EP4531676A1

EP4531676A1 - An algorithm that automates and interprets hearing tests

Info

Publication number: EP4531676A1
Application number: EP22945061.4A
Authority: EP
Inventors: Emre Bayram; Sebahattin ÜNLÜ; Nihat Cengiz PAMUK; Ersin Onur ERDOGAN; Ismet POYRAZ
Original assignee: Aihear Teknoloji AS
Current assignee: Aihear Teknoloji AS
Priority date: 2022-05-30
Filing date: 2022-05-30
Publication date: 2025-04-09
Also published as: EP4531676A4; WO2023234881A1

Abstract

The present invention relates to an artificial intelligence-based algorithm that automatically performs and interprets hearing tests, which allows for more accurate and accurate results by allowing hearing tests to be performed automatically by integrating them into digital systems, and its feature; each of the current situation (M) for the patient (H), depending on perception, consciousness and that gave the existing answers (C) or patient (H), by sent or unsent expects the software to the software for each entry in that case, the system will take action rules already defined to the system in which data in the form of the information that was where the entrance to the base, the identification module (I), the patient (H), for the measurement and evaluation of hearing health pure sound, pure sound masking test module, speech and speech masking test module, SISI test module, ABLE test module, AMBL test module, TEN module, Weber test module, Tone Decay test module, Stenger test module, Tinnitus therapy and Tinnitus test module, MED test module it includes machine learning and artificial intelligence module (2), which contains and was developed as (rule-based system).

Description

AN ALGORITHM THAT AUTOMATES AND INTERPRETS HEARING TESTS

Technical Field:

The present invention relates to an algorithm that automates and interprets subjective hearing tests on both audiometer devices and web and application services in an artificial intelligence-based manner that allows for more accurate and accurate results by allowing hearing tests to be performed automatically by integrating them into digital systems.

State of the Art:

Currently, audiometry tests (hearing tests) are performed that cover a number of measurements to measure hearing function and assess hearing health. Audiometry tests include comprehensive tests performed to determine the hearing abilities of people and guide the diagnosis, treatment and rehabilitation of hearing loss. Audiometry tests are performed using audiometers that have direct contact with the patient and are used manually by clinicians.

Subjective hearing tests, audiometry test, which is a variant of the procedure in the patient with different levels of intensity and frequency of the audio signal depending on the signal and sending the consciousness and perception of the patient against the answers by expert clinicians as a result of evaluation in the form of a report to be submitted to the physician of the findings is carried out. For the correct diagnosis and rehabilitation in subjective hearing tests, it is necessary to act in a coordinated manner between the specialist and the patient.

In the current system, when performing subjective audiometry tests, the patient's responses are received in accordance with the expert's instructions, and all these responses are recorded manually by the specialist. The fact that the specialist needs to perform this operation in detail for many tests creates a large workload. The distraction and concentration disorder caused by excessive workload in the specialist causes the error rates in the test results to increase. At the same time, since the evaluation of the patient's responses depending on his perception may vary from specialist to specialist, no clear inference can be made about the accuracy and accuracy of the results. In addition, the fact that the success of the test results depends on the specialist's knowledge and experience also creates a doubt as to whether accurate results have been achieved for diagnosis and rehabilitation.

In the patent application numbered US20100281982A1 a hearing test and screening system for testing and screening the hearing acuity of a person which includes a computer having built therein a graphic-control program for generating a pure tone signal required for a hearing test, an audio apparatus having a connection port coupled to the computer for receiving the pure tone signal from the computer and outputting a set of test signals, each having 3 frequencies minimum, at a constant sound volume upon receipt of the pure tone signal and a response button for operation by the test person to provide a feedback signal to the computer, and earphones coupled to the audio apparatus for enabling the test person to hear the set of test signals outputted by the audio apparatus.

In the application given above, the system subject to the invention limits the scope of the invention by measuring only pure sound audiometry. In addition, the determination of the minimum measurement threshold value as a result of continuous adjustments by the specialist in the method provided for the hearing threshold measurements required for pure audio audiometry is an indication that the system still works largely dependent on the specialist. This process both causes a waste of time and creates an obstacle to the automation of the system.

In the patent application numbered CN205163084U the utility model discloses a multi - functional audiometer, include: a drive circuit for controlling the controller of audiometer operation, be used for according to predetermined rule start request module startup module, be used for providing the audiometer and detect the request module of function and be used for driving the audiometer operation, the controller pass through drive circuit respectively with startup module, request module connect. Compared with the prior art, the utility model discloses can satisfy to the hearing test demand under the multiple situations, simultaneously, realize better test accuracy. Special detection unit, SISI detection, ABLE (alternative binaural loudness balance) detection, STENGER it is used to provide detection and HIS detection. In the examination method, the Pure Tone hearing threshold of the patient's two ears is determined first.

The modules containing the algorithms used to measure and evaluate hearing health are not specifically included in the application given above. Only a limited number of detection functions are included within the scope of the interrogation module

As a result, the above-mentioned subjective hearing tests, which can overcome the disadvantages of making autonomous, evaluating, interpreting, enabling more precise and accurate results to be obtained, experts and specialists routine process that reduces workload and commitment to challenging, easy, and practical application with timesaving, the margin for error is low, diagnosis, treatment and rehabilitation there is a need for a new technology based on artificial intelligence for the right guide.

Description of the Invention:

This invention is an algorithm that automates and interprets hearing tests, and its feature is that it is a new technology based on artificial intelligence that performs hearing tests autonomously, which allows you to get more accurate and specific results saves time with easy and practical application.

