First demonstrated in the 1950s, one of most common implanted biomedical devices is the pacemaker (1). These devices are among the most highly energy constrained, requiring years of operation on a single battery charge to avoid repeated surgery. The average power consumption of a pacemaker is on the order of five to ten microwatts, and is enabled by the minimal processing requirements and low analog-to-digital converter (ADC) speeds (100–1000 samples per second) (1). Modern pacemakers, however, often include support for cardiac defibrillation, which requires large electric pulses that are both power and energy intensive, resulting in significant circuit design challenges.
Like pacemakers, both hearing aids and cochlear processors are relatively mature biomedical applications. Today, they make front-running strides in power-management (3, 4, 6). Their power consumption ranges down to hundreds of microwatts, and, in some cases, by leveraging high dynamic range ADCs (i.e., 75 dB), they dynamically adjust their power, performance, and ergonomics simultaneously. This agility allows them to easily adapt for individual patients, but, even more impressively, also for the varying acoustic environments each patient faces.
In contrast to pacemakers and hearing aids, neural recording systems and retinal stimulators are emerging applications. Researchers are focused on challenges in the interface and acquisition electronics as well as processing platforms. Nonetheless, many critical challenges are being dealt with by using smart arrays that leverage both material processing innovations and electronics innovations, leading to the integration of more than 100 acquisition channels in a neural recording system (7, 8) and 16 stimulus electrodes in a retinal stimulator (10). Implanted stimulators and recorders offer the potential to revolutionize the treatment of many medical conditions. For example, implanted deep-brain stimulators are used to treat patients with Parkinson's disease (12) and intramuscular stimulation is being investigated for treating paralyzed muscles and limbs (13).
Finally, as biomedical devices become more prevalent, there is increasing need for these devices to support the formation of body-area networks. Body-area networks allow for individual devices to collaborate and communicate with one another. For instance, a finger-mounted pulse oximeter worn by a patient in an operating room can wirelessly send its status to medical personnel to enhance patient monitoring. Electronic devices compatible with body-area networks have widely varying performance requirements depending on the specific application, but power requirements are typically on the order of hundreds of microwatts to milliwatts. Further, for robust communication, the devices must employ standards-compliant wireless communication links.