A new emerging study in Psychophysiology reports that resonance paced breathing—slow breathing guided at about six breaths per minute—rapidly increases functional connectivity within the brain’s central autonomic network (CAN), a set of regions coordinating arousal, interoception, emotion, and cardiovascular control. That 0.1 Hz rhythm is also the standard target used in HRV biofeedback, often called the resonant pace, because it tends to maximize the size and regularity of heart rate oscillations.

It’s easy to underestimate the breath because we do it all day without thinking. But breathing is one of the rare biological “hybrids”: it runs automatically, yet it can be steered deliberately. When we slow breathing to around 0.1 Hz, we tend to engage the arterial baroreflex resonance—a rhythmic loop between heart, blood vessels, and brain that helps stabilize blood pressure and tune autonomic balance. Clinically, resonance paced breathing has been associated with meaningful reductions in stress, anxiety, and depressive symptoms, alongside cardiovascular benefits such as increased heart rate variability (HRV) and improved baroreflex sensitivity.

This is where biofeedback and neurofeedback become especially interesting. Biofeedback uses real-time physiological signals (like HRV, respiration, or electrodermal activity) to help people learn self-regulation through feedback. Neurofeedback applies the same learning logic to brain signals (most often EEG), shaping patterns of neural activity through reinforcement. Both approaches share a common aim: turning unconscious regulation into learnable skill.

What’s been missing, until recently, is a clear picture of what the brain is doing in the moment during resonance breathing. If the heart starts oscillating more strongly at 0.1 Hz, does the CAN—our body–brain “air traffic control” system for arousal—start coordinating differently as well? This study suggests the answer is yes, and that the insula may be a key junction where bodily signals are integrated with attention, emotion, and cognitive control.

Methods

This investigation pooled neuroimaging and self-report data from four separate fMRI studies, all run at the same imaging center with the same scanner and acquisition parameters. The shared aim across these parent studies was to examine how breathing paced at six breaths per minute influences brain function.

Participants

A total of 177 individuals were recruited (primarily young adults), including people with depression, substance use disorder, or no diagnosis. After stringent motion-quality thresholds were applied, 147 participants remained for analysis. The final sample was 55% women, with a mean age of roughly 22 years.

Resonance paced breathing task

Resonance paced breathing (RPB) was standardized to six breaths per minute (0.1 Hz) for all participants. Before scanning, participants received about three minutes of training using a visual pacer (E-Z Air, Thought Technology). The pacing pattern was symmetrical: five seconds inhalation, five seconds exhalation, with no pause. Participants were coached to breathe slowly but comfortably and not too deeply to reduce risk of hyperventilation.

In the scanner, participants completed two brief breathing conditions:

• Natural breathing: a typical resting-state setup with a fixation cross for six minutes.

• RPB: five minutes of paced breathing with the visual pacer.

To enable a direct comparison of connectivity between conditions, analyses used the first five minutes (150 volumes) from the natural breathing run to match the five-minute paced-breathing run.

Physiological monitoring included ECG and a respiration belt (and, in some studies, a finger pulse sensor). Task compliance was verified in real time and again during post-processing. The paced breathing manipulation produced the expected increase in low-frequency HRV power and total HRV power, supporting baroreflex engagement.

Neuroimaging and connectivity analysis

BOLD fMRI data were acquired on a 3T Siemens Trio system. Preprocessing used AFNI with motion correction, scrubbing of outlier time points, nuisance regression (including CSF/white matter components), and bandpass filtering (0.01–0.15 Hz) to focus on low-frequency fluctuations and allow for individual variability around resonance.

The team used a region-of-interest approach focused on 13 CAN nodes defined a priori from prior autonomic neuroimaging work, including cortical (e.g., insula, cingulate, ventromedial prefrontal cortex), subcortical (e.g., thalamus, amygdala), and brainstem entry-point (nucleus of the solitary tract) regions. For each participant and each breathing condition, they computed pairwise correlations among ROI time series (78 ROI pairs), transformed correlations to Fisher’s z, and tested condition differences using linear mixed models with adaptive false discovery rate correction.

Results

Compared to natural breathing, resonance paced breathing increased functional connectivity in 15 ROI pairs within the central autonomic network after correction for multiple comparisons.

The main pattern: insula-centered strengthening

The standout finding was how often the insula appeared in the “increased connectivity” list: 10 of the 15 significant ROI pairs involved an insular subdivision. This mattered because the insula is a primary viscerosensory and interoceptive hub—often described as a region that helps transform bodily signals into subjective feeling states and action tendencies.

