Near-field recordings expose the limits of far-field measures of peripheral vestibular nerve activity: implications...

Vestibular short-latency evoked potentials (VsEPs) provide an objective measure of otolith-driven vestibular nerve activity but remain technically challenging to record when compared with cochlear evoked responses. This difficulty arises from the location of the vestibular system, in close proximity to the cochlea, which generates large biopotentials to similar stimuli. In experimental animal models, VsEPs are commonly recorded using far-field (FF) scalp electrodes, however the extent to which this approach faithfully reflects near-field (NF) recordings remains unclear. Here, we directly compare simultaneous NF facial nerve canal (FNC) VsEPs and FF scalp responses in vivo in the guinea pig in response to interaural bone-conducted vibration (BCV). NF VsEPs were recorded from within the FNC adjacent to the vestibular nerve bundle, while FF responses were recorded from the scalp. NF FNC VsEPs exhibited amplitudes roughly 20-fold larger, with SNR values approximately 10-15 dB higher than FF scalp responses across animals. NF responses revealed a ∼90% reduction during transient hypoxia and an almost complete abolition following tetrodotoxin application to the otolithic macula. In contrast, FF scalp VsEPs underestimated vestibular hypofunction and retained residual responses following neural blockade and death, consistent with non-neural origins. Micro-CT imaging of the guinea pig temporal bone confirmed the close anatomical proximity of the utricular macula to the FNC, providing the anatomical basis for the superior signal amplitude recorded at this site. Together, these findings demonstrate that NF FNC recordings provide a more sensitive and selective measure of peripheral vestibular nerve activity than FF scalp electrodes in the guinea pig, demonstrating the important need for optimising recording location and experimental controls when recording vestibular evoked potentials.