses. Since people’s thresholds to pain in tests like this vary according to pain syndrome, such tests can help with diagnosis.


“Our treatments for chronic pain are very bad,” says Richard E. Harris, a neuroscientist at the University of Michigan’s Chronic Pain and Fatigue Research Center and a co-researcher on the study, which should ultimately help to improve diagnoses and therapies. Today, doctors mostly define pain by where it is: the abdomen, the lower back, the joints. Then they offer up treatments, usually anti-inflammatories or opioids, that too often do nothing to the cells and molecules causing a person to hurt. A recent analysis in the Journal of the American Medical Association found that opioids reduced pain by an average of less than one point on a 10-point scale, across a variety of chronic conditions.

As part of the precision medicine movement and thanks to modern brain-imaging technology, scientists are starting to puzzle out the different types of pain: what causes them, how to diagnose them and how to prescribe treatments to match. It’s an area that is far from settled. As recently as 2017, the International Association for the Study of Pain defined a new pain type, called nociplastic. It’s characterized by the absence of any nerve or tissue damage in the parts that hurt.

Dan Clauw, director of the Michigan pain center, is passionate about helping people with this kind of long-misunderstood pain, which could underpin chronic conditions, such as fibromyalgia, that afflict millions. His blue eyes flash behind spectacles as he describes crisscrossing the globe to educate other physicians about nociplastic pain. He’s wearing a navy blazer and slacks when we meet for lunch between my testing sessions, because he’s just returned from giving a presentation about marijuana and pain. He jokes that his colleagues won’t recognize him out of his usual jeans.

Imaging the brain, along with doing prodding and poking tests of the type I endured, is beginning to point to signatures that explain the problem and suggest solutions. Eventually, this knowledge will help scientists to develop more targeted therapies, so doctors can treat patients better.

Taxonomy of pain

In broad strokes, pain falls into three categories: nociceptive, neuropathic and nociplastic. (“Noci-” is from the Latin for “to do harm.”)

Nociceptive pain results from inflammation or direct damage to tissues. When that torture device squeezes my thumb, for example, pain-sensing nerves notice the pressure and spring into action. They transmit messages to my spinal cord, which sends them on to my brain, telling me “Ouch!”

This kind of discomfort is often short-lived; mine dissipates after I’ve sucked on my thumb for a few moments. Nociceptive pain can also be chronic, though — for example in osteoarthritis, where the cartilage in joints wears away and causes stretching of tendons and ligaments, or through the ongoing inflammation of rheumatoid arthritis.

Neuropathic pain, in contrast, happens when the pain-sensing nerves themselves are damaged or irritated, so that they send inappropriate “Ow!” signals to the brain. It typically results from some injury or disease, such as diabetes or shingles. It can also happen when a nerve is pinched, as in the case of carpal tunnel syndrome, when a nerve in the wrist gets squeezed. It’s often long-lasting, unless the damage is repaired.

And nociplastic, the newly named type, results from no obvious inflammation or injury. Rather, it’s as if the volume knob for pain is turned up way too high, not at the pain site itself but further afield. Nociplastic pain seems to arise in parts of the central nervous system — the brain or spinal cord — that receive, transmit, or process those “Ouch!” signals. These nerves misfire, creating a sensation of pain even though nothing may be wrong. The location of the problem, the central nervous system, is why Clauw prefers to call it “central sensitization.” The classic example is fibromyalgia, which causes pain that seems to stem from muscles, tendons and joints, despite the real problem’s lying in the brain or spinal cord.

A table shows current thinking that delineates three classes of pain and their features — such as whether the pain is generated in the peripheral or central nervous system.

Scientists’ understanding of pain continues to evolve and so do the various terms used to describe it. Ideally, definitions are standardized and reflect the biology underpinning the pain, but the lack of straightforward tests for parsing types of pain makes defining it a challenge. Nociceptive pain involves pain-sensing nerves called nociceptors, which also can be involved in neuropathic pain. A third pain type is believed to arise wholly in the central nervous system. But there can be overlap: Nociceptive and neuropathic pain can, over time, lead to central nervous system-generated pain.

Complicating the picture, a person might have more than one type of pain going on at the same time. In 2012, the journal Pain published a case report of a person with burning, prickling pain on both sides of the body. Treatment with pregabalin, an epilepsy medication that can also address neuropathic pain and central sensitization, relieved pain on the right side of the body, but not the left.

All this pain classifying is more than an academic exercise: It should help guide how to treat people. For example, consider a patient with knee pain. If the issue is nociceptive, anti-inflammatories or knee surgery should help. But if the problem is central, those treatments probably won’t make much difference. A better bet would be medications that can directly influence the misfiring central nervous system. Some antidepressants, for example, act on the brain’s chemical messengers — neurotransmitters — that are involved in pain, altering their signaling to quell the “Ouch” message.

