Distinct electrophysiological profiles of bacterial mechanosensitive channels for sonogenetic actuator selection

Objective.Sonogenetics combines ultrasound stimulation with genetically encoded mechanosensitive (MS) ion channels for cell-targeted neuromodulation. Actuator choice, however, remains largely empirical becausein vivoelectrophysiological response signatures are rarely compared under matched conditions. Here, we conducted an exploratoryin vivobenchmarking of three bacterial MS channels (MscL-G22S, MscL-G22N, and MscS) during transcranial ultrasound stimulation in anesthetized rat primary visual cortex (V1).Approach.local field potentials (LFPs) were recorded via a microelectrode array from V1 expressing AAV-delivered channels during graded ultrasound stimulation (1 MHz;I_spta100-400 mW cm^-2). We quantified baseline activity, ultrasound-evoked potentials (UEPs), trial-to-trial response distributions, and frequency-band power dynamics.Main results.Channel identity shaped both baseline and ultrasound-evoked cortical activity. MscS increased baseline LFP total power (∼2.5 dB vs control,P= 0.0035), whereas MscL-G22S shifted baseline band composition (reduced theta, enhanced gamma). MscL-G22S showed the lowest detectable UEP threshold, producing a detectable N1 at 100 mW cm^-2and an intensity-dependent N1 increase up to ∼2-fold at 400 mW cm^-2(P< 0.0001). Latency depended on both channel and intensity: MscL-G22N responded faster at low intensity, while MscL-G22S accelerated at higher intensities. MscL expression narrowed trial-to-trial response distributions (bimodal to unimodal). Spectrally, MscL-G22N enhanced theta power, whereas MscL-G22S recruited beta-gamma oscillations at high intensity.Significance.Under matched stimulation and expression conditions, bacterial MscL produced distinct network-level response profiles spanning UEP threshold, response timing, trial-to-trial consistency, and oscillatory modulation. These exploratory benchmarks provide quantitative reference data for comparing sonogenetic actuators and may inform actuator selection for closed-loop neuromodulation.