Today's article comes from the Frontiers in Neurology journal. The authors are Koreki et al., from University College London, in the UK. In this paper, they explore whether interoception might explain why some people develop chronic daily migraines while others get them more sporadically.
DOI: 10.3389/fneur.2025.1643260
I've had migraines since I was in high school. I'd tell doctors about them, and they'd barely react. After years of this, one finally asked me: "Hasn't anyone ever sent you to a neurologist?" No, they hadn't. So I went, and it turns out, there are treatments. And today my symptoms are under control. So yeah, my neurologist saved the day. But even he couldn't tell me why these things were happening in the first place. What was causing my migraines, or why they happened as frequently as they did.
That was 8 years ago. And ever since, I've continued to wonder: what makes some of us get migraines, and why does it happen to some people every day, and others only sometimes?
In this paper, researchers think they might have found a clue in your brain's ability to sense what's happening inside your body. On today's episode, we'll walk through how they set up their study, what they measured, how they measured it, and what their results mean. Let's dive in.
Your body is constantly sending signals to your brain. Your heart is beating, your lungs are expanding, your stomach might be rumbling, blood is flowing through your veins. But most of the time, you're not consciously aware of any of this, your brain is just monitoring these signals in the background. Your brain's ability to perceive these internal bodily signals (whether you're aware of them or not) is called interoception. And this is what allows you to know when you're hungry, when you're thirsty, when you need to use the bathroom, or when your heart is racing.
There are three different aspects to interoception that researchers study.
The 'prediction error' concept is crucial to understanding today's research, so let's unpack it. Your brain is a prediction machine, it's constantly generating expectations about incoming sensory information based on past experience. When you walk up stairs, your brain predicts that your heart rate will increase and you'll feel slightly out of breath. When these predictions match the actual signals from your body, everything feels normal. But when there's a mismatch between prediction and reality, you get a prediction error signal that helps your brain update its models. There are two types of interoceptive prediction errors that we can measure.
Now, where do migraines fit into all this? A migraine isn't just a headache. It's a complex neurological condition affecting the trigeminovascular system. During an attack, people often experience interoceptive phenomena: nausea, sensitivity to light and sound, everything just feeling wrong. And before the headache even hits many people experience premonitory symptoms like fatigue, mood changes, or food cravings. These are all signs that the brain's internal monitoring system is acting up. The brain structures involved in migraine pathophysiology overlap significantly with interoceptive networks. The brainstem, insula, and default mode network all play crucial roles in both migraine generation and interoceptive processing. The insula, in particular, is considered the primary interoceptive cortex, integrating signals from throughout the body to create your sense of internal bodily state. Dysfunction in these networks could theoretically affect both pain processing and bodily awareness.
But here's where it gets interesting. While we know a lot about episodic migraines that come and go, we understand much less about chronic migraine. Some people have occasional migraine attacks, others develop chronic daily symptoms that completely disrupt their lives. What makes the difference? The authors of this paper suspected that interoception might hold some clues.
They also noticed something else. Many people with chronic migraines also experience dissociation. Dissociation is when you feel disconnected from your thoughts, emotions, surroundings, or even your own body. It comes in two main flavors: psychoform dissociation, where you feel detached from your mental processes like thoughts and emotions, and somatoform dissociation, where you feel disconnected from your physical body and bodily sensations. The overlap between chronic migraines and dissociation made the researchers wonder if there might be a common underlying mechanism involving the brain's construction of bodily self-representation.
So they designed a study to test whether people with migraines have different interoceptive abilities compared to healthy controls, and whether these differences might help explain why some migraines become chronic.
They recruited 49 participants. Twenty-three were healthy controls without neurological disorders. Twenty-six had migraines. Fifteen of those had chronic migraines (15 or more days per month), 11 had episodic migraines (fewer than 15 days per month). Each person underwent an assessment in a quiet room with no visible clocks, and they wore a finger pulse oximeter connected to a laptop. The first test was a heartbeat tracking task. Participants sat quietly and were asked to count their own heartbeats during six different time windows ranging from 25 to 50 seconds. They had to do this without touching their body or looking at a clock. Meanwhile, researchers measured their actual heart rate using the oximeter. Interoceptive accuracy was calculated as 1 minus the absolute difference between actual and reported heartbeats, divided by the average of actual and reported heartbeats. These scores were then averaged across all trials to create a mean accuracy score. As a baseline, they also completed a time tracking task using the same procedure, where they estimated the duration of the time windows rather than counting heartbeats. This helped distinguish true interoceptive ability from general time estimation skills. In other words they wanted to see if you were good/bad at counting your heartbeat irrespective of whether you were good/bad at counting time in general.
