Catalog specification sheet - Intracortical

Utah Microelectrode Array (UEA)

Scanning electron micrograph of a Utah array with a dense square grid of tapered penetrating silicon electrodes. — Utah electrode array, scanning electron micrograph. 2 mm scale retained.

Unwired 100-electrode research array (2014), not the complete current NeuroPort assembly.

Image credit & source

Fernandez et al., Frontiers in Neuroengineering (2014), Fig. 1A · Frontiers · Figure 1A · CC BY 3.0.

Photographic panel cropped from the original figure; resized and re-encoded. Original scale bars and annotations retained.

Code: C3?
Reviewed: 2026-10-07
Interface: intracortical
Evidence stage: human

Independent, source-linked catalog sheet. Not a manufacturer-issued datasheet, regulatory decision or instructions for clinical use. Human evidence does not establish approval. Source-specific restrictions, conflicts and missing specifications are retained below.

Parallel structural variants

Utah arrays

The slanted array targets different depths in peripheral nerves. It does not supersede the standard cortical Utah array.

Version history and parallel configurations (2)

Utah Microelectrode Array (UEA) Standard silicon-array family for cortical recording and stimulation.
Utah Slanted Electrode Array (USEA) Unequal shank lengths sample nerve fascicles at different depths.

Utah Microelectrode Array (UEA)

A silicon array of penetrating electrodes for recording cortical neural activity. The Utah research family and the NeuroPort clinical product share an array architecture, but their configurations, labeling and permitted uses are not interchangeable.

Primary manufacturer: Blackrock Neurotech. The University of Utah is the array’s origin; the BrainGate consortium is a clinical research user. These are different roles, rather than three interchangeable manufacturers.

The BrainGate pilot used a 96-microelectrode sensor for a human neural cursor and device control.

Core interface specifications

The 12 fields follow the Blackrock Utah table for comparison, not a manufacturer-issued datasheet. Missing values stay blank; inapplicable fields are marked. Configuration-specific details and limits follow below.

Field	Specification and source scope
Electrode Pitch	400 µm; current Utah manufacturer page
Channel Count	16-1024 in the current system/options table; not one array. Overview says up to 96 per array, FAQ 96-128. NeuroPort IFU: 100 physical, 96 connected
Output Connectors	Utah options: Omnetics, CerePort pedestal 128/256, Custom
Output Conn. dimensions L x W x H	Utah page: Omnetics 7 × 37 × 9 mm; CerePort 128 16.5 × 12 × 19 mm and 256 16.5 × 10.7 × 19 mm (height × neck diameter × base diameter). Research IFU instead gives Omnetics 9 mm height, 7 mm base length, 13 mm base width; conflict retained
Standard Electrode Lengths	Utah: 0.5-1.5 mm research; 1.0-1.5 mm clinical. NeuroPort configuration: 1.0 or 1.5 mm
Impedance	Utah page: platinum 20-800 kΩ; SIROF/IrOx 1-80 kΩ; frequency not stated. NeuroPort values and IFU differences retained below
Array Dimensions	Utah table literally says “Customizable from 2 - 12” with no unit. FAQ: 4 × 4 mm footprint, 0.2 mm substrate; not a universal customized layout
Multi-Port Options	1, 2, 3, 4, Utah options table; not equivalent to the electrode count
Metalization	Platinum or sputtered iridium oxide (SIROF). Utah page repeats impedance ranges in this row; those are not material dimensions
Wire Bundle Length	Utah options: customizable 20-130 mm. NeuroPort page: 13 cm long, 0.55 mm wide
Reference and Ground	Utah page: “Ground Source and Selectable Reference Wires”. Assembly pin mapping is configuration-specific, not inferred from a connector name
Insulation	Parylene-C, Utah options table; not a lifetime or MR-safety qualification

