The Aorta’s Hidden Second Pump: How Arterial Stretch May Steer Blood to the Brain

For more than a century, the aorta has been portrayed as a passive highway for blood leaving the heart. A new laboratory study published in the Annals of Biomedical Engineering challenges that picture in a striking way. Researchers at the University of Southern California and the University of California, Merced, built a physiologically accurate artificial circulation and showed that the simple longitudinal stretching and recoiling of the aorta—the up-and-down axial motion the vessel undergoes with every heartbeat—can generate a wave-pumping effect that pushes measurable net flow toward the carotid arteries, the vessels that supply the brain. In some wave conditions the effect drove flow forward at roughly 300 milliliters per minute; in others it actually reversed, pulling blood backward. The finding suggests that the aorta may act less like a pipe and more like a second, auxiliary pump whose performance depends delicately on heart rate, arterial stiffness, and the geometry of the wave field inside the vessel.

The motivation for the work lies in a persistent clinical mystery. Heart failure is well known to impair the heart’s ability to eject blood, but epidemiological studies have also linked it to cognitive impairment, brain injury, and an elevated risk of dementia, including Alzheimer’s disease. Abnormal arterial hemodynamics and cerebrovascular dysfunction are increasingly recognized as part of this heart–brain axis, yet the precise hemodynamic mechanisms connecting a failing heart to an under-perfused brain remain incompletely understood. One overlooked candidate is the coupling between the left ventricle and the proximal aorta. During systole, the contraction of the ventricle displaces the aortic annulus, stretching the ascending aorta along its axis by a considerable distance. That stretch stores elastic energy in the spring-like elements of the aortic wall, energy that is released as the vessel recoils during diastole. Cardiac MRI studies have measured this longitudinal displacement at 10 to 20 millimeters in healthy subjects but only 5 to 9 millimeters in patients with heart failure—a difference the researchers suspected might matter far more than previously appreciated.

To isolate the effect, the team did something clever: they removed the heart from the equation entirely. Their in-vitro circulation system contained no left-ventricle-like pulsatile pump at all. Instead, a one-to-one-scale aortic phantom, fabricated in-house from latex or silicone using a stainless-steel mold that reproduced the major arterial branches and physiological tapering, was mounted in an acrylic box. A computer-controlled stepper motor drove interchangeable spiral cams that pulled the aortic root axially and then released it, mimicking the active stretch of systole and the passive recoil of diastole. The supra-aortic branches were deliberately fixed, replicating the anchoring role of connective tissue in the body, so that the imposed displacement was transmitted primarily through the ascending aorta. By eliminating ventricular pumping, the researchers could attribute any net flow they measured at the carotid artery directly to the longitudinal stretch mechanism rather than to a superposition of cardiac and aortic waves.

The experimental matrix was exhaustive. Aortic phantoms spanning a range of stiffness—quantified by pulse wave velocity, a standard clinical index of arterial stiffening—were subjected to stretching amplitudes of 5 to 20 millimeters, chosen to overlap the values measured by cardiac MRI in healthy people and heart failure patients. Stretching frequencies ranged from 0.5 to 2.5 hertz, corresponding to heart rates of 30 to 150 beats per minute. High-fidelity pressure was recorded with a piezoelectric-tip catheter inserted into the left common carotid artery, while flow was measured with an ultrasound flowmeter clamped onto the adjacent tubing, both sampled at one kilohertz through a PowerLab acquisition system. Water served as the working fluid, a simplification the authors note likely underestimates viscous losses but does not alter the wave dynamics that were the focus of the study. Measurements began only after the system reached a periodic steady state, typically after ten cycles, and were ensemble-averaged over multiple cycles for reliability.

The results were unambiguous: longitudinal stretch and recoil alone generated a wave-pumping effect in the carotid artery, and both the magnitude and the direction of the resulting net mean flow depended strongly on three coupled parameters—stretching frequency, stretching amplitude, and aortic stiffness as reflected in pulse wave velocity. When plotted against frequency alone, the net carotid flow varied substantially, reaching positive values of about 150 milliliters per minute in some conditions and negative values of about 30 milliliters per minute in others, depending on the phantom’s wave speed. Intriguingly, pulse wave velocity by itself did not fully characterize the response; phantoms with similar wave speeds could produce different flows. The unifying variable turned out to be the wave condition number, a dimensionless parameter that compares the characteristic wave propagation timescale with the excitation timescale—in effect, asking whether waves launched by the stretch have time to travel the length of the aorta and interact constructively or destructively within each cycle.

