Illustrative example demonstrating the advantages of the UMAP-based approach. A single electrode records extracellular activity from three different neurons. Although their spike waveforms appear similar, the information each neuron encodes—and their respective firing rates—differs substantially. UMAP reliably distinguishes neurons with low firing rates, even when their activity patterns are sparse but potentially rich in information. In contrast, traditional feature-based spike-sorting methods frequently merge spikes from these low-activity neurons with those of more active ones, thereby losing valuable information and ultimately reducing decoding performance.
Spike sorting can reveal single-neuron activity from extracellular #electrophysiological recordings. This study shows that replacing trad methods with UMAP enhances #SpikeSorting accuracy, efficiency & scalability, paving the way for automated large-scale analysis @plosbiology.org 🧪 plos.io/3KrioSH
