Asynchronous Time-domain Neural Recording Interfaces

Implantable neural recording devices are being used by neuroscientists and neurosurgeons for monitoring the behaviour of particular parts of the brain. Traditionally, these implanted devices were wired and connected to the measurement and computation equipment to process the signals. Such a setup restricts the physical movement of the subject and limits the number and quality of the experiments or monitoring that can be performed. With the advancement of microelectronics manufacturing processes allowing lower power operation while simultaneously occupying a smaller area on chip, and circuit and system level architectural inventions, wireless multi-channel neural recording systems have become available during the last decade. However, the capabilities of neural recording implants are still limited by their power density in order not to harm the surrounding tissue by heat generation. If the operational capabilities of the neural recording implants are to be improved in terms of number of electrodes supported, number of simultaneous recording channels, on-chip signal processing, and operational longevity, new methods of power consumption reduction are required.

Such an ultra-low power neural recording device has multiple uses for both the scientists/researchers and the general public: i) it can be used for experiments that require continuous neural monitoring with minimal energy requirements and implant size, ii) it can be used as part of a point-of-care device for the patients, with the minimum amount of burden to the patient, such as the need to change batteries, weight due to batteries, etc., and iii) it can be used as part of a closed loop system of a bioelectronic system for alleviating certain pathological conditions, such as epilepsy, tinnitus, etc. Furthermore, if such a recording device can be supplemented with additional multi-modal measurement capabilities, it will be immensely useful for the researchers and medical personnel to better diagnose different pathologies.

During the course of this project, the main aim of the research was to find novel ways of power consumption reduction for such multi-modal and multi-purpose implantable bioelectronic devices. Based on our findings and using time-mode operation (a special mode of operation where the delay of a signal represents the information to be processed), a wireless neural recording chip was developed and implemented. The implemented microchip achieves three orders of magnitude lower energy dissipation when compared to the other implementations in the literature. Furthermore, the sensing and measurement capabilities of the proof of concept implementation were improved with the addition of imaging, temperature, and electroanalytical sensors.

During the research activities of the project, it was found out that by employing time-domain signal processing (TDSP) for realizing required functions in neural recording implants, it is possible to achieve up to three orders of magnitude (up to 1000 times less) energy dissipation reduction in neural recording and sensing implants. In addition to the energy dissipation reduction properties of TDSP, it was discovered that by employing the time-domain operation concept, extremely sensitive measurement circuits can be built. A proof of concept microchip was implemented to investigate the ingress of the water molecules and ions when a biomedical implant is placed inside a biological body.

As the proposed time-domain proof of concept, a wireless neural recording microchip was developed and implemented. The implemented microchip achieves three orders of magnitude lower energy dissipation when compared to the other implementations in the literature. Time-domain operation was further utilized to create extremely energy efficient sensors for future multi-modal sensing biomedical implants. The time-domain sensors implemented during the course of the project will allow future bioelectronics implants to measure light (similar to an image sensor), temperature, extremely small electrical currents, and ion concentration and electro-chemical properties of liquids and tissues with the least amount of energy while being implemented using low-cost standard CMOS technologies.

Finally, time-domain operation was used for creating an extremely energy efficient artificial neural network (ANN) implementation, which reduces the energy dissipation by an order of magnitude when compared to the most energy efficient implementation in the literature.

Preliminary results of the project show 10x to 1000x energy dissipation reduction in neural recording implant implementation. These numbers open up the way to self-powered autonomous implants, which will be powered by energy harvesting utilizing the ambient energy sources in the future. A continuously powered and operating system will minimize the burden to the end user opening up the path to different healthcare applications.

During the course of the project, 2 preliminary patents were applied for and granted, and 2 more patent applications will be prepared and submitted. Based on the patents, founding a start-up company to commercialize the outcome of the research is in progress. The start-up will be focusing on ultra-low cost health monitoring systems which dissipate the least amount of energy while giving the least amount of burden to the patient that is being monitored. A low cost, easy to use health monitoring system will have tremendous effects on the general well-being of the public. Such a system will allow everyone to have their own personal health monitoring systems with the least amount of burden to the person's monetary budget. In addition, the envisioned and soon to be realized system will be immensely useful in both hospital and clinical settings. Furthermore, the research output will be employed in academia and research institutes for realizing extremely capable and miniaturized future state of the art biomedical implants.

Periodic Reporting for period 1 - ATiNaRI (Asynchronous Time-domain Neural Recording Interfaces)

Partager cette page

Télécharger