Ultra-high density electrodes improve detection, yield, and cell type identification in neuronal recordings.

To understand the neural basis of behavior, it is essential to sensitively and accurately measure neural activity at single neuron and single spike resolution. Extracellular electrophysiology delivers this, but it has biases in the neurons it detects and it imperfectly resolves their action potentials. To minimize these limitations, we developed a silicon probe with much smaller and denser recording sites than previous designs, called Neuropixels Ultra (NP Ultra). This device samples neuronal activity at ultra-high spatial density (~10 times higher than previous probes) with low noise levels, while trading off recording span. NP Ultra is effectively an implantable voltage-sensing camera that captures a planar image of a neuron's electrical field. We use a spike sorting algorithm optimized for these probes to demonstrate that the yield of visually-responsive neurons in recordings from mouse visual cortex improves up to ~3-fold. We show that NP Ultra can record from small neuronal structures including axons and dendrites. Recordings across multiple brain regions and four species revealed a subset of extracellular action potentials with unexpectedly small spatial spread and axon-like features. We share a large-scale dataset of these brain-wide recordings in mice as a resource for studies of neuronal biophysics. Finally, using ground-truth identification of three major inhibitory cortical cell types, we found that these cell types were discriminable with approximately 75% success, a significant improvement over lower-resolution recordings. NP Ultra improves spike sorting performance, detection of subcellular compartments, and cell type classification to enable more powerful dissection of neural circuit activity during behavior.

Large-scale high-density brain-wide neural recording in nonhuman primates.

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.

High-density electrophysiological recordings in macaque using a chronically implanted 128-channel passive silicon probe.

OBJECTIVE: The analysis of interactions among local populations of neurons in the cerebral cortex (e.g. within cortical microcolumns) requires high resolution and high channel count recordings from chronically implanted laminar microelectrode arrays. The request for high-density recordings of a large number of recording sites can presently only be accomplished by probes realized using complementary metal-oxide-semiconductor (CMOS) technology. In preparation for their use in non-human primates, we aimed for neural probe validation in a head-fixed approach analyzing the long-term recording capability. APPROACH: We examined chronically implanted silicon-based laminar probes, realized using a CMOS technology in combination with micromachining, to record from the primary visual cortex (V1) of a monkey. We used a passive CMOS probe that had 128 electrodes arranged at a pitch of 22.5 µm in four columns and 32 rows on a slender shank. In order to validate the performance of a dedicated microdrive, the overall dimensions of probe and interface boards were chosen to be compatible with the final active CMOS probe comprising integrated circuitry. MAIN RESULTS: Using the passive probe, we recorded simultaneously local field potentials (LFP) and spiking multiunit activity (MUA) in V1 of an awake behaving macaque monkey. We found that an insertion through the dura and subsequent readjustments of the chronically implanted neural probe was possible and allowed us to record stable LFPs for more than five months. The quality of MUA degraded within the first month but remained sufficiently high to permit mapping of receptive fields during the full recording period. SIGNIFICANCE: We conclude that the passive silicon probe enables semi-chronic recordings of high quality of LFP and MUA for a time span exceeding five months. The new microdrive compatible with a commercial recording chamber successfully demonstrated the readjustment of the probe position while the implemented plug structure effectively reduced brain tissue movement relative to the probe.

Neuropixels Data-Acquisition System: A Scalable Platform for Parallel Recording of 10 000+ Electrophysiological Signals.

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.

Fine-scale mapping of cortical laminar activity during sleep slow oscillations using high-density linear silicon probes.

BACKGROUND: The cortical slow (∼1 Hz) oscillation (SO), which is thought to play an active role in the consolidation of memories, is a brain rhythm characteristic of slow-wave sleep, with alternating periods of neuronal activity and silence. Although the laminar distribution of cortical activity during SO is well-studied by using linear neural probes, traditional devices have a relatively low (20-100 µm) spatial resolution along cortical layers. NEW METHOD: In this work, we demonstrate a high-density linear silicon probe fabricated to record the SO with very high spatial resolution (∼6 µm), simultaneously from multiple cortical layers. Ketamine/xylazine-induced SO was acquired acutely from the neocortex of rats, followed by the examination of the high-resolution laminar structure of cortical activity. RESULTS: The probe provided high-quality extracellular recordings, and the obtained cortical laminar profiles of the SO were in good agreement with the literature data. Furthermore, we could record the simultaneous activity of 30-50 cortical single units. Spiking activity of these neurons showed layer-specific differences. COMPARISON WITH EXISTING METHODS: The developed silicon probe measures neuronal activity with at least a three-fold higher spatial resolution compared with traditional linear probes. By exploiting this feature, we could determine the site of up-state initiation with a higher precision than before. Additionally, increased spatial resolution may provide more reliable spike sorting results, as well as a higher single unit yield. CONCLUSIONS: The high spatial resolution provided by the electrodes allows to examine the fine structure of local population activity during sleep SO in greater detail.

A silicon-based neural probe with densely-packed low-impedance titanium nitride microelectrodes for ultrahigh-resolution in vivo recordings.

