Optimizing Transcutaneous Electrical Stimulation Based on Frequency-Dependent Tissue Modeling and Signal Analysis

Transcutaneous electrical stimulation (TES) is limited by cutaneous discomfort caused by unavoidable activation of superficial sensory nerves during current delivery to deep targets. While psychophysical studies have empirically identified waveforms that reduce skin sensation, existing computational models typically employ quasi-static approximations that neglect the pronounced dielectric dispersion of biological tissues, leaving the biophysical mechanisms poorly understood. We developed a finite-element model of the human forearm incorporating the frequency-dependent dielectric properties of skin, fat, muscle, bone, and nerve tissues (DC to 1 MHz), coupled with a linear time-invariant signal-processing framework based on the system transfer function H(f). The model quantitatively reproduced the waveform-dependent sensation trends reported by Hsu et al. We introduced a penetration ratio to quantify deep-to-superficial nerve activation and found that all time-varying waveforms exhibit lower penetration than direct current (DC), revealing a skin-effect-like behavior of electrical current in biological tissues. Moreover, deep and superficial nerve activations co-varied under waveform parameter changes, indicating that waveform optimization alone cannot improve depth selectivity. Electrode spatial configuration was shown to offer a complementary strategy: positioning the active electrode close to the target muscle nerve while avoiding superficial cutaneous nerves, combined with sufficient transverse spacing, enhances depth selectivity. This work bridges psychophysical observations with tissue electrophysiology and provides a computational tool for waveform and electrode optimization in TES.