Implantable Medical Devices: The Power-Constrained Frontier
November 6, 2009
Electronic devices for medicine are a rapidly growing area of technology. In-vivo monitoring and treatment of key biological parameters can greatly assist in managing health and preventing disease. However excess analog power consumption and insufficient power supply prohibit the widespread deployment of implantable medical devices (IMDs) for many applications.
Biological signal channels differ considerably from human-made communications channels, generating new design challenges for IMDs. Furthermore performance requirements of analog circuits for IMDs vary as a function of patient physique, health, device placement, patient activity etc, and thus cannot be known accurately prior to deployment. This talk presents a system-configured analog design approach to address these challenges and that approach is applied to signal acquisition and power delivery for neural sensors in motor prostheses.
An analog-to-digital converter (ADC) array which digitizes the neural signals sensed by an implanted microelectrode array is described. The resolution of each ADC cell is varied according to the neural data content of the signal from the corresponding electrode. Realized in 0.13µm CMOS the ADC achieves a figure of merit of 15fJ per conversion step.
A new method of wireless power transfer is developed for implanted devices which are constrained in size. First, the optimal frequency for wireless power transmission through tissue to area constrained receivers is derived. Second, an adaptive matching scheme which increases the robustness of the link gain to inevitable dielectric, range and alignment variations associated with an IMD is presented. Third, a low voltage rectifier is presented which reduces the voltage drop per stage to considerably less than a threshold voltage. The power receiver implemented in 0.13µm CMOS delivers 120µW at 1.2V DC to the IMD from a 2mm x 2mm on-board receive antenna, through 15mm of tissue.
I will outline some of the new projects in my BioElectronics Group at UC Davis including an implant positioning system (IPS), an ingestible rumen sensor, and neuro-stimulation devices - all of which will utilize adaptive analog circuits and mm-sized implantable power receivers.
Stephen O'Driscoll received the BE in Electrical Engineering from University College Cork, Ireland in 2001. In 1999 and 2000 he worked on microwave circuits for radar at Farran Technology, Ireland. From 2001 to 2003 he was at Cypress Semiconductor, San José, where he designed clock and data recovery phase locked loops. He received the MS degree in 2005 and the PhD degree earlier this year, both in Electrical Engineering Stanford University where he worked in Prof Teresa Meng's lab. He joined the faculty at UC Davis in August of this year where he is conducting research on analog and mixed-signal circuit design, with particular focus on biomedical devices.