Devices

A 10×10 silicon intracortical microelectrode array (typically 96 wired channels): excellent spikes early, but challenging chronic stability.

Device — Intracortical

Utah Microelectrode Array (UEA)

Official site →

BCI · intracortical · penetrating · microelectrode array · Utah Array · UEA

Utah Microelectrode Array (UEA)

One-line verdict: A rigid silicon intracortical microelectrode array that delivers high-quality spikes early, but pays a chronic penalty from penetration trauma, micromotion mismatch, and gliosis-driven signal loss.

Quick tags: Recording · Stimulation · Closed-loop (system-dependent) · Channels: 96–128 · Species: Human/NHP/rodent · First implanted: ~1998

Overview

What it is: The Utah array is a 10×10 bed of silicon shanks inserted into cortex (often motor cortex). It is typically wired out to external amplifiers/recorders via a percutaneous connector in many classic research deployments.

Why it matters: It is the workhorse design behind a large fraction of landmark intracortical human BCI results (cursor control, typing, robotic arm control), and it defines the “spike-first” performance baseline.

Most comparable devices: microwire bundles, other penetrating intracortical arrays, flexible thread-based intracortical systems.

Spec Card Grid

Identity

  • Device name: Utah Microelectrode Array (UEA)
  • Canonical ID: BTSD-0001
  • Inventor / key authors: Richard Normann (University of Utah)
  • Org / manufacturer: Blackrock Neurotech (modern manufacturing)
  • First demonstrated (year): ~1992 (prototype)
  • First implanted (year): ~1998 (human)
  • Species: Human, NHP, rodent
  • Regulatory / trial status: Human research (IDE)
  • Primary use: Recording + stimulation
  • Primary target: Motor cortex (common), other cortical targets

Geometry & Architecture

  • Interface type: Intracortical
  • Penetrating?: yes
  • Form factor: shank array (silicon)
  • Array layout: 10×10 needle bed
  • Footprint (mm): ~4 × 4
  • Insertion depth (mm): ~1.0–1.5 (typical human motor cortex)
  • Shank / lead dimensions: shank width ~80 µm (length above)
  • Site spacing (µm): 400
  • Tip geometry: sharpened silicon
  • Insertion method: pneumatic impactor
  • Anchoring method: percutaneous pedestal / skull-mounted connector (system-dependent)
  • Packaging location: often percutaneous in classic research stacks

Electrode & Channel Physics

  • Channel count: 96–128 (typical wired)
  • Active sites used (vs total): typically 96 recording channels in many deployed arrays
  • Electrode material: platinum / iridium
  • Site area (µm²): ~200–400
  • Impedance @ 1 kHz: ~100–500 kΩ
  • Noise floor / SNR: varies by system; spikes are often strong acutely
  • Recording modality: spikes + LFP
  • Stimulation capability: yes
  • Charge injection limit / safe stim range: system- and electrode-dependent (often not reported uniformly)

Tissue Interface & Bioresponse

  • Target tissue: cortex
  • BBB disruption: high (penetrating)
  • Vascular disruption risk: moderate–high (depends on placement)
  • Micromotion sensitivity: high (rigid silicon vs soft brain)
  • Gliosis / encapsulation: commonly observed in chronic implants
  • Neuron loss (if reported): often reported as substantial within ~100 µm over time
  • Foreign-body response mitigation: coatings/material variants exist, but core rigidity remains
  • Typical failure mode: gradual channel loss / encapsulation, infections related to percutaneous components, connector issues

System Architecture

  • Onboard electronics: none on the array (classic)
  • Data path: tethered / percutaneous in many research systems
  • Telemetry bandwidth: N/A (tethered)
  • Sampling rate: system-dependent
  • Power: external
  • Thermal management: external (classic)
  • Hermeticity: percutaneous connector systems vary
  • MRI compatibility: generally no/unknown unless explicitly specified for a given configuration
  • Surgical complexity: craniotomy + insertion tooling

Performance Envelope

  • Typical yield (acute): high (spike yield commonly strong early)
  • Typical yield (chronic): variable; declines over months
  • Stability over time: often 6–36 months of “good signals” reported in many programs (context-dependent)
  • Longevity (median / max): variable (context-dependent)
  • Revision / explant: explantable; revision surgeries not uncommon in long studies
  • Adverse events (high-level): depends on protocol; percutaneous infection risk exists
  • Notable demos / tasks: cursor control, typing, robotic arm control

Clinical / Preclinical Evidence

  • N implanted subjects / animals: >30 humans reported across programs (order-of-magnitude)
  • Follow-up duration: months to years in research cohorts
  • Indications: paralysis, ALS, stroke (research)
  • Trial registry links: varies by program (to add)
  • Primary outcomes: communication/control task performance
  • Key limitations of evidence: heterogeneous hardware stacks and reporting; chronic performance varies widely

Engineering Verdict

Strengths:

  • strong spike recordings early
  • mature ecosystem across decades of research

Limitations / failure modes:

  • rigid penetrating shanks + micromotion drive chronic signal loss
  • percutaneous connectors introduce infection/maintenance burden

Scaling constraints:

  • wiring/connector complexity
  • chronic biology (gliosis/encapsulation)
  • surgery time and placement constraints

What newer designs try to fix:

  • mechanical mismatch + micromotion
  • percutaneous connector infection route
  • high-channel wiring burden

Simulation Hooks (for BuildTheSimulation)

  • Minimal model to reproduce: rigid shank array in tissue with a time-varying gliosis/encapsulation layer
  • Parameters to expose as sliders: insertion depth, neuron density, encapsulation thickness, impedance drift
  • What outputs to visualize: spike yield, SNR proxy, channel survival vs time

References

  • (Add primary Normann + chronic response + BrainGate methods papers here)