In this study, we developed and validated a single-shank silicon-based neural probe with 128 closely-packed microelectrodes suitable for high-resolution extracellular recordings. The 8-mm-long, 100-µm-wide and 50-µm-thick implantable shank of the probe fabricated using a 0.13-µm complementary metal-oxide-semiconductor (CMOS) metallization technology contains square-shaped (20 × 20 µm2), low-impedance (~ 50 kΩ at 1 kHz) recording sites made of rough and porous titanium nitride which are arranged in a 32 × 4 dense array with an inter-electrode pitch of 22.5 µm. The electrophysiological performance of the probe was tested in in vivo experiments by implanting it acutely into neocortical areas of anesthetized animals (rats, mice and cats). We recorded local field potentials, single- and multi-unit activity with superior quality from all layers of the neocortex of the three animal models, even after reusing the probe in multiple (> 10) experiments. The low-impedance electrodes monitored spiking activity with high signal-to-noise ratio; the peak-to-peak amplitude of extracellularly recorded action potentials of well-separable neurons ranged from 0.1 mV up to 1.1 mV. The high spatial sampling of neuronal activity made it possible to detect action potentials of the same neuron on multiple, adjacent recording sites, allowing a more reliable single unit isolation and the investigation of the spatiotemporal dynamics of extracellular action potential waveforms in greater detail. Moreover, the probe was developed with the specific goal to use it as a tool for the validation of electrophysiological data recorded with high-channel-count, high-density neural probes comprising integrated CMOS circuitry.

Time Multiplexed Active Neural Probe with 1356 Parallel Recording Sites.

We present a high electrode density and high channel count CMOS (complementary metal-oxide-semiconductor) active neural probe containing 1344 neuron sized recording pixels (20 µm × 20 µm) and 12 reference pixels (20 µm × 80 µm), densely packed on a 50 µm thick, 100 µm wide, and 8 mm long shank. The active electrodes or pixels consist of dedicated in-situ circuits for signal source amplification, which are directly located under each electrode. The probe supports the simultaneous recording of all 1356 electrodes with sufficient signal to noise ratio for typical neuroscience applications. For enhanced performance, further noise reduction can be achieved while using half of the electrodes (678). Both of these numbers considerably surpass the state-of-the art active neural probes in both electrode count and number of recording channels. The measured input referred noise in the action potential band is 12.4 µVrms, while using 678 electrodes, with just 3 µW power dissipation per pixel and 45 µW per read-out channel (including data transmission).

A Neural Probe With Up to 966 Electrodes and Up to 384 Configurable Channels in 0.13 $\mu$m SOI CMOS.

In vivo recording of neural action-potential and local-field-potential signals requires the use of high-resolution penetrating probes. Several international initiatives to better understand the brain are driving technology efforts towards maximizing the number of recording sites while minimizing the neural probe dimensions. We designed and fabricated (0.13- µm SOI Al CMOS) a 384-channel configurable neural probe for large-scale in vivo recording of neural signals. Up to 966 selectable active electrodes were integrated along an implantable shank (70 µm wide, 10 mm long, 20  µm thick), achieving a crosstalk of [Formula: see text] dB. The probe base (5 × 9 mm 2 ) implements dual-band recording and a 171.6 Mbps digital interface. Measurement results show a total input-referred noise of 6.4 µ V rms and a total power consumption of 49.1  µW/channel.

Ultra-high-density in-vivo neural probes.

The past decade has witnessed an explosive growth in our ability to observe and measure brain activity. Among different functional brain imaging techniques, the electrical measurement of neural activity using neural probes provides highest temporal resolution. Yet, the electrode density and the observability of currently available neural probe technologies fall short of the density of neurons in the brain by several orders of magnitude. This paper presents opportunities for neural probes to utilize advances in CMOS technology for increasing electrode density and observability of neural activity, while minimizing the tissue damage. The authors present opportunities for neural probes to adapt advanced CMOS technologies and discuss challenges in terms of maintaining the signal integrity and implementing data communication.

Evaluation of intrinsic image algorithms to detect the shadows cast by static objects outdoors.

In some automatic scene analysis applications, the presence of shadows becomes a nuisance that is necessary to deal with. As a consequence, a preliminary stage in many computer vision algorithms is to attenuate their effect. In this paper, we focus our attention on the detection of shadows cast by static objects outdoors, as the scene is viewed for extended periods of time (days, weeks) from a fixed camera and considering daylight intervals where the main source of light is the sun. In this context, we report two contributions. First, we introduce the use of synthetic images for which ground truth can be generated automatically, avoiding the tedious effort of manual annotation. Secondly, we report a novel application of the intrinsic image concept to the automatic detection of shadows cast by static objects in outdoors. We make both a quantitative and a qualitative evaluation of several algorithms based on this image representation. For the quantitative evaluation, we used the synthetic data set, while for the qualitative evaluation we used both data sets. Our experimental results show that the evaluated methods can partially solve the problem of shadow detection.

Social network extraction and analysis based on multimodal dyadic interaction.

Social interactions are a very important component in people's lives. Social network analysis has become a common technique used to model and quantify the properties of social interactions. In this paper, we propose an integrated framework to explore the characteristics of a social network extracted from multimodal dyadic interactions. For our study, we used a set of videos belonging to New York Times' Blogging Heads opinion blog. The Social Network is represented as an oriented graph, whose directed links are determined by the Influence Model. The links' weights are a measure of the "influence" a person has over the other. The states of the Influence Model encode automatically extracted audio/visual features from our videos using state-of-the art algorithms. Our results are reported in terms of accuracy of audio/visual data fusion for speaker segmentation and centrality measures used to characterize the extracted social network.

Three-dimensional face pose detection and tracking using monocular videos: tool and application.

Recently, we have proposed a real-time tracker that simultaneously tracks the 3-D head pose and facial actions in monocular video sequences that can be provided by low quality cameras. This paper has two main contributions. First, we propose an automatic 3-D face pose initialization scheme for the real-time tracker by adopting a 2-D face detector and an eigenface system. Second, we use the proposed methods-the initialization and tracking-for enhancing the human-machine interaction functionality of an AIBO robot. More precisely, we show how the orientation of the robot's camera (or any active vision system) can be controlled through the estimation of the user's head pose. Applications based on head-pose imitation such as telepresence, virtual reality, and video games can directly exploit the proposed techniques. Experiments on real videos confirm the robustness and usefulness of the proposed methods.