Preliminary results of classifying otosclerosis and disarticulation using a convolutional neural network trained with simulated wideband acoustic immittance data.

Current noninvasive methods of clinical practice often do not identify the causes of conductive hearing loss due to pathologic changes in the middle ear with sufficient certainty. Wideband acoustic immittance (WAI) measurement is noninvasive, inexpensive and objective. It is very sensitive to pathologic changes in the middle ear and therefore promising for diagnosis. However, evaluation of the data is difficult because of large interindividual variations. Machine learning methods like Convolutional neural networks (CNN) which might be able to deal with this overlaying pattern require a large amount of labeled measurement data for training and validation. This is difficult to provide given the low prevalence of many middle-ear pathologies. Therefore, this study proposes an approach in which the WAI training data of the CNN are simulated with a finite-element ear model and the Monte-Carlo method. With this approach, virtual populations of normal, otosclerotic, and disarticulated ears were generated, consistent with the averaged data of measured populations and well representing the qualitative characteristics of individuals. The CNN trained with the virtual data achieved for otosclerosis an AUC of 91.1 %, a sensitivity of 85.7 %, and a specificity of 85.2 %. For disarticulation, an AUC of 99.5 %, sensitivity of 100 %, and specificity of 93.1 % was achieved. Furthermore, it was estimated that specificity could potentially be increased to about 99 % in both pathological cases if stapes reflex threshold measurements were used to confirm the diagnosis. Thus, the procedures' performance is comparable to classifiers from other studies trained with real measurement data, and therefore the procedure offers great potential for the diagnosis of rare pathologies or early-stages pathologies. The clinical potential of these preliminary results remains to be evaluated on more measurement data and additional pathologies.

The impact of tympanic membrane perforations on middle ear transfer function.

PURPOSE: Injury or inflammation of the middle ear often results in the persistent tympanic membrane (TM) perforations, leading to conductive hearing loss (HL). However, in some cases the magnitude of HL exceeds that attributable by the TM perforation alone. The aim of the study is to better understand the effects of location and size of TM perforations on the sound transmission properties of the middle ear. METHODS: The middle ear transfer functions (METF) of six human temporal bones (TB) were compared before and after perforating the TM at different locations (anterior or posterior lower quadrant) and to different degrees (1 mm, » of the TM, ½ of the TM, and full ablation). The sound-induced velocity of the stapes footplate was measured using single-point laser-Doppler-vibrometry (LDV). The METF were correlated with a Finite Element (FE) model of the middle ear, in which similar alterations were simulated. RESULTS: The measured and calculated METF showed frequency and perforation size dependent losses at all perforation locations. Starting at low frequencies, the loss expanded to higher frequencies with increased perforation size. In direct comparison, posterior TM perforations affected the transmission properties to a larger degree than anterior perforations. The asymmetry of the TM causes the malleus-incus complex to rotate and results in larger deflections in the posterior TM quadrants than in the anterior TM quadrants. Simulations in the FE model with a sealed cavity show that small perforations lead to a decrease in TM rigidity and thus to an increase in oscillation amplitude of the TM mainly above 1 kHz. CONCLUSION: Size and location of TM perforations have a characteristic influence on the METF. The correlation of the experimental LDV measurements with an FE model contributes to a better understanding of the pathologic mechanisms of middle-ear diseases. If small perforations with significant HL are observed in daily clinical practice, additional middle ear pathologies should be considered. Further investigations on the loss of TM pretension due to perforations may be informative.

Parameter Identification From Normal and Pathological Middle Ears Using a Tailored Parameter Identification Algorithm.

Current clinical practice is often unable to identify the causes of conductive hearing loss in the middle ear with sufficient certainty without exploratory surgery. Besides the large uncertainties due to interindividual variances, only partially understood cause-effect principles are a major reason for the hesitant use of objective methods such as wideband tympanometry in diagnosis, despite their high sensitivity to pathological changes. For a better understanding of objective metrics of the middle ear, this study presents a model that can be used to reproduce characteristic changes in metrics of the middle ear by altering local physical model parameters linked to the anatomical causes of a pathology. A finite-element model is, therefore, fitted with an adaptive parameter identification algorithm to results of a temporal bone study with stepwise and systematically prepared pathologies. The fitted model is able to reproduce well the measured quantities reflectance, impedance, umbo and stapes transfer function for normal ears and ears with otosclerosis, malleus fixation, and disarticulation. In addition to a good representation of the characteristic influences of the pathologies in the measured quantities, a clear assignment of identified model parameters and pathologies consistent with previous studies is achieved. The identification results highlight the importance of the local stiffness and damping values in the middle ear for correct mapping of pathological characteristics and address the challenges of limited measurement data and wide parameter ranges from the literature. The great sensitivity of the model with respect to pathologies indicates a high potential for application in model-based diagnosis.