The wave condition number consolidated the data remarkably well. When the researchers examined net flow as a function of this single parameter, phantoms that had behaved inconsistently as a function of frequency fell onto a coherent trend. At very low wave condition numbers, around 0.01, negligible net flow was generated regardless of how much elastic energy was stored in the stretched aortic wall, indicating inefficient conversion of energy into transport. As the wave condition number rose, flow became increasingly sensitive to stored energy, peaking at intermediate values near 0.06, where elastic energy of about 125 millijoules produced net flows approaching 300 milliliters per minute. At still higher wave condition numbers, meaningful net flow appeared even at low stored energies of roughly 20 millijoules, suggesting that favorable wave states dramatically improve the efficiency with which stretch energy is converted into forward transport. A sensitivity analysis confirmed the shift: in the low wave condition number regime, wave dynamics limit the conversion, whereas in the high regime the stored elastic energy itself becomes the dominant determinant of flow.

Building on these observations, the team formulated a physics-informed reduced-order model in which the cycle-averaged carotid flow equals the stored elastic energy—an expression derived from the Moens–Korteweg equation linking wave speed to wall elasticity—multiplied by a quadratic function of the wave condition number. Fitted by nonlinear least squares to all 144 experimental conditions, the model showed strong agreement with the measured data, with a correlation coefficient of 0.827 and a p-value below 0.0001. The largest deviations occurred in the low-flow and retrograde regimes, which the authors attribute to the model capturing the dominant wave-pumping contribution rather than the full spectral richness of the waveform; frequency-domain analysis showed much weaker agreement for individual harmonics, which are shaped by reflections, impedance mismatches, and damping not represented in the reduced formulation. The model also revealed that for each level of aortic stiffness there exists an optimal stretching frequency that maximizes carotid flow, and that this optimum shifts as stiffness changes—a coupled signature of impedance-pump behavior previously described in compliant tubes and embryonic heart tubes.

The appearance of retrograde flow deserves particular attention. At low excitation frequencies, the wave-pumping contribution is weak, allowing a local suction effect produced during the active stretching phase to dominate and drag blood backward, away from the brain. As frequency increases, favorable wave interactions amplify the pumping contribution and flip the net flow forward. This bidirectionality means that the same aortic motion can either assist or hinder cerebral perfusion depending on the wave state of the system—a sobering thought given that aging stiffens the aorta, raises pulse wave velocity, and reduces longitudinal stretching, all of which would shift the wave condition number and potentially move a patient away from optimal pumping conditions. In heart failure, where aortic annular displacement is already halved, the diminished elastic energy storage would be expected to compound the problem, offering a mechanistic bridge between cardiac dysfunction and the reduced cerebral perfusion increasingly documented in these patients.

The authors are careful to frame the work as evidence of an isolated mechanism rather than a direct estimate of its contribution in living humans. Their phantoms, made of latex and silicone, lack the nonlinear, anisotropic, and viscoelastic character of real arterial tissue; the system imposes only longitudinal motion, omitting the radial and circumferential expansion that ventricular ejection produces in vivo; the downstream cerebral vasculature is simplified; and cerebral autoregulation—the brain’s own ability to maintain flow—is absent. Water, rather than a blood-mimicking fluid, circulated through the rig. The optimal frequencies identified by the model should therefore not be read as heart rates a clinician could prescribe. Still, the framework has a practical appeal: the wave condition number can be estimated non-invasively from the arterial pressure waveform alone, as prior work by the same group has demonstrated, raising the possibility of a bedside index for characterizing cardiovascular wave dynamics and their influence on cerebral perfusion.

If validated clinically, the implications could reach from cardiology to neurology. A measurable, modifiable mechanism linking aortic biomechanics to brain blood flow would give researchers a new target for understanding why heart failure and arterial stiffening accelerate cognitive decline, and potentially a new lever for intervention—whether through therapies that restore aortic compliance, pacing strategies that tune the wave state, or diagnostics that flag patients whose aortas have drifted into flow-unfavorable wave conditions. For now, the study’s most vivid contribution is conceptual: the aorta is not merely a conduit but an active participant in circulation, a muscular-elastic structure whose longitudinal dance with every heartbeat helps decide how much blood reaches the brain. The second heart, it seems, has been hiding in plain sight along the body’s largest artery all along.

Subject of Research: The role of aortic longitudinal stretch and recoil in wave-driven cerebral blood flow

Article Title: Aortic Longitudinal Stretch and Recoil Affect Cerebral Perfusion: Insights from a Physiologically Accurate In-Vitro Experiment Model

Article References: Geng, H., Aghilinejad, A., Bilgi, C., Nation, D. A., & Pahlevan, N. M. (2026). Aortic Longitudinal Stretch and Recoil Affect Cerebral Perfusion: Insights from a Physiologically Accurate In-Vitro Experiment Model. Annals of Biomedical Engineering. https://doi.org/10.1007/s10439-026-04408-y

Image Credits: AI Generated

DOI: 10.1007/s10439-026-04408-y

Keywords: aorta, cerebral perfusion, wave pumping, hemodynamics, arterial stiffness, pulse wave velocity, wave condition number, heart failure, carotid flow, in-vitro model, biomechanics, impedance pumping