Identity

Field	Value and source scope
Device	Utah Microelectrode Array research family; NeuroPort is the human clinical configuration
Manufacturer	Blackrock Neurotech (formerly Blackrock Microsystems)
Interface class	Intracortical, penetrating silicon microelectrode array
Origin	University of Utah
First demonstrated	Not restated without a cited source; the origins entry covers early manufacturing history
First human implant	Earlier unsourced dates were removed rather than kept
Species studied	Human (NeuroPort programs) and non-human primate (2013 Barrese cohort)
Regulatory status	NeuroPort April 2022 IFU: temporary recording and monitoring for less than 30 days; research arrays are investigational
Function	Recording; the research page also describes stimulation. The NeuroPort labeling conflict is kept below
Target tissue	Cerebral cortex

Geometry and architecture

Field	Value and source scope
Interface type	Penetrating intracortical array
Array layout	NeuroPort: 10 x 10 grid; research family configurable
Electrode count	100 physical, 96 connected (NeuroPort IFU); conflicting manufacturer marketing counts are preserved in Published geometry
Pitch	400 µm, current Utah and NeuroPort pages
Electrode lengths	0.5-1.5 mm research; 1.0-1.5 mm clinical
Shank width / thickness	Utah FAQ gives a 4 x 4 mm footprint, 0.2 mm substrate
Tip and exposed site geometry	Not reconstructed
Contact coating	Platinum or sputtered iridium oxide (SIROF)
Insulation	Parylene-C, Utah options table
Insertion method	Configuration-dependent tooling; specifics not reported
Anchoring and fixation	Percutaneous pedestal or connector; array fixation detail not reported

Electrode and channel physics

Field	Value and source scope
Exposed site area
Electrode material	Silicon shank with platinum or SIROF contact metallization
Impedance (with measurement frequency)	NeuroPort page: platinum 100-800 kΩ, SIROF at most 50 kΩ at 1 kHz; Utah table: platinum 20-800 kΩ, SIROF/IrOx 1-80 kΩ with frequency unstated; 2022 IFU: 100-800 kΩ platinum, 1-80 kΩ iridium oxide. Source-specific figures kept, not merged
Noise floor	Acquisition-system dependent
Recording modality	Single- and multi-unit spiking and local field potentials in published use
Sampling rate	Configuration dependent; set by the external acquisition system
Stimulation capability	Research arrays: described by the manufacturer. NeuroPort IFU: recording-only labeling; conflict retained in Recording and stimulation
Charge injection limit
Reference and ground	”Ground Source and Selectable Reference Wires”, Utah page

Tissue interface and bioresponse

Field	Value and source scope
Target tissue	Cerebral cortex
Insertion trauma / BBB disruption	Inherent to penetrating insertion; no quantitative value
Vascular disruption risk	Qualitative risk of penetrating placement; not reported quantitatively
Micromotion sensitivity	Rigid silicon array in soft tissue; qualitative concern, no reviewed quantitative source
Gliosis and encapsulation	Meningeal reactions and insulation degradation documented in the 2013 macaque failure cohort
Neuron loss near sites
Foreign-body response mitigation
Typical failure modes	Connector failures, meningeal reactions, insulation degradation and gradual channel loss in the 2013 cohort

System architecture

Field	Value and source scope
Onboard electronics	None on the array; passive electrodes routed to a percutaneous connector
Data path	Wired percutaneous connection to external acquisition hardware
Telemetry bandwidth	Not applicable; wired
Sampling rate	Configuration dependent
Power	External
Thermal management	External
Packaging and hermeticity	Not a hermetically sealed implant; percutaneous connector
MRI compatibility
Surgical complexity	Craniotomy, array insertion and percutaneous pedestal placement
Output connectors	Omnetics, CerePort pedestal 128/256, custom (Utah options)

