Machine-learning based classification of middle-ear fixation and separation using sweep frequency impedance information reflecting middle-ear dynamics.

The sweep frequency impedance (SFI) meter is an apparatus that delivers a frequency-sweeping sound into the ear canal and evaluates dynamic characteristics of the middle ear based on changes in sound pressure in the ear canal. We have renewed the SFI meter for detecting sound pressure variations with higher signal-to-noise ratio. Our previous studies proposed a potential to utilize it for diagnosis of ossicular fixation and separation using a two-dimensional (2D) mobility map of the middle ear derived from two features extracted from sound pressure variations. However, a concrete classification criterion in this 2D information is necessary for clinical applications since the diagnostic process is assumed to be challenging due to pronounced interphasic overlaps of the map. We conducted the SFI measurement with normal and impaired ears using the renewed SFI meter and proposed a machine-learning-based prediction method of middle-ear dysfunctions in 2D characteristics. We showed that ossicular chain fixation was predicted with an accuracy of 0.8 and an area under the receiver operating characteristic curve (AUC) of 0.86 and ossicular chain separation with an accuracy of 1.0 and AUC of 1.0. The proposed method has the potential to predict the middle-ear dysfunctions more accurately than a conventional tympanometry method.