Simulations of low-frequency vibration pattern at the inner ear for activation of the vestibular system

Vestibular stimulation is a promising method for mitigating motion sickness and assessing vestibular function. However, conventional methods such as galvanic, caloric, or bone conduction (BC) stimulation lack the spatial selectivity needed to replicate the precise inputs of natural head movements. In this study, a novel approach is proposed using multiple BC transducers with coordinated magnitude and phase to generate targeted vestibular motions. A validated finite element model of the human head was employed to simulate vestibular responses to stimuli from four transducer locations. A genetic algorithm was then used to optimize the magnitude and phase of each input to produce desired one-dimensional (1D), two-dimensional (2D), and rotational vestibular motions. The results demonstrate that the optimized BC inputs can generate 1D linear, 2D planar, and rotational motions with high directional accuracy. The deviations from the target axes were consistently below the known psychophysical thresholds for motion perception. This computational study confirms the feasibility of using multi-site BC stimulation to achieve precise and configurable vestibular actuation, offering a significant improvement over existing techniques.