Search

High-density, integrated silicon electrodes have begun to transform systems neuroscience, by enabling large-scale neural population recordings with single cell resolution. Existing technologies, however, have provided limited functionality in nonhuman primate species such as macaques, which offer close models of human cognition and behavior. Here, we report the design, fabrication, and performance of Neuropixels 1.0-NHP, a high channel count linear electrode array designed to enable large-scale simultaneous recording in superficial and deep structures within the macaque or other large animal brain. These devices were fabricated in two versions: 4416 electrodes along a 45 mm shank, and 2496 along a 25 mm shank. For both versions, users can programmatically select 384 channels, enabling simultaneous multi-area recording with a single probe. We demonstrate recording from over 3000 single neurons within a session, and simultaneous recordings from over 1000 neurons using multiple probes. This technology represents a significant increase in recording access and scalability relative to existing technologies, and enables new classes of experiments involving fine-grained electrophysiological characterization of brain areas, functional connectivity between cells, and simultaneous brain-wide recording at scale.

An ultra-small-area, low-power analog front-end (AFE) for high-density neural recording is presented in this paper. It features an 11-bit incremental delta-sigma analog-to-digital converter (σ ADC) enhanced with an offset-rejecting event-driven input biasing network. This network avoids saturation of the ADC input caused by leakage of the input-coupling capacitor implemented in an advanced technology node. Combining AC-coupling with direct data conversion, the proposed AFE can tolerate a rail-to-rail electrode offset and achieves a good trade-off between power, noise, bandwidth, input impedance, and area. Fabricated in a 22-nm fully-depleted silicon on insulator (FDSOI) process, the design occupies an active area of

The recording of biopotential signals using techniques such as electroencephalography (EEG) and electrocardiography (ECG) poses important challenges to the design of the front-end readout circuits in terms of noise, electrode DC offset cancellation and motion artifact tolerance. In this paper, we present a 2nd-order hybrid-CTDT Δ∑-∑ modulator front-end architecture that tackles these challenges by taking advantage of the over-sampling and noise-shaping characteristics of a traditional Δ∑ modulator, while employing an extra ∑-stage in the feedback loop to remove electrode DC offsets and accommodate motion artifacts. To meet the stringent noise requirements of this application, a capacitively-coupled chopper-stabilized amplifier located in the forward path of the modulator loop serves simultaneously as an input stage and an active adder. A prototype of this direct-to-digital front-end chip is fabricated in a standard 0.18-μm CMOS process and achieves a peak SNR of 105.6 dB and a dynamic range of 108.3 dB, for a maximum input range of 720 mVpp. The measured input-referred noise is 0.98 μVrms over a bandwidth of 0.5–100 Hz, and the measured CMRR is >100 dB. ECG and EEG measurements in human subjects demonstrate the capability of this architecture to acquire biopotential signals in the presence of large motion artifacts.

This paper presents a scalable neural recording analog front-end architecture enabling simultaneous acquisition of action potentials, local field potentials, electrode DC offsets and stimulation artifacts without saturation. By combining a DC-coupled $\Delta $ - $\Delta \Sigma $ architecture with new bootstrapping and chopping schemes, the proposed readout IC achieves an area of 0.0077 mm2 per channel, an input-referred noise of 5.53 ± 0.36 $\mu \text{V}_{\mathrm{ rms}}$ in the action potential band and 2.88 ± 0.18 $\mu \text{V}_{\mathrm{ rms}}$ in the local field potential band, a dynamic range of 77 dB, an electrode-DC-offset tolerance of ±70 mV and an input impedance of 663 $\text{M}\Omega $ . To validate this neural readout architecture, we fabricated a 16-channel proof of-concept IC and validated it in an in vitro setting, demonstrating the capability to record extracellular signals even when using small, high-impedance electrodes. Because of the small area achieved, this architecture can be used to implement ultra-high-density neural probes for large-scale electrophysiology.