Several strengthened connections linked the thalamus to insular regions. For example, thalamus connectivity increased with the left anterior insula and right anterior insula, suggesting enhanced relay and integration of afferent bodily information during paced breathing. There were also increases between anterior insula and posterior insula, consistent with more coordinated transfer along the interoceptive hierarchy.

Connectivity bridges to attention, emotion, and control

RPB increased connectivity among nodes that sit at crossroads between autonomic regulation and higher-order processing:

• Thalamus showed increased connectivity with multiple regions, including the mid-cingulate cortex, right angular gyrus, right amygdala, and right fronto-insular cortex.

• Right angular gyrus connectivity increased with bilateral anterior insula, left posterior insula, and mid-cingulate cortex—an interesting signature because the angular gyrus is often implicated in integrating multisensory information and shifting attention.

• Right amygdala connectivity increased with the left anterior insula and mid-cingulate cortex, aligning with the idea that paced breathing may influence the coupling of salience/interoceptive processing with emotional arousal circuitry.

• Additional increases involved right anterior insula coupling with a medial prefrontal/cingulate node, and coupling between ventral posterior cingulate and right fronto-insular cortex.

Quality checks and moderators

Because paced breathing can increase head motion, the authors evaluated motion–connectivity correlations for the 15 significant ROI pairs. Correlations were generally small, and only one pair showed a statistically notable difference between conditions.

Exploratory moderation analyses indicated the connectivity increases did not systematically vary with age, sex, perceived stress, affect, anxiety, or depression symptoms once correction procedures were applied.

Discussion

This study offers a helpful piece of the puzzle for anyone interested in arousal regulation: resonance paced breathing didn’t just shift peripheral physiology—it was accompanied by measurable, real-time reorganization of connectivity within the brain’s central autonomic network.

One practical way to think about the CAN is as the nervous system’s “coordination committee.” It keeps tabs on internal state (heart, breath, blood pressure), evaluates what the environment demands, and helps decide whether we should mobilize, freeze, focus, or recover. The present results suggest that when breathing is paced at the baroreflex resonance frequency, this committee starts communicating more tightly—particularly through insular pathways.

The insula-heavy pattern fits a coherent physiological story. Resonance breathing increases the amplitude of HR oscillations by synchronizing respiratory sinus arrhythmia with the baroreflex cycle. Bigger, cleaner oscillations mean stronger rhythmic afferent signaling to the brainstem and onward to thalamic and cortical hubs. When the insula shows increased coupling with posterior insula, thalamus, angular gyrus, cingulate, and amygdala, it hints at a system that is not merely “calming down,” but actively integrating bodily rhythms with attention, salience detection, and executive control.

Clinically, this matters because many common struggles—anxiety surges, craving spikes, insomnia spirals, irritability, rumination—share a feature: the body’s arousal machinery becomes sticky. People often describe it as “I know I’m safe, but my body doesn’t agree.” The value of resonance breathing is that it gives the body a rhythmic input that can be applied in the moment, without needing perfect insight or lengthy preparation.

The authors also raise an important precision-medicine question: not everyone responds equally to the same intervention. Their prior modeling work (and broader clinical experience across interventions) suggests a meaningful minority may show limited physiological shifts during resonance breathing. If acute CAN connectivity changes can be shown to predict longer-term benefit from HRV biofeedback or resonance breathing practice, that would be a major step toward matching the right tool to the right person.

There are also several caveats worth keeping in view. The data came from four different studies, and the paced breathing condition followed other task contexts that differed across studies. The control condition did not include a visual pacer, so some of the thalamic effects could partly reflect visual attention processes. Brainstem signal quality at 3T can be challenging, which may help explain why nucleus-of-the-solitary-tract connectivity differences were not robust. And, uniquely for this paradigm, some physiological “noise” is arguably part of the signal—making it difficult to fully separate mechanical cardiorespiratory contributions from neural mechanisms.

Even with those limitations, the overall pattern supports a clinically intuitive conclusion: paced breathing at resonance does not simply create a subjective sense of calm; it appears to tighten coordination among brain regions that translate body state into emotion, attention, and behavioral control.

A broader theme emerges here that connects neatly with neurofeedback practice: the most powerful self-regulation interventions often work by improving communication—between systems, between networks, and between levels of the hierarchy. Resonance paced breathing may be one of the more accessible ways to start that conversation, especially when paired with other skills that support attention, metacognition, and cognitive flexibility.

