Definition
Microelectrode array: A microelectrode array (MEA) is a device with many microscopic electrodes, usually on a glass or silicon surface, used to record electrical activity from neurons or heart cells and to stimulate them. Cultured neurons or brain organoids sit on the array, and each electrode picks up voltage changes from nearby cells without piercing them.
Also known as: MEA, multielectrode array, multi-electrode array
A microelectrode array (MEA) is a device carrying many microscopic electrodes, usually on a glass or silicon base, that records electrical activity from living cells and can also stimulate them. Neurons or brain organoids sit on the array, and each electrode senses voltage changes from nearby cells without piercing them, over days to months.
Key takeaways
- MEAs record extracellular signals: spikes from individual neurons and slower field potentials from groups of cells.
- Most in vitro MEAs can also stimulate, which makes them the standard two-way interface for wetware computing.
- Thomas et al. described an early miniature MEA for cultured cells in Experimental Cell Research (1972).
- Arrays range from tens of passive electrodes to CMOS chips with thousands, known as high-density microelectrode arrays.
- The main limits are spatial coverage, signal quality, stimulation artifacts, and reaching inside 3D tissue.
What is a microelectrode array?
An MEA is a cell culture surface with a built-in grid of electrodes, each a few to tens of micrometers across, connected to amplifiers. Cells are plated or placed directly on top. When a neuron near an electrode fires an action potential, ions flow across its membrane, creating a tiny voltage change in the surrounding fluid that the electrode detects.
Because the electrodes do not penetrate the cells, the same culture can be recorded repeatedly for weeks or months. That makes MEAs suited to studying how networks develop and change, a core need in electrophysiology of cultured neurons and brain organoids. Spira and Hai reviewed MEA technologies for neuroscience and cardiology in Nature Nanotechnology (2013).
How does a microelectrode array work?
Recording
- Sensing. Each electrode picks up the extracellular potential near it, a mix of spikes from close neurons, background activity and noise.
- Amplification and filtering. Signals of tens to hundreds of microvolts are amplified. A high-pass filter isolates spikes, and a low-pass filter isolates local field potentials.
- Digitization. Signals are sampled thousands of times per second per channel so that spike shapes are preserved.
- Spike detection and sorting. Software detects threshold crossings and may use spike sorting to assign spikes to individual neurons.
Stimulation
Applying a brief current or voltage pulse to an electrode can trigger activity in nearby cells. By choosing which electrodes to stimulate, when, and how strongly, experimenters can deliver patterned inputs. When stimulation depends on the recorded response in real time, the system is performing closed-loop stimulation.
What signals does an MEA capture?
- Spikes, sharp deflections from single neurons, used to study firing rates and timing.
- Network bursts, near-simultaneous firing across many electrodes, a hallmark of cultured networks (see network bursts).
- Local field potentials, slower oscillations reflecting summed synaptic activity. Trujillo et al. reported in Cell Stem Cell (2019) that oscillatory activity recorded from cortical organoids on MEAs grew more complex over months.
A worked example: a simple MEA stimulation experiment
To make the workflow concrete, here is what a basic open-loop experiment on a cultured network might look like:
- Record ten minutes of spontaneous activity to establish a baseline firing rate and burst pattern on each electrode.
- Choose one electrode as the stimulation site.
- Deliver a short train of pulses from that electrode at fixed intervals.
- For each pulse, collect the spikes recorded on all other electrodes in a short window afterward, excluding the artifact period.
- Compare evoked responses with the baseline to map which parts of the network respond, and how quickly.
Changing step 3 so that the next stimulus depends on the last response turns the same setup into a closed-loop experiment.
Types of microelectrode arrays
| Type | Electrode count | Construction | Strengths | Trade-offs |
|---|---|---|---|---|
| Passive MEA | Tens to a few hundred | Metal electrodes on glass, wired to external amplifiers | Simple, transparent, well established | Sparse coverage, each electrode needs its own wire |
| High-density (CMOS) MEA | Thousands to tens of thousands | Electrodes integrated with amplifiers on a silicon chip | Subcellular resolution, large coverage | Opaque chip, data volume, cost |
| Multiwell MEA | Small arrays in many wells | Plate format | Parallel experiments, drug screening | Few electrodes per well |
| 3D or flexible MEA | Varies | Protruding, mesh or shell electrodes | Reaches inside tissue such as organoids | Less mature, harder to fabricate |
Why do microelectrode arrays matter for biocomputing?
