Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-10617334-B2
Publication Date
2020-04-14
Expiration Date
Abstract
An optimization system for testing a patient's hearing comprising a controller, right and left ear pieces, and a memory. The controller is configured to provide a series of tones to the right and left ear pieces and receive feedback from the patient between each tone provided, indicating that the respective tone was detected in one of the right ear piece and the left ear piece or that the respective tone was not detected. The controller is configured to also generate a data point on an audiogram after receiving each feedback and, after each data point is generated, compute a statistical distribution based on the generated data points, wherein the statistical distribution includes a lowest confidence level. A subsequent tone is selected, each subsequent tone provided in the series of tones being a tone represented at the lowest confidence level in statistical distribution at the time of selection.
Core Innovation
The invention describes an optimization system for testing a patient’s hearing. An auditory-device processor/controller provides a series of tones to an ear piece and receives patient feedback indicating whether each tone was detected or not detected. After each feedback, the system generates a data point on an audiogram based on the respective feedback.
After generating the audiogram data points, the controller computes a statistical distribution that includes a lowest confidence level. The system then selects a subsequent tone represented at the lowest confidence level in the statistical distribution at the time of selection. This selection process is intended to efficiently estimate hearing thresholds using the evolving audiogram data.
In additional embodiments, the system is further configured to produce hearing-thresholds graph outputs, including an upper curve and a lower curve, derived from the series of tones and patient feedback. The disclosed approach also includes using a Gaussian process for computing the statistical distribution. The system may use right ear piece and left ear piece handling to generate corresponding right and left audiogram data points.
The broader disclosed system further includes classifying environmental sounds and selecting reference sound profiles to pair with stored parameter sets for automatically adjusting device parameters. The disclosed sound classification uses wavelet scattering transform, Fourier transform, spectrogram-based approaches, and clustering such as SVM and K-clustering. Geolocation filtering is also described as an optional input for selecting reference sound profiles.
Claims Coverage
The partial document provides two independent claims with additional dependent refinements. Across the independent claims, the main inventive features are tone-based hearing testing with patient-detection feedback, audiogram data-point generation, computation of a statistical distribution including a lowest confidence level, and selecting the next tone represented at that lowest confidence level.
Tone series with patient detection feedback
Providing a series of tones to an ear piece; receiving feedback from the patient between each tone provided, wherein each feedback indicates that the respective tone was detected or not detected.
Audiogram data points from feedback
Generating a data point on an audiogram after receiving each feedback, wherein each data point is based on the respective feedback.
Statistical distribution with lowest confidence level
After each data point is generated, compute a statistical distribution based on the generated data points, wherein the statistical distribution includes a lowest confidence level.
Select subsequent tone at lowest confidence level
Selecting a subsequent tone to provide in the series of tones, wherein each subsequent tone provided in the series of tones is a tone represented at the lowest confidence level in statistical distribution at the time of selection.
Controller-driven hearing test system architecture
An optimization system for testing a patient's hearing comprising a controller, an ear piece in audible communication with the controller, and a memory including instructions that cause the controller to perform the tone presentation, feedback receiving, audiogram data point generation, statistical distribution computation with a lowest confidence level, and selection of a subsequent tone at the lowest confidence level.
Method steps for hearing testing using lowest confidence selection
A method of testing a patient's hearing comprising providing a series of tones to an ear piece; receiving feedback from the patient between each tone provided; generating a data point on an audiogram after receiving each feedback; after each data point is generated, computing a statistical distribution including a lowest confidence level; and selecting a subsequent tone represented at the lowest confidence level in the statistical distribution at the time of selection.
Both independent claims center on an adaptive hearing test that uses patient detection and non-detection feedback to build audiogram data points, computes a statistical distribution with a lowest confidence level, and selects the next tone represented at that lowest confidence level at the time of selection. The dependent claim refinements add Gaussian-process computation, hearing-thresholds graph outputs with upper and lower curves, and optional right and left ear piece handling.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?