Driven by needs in neuroscientific research, future neural interface technologies demand integrated circuits that can record a large number of channels of neural signals in parallel while maintaining a miniaturized physical form factor. Using conventional methods, it is challenging to reduce circuit area while maintaining the high dynamic range, low noise, and low power consumption required in the neural application. This paper proposes to address this challenge using a VCO-based continuous-time delta-sigma modulator (CTDSM) circuit, which can record and digitize neural signals directly without the need for front-end instrumentation amplifiers and anti-aliasing filters, which are limited by the abovementioned circuit-area performance tradeoff. Thanks to the multi-level quantization and intrinsic mismatch-shaping capabilities of the VCO-based approach, the proposed first-order CTDSM can achieve comparable electrical performance to a higher-order CTDSM while offering further area and power reductions. We prototyped the circuit in a 22-channel test chip and demonstrate, based on the chip measurement results, that the proposed modulator occupies an area of 0.00426 mm2 while achieving input-referred noise levels of 6.26 and 3.54 µVrms in the action potential (AP) and local field potential (LFP) bands, respectively. With a 77.6 dB wide-dynamic range, the noise and total harmonic distortion meet the requirements of a neural interface with up to 149 mVpp input AC amplitude or up to ±68 mV DC offsets. We also validated the feasibility of the circuit for multi-channel recording applications by examining the impact of cross-channel VCO oscillation interferences on the circuit noise performance. The experimental results demonstrate the proposed architecture is an excellent candidate to implement future multi-channel neural-recording interfaces.

Most of us take for granted the complex functions that our brains perform every day to enable us to think, move, and sense, but there are many diseases and disorders that affect them and lead to their partial or total loss. According to the World Health Organization, approximately 1.3 billion people worldwide are visually impaired (in 2018), more than 466 million people have disabling hearing impairment (2019), over 10 million people live with Parkinson's disease (2018), roughly 50 million people have epilepsy (2019), and nearly 50 million people live with dementia (2019) (see Figure 1). Those numbers, a high percentage of which correspond to developing countries, are expected to grow rapidly in the coming years. In 2017, the estimated annual cost of the major neurological diseases reached USd800 billion in the United States alone [1], which clearly indicates that they are rapidly becoming a major financial burden to healthcare in our society.

Recording many neurons for a long time The ultimate aim of chronic recordings is to sample from the same neuron over days and weeks. However, this goal has been difficult to achieve for large populations of neurons. Steinmetz et al. describe the development and testing of Neuropixels 2.0. This new electrophysiological recording tool is a miniaturized, high-density probe for both acute and long-term experiments combined with sophisticated software algorithms for fully automatic post hoc computational stabilization. The technique also provides a strategy for extending the number of recorded sites beyond the number of available recording channels. In freely moving animals, extremely large numbers of individual neurons could thus be followed and tracked with the same probe for weeks and occasionally months. Science , this issue p. eabf4588

The recording of biopotential signals using techniques such as electroencephalography (EEG) and electrocardiography (ECG) poses important challenges to the design of the front-end readout circuits in terms of noise, electrode DC offset cancellation and motion artifact tolerance. In this paper, we present a 2

To study the dynamics of neural processing across timescales, we require the ability to follow the spiking of thousands of individually separable neurons over weeks and months, during unrestrained behavior. To address this need, we introduce the Neuropixels 2.0 probe together with novel analysis algorithms. The new probe has over 5,000 sites and is miniaturized such that two probes plus a headstage, recording 768 sites at once, weigh just over 1 g, suitable for implanting chronically in small mammals. Recordings with high quality signals persisting for at least two months were reliably obtained in two species and six different labs. Improved site density and arrangement combined with new data processing methods enable automatic post-hoc stabilization of data despite brain movements during behavior and across days, allowing recording from the same neurons in the mouse visual cortex for over 2 months. Additionally, an optional configuration allows for recording from multiple sites per available channel, with a penalty to signal-to-noise ratio. These probes and algorithms enable stable recordings from >10,000 sites during free behavior in small animals such as mice.

