FAQ

NeuroNexus makes silicon neural probes — 210+ catalog designs in 2,000+ variants — and thin-film polymer grid arrays, 50+ catalog designs, for penetrating and surface recording, plus specialty electrodes for stimulation, optogenetics and drug delivery.

Start from whatever is fixed first: browse electrodes by target, modality, or signal, or use the Product Finder to go from your requirements to a matched configuration. Solution Bundles pair probes, acquisition, and accessories that work together as a starting point you customize, and application scientists will help you equip the study end to end.

Yes. NeuroNexus probes work with third-party acquisition systems, including Open Ephys and Intan, through standard connectors or a NeuroNexus adapter that mates a probe package to hardware it was not terminated for; Radiens reads those systems' recordings.

A probe package is the connector assembly built onto the electrode array: the interconnect from the bond pads to the connector (a rigid board, a flexible cable, or none where the connector sits on the assembly), the connector a headstage mates to, the covering over the bond region, and the ground and reference wires. A finished probe is the array plus the package, the same array is ordered on different packages, and the package is built on at assembly, so it is decided once. It is selected by the experimental preparation (acute, chronic, freely moving), what the subject can bear, and plug-in compatibility with the headstage; a package name reads as series plus channel count (A32, CM32, H64).

An H-style (Hybrid) package mounts a silicon probe to one end of a thin-film polyimide ribbon cable, with the other end of the cable attached to a PCB and connector. The assembly mechanically uncouples the probe from the connector; the thin-film cable is very flexible and can even twist slightly. The "H" stands for hybrid assembly, the hybrid of a thin-film silicon probe and a thin-film polymer component, silicon and polymer. It is the chronic package for a floating implant, a microdrive or connector stand-off, at 16 to 128 channels on Omnetics, Hirose or MRI Omnetics connectors.

Yes. It is a configuration of the NeuroNexus connector-pair adapter: input connector Omnetics 32, output connector Omnetics 16 (SKU adpt_nnx--pair.om32.om16, NetSuite item Adpt.Omnetics32-Omnetics16). A 2× version splits one Omnetics 32 onto two Omnetics 16 (adpt_nnx--pair.om32.om16.x2), and the reverse direction, Omnetics 16 to Omnetics 32, is adpt_nnx--pair.om16.om32. Choose the pairing on the adapter's product page and request a quote from the cart.

They are NeuroNexus package connector codes. A16 is the 16-channel A-series package on a DIP-16 (Mill-Max) connector; A32 is the 32-channel A-series package on a Samtec 40-pin connector; A64 is two Samtec 40. Z16, Z32 and Z64 are Hirose ZIF connectors for TDT ZIF-Clip headstages. OM16, OM32 and OM64 are Omnetics nano connectors with 16, 32 and 64 channels; OM8 is the 8-channel Omnetics. CQ4 and Q4 are the edge connectors on Q-Trode packages. An adapter name reads input to output: A32 → Omnetics 32 takes an A32 package and presents an Omnetics 32.

With the same 177 µm² sites: the Poly3 25s designs, A1x32-Poly3-5mm-25s-177 and A1x32-Poly3-10mm-25s-177, three staggered columns with 25 µm between neighboring sites and a site every 12.5 µm along the shank. There is no 32-site Poly2 design with closer spacing than 50s (A1x32-Poly2-3mm, 5mm and 10mm-50s-177: 50 µm between neighbors, a site every 25 µm along the shank), and A1x32-Poly3-6mm-50-177 is 50 µm with unstaggered columns. At other site areas, the Poly5 designs are tighter still: A1x32-Poly5-6mm-35s-100 (17.5 µm along the shank, 98 µm² sites) and A1x32-Poly5-6mm-40s-225 (20 µm, 220 µm² sites). If 64 sites are acceptable, V1x64-Poly3-10mm-25s-177 (12.5 µm along the shank, on a Vector support body) and A1x64-Poly2-6mm-23s-160 (23 µm) are the catalog options; the package and headstage change with the site count. In a name, the number before the site area is the spacing between neighboring sites and an s means the columns are staggered; the along-shank pitch is on the datasheet.

Substrate thickness sets bending stiffness, the tissue the shank displaces, handling, and how deep a probe can be driven without buckling. Catalog designs come at 15 and 50 µm: the thinner substrate displaces less tissue and bends more; the thicker holds a straight trajectory on long insertions. For deep targets in large animals the Vector Array mounts a 10 to 15 mm silicon array on a rigid 70 or 110 mm support body. The process also reaches 5 µm, a thickness at which less tissue encapsulation has been reported (Seymour and Kipke, 2007); that is a capability discussed per design, not a catalog line.

Yes. NeuroNexus silicon probes are used in acute and chronic studies, and the published chronic record runs to months: Okun et al. (2016) reported long-term recordings with immobile silicon probes in mouse cortex, Haghighi et al. (2025) implanted four-shank silicon arrays in rat somatosensory cortex for 16 weeks, and Mueller et al. (2025) used a 16-site NeuroNexus probe as the silicon comparator in a 12-week chronic study in rat. The amorphous silicon carbide insulation is the material choice behind chronic lifetime, and the package decides the implant method.

Yes. The catalog spans the site counts, shank lengths, geometries and thicknesses the range of animal models needs. Published examples run from Drosophila (Van De Poll and van Swinderen, 2025, long-term multichannel recordings across sleep and wake) to macaques (Thirunavukkarasu et al., 2026, laminar probes in the supplementary eye field; Perich et al., 2026, deep brain structures with Vector Arrays). The citation database holds the rest.

