2 shows two conventional architectures of implantable sensing systems, which typically consist of one or multiple channels of analog front-end (AFE), ADC, digital signal processing (DSP), memory, and a wireless transmitter (TX). Since there is no sufficient volume for a battery, the electronic system should consume low energy well below 100 μW, to enable wireless power transfer.įig. A nerve implant with a volume in the millimeter-scale is highly preferred. To avoid nerve tissue damages, such nerve implants should have strict volume and energy constraints. In order to have high spatial selectivity, neural implants for peripheral nerves should be placed very close to the surface (or inside) of the nerve, as illustrated in Fig. This increases the energy consumption not only in wireless transmission, but also in local processing, storage, and transportation of the data.Ĭoncept illustration of neural recording of peripheral nerves, and the conceptual illustration of nerve conduction velocity measurement. To achieve such temporal precision, the analog-to-digital converters (ADCs) in a conventional neural recording system need to have a sampling rate of 10’s of kSample/s (kSps), which is 10-100× higher compared to the sampling of other electrocardiogram (ECG) signals. Better temporal precision of the recording allows the volume of the nerve implant (e.g., nerve cuff) to be further miniaturized. To achieve high accuracy of NCV with a miniature nerve implant, temporal precision of the recording should be in the order of 10’s of μs, since NCV of a myelinated nerve can be up to 120 meter/s. This precision is especially challenging for NCV studies, which measures the time difference between peaks of two CAPs recorded from two locations on the same nerve, as shown in Fig. The requirement on temporal precision for such measurements is strict since the CAPs typically last for approximately one millisecond. Next, the CAP peak-to-trough duration and nerve conduction velocity (NCV) are widely used diagnostic tools for various neuropathies. Decoding of the firing pattern of afferent compound action potentials (CAPs), the result of summation of many APs from the individual axons in a nerve trunk, holds the promise for indirect sensing of clinically relevant information, e.g., inflammation status or glucose levels, which can be employed in future electroceutical closed-loop applications. Nerve ENG provides rich clinical information for diagnosis and can be the source of modulating human health as electroceuticals. The electroneurogram (ENG) can be measured with a nerve cuff or a neural probe surrounding or penetrating the peripheral nerves, respectively. THE peripheral nervous system (PNS) can be seen as a “highway” for propagating neuron firings, i.e., action potentials (AP), for the bidirectional communication between the central nervous system (CNS) and various organs. The presented NSS also extracts temporal features of compound action potential signals with 10-μs precision. The prototype is fabricated in 40-nm CMOS occupying a 0.32-mm 2 active area and consumes in total 28.2 μW and 50 μW power in feature extraction and full diagnosis mode, respectively. An event-driven pulse-based body channel communication (Pulse-BCC) with serialized address-event representation encoding (AER) schemes minimizes transmission energy and form factor. A fully synthesized spiking neural network (SNN) extracts temporal features of compound action potential signals consumes only 13 μW. A clockless level-crossing (LC) ADC with background offset calibration has been employed to reduce the data rate, while maintaining a high signal to quantization noise ratio. The proposed NSS consists of three sub-circuits. The proposed NSS employs event-based sampling which, by leveraging the sparse nature of electroneurogram (ENG) signals, achieves a data compression ratio of >125×, while maintaining a low normalized RMS error of 4% after reconstruction. This paper presents a bio-inspired event-driven neuromorphic sensing system (NSS) capable of performing on-chip feature extraction and “send-on-delta” pulse-based transmission, targeting peripheral-nerve neural recording applications.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |