In February 2019, Jane Green enrolled on an experimental clinical trial targeted at autistic individuals with anxiety disorder. Green, who is in her mid-50s, is an autism activist and educator. She describes her autism as something that enables her to be constantly “creative and buzzy” but makes her prone to irritation when things don’t get done. “People call it obsessive,” she says. “I call it passionate. I’m like a very busy duck swimming against a strong current, all the time.”
Green also has Hypermobile Ehlers-Danlos syndrome, a rare genetic condition that degrades collagen, the glue that binds together skin, bones, muscle and the internal organs. Although she has only been diagnosed recently, she has been afflicted with painful joint dislocations and sprains for as long as she can remember. “Doctors used to call me bendy,” she says. One day, she recalls, she was walking across a car park and her ankle just kind of “fell out”. People who suffer from hypermobility also have stretchy and delicate skin and often have immune complications. They are also seven times more likely to be diagnosed as autistic.
On the first day of the trial, which took place at the Brighton and Sussex Medical School, Green was welcomed by a young postdoctoral researcher by the name of Lisa Quadt. The trial was advertised as an “innovative psychological therapy” called ADIE (which stands for Aligning Dimensions of Interoceptive Experience), which the researchers hoped could reduce some forms of anxiety in autistic individuals for whom standard pharmacological treatments had proven ineffective.
Quadt explained that severe anxiety in autistic adults is very common and could be caused by overreactions to sudden physical sensations. These weren’t mental forms of anxiety, like rumination or depression, but rather fully embodied ones, like panic attacks, when a twinge in the chest and an acceleration of heartbeat can trigger the feeling that your heart is about to stop and that you’re about to die. The therapy would hopefully help participants by attuning them to these bodily sensations and, as a result, empowering them to control them.
Green described to Quadt how she would often feel assailed by sensations that would emerge unexpectedly and rapidly overwhelm her. She described a particular pattern as the “woosh”, a feeling akin to being inside a falling elevator. “It just goes haywire. It just goes crazy,” she says. She feels her body being pumped with adrenaline and histamines, leaving her gasping for air and enveloping her skin with a red, itchy rash that feels so hot “you could cook an egg on it”. “Once I had a crisis so bad I ended up in hospital and couldn’t eat for months,” she says. “It made me really anxious because everything is out of control.”
As part of her initial assessment, Green was asked to fill out questionnaires and do a few tests. One test was a heartbeat tracking task. This is a simple test, which involves counting the number of heartbeats in a given interval of time, which varies between 20 to 45 seconds, for a total of six trials. Participants have to try sensing their heart internally, rather than physically feeling for a pulse. They are also connected to a pulse oximeter which records their actual heartbeats.
The task is a test of how good people are at detecting their own heartbeats. Those performing it tend to underestimate the number of beats. Individuals with a slower heart rate, like athletes, are usually more accurate. When Quadt asked how many she had counted, Green had no idea, so she just made a wild guess. “I’m quite competitive and I wanted to get it right, but I just didn’t really understand it,” she says. “I tried to find a pattern but the times kept changing, so I couldn’t, so I just kept guessing.” She had no idea how she was supposed to feel her heart and began to feel irritated. She thought to herself: “Who invented this awful thing?”
Sarah Garfinkel describes herself as an emotional person. Her emotional range is seldom moderate: when she’s happy, she’s really happy; when she’s sad, she’s very sad. For her, even the emotions evoked by mundane occurrences find distinct bodily expressions: she’s the type of person who literally jumps when watching horror movies, who visibly tenses up in pain during awkward social situations. “I’d probably be the worst therapist in the world, because when someone’s having a hard time, I just sit there and weep,” she says. “I’m useless. But that also means I share other people’s joy and pain and that’s a nice thing. It’s what makes us human and what connects us.”
In 2004, Garfinkel embarked on a career as a neuroscientist. Her PhD project, at the University of Sussex, was an investigation on the effects of alcohol on memory. “The hypothesis was that alcohol only affected explicit memory, leaving implicit memories intact,” she says (implicit memories are things you remember unconsciously, like riding a bike, whereas explicit memories are when you consciously try to memorise something, such as reciting a list of words). “On my first experiment, I got everyone so hammered, I basically knocked out all of their memories, implicit and explicit. It was not very profound.”