Brendan’s perspective

If you spend enough time doing neurofeedback, you develop a slightly unfair advantage: you get to watch the nervous system try to regulate itself in real time. This paper adds a useful missing camera angle. During resonance paced breathing (RPB), the central autonomic network (CAN) shows tighter coordination, and the insula shows up repeatedly as a key connector, pairing more strongly with thalamus, cingulate, angular gyrus, and limbic regions. In other words, when the breath becomes rhythmic at 0.1 Hz—the same six-breaths-per-minute resonant pace used routinely in HRV biofeedback—the brain’s “body–state management team” seems to hold a better meeting. That’s especially exciting for EEG-neurofeedback, because the insula is a deep structure that sits well beyond the reach of our scalp electrodes; breath-driven autonomic entrainment may be one of the cleanest ways to influence insula-centered networks while we train what we can measure at the scalp.

In practical terms, I read these findings as permission to treat RPB as more than a relaxation warm-up. It can be a state-setting tool that helps shape what happens during neurofeedback: cleaner baselines, more predictable arousal, and a more stable platform for learning. Below are four ways I would translate this into day-to-day practice.

1) Stacking resonance breathing with EEG training

The simplest stack is: breath first, brain second. Five minutes of RPB at the start of a session often functions like tuning an instrument before a performance. You are not trying to “fix” anything in five minutes; you are creating a more coherent physiological backdrop so the brain has less chaos to compensate for while learning.

A very common pairing is RPB followed by SMR training (12–15 Hz) over sensorimotor cortex. If the person presents with hyperarousal, irritability, insomnia, trauma-related startle, or the kind of anxiety that lives in the shoulders and jaw, I will often reward SMR at C3, C4, or Cz, while inhibiting excessive slow activity (often 4–7 Hz) and high-frequency tension (often 22–30 Hz), with thresholds individualized to their baseline and symptoms. RPB first can reduce the “physiological static,” so SMR becomes less of a wrestling match.

If the person presents more with cognitive overdrive and difficulty downshifting, I may stack RPB with posterior alpha training (8–12 Hz) at POz, again individualized. The aim is not to make someone sleepy; it is to make relaxation accessible on demand. A nice clinical tell is whether the person can keep the paced breathing comfortable while alpha gently increases. When the breath feels forced or the person gets lightheaded, the solution is usually not “try harder,” but “breathe shallower,” or even widen the pacing slightly.

There is also a reverse stack that can be useful: neurofeedback first, RPB second. When someone is learning a new EEG skill (say, sustaining SMR without drifting into drowsiness), finishing with three to five minutes of RPB can consolidate the session and help transfer the skill into a calmer body state. Think of it as saving the file before you close the program.

2) Designing for arousal “stickiness”

Many symptoms are not a lack of regulation, but a problem with transitions. The system gets stuck. An anxious system gets stuck in threat. A depressed system gets stuck in low-energy rumination. A craving-prone system gets stuck in “go get it now.” RPB is valuable here because it offers an external rhythm that can help unstick state changes.

A practical session arc for “sticky” arousal is:

• Settle: 3–5 minutes RPB with a simple cue such as “quiet, low, comfortable.” The goal is a steady rhythm, not deep breathing.

• Stabilize: 10–15 minutes of SMR training (12–15 Hz) at C3, Cz or C4, especially when impulsivity, sleep fragmentation, or sensory hyperreactivity are part of the picture.

• Soften: 10–15 minutes of posterior alpha (8–12 Hz) at POz, especially when tension and vigilance dominate.

This is not a universal recipe. It is a framework that can be bent. Some people need a longer settling phase. Some need to skip the soften phase because alpha pushes them toward dissociation or drowsiness. Some need a brief “mobilize” element at the end (for example, brief beta support at Fz or Cz) so they do not leave the office feeling like they were wrapped in a weighted blanket and forgotten.

Between sessions, I like prescribing micro-doses of RPB: two to three minutes, two to five times per day, anchored to real-life moments (before sleep, before a meeting, after an argument, right after a craving spike). The paper’s emphasis on in-the-moment network modulation supports this “small and often” approach. It is the difference between owning a fire extinguisher and actually keeping it where you can reach it.