Contribution of the flexible incudo-malleal joint to middle-ear sound transmission under static pressure loads.

The incudo-malleal joint (IMJ) in the human middle ear is a true diarthrodial joint and it has been known that the flexibility of this joint does not contribute to better middle-ear sound transmission. Previous studies have proposed that a gliding motion between the malleus and the incus at this joint prevents the transmission of large displacements of the malleus to the incus and stapes and thus contributes to the protection of the inner ear as an immediate response against large static pressure changes. However, dynamic behavior of this joint under static pressure changes has not been fully revealed. In this study, effects of the flexibility of the IMJ on middle-ear sound transmission under static pressure difference between the middle-ear cavity and the environment were investigated. Experiments were performed in human cadaveric temporal bones with static pressures in the range of +/- 2 kPa being applied to the ear canal (relative to middle-ear cavity). Vibrational motions of the umbo and the stapes footplate center in response to acoustic stimulation (0.2-8 kHz) were measured using a 3D-Laser Doppler vibrometer for (1) the natural IMJ and (2) the IMJ with experimentally-reduced flexibility. With the natural condition of the IMJ, vibrations of the umbo and the stapes footplate center under static pressure loads were attenuated at low frequencies below the middle-ear resonance frequency as observed in previous studies. After the flexibility of the IMJ was reduced, additional attenuations of vibrational motion were observed for the umbo under positive static pressures in the ear canal (EC) and the stapes footplate center under both positive and negative static EC pressures. The additional attenuation of vibration reached 4~7 dB for the umbo under positive static EC pressures and the stapes footplate center under negative EC pressures, and 7~11 dB for the stapes footplate center under positive EC pressures. The results of this study indicate an adaptive mechanism of the flexible IMJ in the human middle ear to changes of static EC pressure by reducing the attenuation of the middle-ear sound transmission. Such results are expected to be used for diagnosis of the IMJ stiffening and to be applied to design of middle-ear prostheses.

Model-based hearing diagnostics based on wideband tympanometry measurements utilizing fuzzy arithmetic.

Today's audiometric methods for the diagnosis of middle ear disease are often based on a comparison of measurements with standard curves, that represent the statistical range of normal hearing responses. Because of large inter-individual variances in the middle ear, especially in wideband tympanometry (WBT), specificity and quantitative evaluation are greatly restricted. A new model-based approach could transform today's predominantly qualitative hearing diagnostics into a quantitative and tailored, patient-specific diagnosis, by evaluating WBT measurements with the aid of a middle-ear model. For this particular investigation, a finite element model of a human ear was used. It consisted of an acoustic ear canal and a tympanic cavity model, a middle-ear with detailed nonlinear models of the tympanic membrane and annular ligament, and a simplified inner-ear model. This model has made it possible for us to simulate pathologies like the stiffening of ligaments or joints, because we can simply change the corresponding mechanical parameters of the model. On the other hand, it is also possible to identify pathologies from measurements, by analyzing the parameters obtained by a system identification procedure. This reduces the number of required model parameters through sensitivity studies and parameter clustering. Uncertainties due to the lack of knowledge, subjectivity in numerical implementation and model simplification are taken into account by the application of fuzzy arithmetic. The most confident parameter set can be determined by applying an inverse fuzzy method on the measurement data. The principle and the benefits of this model-based approach are illustrated by the example of a two-mass oscillator, and also by the simulation of the energy absorbance of an ear with malleus fixation, where the parameter changes that are introduced can be determined quantitatively through the system identification.

Experimental study on admissible forces at the incudomalleolar joint.