Four years later, she moved to the University of Michigan to study the emotional processing of fear memories in war veterans from the Iraq and Afghanistan wars who suffered from post-traumatic stress disorder. “They would often have memory flashbacks. If there was a loud noise, they jumped, as if they were under attack,” she says. She was intrigued by the observation that, during these flashbacks, whereas many veterans would start sweating and their hearts would race, others would just go numb, as if disconnected from their bodies. “This wasn’t about the external world. They were clearly carrying it with them internally,” she says. When she made this observation to a fellow researcher, she was told to discard the data related to the non-responders. “I didn’t like that,” she says. “There was something happening in their body that was interacting with their brain and contributing to these different symptoms.”
Suspecting that mainstream neuroscience couldn’t provide an adequate explanation for what she was observing, Garfinkel found herself captivated by a new area of research called interoception. In contrast to exteroception – the collection of senses, from vision to smell, that allows us to scan and palpate the external world – interoception is about the perception of our visceral world. It encompasses the array of biological sensors that permeate our internal organs – the heart, the gut, the lungs – and continuously track the minute variations of temperature, pressure and chemistry within. This stream of biological information constantly flows from the body to the brain, often barely perceptible, seldom caught in the spotlight of consciousness. But when it does emerge, we register its undeniable physicality: the churning gut, the sweaty palms, the galloping heart, the shallow breath. These are the sensorial signatures emotions are made of.
Garfinkel was particularly fascinated with the research of a British psychiatrist called Hugo Critchley. Like many other researchers in the field, Critchley devoted his studies to the heart, an instrument of choice in the study of interoception because of its distinct and rhythmic beat, easy to detect and measure. In 2004, he published one of the most influential reports in the field, a study of how visceral information activated a region in the brain called the insula. Using the heartbeat-tracking task as a gauge for how accurate people were in detecting their bodily sensations – a measure called interoceptive accuracy – Critchley showed that the more accurate someone was at counting their heartbeats, the higher the activation and grey matter volume of the insula.
This finding resonated with Garfinkel, as she too had seen the same sort of hyperactivation of the insula in her PTSD patients. “He was really a pioneer in measuring the signals from the body and the brain and integrating the two together. That just made sense to me,” she says.
Neuroscientist Sarah Garfinkel in her lab office at the Sackler Centre for Consciousness Science
In 2011, Garfinkel managed to find a position under Critchley, at the Sackler Centre for Consciousness Science at the University of Sussex. At the time, Critchley was investigating how the state of the body could influence mental processes. This was informed by a theory of emotion that can be traced back to American psychologist William James, who first proposed that the sensing of bodily changes was emotion itself. According to James, we don’t cry because we are sad, but we are sad because we cry; the heart doesn’t pound because we are afraid, but we are afraid because of the pounding heart. His theory, he wrote in 1884, was “that the bodily changes follow directly the perception of the exciting fact, and that our feeling of the same changes as they occur is the emotion.”
Garfinkel explains: “If we see a snake, our hearts won’t beat faster because we are scared. Seeing the snake will increase our heartbeat and when that’s registered in the brain, that’s what leads to the feeling of fear. He reversed the causality.”
Critchley was particularly interested in the effect that cardiovascular arousal has on the brain. As the heart pulsates, blood is injected into the aorta, extending the arterial wall and stimulating pressure-sensitive sensors called baroreceptors attached to it. These relay information about blood pressure to the brain, activating it in proportion to how strong and fast the heart is beating.
Experiments had already been conducted that showed that if, for instance, you gave someone an electric shock in time with the beat of their heart (as opposed to between heartbeats) they would perceive it as less painful. “Pain is dampened down when the heart and the brain are in active communication,” Critchley says.
He and Garfinkel discovered that this inhibitory effect also affected a cognitive process such as memory. In their study, “What the Heart Forgets”, they showed that when people were presented with a list of words to remember, they tended to forget the words that had been shown in synchrony with the beating of their heart.
In subsequent research they found that, on the other hand, cardiac signals actually boosted the perception of fear. In that study, they flashed pictures of fearful faces to participants, either in synch or in between heartbeats. When they asked people how intense they found the faces, they would systematically judge faces as more fearful if their presentation happened to coincide with their heartbeats.
To Garfinkel, these findings clarified the biological rationale for the influence that cardiovascular arousal seemed to have on the brain. “For instance, when you’re feeling threatened, you need to be super alert to threats in the environment,” she explains. “Having good memory recall or being aware of the pain in your foot isn’t necessarily helpful at that particular time.”
They also illustrated how our experience and perception of the world fluctuates to the tune of our heartbeat. “They brought centre stage the fact that the most important thing for the brain is the body, the vehicle that houses the brain,” says UCL’s Karl Friston, one of the world’s leading neuroscientists. “It is remarkable that, over a timescale of several hundred milliseconds, our physiological state determines in a fundamental way how we experience the world as a sentient creature. You won’t find anything like that in neuroscience in the 20th century.”