3) Interoception as a trainable skill

The insula-heavy pattern in this study matters because it suggests that RPB may support integration of bodily signals with attention and emotional meaning. Clinically, that is the heart of interoception: noticing what the body is doing, without overreacting to it.

There is a common trap here. For some people (especially those with panic, trauma histories, or health anxiety), paying attention to the body can feel like turning up the volume on fear. So interoception has to be titrated. I will often start with neutral, concrete anchors: the feeling of air at the nostrils, the movement of the lower ribs, the contact of feet with the floor. During paced breathing, the instruction is “track one sensation, lightly.” If anxiety rises, we widen the focus to include the room, sound, or vision, and we return to the breath later.

Neurofeedback can support this by making state shifts observable without forcing interpretation. For instance, pairing RPB with SMR training can help someone feel what “steady” means somatically: less bracing, fewer micro-startles, a quieter jaw. Pairing RPB with posterior alpha training can help someone experience the difference between relaxation and collapse. Over time, the goal is a person who can say, “My body is climbing into high gear,” and then actually knows what to do with that information.

4) Why the insula matters even if we can’t place an electrode on it

We cannot place an EEG electrode on the insula and call it a day. That’s exactly why it’s so appealing to couple EEG-neurofeedback with resonance paced breathing: the breathing intervention can tug on insula-centered autonomic circuitry from the inside while EEG targets help sculpt the cortical networks that coordinate with it. But we can influence the networks it talks to, and we can design training around functions that the insula helps coordinate: salience detection, interoceptive integration, and switching between internal and external modes.

Three EEG targets often make sense as indirect levers:

• Sensorimotor stability: SMR (12–15 Hz) at C3/C4/Cz for inhibitory control, sensory gating, and sleep stability. When the system is overreactive, improving stability at the sensorimotor level can reduce the amount of “alarm traffic” moving through the wider network.

• Midline coordination: training around fronto-central midline sites (Fz/FCz/Cz) with careful, individualized goals can support cognitive control and error monitoring, functions tightly linked with cingulate circuitry that interacts with autonomic regulation. The exact frequency target depends on the person’s baseline and symptoms; the principle is to support flexible control without pushing into effortful strain.

• Posterior downshifting: posterior alpha (8–12 Hz) at POz for relaxation and cognitive quieting, especially when rumination and hypervigilance dominate.

The paper also highlights a practical design lesson: the control condition did not include a visual pacer, so some thalamic effects may reflect attention to visual pacing as well as breathing. In clinic, this is actually a feature if we use it well. Pacing is not only a respiratory intervention; it is a gentle attention intervention. If attention is the steering wheel and physiology is the engine, RPB makes it easier to keep the car in the lane.

A final “research versus clinic” note: this study used a fixed rate (six breaths per minute) and brief exposure. Clinically, resonance frequency can vary between individuals, and what matters most is comfort, symptom response, and physiology (for example, avoiding overbreathing). Some clients benefit immediately, some need several sessions to feel it, and a notable minority do not show the same physiological shift. The encouraging part is that this paper gives us a plausible neural rationale for why the breath can be such a powerful entry point. It is not just calming; it is coordinating. And coordination is often the first step toward a nervous system that can change its mind.

Conclusion

This study provides compelling evidence that resonance paced breathing at 0.1 Hz is associated with immediate increases in functional connectivity across the central autonomic network, especially through the insula, thalamus, and other integrative hubs. In plain terms: slow, rhythmic breathing appears to strengthen the brain’s moment-to-moment coordination of bodily signals with attention and emotion.

The clinical promise is twofold. First, these results support resonance breathing as an accessible, scalable strategy for rapid arousal modulation—something that can be used both inside a structured intervention and in the messy reality of daily life. Second, acute connectivity shifts may ultimately help identify biomarkers that predict who will benefit most from longer courses of HRV biofeedback and related breathing-based therapies.

Breathwork is sometimes marketed like magic; it isn’t. But it is one of the most direct ways we can reach into the autonomic nervous system using a voluntary lever. This paper adds a valuable neural layer to that story: when the breath slows into resonance, the brain’s autonomic network doesn’t just react—it coordinates. And that coordination is exactly where durable self-regulation starts.

References

Bates, M. E., Lesnewich, L. M., Pawlak, A. P., Buckman, J. F., & Gohel, S. (2026). Functional connectivity within the central autonomic network increases during resonance paced breathing at 0.1Hz. Psychophysiology, 63, e70263. https://doi.org/10.1111/psyp.70263