HYPOTHESIS: The forces that cause rupture of the incudomalleolar joint during the fixation of stapedial prostheses can be determined by means of load-deflection measurements at the long process of the incus. As in other tissues, 3 ranges of forces can be defined: micro rupture, rupture, and short-term maximum. BACKGROUND: A crucial step in stapes surgery is the attachment of the stapedial prosthesis to the long process of the incus. It is unknown which forces occur during the crimping process that increase the risk of damage to the incudomalleolar joint or incus luxation. The goal of this study was to assess the admissible range of forces at the long process of the incus that would be tolerable before damaging the structures and to compare them with the forces occurring during surgery. METHODS: Load-deflection curves in the lateral-medial and anterior-posterior direction were measured in 9 freshly frozen or fresh temporal bones. The force was measured with a load cell, and displacement was taken from the encoder information of the electrically driven translation stage on which the load cell was mounted. The long process of the incus was coupled to the load cell via a customized needle. We also monitored with video recordings for visual confirmation of findings. RESULTS: The rupture force at which the middle ear was found to be severely injured was 894 (724-1018) mN in the anterior-posterior direction and 695 (574-771) mN in the lateral-medial direction. Micro-ruptures occurred at forces around 568 (469-686) mN in the anterior-posterior direction and in the lateral-medial direction at 406 (254-514) mN. Short-term maximum forces of 1,321 (1,051-1,533) mN were measured in the anterior-posterior direction and 939 (726-1,132) mN in the lateral-medial direction. CONCLUSION: Rupture forces of the incudomalleolar joint could be defined with high accuracy. These results were used to calculate risks of incus luxation or subluxation during stapes surgery. Compared with the use of clip and SMA prostheses, the risk of damage from a crimping procedure is significantly higher.

Contribution of complex stapes motion to cochlea activation.

Classic theories of hearing have considered only a translational component (piston-like component) of the stapes motion as being the effective stimulus for cochlear activation and thus the sensation of hearing. Our previous study (Huber et al., 2008) qualitatively showed that rotational components around the long and short axes of the footplate (rocking-like components) lead to cochlear activation as well. In this study, the contribution of the piston-like and rocking-like components of the stapes motion to cochlea activation was quantitatively investigated with measurements in live guinea pigs and a related mathematical description. The isolated stapes in anesthetized guinea pigs was stimulated by a three-axis piezoelectric actuator, and 3-D motions of the stapes and compound action potential (CAP) of the cochlea were measured simultaneously. The measured values were used to fit a hypothesis of the CAP as a linear combination of the logarithms of the piston-like and rocking-like components. Both the piston-like and rocking-like components activate cochlear responses when they exceed certain thresholds. These thresholds as well as the relation between CAP and intensity of the motion component were different for piston-like and rocking-like components. The threshold was found to be higher and the sensitivity lower for the rocking-like component than the corresponding values for the piston-like component. The influence of the rocking-like component was secondary in cases of piston-dominant motions of the stapes although it may become significant for low amplitudes of the piston-like component.

Errors in measurement of three-dimensional motions of the stapes using a laser Doppler vibrometer system.

Previous studies have suggested complex modes of physiological stapes motions based upon various measurements. The goal of this study was to analyze the detailed errors in measurement of the complex stapes motions using laser Doppler vibrometer (LDV) systems, which are highly sensitive to the stimulation intensity and the exact angulations of the stapes. Stapes motions were measured with acoustic stimuli as well as mechanical stimuli using a custom-made three-axis piezoelectric actuator, and errors in the motion components were analyzed. The ratio of error in each motion component was reduced by increasing the magnitude of the stimuli, but the improvement was limited when the motion component was small relative to other components. This problem was solved with an improved reflectivity on the measurement surface. Errors in estimating the position of the stapes also caused errors on the coordinates of the measurement points and the laser beam direction relative to the stapes footplate, thus producing errors in the 3-D motion components. This effect was small when the position error of the stapes footplate did not exceed 5 degrees.

Complex stapes motions in human ears.

It has been reported that the physiological motion of the stapes in human and several animals in response to acoustic stimulation is mainly piston-like at low frequencies. At higher frequencies, the pattern includes rocking motions around the long and short axes of the footplate in human and animal ears. Measurements of such extended stapes motions are highly sensitive to the exact angulation of the stapes in relation to the measurement devices and to measurement errors. In this study, velocity in a specific direction was measured at multiple points on the footplates of human temporal bones using a Scanning Laser Doppler Vibrometer (SLDV) system, and the elementary components of the stapes motions, which were the piston-like motion and the rocking motions about the short and long axes of the footplate, were calculated from the measurements. The angular position of a laser beam with respect to the stapes and coordinates of the measurement points on the footplate plane were calculated by correlation between the SLDV measurement frame and the footplate-fixed frame, which was obtained from micro-CT images. The ratios of the rocking motions relative to the piston-like motion increased with frequency and reached a maximum around 7 kHz.A novel method for quantitatively assessing measurements of complex stapes motions and error boundaries of the motion components is presented. In the frequency range of 0.5 to 8 kHz, the magnitudes of the piston-like and two rocking motions were larger than estimated values of the corresponding upper error bounds.