- Artificial Heart with Artificial Intelligence -
Overview
Heart failure is the leading cause of death in humankind. To tackle this disease, the United States launched a national project in the 1960s—second only to the Apollo program—called the “Challenge to God,” which marked the beginning of artificial heart development. Today, our research group is working to develop an artificial heart equipped with artificial intelligence, aiming to achieve better survival rates and improved quality of life (QOL) for patients.
Since artificial hearts must be implanted inside the human body, the number of sensors that can be used is extremely limited. To address this challenge, we focus on the latest magnetically levitated artificial heart technology, applying the magnetic levitation system itself as a vital sensor. By combining this with machine learning, we are creating new functionalities for next-generation artificial hearts.
Heartbeat-Synchronized Control
Magnetically levitated artificial hearts typically operate their impellers—the rotating blades that pump blood—at a constant speed. In other words, the blood flow generated by such artificial hearts is non-pulsatile.
In contrast, blood in the human body naturally circulates with the pulsation of the heart. There has been debate over whether the absence of pulsation caused by artificial hearts may lead to various complications. For example, continuous flow can prevent the aortic valve from opening, thereby increasing the risk of aortic valve insufficiency.
Studies have also reported that generating pulsation in sync with the heartbeat can promote the recovery of the diseased heart’s function.
This is where heartbeat-synchronized control becomes important. Conventional approaches used electrocardiograms (ECG) to measure the heart’s rhythm and fed this data back into the artificial heart’s motor speed control system to achieve synchronization. However, in clinical practice, it is unrealistic for patients to wear ECG sensors continuously in daily life.
In our laboratory, we are developing a new technology that detects subtle pulsations of the diseased heart through the magnetically levitated impeller itself. By using machine learning algorithms embedded in the artificial heart, we can determine the timing of pulsations accurately, achieving heartbeat-synchronized control without relying on external sensors such as ECG.
Thrombosis Detection and Prevention Using Magnetic Bearings and Motors
Since the materials that compose artificial hearts are artificial, blood contact inevitably leads to thrombus (blood clot) formation. Although various anti-thrombogenic coatings and blood-compatible materials have been developed, these technologies alone cannot completely prevent thrombosis.
Therefore, in clinical settings, anticoagulant drugs are administered to patients to reduce blood clotting. However, this also increases the risk of complications such as cerebral hemorrhage.
To address this, our laboratory is developing a technique that suppresses blood adhesion to material surfaces by vibrating the magnetically levitated impeller at specific frequencies and trajectories with micro-scale displacements. This prevents thrombosis formation. Furthermore, by analyzing the vibration response of the impeller during this excitation, we are developing a technology to detect thrombosis at an ultra-early stage inside the pump.
