Cancelling out Skull-Induced Aberrations: Analysis of Acoustic Metamaterials Using Neural Networks

During investigations on the use of acoustic metamaterials for the correction of skull-induced aberrations on the ultrasonic beam for neurological applications, inconsistencies in the results reported in the literature have been found. Stemming from these discrepancies, an analytical model is introduced in this study in an attempt to model the behavior of the metamaterial by using the Method of the Reciprocal Theorem. Its inefficacy in explaining the observed phenomenon is proved and, consequently, an engineering solution is proposed that allows for the mathematical modelling of the observed discrepancies with a precision of approximately 90% by means of a new Neural Network architecture called Convolutional Modelling Network.