OPTICAL CALIBRATION OF EXTRACELLULAR SPIKE WAVEFORM RECORDINGS IN AWAKE ANIMALS WITH A GENETICALLY ENCODED VOLTAGE INDICATOR

Extracellular electrophysiology is one of the most widely used techniques in neuroscience to probe neuronal population activity with single-cell resolution. However, the identification of individual neurons relies on spike waveform detection and classification, an approach whose validation is limited by the scarcity of in vivo ground-truth datasets. Existing benchmark experiments typically rely on simultaneous intracellular and extracellular recordings, which may be biased by the constraints of in vivo electrophysiology and leave unresolved questions regarding the detectability of neurons at different distances from the recording electrode.
To overcome these limitations, we developed a novel methodology to systematically calibrate extracellular action potential recordings using genetically identified and spatially localized neurons imaged in awake animals. We expressed the genetically encoded voltage indicator JEDI2P in cortical pyramidal neurons and distinct interneuron subtypes, and established an experimental strategy enabling simultaneous fast voltage imaging and extracellular recordings with silicon probes while minimizing photoelectric artifacts. This approach allowed us, when possible, to match extracellularly recorded spikes to optically detected action potentials from neurons located at varying distances from the probe.
Our initial observations indicate that only neurons located within approximately 25 µm of the probe generate extracellular action potentials exceeding 30 µV in amplitude. Moreover, the relationship between spike amplitude and distance is highly variable, with some neurons producing detectable signals at this distance while others do not. These results suggest that extracellular recordings predominantly sample neurons in the immediate vicinity of the probe and that a substantial fraction of nearby active neurons remains undetected.