Performance envelope

Field	Value and source scope
Acute yield	Not given as a single figure
Chronic yield	2013 macaque cohort: recording durations 0-2,104 days, median 182 days; 62 of 78 arrays failed completely
Stability over time	Variable across the cohort; see Long-term evidence and failures
Longevity	Manufacturer claims more than eight years in one patient; a single-patient marketing claim, not a cohort result
Revision and explant experience
Adverse events	Percutaneous infection risk is noted qualitatively; no rate
Notable demonstrations	BrainGate human cursor and device control (BrainGate pilot)

Clinical and preclinical evidence

Field	Value and source scope
Human subjects	Documented in the BrainGate pilot and subsequent programs; a single audited participant count is not pinned here
Preclinical cohort	78 arrays in 27 rhesus macaques, 2013 failure analysis
Follow-up duration	Up to 2,104 days in the macaque cohort; human durations vary by program
Indications	Research use for paralysis and motor impairment
Trials and registries	BrainGate pilot entry; registry identifiers not pinned in this sheet
Primary outcomes	Cursor, communication and device control in published studies
Key limitations	Heterogeneous configurations, variable chronic performance, and manufacturer-versus-study figure conflicts that are preserved rather than merged

Engineering tradeoffs

Aspect	Assessment
Strengths	Decades of research use, direct cortical access, demonstrated chronic human recording, configurable research family
Limitations	Rigid penetrating silicon, percutaneous connector burden, chronic signal loss, configuration and labeling inconsistencies
Scaling constraints	Per-channel wiring and connector complexity, chronic tissue response, surgical placement time

Published geometry

Field	Source and configuration
Electrode pitch	400 µm, current Utah and NeuroPort product pages
Utah research lengths	0.5 to 1.5 mm, current product page
NeuroPort lengths	1.0 or 1.5 mm, current product page and April 2022 IFU
NeuroPort layout	10 x 10, current product page
NeuroPort electrode count	100 physical electrodes, 96 connected to the percutaneous connector, April 2022 IFU
Utah substrate	4 x 4 mm footprint and 0.2 mm thickness, current Utah FAQ
Tip coatings	Platinum or sputtered iridium oxide film (SIROF), current product pages
NeuroPort wire bundle	13 cm long, 0.55 mm wide, current product page

The manufacturer does not give one consistent electrode count for the whole Utah family. Its current overview says “up to 96 electrodes per array”; the FAQ says “up to 128”, “100 - 128” microneedles and “96 - 128” electrodes. Those figures are preserved rather than treated as the same configuration. Connector names such as CerePort 128 do not, by themselves, establish the number of connected recording sites.

The 1991 manufacturing geometry describes a 4.2 mm substrate. That is historical geometry, not a replacement for the current manufacturer’s 4 mm FAQ figure.

Recording and stimulation

The Utah research page describes recording and stimulation. NeuroPort labeling needs its own distinction: the April 2022 NeuroPort Electrode IFU lists temporary recording and monitoring for less than 30 days, and says this recording device should not be used in applications involving stimulation. The current NeuroPort marketing page describes recording and stimulation. This entry does not erase that conflict or treat a marketing page as clinical authorization. Long-term research use and individual clinical protocols are separate from the IFU’s labeled use.

The NeuroPort product page lists platinum impedance of 100 to 800 kΩ and SIROF impedance of at most 50 kΩ at 1 kHz. The 2022 IFU lists 100 to 800 kΩ for its platinum configurations and 1 to 80 kΩ for its iridium-oxide configurations. These are source- and configuration-specific figures, not one generic Utah impedance range.

Long-term evidence and failures

The 2013 retrospective failure study followed 78 arrays in 27 rhesus macaques. Recording duration ranged from 0 to 2,104 days, with a median of 182 days across all arrays. Sixty-two failed completely; nine experiments ended electively and seven were still active at study close. Connector failures, meningeal reactions and insulation degradation were important findings. These are cohort results, not a service-life warranty or a direct estimate of modern human implant survival.

