Adaptive signal acquisition and power delivery for implanted medical devices

Implantable medical devices are a vitally important area of technology. In-vivo monitoring and treatment of key biological parameters can greatly assist in managing health and preventing disease. Implanted devices typically perform some combination of bio-signal acquisition; signal processing; actuation; and telemetry, all of which consume power. Supplying adequate power and minimizing dissipation are critical constraints in all implanted devices.;The driving application for this work is an implanted neural sensor. Motor neuroprosthetics seek to restore motor function to the paralyzed by sensing neural signals, decoding intended movement and relaying that movement to artificial limbs and actuators. In order to realize an implantable prosthetic processor, which acquires and compresses the neural signals, acute challenges lie in digitizing the neural signals within the power budget and delivering power to the implanted circuitry.;Extremely low power consumption of analog-to-digital converter (ADC) can be attained through integrating signal analysis, system architecture, and thorough analog design. An ADC array which digitizes the neural signals sensed by a microelectrode array is described. The ADC array consists of 96 variable resolution ADC base cells. The base ADC has been implemented as a 100kS/s successive approximation ADC whose resolution can be varied from 3 to 8-bits with corresponding power consumption of 0.23micro W to 0.90microW, achieving an effective number of bits (ENOB) of 7.8 at the 8-bit setting. The resolution of each ADC cell in the array is adjusted according to the neural data content of the signal from the corresponding electrode. Resolution adaption reduces power consumption by a factor of 2.3 whilst maintaining an effective 7.8-bit resolution across all channels. The ADC cell occupies 0.07mm2 in 0.13microm CMOS.;A wireless power transfer system for implanted medical devices which uses antenna area 100 times smaller than previous designs is presented. The optimum frequency for wireless power transmission through several cm of tissue to mm-sized antennae is demonstrated, theoretically and empirically, to be in the low GHz. Simultaneous conjugate matching is introduced which achieves 15dB higher link gain than resonant tuning for this link and adaptive matching circuits desensitize the link to placement inaccuracy and tissue variation. For a 1 mm placement inaccuracy the link gain with static matching would fall by 3.7dB but with adaptive matching the gain reduces by only 0.2dB. A high efficiency current-steering rectifier and regulator are developed which together with a 4mm 2 antenna deliver 140microW at 1.2V DC from a 4cm2 transmit antenna and a 250 mW, 915MHz source through 15mm of tissue. The rectifier achieves a 65% efficiency for a 750mV amplitude received input. The power receiver IC incorporating the adaptive receive match, rectifier and regulator occupies 0.37mm 2 in 0.13microm CMOS.;Our ADC array consumes 20dB less power than the previous best reported neural ADC for the same bandwidth and resolution and the power link delivers over 10dB more power to a 4mm 2 implanted antenna than previous technology. Together these innovations overcome the 30dB gap between the power which was needed for the digitization portion and the power which could be delivered, to enable implanted prosthetic processors. Furthermore these technologies demonstrate that combining bio-signal analysis, adaptive signal processing and reconfigurable circuits can offer orders of magnitude improvement in performance and enable new ranges of miniaturized implanted devices.