Utah Array (Intracortical Microelectrode Array)
A bed of 96 silicon needles, each thinner than a hair, eavesdropping on individual neurons.
The Utah Array is a small (4 mm × 4 mm) grid of 96 silicon spikes, each tipped with a metal electrode, that is pneumatically inserted into the outer layer of the brain (the cerebral cortex). Each electrode can record the electrical signals fired by one or a few individual neurons nearby, or deliver tiny electrical pulses to stimulate those neurons. When connected to a computer, patients with paralysis have used it to move a cursor, control a robotic arm, or produce handwriting — using only their thoughts.
The array is etched from a single piece of silicon. The 96 electrodes are arranged in a 10×10 grid (with four corner positions unused), each tapering to a sharp tip coated with platinum or iridium oxide to improve electrical contact. Lengths range from 0.5 to 1.5 mm, tuned to target specific cortical layers. When a nearby neuron fires an action potential, the tiny electric field it produces creates a measurable voltage on the nearest electrode tip — typically tens to hundreds of microvolts. These signals are amplified, digitized, and processed by an external computer to decode the intended movement. The array is inserted in about 400 microseconds using a pneumatic actuator to minimize tissue damage.
In plain terms
Imagine pressing a tiny bed of 96 micro-needles — each thinner than a hair and shorter than your pinky fingernail — gently into the surface of the brain. Each needle tip can hear the electrical whisper of a single brain cell firing, like pressing your ear against 96 different walls to hear what's being said in each room of a building.
Also called: Utah Intracortical Electrode Array (UIEA), NeuroPort Array, silicon microelectrode array.
Why it matters
- The Utah Array is the only FDA-cleared high-channel intracortical microelectrode platform with demonstrated long-term safety in humans, with data from more than 30,000 patient-days of implantation
- BrainGate clinical trials showed patients controlling cursors, typing, and even feeling touch again using Utah Array signals — the first direct evidence that single-neuron recordings can reliably decode motor intent over years
- The 96-channel density provides enough parallel recordings to achieve meaningful decoding performance; larger systems aim to scale this to thousands of channels
- Utah Array recordings seeded the field of brain-computer interfaces and remain the primary benchmark against which new electrode technologies are compared
Where it lives
- BrainGate / BrainGate2 clinical trials — 14 participants with paralysis implanted 2004–2021, up to 7.6 years of recordings per participant
- NeuroPort system (Blackrock Neurotech) — the commercially manufactured Utah Array, used in >1,000 research labs worldwide
- Neuralink's early prototypes studied Utah Array data to define performance benchmarks for their thread-electrode approach
Key numbers
- Electrode count: 96 active channels (10×10 grid, 4 corners unused)
- Array footprint: 4 × 4 mm (Slightly smaller than a pencil eraser)
- Electrode length (clinical): 1.0–1.5 mm (Targets upper cortical layers)
- Insertion time: ~400 µs (Pneumatic actuator to minimize tissue bounce)
- Max patient-days (BrainGate): >30,000 (Across 14 participants, 2004–2021)
Surprising facts
- In the BrainGate trials (2004–2021), 20 Utah Arrays recorded from 14 participants over 2,319 sessions spanning up to 7.6 years — providing the longest and most comprehensive human intracortical recording dataset ever collected
- Arrays recorded spiking activity on an average of 35.6% of electrodes across the study period, with only a 7% decline over the enrollment period — better longevity than similar devices tested in non-human primates
- Some arrays remained functional for over 980 days (more than 2.6 years) inside a human brain without removal
- The Utah Array has accumulated over 20,000 peer-reviewed citations and is used in more than 1,000 labs worldwide
Open questions
- Some arrays in BrainGate showed increasing signal amplitude: Some arrays in BrainGate showed increasing signal amplitude for the first 30–40 days post-implant before declining — the opposite of naive expectations — possibly because initial glial encapsulation actually improves electrode contact before eventually degrading it; the mechanism is still under investigation
- Three of 14 BrainGate arrays reached decoding signal-to-noise: Three of 14 BrainGate arrays reached decoding signal-to-noise ratios approaching those of able-bodied computer mouse control — despite being implanted for years in brains with severe neurological injury — suggesting the cortex retains its motor-coding capacity far longer and more robustly than previously assumed
A short history
The Utah Intracortical Electrode Array was invented by Richard Normann and colleagues at the University of Utah in the late 1980s and early 1990s. The foundational chronic-recording study in cats was published in 1999 (Campbell et al., Journal of Biomechanical Engineering). The first human implantation for BCI purposes occurred in 2004 as part of the BrainGate trial at Brown University, led by John Donoghue. In 2012, a BrainGate participant named Jan Scheuermann used a Utah Array-based system to control a robotic arm to feed herself chocolate for the first time in years.
Sources
- Rubin et al., 'Interim Safety Profile from the BrainGate Feasibility Study', Neurology (2023) (https://www.neurology.org/doi/10.1212/WNL.0000000000201707)
- Wairagkar et al., 'Long-term performance of intracortical microelectrode arrays in 14 BrainGate participants', medRxiv (2025) (https://pmc.ncbi.nlm.nih.gov/articles/PMC12236888/)
- Blackrock Neurotech — Utah Array product page (https://blackrockneurotech.com/products/utah-array/)
- Campbell et al., 'Chronic recording capability of the Utah Intracortical Electrode Array', PubMed (1998) (https://pubmed.ncbi.nlm.nih.gov/10223510/)