Heartbeat Evoked Potential: The Brain’s Interoception Signal

The heartbeat evoked potential is an EEG signal time-locked to the R-peak of each heartbeat that reveals how attentively your brain is processing the body it lives inside. A larger HEP means the cortex is registering each heartbeat as a meaningful signal. An attenuated HEP means the brain has deprioritized internal body input, the measurable neural signature of being cognitively present while somatically absent.

This article is part of our hub on self-awareness and interoception, where the brain’s reading of the body is examined.

What does the heartbeat evoked potential measure?

The heartbeat evoked potential measures the cortical electrical response your brain generates to each individual heartbeat. Researchers extract it by averaging EEG activity time-locked to the R-peak of the electrocardiogram, with a typical analysis window of 200–600 ms after the R-peak. Larger HEP amplitudes track stronger cortical processing of cardiac afferent signals.

The signal sits inside a much larger EEG record. Each heartbeat sends afferent volleys through the vagus and the spinal viscerosensory pathways into the brainstem, then up to insular and cingulate cortex, where they are integrated with everything else the brain is doing. The HEP is the cortical fingerprint of that integration, small, several microvolts, statistically detectable only after averaging across many heartbeats. A 2021 meta-analysis by Coll and colleagues, synthesizing 45 studies, established that HEP amplitude is reliably modulated by attention to the heartbeat, by physiological arousal, and by clinical status.

Where the signal comes from inside the brain is the part that matters most for what it means. Park, Bernasconi, Salomon, Tallon-Baudry, Spinelli, and colleagues (2017), using direct intracranial electrode recordings, localized HEP generators primarily to the insula and the operculum. Smaller contributions came from the amygdala and fronto-temporal cortex. The same regions that build emotional awareness are the ones generating the HEP.

In my practice, I consistently observe that the burnt-out executive in their early fifties does not notice their own heartbeat under any normal condition. The cortex has stopped flagging the signal as worth attending to, not as a deficit, but as an adaptation to years of training attention exclusively outward.

“The HEP is a fingerprint of attention. The brain that produces a small one is not broken. It is working exactly as it has been trained to work.”

The 2023 Engelen, Solcà, and Tallon-Baudry review in Nature Neuroscience placed the HEP inside the broader story of interoceptive rhythms (cardiac, respiratory, and gastric) entraining cortical activity throughout the brain. The HEP is the EEG signature of cardiac integration. The same architecture handles breath and gut.

Can you train yourself to feel your heartbeat?

You can train yourself to feel your heartbeat, and doing so produces measurable changes in the cortical signal that registers it. Interoceptive accuracy training, sustained directed attention to internal cardiac sensations, increases HEP amplitude over fronto-central electrodes and recruits anterior insular cortex activity. The signal moves because the brain reweights what counts.

The classic behavioral measure is the heartbeat-counting task. Garfinkel and colleagues, in 2014, distinguished three dimensions of interoception: accuracy (objective performance on the counting task), sensibility (self-reported attention to body signals), and awareness (the metacognitive correspondence between the two). Their three-dimensional model is what most current research builds on. Attention training improves all three dimensions, with HEP amplitude tracking the cortical side of the change.

Recent work has tightened the picture. Zaccaro, della Penna, Mussini, Parrotta, Perrucci, and colleagues (2024), in iScience, showed that HEP amplitude rises specifically during cardiac interoceptive attention tasks over fronto-central electrodes, with the increase concentrated on heartbeats that occur during exhalation. Respiratory phase modulates the cortical processing of cardiac signals. This is mechanistically useful: a slow exhale is the moment the cortex is most receptive to the heartbeat it has been ignoring.

In my practice, the young professional in her early thirties who has spent ten years training her attention to deliverables can usually find her heartbeat within the first session. The condition is narrow: only when I direct her to it explicitly, only with her hand on her chest, only with her eyes closed. The capacity returns quickly. The default does not. The cortex has to be retrained to flag the signal at baseline, not just on demand.

It connects to the wider study of stress resilience and regulation that frames how the nervous system stays steady.

The training works because the architecture is plastic. Repeated, directed attention reorganizes which inputs the cortex weights. The HEP is the cortical record of that reweighting, visible in milliseconds.

How does interoception affect emotional awareness?

Interoception affects emotional awareness because emotions are, mechanistically, the brain’s interpretation of bodily signals filtered through prediction. The brain runs a generative model of what the body should be doing and updates the model when actual signals deviate. The anterior insular cortex is where prediction meets afferent reality, and when that comparator is quiet, emotional experience flattens.

This is the core of the Seth and Friston interoceptive inference framework, developed in 2016 in Philosophical Transactions of the Royal Society B, which extended Seth’s earlier 2013 work. Their account places emotion downstream of active interoceptive inference: subjective feelings arise from the brain’s predictive models of bodily afferents, with the anterior insula serving as the locus where predictions are compared against incoming signals. When the prediction-error signal is precise, emotions register clearly. When precision drops, the signal does not reach awareness.