In organoid intelligence, the MEA is the input and output device. It is how a computer “talks” to living neurons and “listens” to their answer. Kagan et al. reported in Neuron (2022) that their DishBrain system used an MEA to encode a simplified Pong game as stimulation and to read out the cells’ activity as paddle movements. Jordan et al. (2024) described a remotely accessible platform in which organoids on MEAs are controlled through software over the internet.
The quality of the interface sets an upper limit on what a biocomputer can do. More electrodes mean more input channels and more detailed readout for neural encoding and decoding. For a deeper walk-through, see microelectrode arrays explained and how neurons learn in a dish.
Microelectrode array vs other recording methods
- Patch clamp gives detailed intracellular recordings from one cell at a time, but is invasive and short-lived.
- Calcium imaging uses fluorescent indicators to watch many cells, but is slower than electrical recording and needs optical access.
- Implanted probes record in living animals or people, but are not used for cultured organoids.
- MEAs trade single-cell detail for long-term, non-invasive, many-channel recording plus stimulation, which is why they dominate in vitro network studies.
What are the limitations of microelectrode arrays?
- Coverage. Passive arrays have electrodes spaced far apart, so many neurons go unrecorded.
- Surface contact. A flat array touches only the bottom of a 3D organoid. Neurons deeper inside are hard to reach.
- Ambiguous signals. One electrode can pick up several neurons, and one neuron can appear on several electrodes, which spike sorting must untangle.
- Stimulation artifacts. Stimulation pulses create large electrical transients that can mask responses immediately afterward.
- Data volume. Many channels sampled at high rates produce large datasets that need efficient storage, ideally in open formats such as Neurodata Without Borders.
Where is MEA technology heading?
Trends include higher electrode density on CMOS chips, 3D and flexible electrodes that wrap around or penetrate organoids, integration with microfluidics for long-term culture, and real-time software that closes the loop with low latency. Each of these widens the channel between silicon and living tissue.
Software for running stimulation protocols and closed-loop experiments on MEAs is the focus of our work at Organoid Programming.
Frequently asked questions
What does an MEA measure?
An MEA measures extracellular voltage, the small electrical changes in the fluid around cells when they are active. When a nearby neuron fires an action potential, the electrode records a brief spike, typically tens to hundreds of microvolts. Slower signals, such as local field potentials reflecting the summed activity of many cells, can be recorded too.
Can a microelectrode array stimulate neurons as well as record them?
Yes, many MEAs can do both. Passing a small current or voltage pulse through an electrode can trigger activity in nearby neurons. This two-way capability is essential for closed-loop experiments, where software reads the network’s activity and decides what stimulation to deliver next, as in the DishBrain experiments reported by Kagan et al. (2022).
What is the difference between an MEA and patch clamp?
Patch clamp uses a glass pipette sealed onto a single cell to record its electrical activity with very high precision, but only for one or a few cells at a time and usually for hours. An MEA records from many cells at once, non-invasively and repeatedly over days or months, but with less detail about each individual cell.
Are MEAs used in the brain or only in the lab?
Both, though the term MEA most often refers to in vitro devices used with cultured cells, tissue slices and organoids. Implantable electrode arrays, such as the Utah array or Neuropixels probes, are related technologies used in animals or people. They share principles with in vitro MEAs but differ in shape, materials and safety requirements.
Why are microelectrode arrays important for organoid intelligence?
An organoid on its own cannot receive inputs or report outputs that a computer understands. The MEA is the interface that makes organoid intelligence possible: it converts digital commands into stimulation and converts neural activity back into digital data. The number, density and placement of electrodes limit how much information can flow in each direction.
Sources
- Thomas et al. (1972). A miniature microelectrode array to monitor the bioelectric activity of cultured cells. Experimental Cell Research.
- Spira & Hai (2013). Multi-electrode array technologies for neuroscience and cardiology. Nature Nanotechnology.
- Kagan et al. (2022). In vitro neurons learn and exhibit sentience when embodied in a simulated game-world. Neuron.
- Trujillo et al. (2019). Complex oscillatory waves emerging from cortical organoids model early human brain network development. Cell Stem Cell.
- Jordan et al. (2024). Open and remotely accessible Neuroplatform for research in wetware computing. Frontiers in Artificial Intelligence.
