Ultrahigh-Density 256-Channel Neural Sensing Microsystem Using TSV-Embedded Neural Probes

Design of a low-noise low-voltage amplifier for improved neural signal recording

Design of amplifier circuits with low-noise operable at low-power to be used, especially for implantable neural interfaces, remains a huge challenge. This research paper presents the design of a low-noise low-voltage neural recording amplifier suitable for amplifying local field potentials and extracellular action potentials so as to meet the end requirement of an implantable neuro-medical system. Critical performance parameters of the smaller circuit blocks of the complete neural amplifier architecture have been found with the help of detailed mathematical analysis and then verified by the simulations conducted using 0.18 µm 4M1P foundry Semi-conductor Laboratory N-well process. The neural amplifier design proposed in this paper passes neural signal of interest with a mid-band gain of 49.9 dB over a bandwidth of 5.3 Hz-8.6 kHz, draws only 11.5 µW of power from ±0.9 V supply voltage, and exhibits an input-referred noise of 2.6 µV_rms with a noise efficiency factor of 2.27. The area consumed by the proposed neural amplifier architecture is 0.192 mm². The complete circuit design carried out in this paper should prove to be useful in equipment for the diagnosis of neurological disorders.

Microscope-FTIR Spectrometry Based Sensor for Neurotransmitters Detection

In this work, we present a new sensing approach for aqueous samples based on the microscope-FTIR spectrometer and applied for neurotransmitters. Our contribution in this work consists of a new sample handling system for the microscope-FTIR spectrometer based on a total reflective mirror, a heated hydrophobic layer for solvent removal/evaporation and sample confinement and a microfluidic system that handles sample injection unlike standard sample handling system which was based only on a total reflective mirror. In addition, another part of our contribution consists of proposing a new algorithm to extract molecular composition of the solution with high estimation ratios and based on the analysis of detected peaks on IR spectra. The data acquired from the microscope-FTIR spectrometer was analyzed by a newly developed algorithm to identify each neurotransmitter in homogeneous and non-homogeneous solutions with high selectivity. We used six neurotransmitter molecules (Dopamine hydrochloride, L-Ascorbic acid, Acetylcholine chloride, y-Aminobutyric, Glycine and L-Glutamic acid). The results obtained based on the algorithm developed showed that, using the new system, the six neurotransmitters can be identified in homogeneous and mixture solutions with an estimation ratio range of 88.8%-100% for Dopamine hydrochloride, 80%-100% for L-Ascorbic acid, 75%-100% for Acetylcholine chloride, 75%-100% for L-Glutamic, 77.7%-100% for y-Aminobutyric and 75%-100% for Glycine.

Compact and Low-Power Neural Spike Compression Using Undercomplete Autoencoders

Implantable microsystems that collect and transmit neural data are becoming very useful entities in the field of neuroscience. Limited by high data rates, on-chip compression is often required to transmit the recorded data without causing power dissipation at levels that would damage sensitive brain tissue. This paper presents a data compression system designed for brain-computer interfaces (BCIs) based on undercomplete autoencoders. To the best of our knowledge, the proposed system is the first to achieve an average spike reconstruction quality of 14-dB signal-to-noise-and-distortion ratio (SNDR) at a 32× compression ratio (CR), 18-dB SNDR at a 16× CR, 22-dB SNDR at an 8× CR, and 35-dB SNDR at a 4× CR of neural spikes. The spike detection and autoencoder-based compression modules are designed and implemented in a standard 45-nm CMOS process. The post-synthesis simulation results report that the compression module consumes between 1.4 and 222.5 [Formula: see text] of power per channel and takes between 0.018 and 0.082mm² of silicon area, depending on the desired CR and number of channels.