The HEP is one of the most direct cortical readouts of this precision. Petzschner and colleagues, in 2018, demonstrated that pure attentional focus modulates HEP amplitude in the 524–620 ms window, interoceptive attention raises the signal, exteroceptive attention lowers it. The same person, the same heartbeat, different cortical response depending on which way the attention is pointing. Al and colleagues, in 2020, in PNAS, extended this by showing that pre-stimulus HEP amplitude predicts later somatosensory detection bias, when the cortex is attending inward, it detects internal events more sharply and external events less sharply.

In my practice, I work with composite cases where a partner managing a complex family system has lost emotional granularity not because anything is broken, but because their attention has been continuously externalized for years. The interoceptive precision drops. The emotions do not arrive in differentiated form. They arrive as a generalized arousal that reads, from the inside, as nothing in particular, until it reads as too much. The HEP literature predicts exactly this pattern.

Why can’t some people feel their heartbeat?

Some people cannot feel their heartbeat because the cortical processing of cardiac afferents has been systematically deprioritized, the brain is not ignoring the signal at the periphery, it is ignoring it at the cortex. Clinical and subclinical populations who report somatic disconnection show attenuated HEP amplitudes. The signal arrives. The brain stops flagging it.

The clearest evidence comes from research where the architecture is most visible. Müller, Schulz, Andermann, Gäbel, Gescher, and colleagues (2015), in JAMA Psychiatry, recorded HEPs in individuals carrying the borderline personality pattern and demonstrated significantly reduced HEP amplitudes compared with healthy controls. Critically, the HEP amplitude reductions correlated with gray matter volume in the left anterior insula and bilateral dorsal anterior cingulate cortex. The structural and functional findings converged on the same regions Park et al. localized the signal to.

The resting brain’s role in self-awareness is explored in how the default mode network builds self-awareness.

Flasbeck and colleagues, in 2020, replicated and extended the finding: HEP amplitudes correlated with alexithymia scores over frontal electrodes. The same architectural pattern shows up in restrictive-eating presentations. Cambi, Solcà, Micali, and Berchio, in 2023, found altered HEP cortical representation in anorexia nervosa during ordinary resting states, right-sided hypoactivation in interoceptive regions, particularly anterior cingulate and orbitofrontal cortex. Across these clinical pictures, the pattern is the same: the cortex is muted to its own body.

The non-clinical version of the same architecture is more common than the clinical one. Todd, Cardellicchio, Swami, Cardini, and Aspell, in 2021, in Cortex, showed that weaker implicit interoception, indexed by lower HEP amplitude, tracks more negative body image in non-clinical samples. The pattern shows up outside research samples too. It requires sustained attentional habits that train the cortex to weight external input over internal, nothing more specific than that.

In my practice, I see the same architecture in someone managing a complex household, a board commitment, and a family system with multiple dependents. They are functioning at a high level. They cannot tell me what they are feeling. The interoceptive signal has gone quiet under chronic load, and the cortex has reorganized around the absence.

What does an attenuated HEP look like in everyday life?

An attenuated HEP looks, in everyday life, like the cognitively present but somatically absent pattern. Executive function, task performance, and articulation remain fully intact. What is missing is the felt sense of where the body is, or what emotional weight is accumulating, until the weight forces itself into awareness as a symptom.

The applied bridge is direct. Müller and colleagues’ 2015 finding that HEP amplitudes negatively correlate with self-reported emotional dysregulation, and partially recover with symptom remission, demonstrates the signal is not just a marker of static deficit, it tracks state changes in the architecture as the architecture moves. Kim, Joss, Marin, Anzolin, Gawande, and colleagues, in 2024, designed an interoceptive compassion-based training protocol with HEP amplitude as the primary mechanistic outcome for individuals presenting depressive symptoms and attenuated interoceptive awareness. The signal is treatable as a movable index of recovery, not a fixed trait.

What this means for an individual presenting attenuated HEP is that the architecture is plastic. Real-Time Neuroplasticity™ in this context refers specifically to insular cortex use-dependent reorganization driven by sustained interoceptive attention. The cortex reweights what it processes based on what attention selects. Live, repeated, directed attention to cardiac and bodily sensations under ordinary conditions, not as a separate practice, but woven into the moments where the architecture has been bypassing the body, drives the reweighting that the HEP measures.

In my practice, the burnt-out executive who cannot identify what they feel until the body interrupts them with a symptom is presenting the everyday correlate of attenuated cortical processing of internal signal. The intervention is not to do more thinking about feelings. It is to retrain the cortex to weight interoceptive input again, the precise mechanism the HEP literature has been mapping in research populations for two decades. The signal is a window into the architecture. The architecture is the thing that moves.

The capacity to name feelings precisely is detailed in building emotional granularity in the brain.

What the First Conversation Looks Like

The first conversation is unhurried. You describe what has been carrying you. The moments where colleagues read you as steady while something else was happening underneath. The symptoms that arrive without the warning that should have preceded them. The feeling of being everywhere except inside the body that is doing the work. I listen for the structural pattern beneath the description: which signal has been muted, which compensation has been carrying the weight, where the live edge sits where the architecture is most movable. I work as a neuroscientist, not as anything that has come before. By the end of the first hour, you typically know whether the pattern in your brain is what we both think it is, and what the first thirty days of working together would look like in practice. There is no homework. There is the work itself.

Frequently Asked Questions