As a teenager Garfinkel used to spend her summers helping her mother, who had given up a career in law to work in a nursery in central London, running a crèche for autistic children. “I was very moved by working with these kids: they were bright and kind and quirky,” Garfinkel says. “I think there’s also honesty to autistic individuals that I really respect.” The experience left her with the indelible idea that the characterisation of autistic people as individuals who lacked empathy and had little need for social interaction was a fundamental misunderstanding.
As a researcher, she began working with autistic people in 2013 at the Neurobehavioural Clinic in Brighton. One of the first things she noticed was that many people she worked with seemed to struggle with sensory overload, an overwhelming sensation that the external world was too intense, that lights are too bright, noises too loud. They also seemed to have problems identifying their own emotions. Often they would forget to eat because they didn’t know if they were hungry. “They might acknowledge that they don’t feel right, but they don’t understand if they feel angry or sad or anything at all. They can’t work it out,” Garfinkel says.
If autistic people were indeed swamped or perplexed by their own bodily sensations, that had obvious implications for their ability to understand the emotions of others. “You can imagine that gets exhausting and quite overwhelming,” she says. “I think that does make you vulnerable to having emotions that might not fit the situation or that might not make sense to other people or yourself.”
Loneliness, for instance, is four times higher in autistic individuals. “We asked them how they feel about it and it’s something that really upsets them,” Garfinkel says. “They actually long for social connection, but they just don’t know how, or have experienced so much rejection that they don’t try anymore.”
Empathy is deeply rooted in interoception: the ability we have to detect our own emotions determines our capacity to sense the emotions of others. In response, our bodies often mirror other people’s bodily changes, reenacting those same emotions on a visceral level. In a study conducted at Hugo Critchley’s lab, researchers showed that when looking at photos of sad faces, the viewer’s pupils would shrink in response. A study of fire-walking rituals in Spain found that the variations of the heart-rate of the person running across hot coals were closely mirrored by their partner, who was watching from the audience.
In 2016, Garfinkel came across a study that tested the empathy response of autistic individuals to other people’s pain. The participants were presented with photographs of painful situations – for example, a hand about to be stabbed – and asked to judge if the person depicted was in pain or not, while their brain activity and skin conductance was measured. The brain scans showed that autistic individuals had reduced response when compared with the neurotypical group, leading the authors to conclude that they presented an empathy deficit.
However, Garfinkel noticed that the graph depicting their skin conductance response was actually orders of magnitude higher. “If you read the paper, it talks about the brain differences but it glosses over the body differences,” Garfinkel says. And their body responses told a completely different story.
Garfinkel wondered what was different about the autistic individuals’ interoceptive abilities that affected their interpretation of bodily sensations. That year, with Critchley, she ran a large-scale study with 80 individuals. Each participant was given two different sets of interoceptive tests. One included the heartbeat-tracking task that tested the participants’ interoceptive accuracy. The second was a self-report questionnaire about their awareness of bodily sensations, covering everything from stomach pains to urge to defecate. This was a subjective measure that Garfinkel called interoceptive sensibility. Traditionally, researchers used these terms interchangeably, assuming they were synonyms. “It was a mess,” Garfinkel says. “They were calling everything ‘interoceptive awareness’ regardless of whether they were using a questionnaire or a brain scan.”
When she compared the results, she found there was no correspondence between how good individuals were at assessing their interoceptive abilities and how good they thought they were. In other words, being an accurate heartbeat detector did not necessarily mean that an individual had any insight into their ability. “This gave us a systematic way of thinking about meta-cognitive awareness in a very specific way, rather than a general arm-wavy way,” Critchley says. “It was about how much insight you have in your ability.”
More importantly, this also gave the researchers a novel suggestion as to why autistic individuals may struggle to process emotions. In a subsequent study, Garfinkel found that while autistic adults generally had poor interoceptive accuracy when compared to a neurotypical group, they tended to rate their own abilities highly. “There was this mismatch between how good they thought they were and how good they actually are,” Garfinkel says. This discrepancy, she found, caused problems like anxiety. “If you think you have very good interoception, but actually your body is sending signals that you’re not able to correctly identify, then that’s associated with high anxiety.” This led her to conclude that if she could somehow reduce this discrepancy, she might be able to treat their anxiety. The question was how.
When Garfinkel first tried the heartbeat-counting task, soon after she joined Critchley’s lab, she assumed she would perform well. “Turns out I was terrible,” she recalls. “It was actually hard. It raised intriguing questions about why I think I’d be better. Just because I’m quite emotional doesn’t mean I have precision into what my body’s doing.”