A specification describes a design; a probe is a design that has been through a process and a package. Site size, trace width, dielectric integrity, bond quality, and encapsulation all have tolerances that never appear on a spec sheet and each changes what the probe records. NeuroNexus controls the four determinants of quality, design, fabrication process, materials, and package assembly, on one proprietary process, so every probe of a design is identical rather than similar. The determinants and how to check them are in What Makes a Thin-Film Electrode Good.

Every catalog design has a documented datasheet, and every probe of a design is made under the same batch conditions on the same proprietary process. Arrays are inspected after release from the wafer, and after packaging every site is impedance-tested; that factory measurement is recorded and is the baseline a laboratory compares against on first use and after every session. Site yield and impedance consistency are tracked from probe to probe and lot to lot. NeuroNexus controls the design, the process, and the intellectual property end to end.

Measure the impedance of every site in saline before mounting and compare the distribution with the probe's own baseline: the factory record on first use, your last measurement afterward. A good array reads tightly, for 177 µm² iridium sites mostly between 200 kΩ and 1 MΩ at 1 kHz; a site far above the cluster has a broken trace, an open bond, or a contaminated surface, one far below has a short. Inspect under a microscope for clean, uniform sites and intact bonds, and keep a per-site log across sessions. Silicon Probes: Setup, Handling, Cleaning and Care walks through it.

NeuroNexus technology has appeared in more than 5,000 peer-reviewed publications across neuroscience and bioengineering, in more than 1,000 laboratories on six continents, in species from fruit flies to non-human primates and in acute and months-long chronic preparations. For a study it means the probe geometry, materials, and packages have been used and reported under conditions close to yours, that reviewers know the devices, and that NeuroNexus probes are the comparator other groups measure new probe designs against. The NeuroNexus scientific publication database lists the publications by product family.

Application Scientists help match the probe geometry and package to the preparation before the order and remain an inquiry or a phone call away afterward; NeuroNexus engineers and scientists hold more than 150 years of cumulative neural-interface design experience and run NIH-funded research programs in-house. Every order is backed by the documentation on this site: surgical guides for acute and chronic preparations, setup, handling, cleaning, and care procedures, a glossary of every design parameter, connectivity and compatibility guides, and troubleshooting. Custom geometries, materials, and packaging are designed with the laboratory and built on the same process with the same quality controls.

No. The thin-film grid sits on the intact skull over the recording area, with no bone removed. Bone screws placed clear of the grid anchor the implant and provide the ground and reference, and the grid is cemented at its edges and at the cable exit.

It records for weeks to months, so a biomarker can be followed through development, or a treatment measured against a baseline, on the same implant. Impedance that rises over weeks usually means the cement is lifting at an edge.

The 30- and 32-site grids end in an H32 Hybrid package that mates a 32-channel headstage such as SmartLink 32 or X3R 32; the 16-site grid ends in an H16 for SmartLink 16. Each runs on Smartbox Pro or XDAQ, with Allego for acquisition and Radiens for analysis.

Yes. Jonak et al. (2018) describe a reusable 30-channel technique for EEG in awake, freely moving mice on the NeuroNexus mouse grid. Handling and cleaning between uses are in Thin-Film Grid Electrodes: Setup, Handling, Cleaning and Care.

No. Neuropixels probes are made and sold by imec. NeuroNexus offers the equipment and software a Neuropixels rig runs on: XDAQ acquisition, the Kontex NP Headstage, the Commutator PRO and Radiens. It also offers two alternatives to the probe: SiNAPS, the other active-pixel probe, and dense passive silicon probes.

Yes, two. SiNAPS is the other active-pixel probe: it digitizes on the shank and records every one of its 256 or 1,024 sites at once, where Neuropixels records 384 chosen from 960 to 5,120 sites (1,536 on the Quad Base). Dense passive silicon probes, such as Buzsáki and polytrode layouts, are functionally comparable for some applications and stimulate on the recording sites.

SiNAPS records every site at once, 1,024 on four or eight shanks, against 384 for Neuropixels 1.0 and 2.0 and 1,536 for the 2.0 Quad Base, which choose their channels from 960 to 5,120 sites. Both are 12-bit, except Neuropixels 1.0 at 10-bit, with comparable noise. Neuropixels has the thinner, 10 mm shank and samples at 30 kS/s against 20. Neither stimulates.

Not through the Neuropixels probe, which records only. Stimulation goes through a passive probe with stimulation-rated sites on an X-Series record-and-stimulate headstage. The Commutator PRO carries both on one torque-compensated joint, up to four Neuropixels ports and a passive probe at up to 64 record-and-stimulate channels, into XDAQ, so a freely moving animal is stimulated while Neuropixels records.

Every XDAQ has a Neuropixels mode. The XDAQ ONE+ takes up to four Neuropixels probes, 384 channels each, beside its passive channels; the CORE2 and ALL-IN-ONE record up to 1,024+ Neuropixels channels. XDAQ in NP mode runs on Kontex's acquisition software, which uses the Open Ephys plugin for the Neuropixels interface, and the Kontex NP Headstage connects Neuropixels 1.0 and 2.0 probes.

Yes. The Commutator PRO has up to four Neuropixels ports. A motor, driven by a gyroscope on the animal or a Hall sensor on the rotor, cancels the cable torque, so a freely moving animal supplies no force and the recording stays stable. Its HDMI port carries a passive probe at the same time, and an optional 5 mm passthrough carries a fiber or fluid line.