What is Single-Unit Recording?

The electrophysiological technique of isolating and recording the action potentials (spikes) of individual neurons using intracortical microelectrodes.

What category does Single-Unit Recording belong to in BCI?

Single-Unit Recording is a Recording Modalities concept in brain-computer interface science.

Single-Unit Recording — BCI Glossary

Single-unit recording refers to the detection and isolation of action potentials from individual neurons. An intracortical microelectrode positioned close to a neuron's cell body or axon can detect the extracellular voltage deflection produced when that neuron fires. When the electrode is close enough and the signal-to-noise ratio is sufficient, the recorded waveform can be attributed to a single identifiable neuron — a "single unit."

Technique

Single-unit isolation requires several conditions:

Electrode proximity: The recording tip must be within approximately 50-100 micrometers of the neuron to detect its spike above the noise floor
High impedance electrodes: Smaller electrode tips (higher impedance) provide better spatial selectivity, detecting activity from a smaller volume of tissue
Spike sorting: When multiple neurons are detected on the same electrode, their waveforms must be separated computationally using amplitude, shape, and timing features

Importance for BCI

Single-unit recordings provide the highest-resolution view of neural activity available. Each isolated neuron's firing rate and timing carry specific information about motor intent, sensory experience, or cognitive state. The earliest BrainGate demonstrations relied on single-unit tuning curves — measuring how each neuron's firing rate varied with intended movement direction — to build population vector decoders.

Single Units vs. Multi-Unit Activity

In practice, many BCI systems use multi-unit activity (MUA) — the aggregate spiking of multiple neurons detected on a single electrode — rather than isolated single units. MUA is easier to extract (no spike sorting required), more robust to electrode drift, and often provides comparable decoding performance for cursor control tasks. Modern BCI decoders, particularly deep learning approaches, can leverage both single-unit and multi-unit signals effectively.

Longevity Challenges

Maintaining stable single-unit recordings over months to years is one of the primary challenges in chronic BCI implants. Micromotion between the electrode and brain tissue, glial scarring, and neuronal death near the electrode tip all contribute to gradual loss of detectable single units. The Utah Array typically yields 30-50 isolatable single units at implant, declining to 10-20 within the first year. This degradation has motivated the development of flexible electrodes and alternative signal modalities (LFP, ECoG) that may be more stable long-term.