LFP-LOC: an LFP power-based method for validating the anatomical placement of high-density neural probes in rodents

High-density CMOS-based neural probes provide unprecedented spatiotemporal resolution for in-vivo electrophysiology, yet accurate validation of implant position remains challenging. Here we present LFP-LOC, a simple and interpretable method for intraoperative validation and refinement of probe anatomical location based on the spatial distribution of local field potential (LFP) power. Using spontaneous activity recordings performed in rodents, we compute power spectral densities in canonical LFP bands and apply dimensionality reduction and clustering to identify electrodes with shared spectral signatures. Across multiple implant sites, probe technologies, electrode layouts, and experimental conditions, the resulting clusters consistently align with anatomical boundaries. Applied to high-density probes with up to 1,024 electrodes/channels and sub-30 μm pitch, power features converge within approximately 20 s of recordings, allowing online intraoperative assessment. By leveraging the robust relationship between LFP power and brain structure, LFP-LOC enables rapid validation and adjustment of probe placement during surgery, complements histological validation, and may facilitate mesoscale electrophysiological studies.