Multi-electrode arrays (MEAs) are a candidate technology to screen cardiotoxicity in vitro because they enable noninvasive recording of cardiac beating rate, electrical field potential duration, and other parameters. In this paper, we present an active MEA chip featuring 16 384 electrodes, 1024 simultaneous readout channels, and 64 stimulation units (SUs) to enable six different cell-interfacing modalities: extracellular and intracellular recording, current and voltage stimulation, fast impedance motoring, and impedance spectroscopy (IS). The chip was designed and fabricated in a 0.13- $\mu \text{m}$ Al CMOS technology, and the electrodes were fabricated on top of the silicon substrate using biocompatible TiN. Measurement results show a total input-referred noise of $7.5 \pm 0.6~\mu \text{V}_{\mathrm{rms}}$ and $12.0 \pm 2.4~\mu \text{V}_{\mathrm {rms}}$ for the action potential and full bands, respectively, with a total power consumption of 95 mW for 1024 readout channels. We have experimentally validated the CMOS MEA in in vitro experiments, demonstrating all the different modalities. This novel platform will potentially enable high-throughput electrical activity monitoring and drug screening of cardiomyocytes.

Large-scale neural recording requires more advanced and denser brain interfaces. This paper describes the design and fabrication of the CMOS-based Neuropixels neural probe, which integrates a high-density micro-electrode array and a high channel count to enable large-scale electrophysiology in small animals. Miniaturization and scalability aspects are also discussed here.

We review recent progress in neural probes for brain recording, with a focus on the Neuropixels platform. Historically the number of neurons’ recorded simultaneously, follows a Moore’s law like behavior, with numbers doubling every 6.7 years. Using traditional techniques of probe fabrication, continuing to scale up electrode densities is very challenging. We describe a custom CMOS process technology that enables electrode counts well beyond 1000 electrodes; with the aim to characterize large neural populations with single neuron spatial precision and millisecond timing resolution. This required integrating analog and digital circuitry with the electrode array, making it a standalone integrated electrophysiology recording system. Input referred noise and power per channel is 7.5µV and

Although MOS-based pseudo-resistors are capable of achieving the ultra-high resistance required to design small-area neural amplifiers, they suffer from substrate leakage currents which become dramatically higher when the devices are exposed to light. This can cause unwanted amplifier offsets and hinders the usability of these devices in the emerging optoelectrical neural applications. A solution is proposed to compensate the pseudo-resistor leakage current using a dummy deep N-well device. Experimental results show that neural amplifiers implemented with the compensated pseudo-resistor become more robust to photo-induced leakage currents.

Although CMOS fabrication has enabled a quick evolution in the design of high-density neural probes and neural-recording chips, the scaling and miniaturization of the complete data-acquisition systems has happened at a slower pace. This is mainly due to the complexity and the many requirements that change depending on the specific experimental settings. In essence, the fundamental challenge of a neural-recording system is getting the signals describing the largest possible set of neurons out of the brain and down to data storage for analysis. This requires a complete system optimization that considers the physical, electrical, thermal and signal-processing requirements, while accounting for available technology, manufacturing constraints and budget. Here we present a scalable and open-standards-based open-source data-acquisition system capable of recording from over 10,000 channels of raw neural data simultaneously. The components and their interfaces have been optimized to ensure robustness and minimum invasiveness in small-rodent electrophysiology.

Large-scale in vivo electrophysiology requires tools that enable simultaneous recording of multiple brain regions at single-neuron level. This calls for the design of more compact neural probes that offer even larger arrays of addressable sites and high channel counts. With this aim, we present in this paper a quad-shank approach to integrate as many as 5,120 sites on a single probe. Compact fully-differential recording channels were designed using a single-gain-stage neural amplifier with a 14-bit ADC, achieving a mean input-referred noise of 7.44 μVrms in the action-potential band and 7.65 μVrms in the local-field-potential band, a mean total harmonic distortion of 0.17% at 1 kHz and a mean input-referred offset of 169 μV. The probe base incorporates 384 channels with on-chip power management, reference-voltage generation and digital control, thus achieving the highest level of integration in a neural probe and excellent channel-to-channel uniformity. Therefore, no calibration or external circuitry are required to achieve the above-mentioned performance. With a total area of 2.2 × 8.67 mm2 and a power consumption of 36.5 mW, the presented probe enables full-system miniaturization for acute or chronic use in small rodents.