Since then, however, she noticed something else. Every time she performed the task, be it to test the equipment or as simple curiosity, she kept getting better at it. “People sometimes don’t know what to focus on. When I started I kept trying to focus on my chest and actually people don’t necessarily feel their heartbeat in their chest. They can feel it in other parts of their body. Once you’re aware of that, then you can start focusing on other parts and maybe then you’ll be more sensitive to a signal,” she says.
This told her that while many who studied interoceptive accuracy seemed to consider it as a fixed, immutable trait, there was a chance that it could actually be a flexible, trainable one. And if she could train autistic individuals to have better precision and understanding of their body, then maybe she could also reduce their anxiety.
The ADIE clinical trial was launched in 2017. It recruited more than 120 people, and each completed eight training sessions over several weeks. Participants were required to perform not only the heartbeat-counting task, but another one called the heartbeat discrimination task, during which they were played a series of beeps and had to say whether they were in synchrony with their heartbeat or not. To assist in the training, after each test Lisa Quadt would give feedback on their performance. They would also be asked to do some mild exercise before the sessions, to make their heartbeat more perceptible. The researchers encouraged participants to find out what worked for them. “We would tell them to try to say to themselves that this was just their heart and that there was nothing to worry about,” Quadt says. “It was important that they stopped catastrophising.”
After only four sessions, Quadt noticed that although the results were encouraging, the participants still seemed unaware of the progress they were making. “They would say things like ‘This is not doing anything for me’ or ‘This is stupid, I don’t know why I’m here.’ One person got more anxious. I don’t think they realise how much they’ve improved,” says Quadt.
Jane Green was no exception. “Even her first assessment was good,” Quadt recalls. But Green didn’t believe her. “I had no idea what I was doing,” she says. “I was just guessing pretty much towards the middle of it.” After a while, however, she realised that she could sense her heartbeat under her chin. Near the end of the trial, she could sense her heartbeat at will.
What difference this made to her was revealed one day, during a work meeting. “I had two people that began being quite aggressive. I didn’t know them. They interrupted the meeting and abused me verbally,” she recalls. She began feeling the “woosh”, but told herself “I know what’s happening. My heart’s going to go crazy.” She visualised it like a fountain of stuff being ejected and her squashing it down. “At the end, I got a bit itchy, but nothing major,” she says. “That made me see that this was like an armour to get through life.”
According to preliminary data from the study, interoceptive training resulted in marked reduction in anxiety symptoms. This is corroborated by some of the feedback from the participants at the end of the trial: “As the inner channel gets clearer, the outer channel gets more quiet.” “When I notice the impacts of anxiety on my body I am more aware of them and am able to reassure myself that it is just a physical reaction. I am better at taking deep breaths and trying to slow down my breathing and heart rate, rather than letting it spiral.” “I believe it has increased my awareness of hunger and as a result I remember to eat/drink/go to the toilet.”
Garfinkel considers these answers a unique insight into the psychology of interoception, the fine balance between what the body is doing and what the brain is aware of. “I feel like our senses are like a seesaw between the outside world and the inside world, trying to find an equilibrium” she says.
The findings of the ADIE trial also open the possibility for a novel approach to psychiatric conditions. Following her research in autism, Garfinkel has extended her studies to include conditions such as schizophrenia and PTSD. “Schizophrenia fascinates me because it’s associated with a very regular beating heart,” she says. “What’s wonderful about the heart is that it doesn’t beat regularly. It is associated with patterned responses that can be elicited by different scenarios, and actually, that’s a healthy thing. With schizophrenia, it’s like the heart is disconnected from everything.”
She points out that, after researchers observed that patients with mental health disorders who also took statins made fewer hospital visits, an ongoing clinical trial was launched, looking at the effects of blood pressure medication for people with post-traumatic stress disorder. “We’ve been bombarding the brain with drugs for depression and schizophrenia, and we haven’t seen much progress in 20 years,” she says. “I personally hope that the future of mental health conditions will be these body-based, peripheral physiology-based interventions.” It is time, she hopes, we start listening to our hearts.
More great stories from WIRED
💿 Meet the Excel warriors saving the world from spreadsheet disaster
🚀 Inside Gravity’s daring mission to make jetpacks a reality
🧥 Don’t get cold this winter. How to the get the most out of your home heating
🔊 Listen to The WIRED Podcast, the week in science, technology and culture, delivered every Friday
👉 Follow WIRED on Twitter, Instagram, Facebook and LinkedIn