The current NeuroPort page claims recording and stimulation for more than eight years in one patient. That manufacturer claim is not the same evidence as the historical macaque cohort. The older study’s approximately eight-year complete-signal-loss estimate was a prediction from trends, not an observed eight-year survival result.

Model limits

The site’s Utah model is a reference array geometry, not a reconstruction of every sold configuration or a clinical implant assembly. The sources above do not specify a complete current shank profile, tip exposure geometry or every packaging component. No new geometry is inferred here.

Sources

Blackrock Neurotech. Utah Array, current product page. Read 7 October 2026.
Blackrock Neurotech. NeuroPort Electrode 96, current product page. Read 7 October 2026.
Blackrock Microsystems. NeuroPort Electrode IFU, revision 3.00, April 2022.
Barrese JC et al. Failure mode analysis of silicon-based intracortical microelectrode arrays in non-human primates. Journal of Neural Engineering, 2013.
Blackrock Microsystems. Research Arrays IFU, LB-0514 revision 5.00, November 2020; connector dimensions differ from the current web table.

Recorded interface · Recorded from 96-channel Utah arrays in BrainGate studies. The viewer uses author-published session summaries; it does not reconstruct voltage waveforms.

Published array-session measurements

Utah arrays · recording over time

Measured data

Explore electrode yield, impedance and waveform summary metrics across 20 arrays in 14 BrainGate participants. Each point is a published array-session measurement.

Loading study measurements…

Open the complete Recording Library →

Models and reference visualizations

3D GEOMETRY · REFERENCE MODEL

Utah Array · 10 × 10 · 1.5 mm

Electrode-bearing structure · dimensions in millimeters

Loading 3D viewer…

Physical shanks100 (10 × 10)

Wired channelsTypically 96; map not assigned

Pitch400 µm center to center

Substrate4 × 4 × 0.2 mm

Shank length1.5 mm selected variant

Model fidelity. Published pitch, footprint, substrate thickness and selected shank length. Maximum shaft radius (40 µm), the smoothly tapering profile with basal flare, and the 80 µm tip region are illustrative, not measured contact area. The taper follows the source silhouette without claiming a recovered fabrication profile. All 100 physical shanks are shown; no acquisition channel map is inferred. Leads and connector are omitted.

Color, transparency and surface finish approximate the references. Use Highlight contacts to inspect electrode locations.

Download 3D model (.glb)

Sources & reference images

Dimensions come from the linked specifications; images guide shape only. Reference images remain on their owners’ websites.

JSON includes physical site coordinates, unknown channel mappings and approximation notes. GLB uses meters. These models do not predict recording or stimulation performance.

REAL HUMAN DATA · BRAINGATE2

From an array to a real signal

Explore attempted handwriting recorded from two 96-channel Utah arrays in participant T5's motor cortex. These are measured recordings, not simulated signals.

Loading the recording excerpts...

Source, license & limitations

Willett, F. R., Avansino, D. T., Hochberg, L. R., Henderson, J. M. & Shenoy, K. V. (2021). High-performance brain-to-text communication via handwriting. BrainGate2 study participant T5.

Dryad dataset · doi:10.5061/dryad.wh70rxwmv · CC0 public-domain dedication and archive mirror. Three excerpts from the December 18, 2019 session; original 10 ms counts retained without invented spike timing. Only the rate trace is averaged.

Threshold crossings were detected at −3.5×RMS, per the dataset documentation. These files do not contain raw voltage or exact spike timestamps. The page does not run a decoder, reconstruct handwriting, or claim these samples represent all implants or participants. Channel order follows the dataset, not the 3D model's physical contact mapping.

REAL HUMAN DATA · LIVE DECODE IN YOUR BROWSER

Watch letters emerge from the signal

A decoder trained on four of T5's calibration sessions scores every 20 ms slice of these held-out recordings against 31 characters. Pick a moment to inspect its top five guesses.

Loading the decoder...

Devices