Background: Uniform current density electrodes can reduce the risk of electrode corrosion and tissue damage, but existing geometries do not suit thin planar leads. Objectives: This study examines current density distribution (CDD) on lowprofile electrode designs and proposes an approach to obtain optimal geometries with near-uniform CDD requiring reduced recession depths. Methods: The electrodes are defined as parametric curves, analyzed using the quasi-static approach, and optimized using the downhill simplex method. Results: Obtained electrode designs with near-uniform CDD remain low profile and require twice as less recession depths comparing to reference designs. Conclusion: These electrode shapes are suitable for miniature high-resolution planar DBS leads

Multi-electrode arrays, both active or passive, emerged as ideal technologies to unveil intricated electrophysiological dynamics of cells and tissues. Active MEAs, designed using complementary metal oxide semiconductor technology (CMOS), stand over passive devices thanks to the possibility of achieving single-cell resolution, the reduced electrode size, the reduced crosstalk and the higher functionality and portability. Nevertheless, most of the reported CMOS MEA systems mainly rely on a single operational modality, which strongly hampers the applicability range of a single device. This can be a limiting factor considering that most biological and electrophysiological dynamics are often based on the synergy of multiple and complex mechanisms acting from different angles on the same phenomena. Here, we designed a CMOS MEA chip with 16,384 titanium nitride electrodes, 6 independent operational modalities and 1,024 parallel recording channels for neuro-electrophysiological studies. Sixteen independent active areas are patterned on the chip surface forming a 4 × 4 matrix, each one including 1,024 electrodes. Electrodes of four different sizes are present on the chip surface, ranging from 2.5 × 3.5 μm2 up to 11 × 11.0 μm2, with 15 μm pitch. In this paper, we exploited the impedance monitoring and voltage recording modalities not only to monitor the growth and development of primary rat hippocampal neurons, but also to assess their electrophysiological activity over time showing a mean spike amplitude of 144.8 ± 84.6 μV. Fixed frequency (1 kHz) and high sampling rate (30 kHz) impedance measurements were used to evaluate the cellular adhesion and growth on the chip surface. Thanks to the high-density configuration of the electrodes, as well as their dimension and pitch, the chip can appreciate the evolutions of the cell culture morphology starting from the moment of the seeding up to mature culture conditions. The measurements were confirmed by fluorescent staining. The effect of the different electrode sizes on the spike amplitudes and noise were also discussed. The multi-modality of the presented CMOS MEA allows for the simultaneous assessment of different physiological properties of the cultured neurons. Therefore, it can pave the way both to answer complex fundamental neuroscience questions as well as to aid the current drug-development paradigm.

Deep brain stimulation is an established surgical treatment for several neurological and movement disorders, such as Parkinson's disease, in which electrostimulation is applied to targeted deep nuclei in the basal ganglia through implanted electrode leads. Recent technological improvements in the field have focused on the theoretical advantage of current steering and adaptive (closed-loop) deep brain stimulation. Current steering between several active electrodes would allow for improved accuracy when targeting the desired brain structures. This has the additional benefit of avoiding undesired stimulation of neural tracts that are related to side effects, e.g., internal capsule fibres of passage in subthalamic nucleus deep brain stimulation. Closed-loop deep brain stimulation is based on the premise of continuous recording of a proxy for pathological neural activity (such as beta-band power of measured local field potentials in patients with Parkinson's disease) and accordingly adapting the used stimulus parameters. In this study, we investigate the suitability of an existing highresolution neurorecording probe for high-precision neurostimulation. If a subset of the probe's recording electrodes can be used for stimulation, then the probe would be a suitable candidate for closed-loop deep brain stimulation. A finiteelement model is used to calculate the electric potential, induced by current injection through the high-resolution probe, for different sets of active electrodes. Volumes of activated tissue are calculated and a comparison is made between the highresolution probe and a conventional stimulation lead. We investigate the capability of the probe to shift the volume of activated tissue by steering currents to different sets of active electrodes. Finally, safety limits for the injected current are used to determine the size of the volume in which neurons can be activated with the relatively small electrodes patches on the highresolution probe.