Nondrug treatments such as acupuncture and cognitive behavioral therapy also may help because they influence how the brain perceives pain. Acupuncture boosts availability of brain receptors that respond to the body’s natural painkillers. A recent analysis in JAMA Internal Medicine of more than 6,000 people taking opioids found that treatments such as meditation, hypnosis and cognitive behavioral therapy reduced pain and diminished the drug doses needed to control it.

Grand Central Station

Though the term “nociplastic” is new, Clifford Woolf, a neurobiologist at Boston Children’s Hospital and Harvard Medical School, first proposed the concept in 1983. Yet the idea has been slow to catch on. In the 1990s, when Clauw began studying fibromyalgia, it was a disease so vague, so puzzling, that some physicians simply denied its existence.

Today, fibromyalgia is more likely to be accepted as a real condition. But many doctors still don’t appreciate how centralized problems might underlie pain even when the symptoms look nociceptive or neuropathic, Clauw says. The distinctions between pain types are not clean: If left untreated, nociceptive pain may sensitize the nervous system, turning a temporary problem into chronic, nociplastic pain, for example. Clauw and his Michigan colleagues believe that central sensitization shows up in myriad conditions, from irritable bowel syndrome to chronic pelvic pain to dry eye disease. And in the study I’ve signed up for, they want to clarify how often this happens and how doctors might detect it in patients who show up begging for pain relief.

To that end, the team has enrolled people with three different pain disorders that seem, on the surface, to be nociceptive or neuropathic. The scientists will test their pain before and after standard treatments. If the pain is in fact central, the treatments shouldn’t work — a disappointment for the participants, but one that might eventually lead to better understanding and treatment for them and others like them.

Two categories of subjects have what looks like nociceptive pain: those with osteoarthritis of the hip, who will receive a hip replacement, and those with inflammatory rheumatoid arthritis, who will be treated with modern medications. A third group, people with carpal tunnel syndrome, represent neuropathic pain and will get surgery to relieve the pressure on the nerve.

But if Clauw and his crew are right, then some of these people will really be suffering from central sensitization, instead of or in addition to the nociceptive or neuropathic problem. Two control groups will help tease that out: People with fibromyalgia will show the researchers what pure central sensitization looks like, and those like me, with no chronic pain, will represent the non-central state.

A graphic depicts the types of situations and settings, from clinical trials to treating comatose patients to courts of law, that would benefit from having clear biological indicators of the various types of pain.

The primary way that physicians measure pain today is to ask someone how much they’re hurting. Identification of biomarkers from, for example, brain imaging or blood tests could provide more objective measures of pain that would offer benefits in a variety of settings.

Once all the data are in, the researchers hope that pain features shared by the people with fibromyalgia and the others whose treatments don’t work will reveal a potential signature for central sensitization.  

The challenge is that there’s no simple blood test or X-ray that will distinguish one type of pain from another. “There’s no single measure that, by itself, will represent pain,” says Woolf, author of a paper in the Annual Review of Neuroscienceabout pain caused by problems in the sensory machinery. “We need a composite.”

Hurts so good

To build that composite, scientists must resort to a variety of indirect measures, including responses to the pokes and prods being inflicted on me and other subjects.

This particular piece of the picture, called quantitative sensory testing or QST, measures the threshold at which a person can feel a given sensation — such as pressure, heat or cold — and when that sensation becomes painful. This can reveal how a person’s nervous system deals with pain, and how that system might be off-kilter. Specific defects in nerves lead to specific changes in pain responses, helping scientists to distinguish one pain type from another.

It’s simple, but revealing. For example, in the case of the thumb-press test, a person with fibromyalgia would probably start to feel pain at around four pounds of pressure. Clauw, who has no chronic pain of any stripe and is relatively pain-insensitive, says that he can handle up to about 18 pounds of pressure before it becomes uncomfortable. The average person would probably start to feel bothered at around eight pounds.

Or take a test where I’m poked in the forearm with a needle. The device retracts into the handle like a Hollywood special-effects knife, so it doesn’t pierce my skin, but it doesn’t feel great — I rate it a 7 out of 100. Then I get 10 pokes in quick succession. That hurts more, at 32. This is a normal response, but if I had central sensitization, I would likely have found the 10-poke series much more painful.

In addition to sorting out nociceptive or neuropathic from centralized pain, QST also seems able to reveal subtypes. In research published in 2017, three European consortia performed QST on 900 people with diverse pain conditions, all considered to be neuropathic. The testing separated the subjects into three clusters, and the study authors predicted that each would be suited to different treatments.

Schematic shows an idealized approach to treating people with therapies that are appropriate for their pain subtype. First, an individual’s pain symptoms are assessed, then people with similar kinds of pain are sorted into clinical trials. Those trials yield treatments that work for a given sub-type of pain, and future patients benefit by receiving treatment for their subtype.

Better-defined markers for different types of pain could radically improve pain management. As shown, it would allow patients to be sorted into clinical trials that would reveal the best treatments for each pain subtype. Results of those trials would help physicians treat individual patients more effectively.