The second test was a heartbeat discrimination task that had 20 trials. In each trial, participants listened to ten tones at 440 Hz and had to judge whether the tones were synchronized with their heartbeat or not. The tones were either triggered by the onset of the finger pulse waveform, occurring about 250 milliseconds after the R-wave for the synchronous condition, or presented at a fixed delay of 550 milliseconds after the R-wave for the asynchronous condition. Then, after each trial, participants rated their confidence in their judgment using a visual scale ranging from "total guess" to "complete confidence." Interoceptive accuracy was calculated as the ratio of correct to incorrect judgments.
The authors also measured interoceptive sensibility using the Porges Body Perception Questionnaire Awareness Scale, which asks people to rate their awareness of various bodily sensations like heartbeat, breathing, muscle tension, and temperature changes. This captures people's subjective beliefs about their interoceptive abilities.
From these measures, they calculated the two types of prediction errors I mentioned earlier.
So, at the end of the day, what did they find?
The results were consistent with their hypotheses. People with migraines showed significantly poorer interoceptive accuracy on the heartbeat tracking task (the first test) compared to healthy controls. The median accuracy score was 0.50 with an interquartile range of 0.43 for the migraine group versus 0.78 with an interquartile range of 0.26 for controls. Importantly, there was no difference between groups on the time tracking control task, suggesting that the deficit was specific to interoceptive rather than general temporal processing.
Interestingly though, performance on the heartbeat discrimination task (the 2nd test) didn't differ significantly between groups, though there was a trend in the same direction. This suggests that the heartbeat tracking task, which requires sustained attention to internal signals over longer periods, might be more sensitive to interoceptive deficits in migraine than the briefer discrimination task.
But here's the twist: despite being worse at actually sensing their heartbeats, people with migraines reported much higher interoceptive sensibility. The median sensibility score was 110 with an interquartile range of 52 for people with migraines versus just 39 with an interquartile range of 14 for controls. So in other words, they had poor objective performance but high subjective confidence in their interoceptive abilities. This created large interoceptive trait prediction errors for both of the tests.
To ensure these differences weren't just due to other factors like anxiety or depression, the researchers ran statistical models controlling for these variables. Even after accounting for differences in anxiety, depression, and dissociation scores between groups, the differences in interoceptive trait prediction error remained highly significant. The researchers also found higher levels of both psychoform and somatoform dissociation in people with migraines, along with increased anxiety and depression.
But the most intriguing findings emerged when they compared people with chronic versus episodic migraines within the patient group. While there were no significant differences between chronic and episodic patients in basic interoceptive accuracy, sensibility, or trait prediction errors, people with chronic migraines showed something unique: significantly greater interoceptive state prediction error. And this difference remained significant even after controlling for age and somatoform dissociation scores in statistical models.
So why might these findings matter? The researchers propose that these interoceptive abnormalities might help explain how episodic migraines transform into chronic ones, and they frame their interpretation within predictive coding theory.
Predictive coding is a framework for understanding how the brain processes sensory information. Rather than passively receiving signals from the world and body, your brain actively generates predictions about incoming sensory data based on prior experience. When actual signals match predictions, you get a sense of normal experience. When there are mismatches, prediction error signals help update the brain's models. Applied to interoception, this means your brain is constantly predicting what internal bodily signals should feel like. In healthy individuals, these predictions are generally well-calibrated, leading to accurate bodily awareness and appropriate responses to internal states. But what happens when this system becomes miscalibrated? The researchers suggest that in people with migraines, there might be persistently imprecise interoceptive prediction signals. The trait prediction errors they observed might reflect a chronic mismatch between subjective beliefs about interoceptive ability and the actual precision of bottom-up interoceptive signals.
This could arise because migraine attacks themselves are intensely interoceptive experiences involving pain, nausea, and various bodily sensations. Repeated exposure to these attacks might lead people to develop heightened attention to bodily signals and increased confidence in their ability to detect them. However, the actual sensitivity of their interoceptive networks might not increase correspondingly, perhaps due to structural or functional changes in interoceptive brain regions documented in other studies.
These findings might also connect to medication overuse, which is common in chronic migraine and can perpetuate the condition. Frequent analgesic use might dampen bottom-up interoceptive signaling pathways, reducing opportunities for the brain to recalibrate its predictions based on actual bodily states. This could lead to over-reliance on maladaptive top-down predictions of pain, creating a cycle where medication withdrawal leads to heightened prediction errors and increased medication use.
If you want to dive deeper into predictive coding, or explore the statistical analyses, or see the correlation matrices between body awareness measures and clinical variables, I'd highly recommend downloading the full paper. The authors also include the questionnaire items they used to assess dissociation and body awareness confidence.