The first cluster was characterized by deficits in sensation to touch, heat, or pokes that would normally be painful. This suggests that central sensitization might be behind the pain in some of these people, says study coauthor Nadine Attal, a pain specialist at the Assistance Publique-Hôpitaux de Paris. Opioids, antiepileptics or antidepressants (used for their effects on pain nerves, not mood) might help, because they act in the brain.

The second group was defined by extreme sensitivity to hot and cold — like skin when it’s sunburned, which puts pain-sensing nerves on high alert. For this kind of neuropathic pain, local, numbing medications such as lidocaine, Botox or capsaicin (a therapeutic substance from hot peppers) might be the right choice.

People in the third group were particularly sensitive to pressure and pinpricks, and its members often reported pain akin to burning or electrical shock. This was a more complex group, Attal says; she thinks topical medications or antiepileptics might help. But now that researchers have the categories better defined, they can directly test medications to find what truly works best for each.

Pictures of pain

Looking at the brain in pain also can help scientists distinguish pain types, although the answers aren’t clear-cut. There’s no one, lone spot where pain lights up the brain, says Sean Mackey, chief of the division of pain medicine at Stanford University in California. Rather, the pain response is distributed across a circuit that encompasses several brain areas.

In the afternoon of my day as a pain-study subject, I’m led to the university’s North Campus for an MRI. The technician slides me into a gray, General Electric-branded, upright donut about the size of a golf cart. The outside is festooned with frolicsome animal stickers (many subjects from other studies are children), but these do nothing to allay the discomfort of lying perfectly still with my head in a vise for an hour and a half.

As I lie there, listening to the scanner’s inharmonious beeps, rumbles and alien-laser-gun sounds, I’m not thinking of anything in particular. Nonetheless, certain parts of my brain tend to draw blood at the same time, suggesting that they’re acting in sync. These are called networks.

MRI scans of the brain from several angles show heightened activity in brain regions that have been associated with pain from fibromyalgia.

Roughly half of people with rheumatoid arthritis experience pain even when using medications that control the inflammation. MRI scans of some of these patients reveal amped up connectivity between two brain regions, the default mode network and insula. This brain connectivity also has been found in people with fibromyalgia, a chronic pain condition with roots in the central nervous system. The discovery suggests that rather than inflammation alone, a dysfunctional central nervous system can also play a role in the pain of rheumatoid arthritis.


One that Harris and colleagues are particularly interested in is called the default mode network. It turns on when I’m at rest and my mind wanders to topics involving myself: what I had for breakfast, perhaps, or what I’m planning for tonight once my day of pain is over.

Another network they’re watching is the salience network, which lights up when a person notices a new sensation — say, the squeezing of their thumb — to determine which sensations are worth responding to. It includes the insula, a pyramid-shaped bit of brain that Mackey and others have linked to pain.

Normally, the insula and the default mode network are unlikely to act at the same time. But Harris and colleagues discovered that in people with fibromyalgia, they were much more likely to flash in synchrony.

That makes sense, says Rob Edwards, a pain psychologist at Harvard Medical School and Brigham and Women’s Hospital in Boston. For someone living with chronic pain, the pain can become a core part of their identity. “The salience-related threat intrudes on, and even takes over, the way that you think about yourself,” he says.

Painkillers, personalized

It may be possible to undo that intrusion, though. Edwards is currently testing cognitive behavioral therapy, or CBT, in people with fibromyalgia. In no way is he suggesting that their pain, or any pain, is imaginary, but therapy can help people deal with pain better and even reduce it. “It’s all about enforcing a sense of control and mastery,” says Bob Kerns, a pain psychologist at Yale University in New Haven, Connecticut, who coauthored a paper in the Annual Review of Clinical Psychology on psychological treatment for chronic pain.

In the study so far, CBT seems to be disentangling the salience and default mode networks in some people with fibromyalgia. Edwards predicts those people will also experience pain relief.

Being able to forecast who will benefit from a given treatment could make a huge difference not just for individual patients, but also in clinical trials for new pain-relief drugs. If scientists test a pain drug on 100 people, but only a fraction of those subjects actually have the pain mechanism the drug can treat, the medicine will look like a flop — even if it’s a superstar for a particular subset of patients. This has “almost certainly” happened in past trials, Woolf says.

Mackey envisions a future in which pain patients can be tested for the underlying problem, perhaps with the same kinds of tests I underwent at the University of Michigan, plus many more assessments. For example, scientists are analyzing nerve endings in small skin samples from pain patients, and others aim to tease out the role of genetics in chronic pain. Simple questionnaires can also help to identify pain types, all with this goal of prescribing medications tailored for a person’s specific flavor of misery.

Medicine isn’t quite there yet — in fact, only 10 years ago Mackey would have called that scenario science fiction. “Stay tuned,” he says, “because it’s no longer science fiction. . . . We’re going to get there.”

As required by the University of Michigan Institutional Review Board, Amber Dance was compensated $275 for her participation in the study at the Chronic Pain and Fatigue Research Center. She donated that amount to the American Chronic Pain Association.