The invention is based on the integration of hearing tests into digital systems and the measurement and evaluation of these tests autonomously in order to achieve all the objectives mentioned above and which will emerge from the detailed description below.

The application and functioning of hearing tests in the system subject to the invention is designed in accordance with the information provided by field experts and supported by the scientific literature. The invention provides for the autonomous realization of the hearing test process, not allowing errors and illusions that may arise from the lack of knowledge and experience of specialists to affect the test result.

By integrating the invention into audiometers, it is possible to reduce the routine and laborious workload of the personnel who carry out the hearing test. In this way, the lack of attention and concentration caused by excessive workload on the specialist does not affect the accuracy of the test results.

By integrating the system subject to the invention into audiometers, the need for an experienced specialist in the field of autonomous realization of the routine process regardless of the specialist who conducts the hearing test and the external dependence of the system are reduced.

Instead of applying the test measurements in the invention on the basis of subjective inferences based on the expert's own request, applying them based on proven procedures ensures that the process proceeds reliably and decisively without any doubt about its accuracy.

The system of the invention, depending on the patient's perception measurements during the test, the response retrieved from the source data in line with the literature, interpretation, evaluation, and hearing health for any doubt about the correctness of a result with progress in the process provides a reliable and stable manner.

The process of the invention in accordance with both perform hearing tests and test results of an autonomous low margin of error the source data retrieved from the literature in accordance with autonomous interpretation, evaluation, and diagnosis and rehabilitation for hearing health information is obtained for a result with the appropriate guide.

The characteristic structural features and all the advantages of the system of the invention outlined below and shapes, and these shapes by making references to the detailed written description will be understood more clearly and, therefore, must be made in consideration of the evaluation of these figures and detailed description.

Description of the Figures:

The invention will be described with reference to the accompanying figures, so that the features of the invention will be more clearly understood and appreciated, but the purpose of this is not to limit the invention to these certain regulations. On the contrary, it is intended to cover all alternatives, changes and equivalences that can be included in the area of the invention defined by the accompanying claims. The details shown should be understood that they are shown only for the purpose of describing the preferred embodiments of the present invention and are presented in order to provide the most convenient and easily understandable description of both the shaping of methods and the rules and conceptual features of the invention. In these drawings;

Figure 1 The subject of the invention is a schematic view of the elements contained in the system.

The figures to help understand the present invention are numbered as indicated in the attached image and are given below along with their names.

Description of References:

1. Identification Module

2. Artificial Intelligence Module

A. Medical History

M. The Current Situation

C. Answers

K. Defined Rules

T. Reaction

U. Stimulus

H. Patient Description Of the Invention:

Invention; the current status of each (M) for the patient's (H) answers that gave the perception depending on the existing awareness and (C) the patient or (H) to the software by software that expects sent or unsent case for each entry, the system will take action rules already defined to the system in which data in the form of the information that was where the entrance to the base, the identification module (1), the patient (H) the measurement and evaluation of hearing health for pure sound, pure sound masking test module, speech and speech masking test module, SISI test module, ABLB test module, AMBL test module, The TEN test module, Weber test module, Tone Decay test module, Stenger test module, Tinnitus therapy test module it includes machine learning and/or artificial intelligence module (2), which includes the MLD test module and was developed as a rule-based system (rule-based system).).

The invention includes a machine learning and/or rule-based artificial intelligence module (2) and machine learning methods in which outputs are obtained by interpreting these data as a result of the introduction of predefined rules into the system database by the identification module (1).

The invention includes a machine learning and/or rule-based artificial intelligence module (2) that allows test modules to autonomously initiate, terminate, and interpret the results of tests in accordance with the stimulation and patient (H) reaction (T) of the relevant tests.

The invention includes a rule-based artificial intelligence module (2) that interprets test results by the MLD test module and the Tinnitus test module that it contains.

The invention includes a machine learning and/or rule-based artificial intelligence module (2) that interprets data in a combined way as a result of recording predefined rules in the system by a specialist in the identification module (1) along with data received in accordance with the existing answers (C) given by the patient (H) depending on his consciousness and perception. The invention includes a machine learning and/or artificial intelligence module (2) in which medical history (A) information is entered by a specialist, from which the patient's (H) history is obtained, in order to assist in diagnosis after various questions have been asked to the patient (H) before starting the testing process.

The invention includes a machine learning and/or rule-based artificial intelligence module (2) that performs the process of assessing the patient's hearing according to the defined Rules (K) contained in the identification module (1) with medical history (A) information entered into the system about the patient (H).

The invention is achieved by transmitting the first stimulus (U) to the patient (H) the start of the testing process is also from the patient (H) the stimulus (U) in reaction (T) to the expected phase of alert type, intensity, frequency, duration, modulation, continuity, orientation, or identification of necessary changes in regard to a different module (1) in the defined rules (K) that makes it according to a machine learning and/or rule-based artificial intelligence module (2).

The invention provides information about the medical history (A) entered about the patient (H) before starting the test (may not be received), the push-button state corresponding to the patient's (H) answers (C) from the very beginning of the test until the current state (M), the current state (M) in the test (type, intensity, frequency of sound, how many ms the sound goes to which ear, etc.) and includes machine learning and/or artificial intelligence module (2), which performs the operations of checking the patient's (H) answers to the current situation (M).

Detailed Description of The Invention:

The present invention comprises an algorithm for integrating the implementation and operation of hearing tests into digital systems and autonomously performing measurement and evaluation of these tests. The machine learning and/or artificial intelligence module (2), developed as a rule-based system (rule-based system), is included in the system that is the subject of the invention. With this rule-based system, information from the environment is stored in the database and outputs are obtained by interpreting these data in accordance with the rules defined in advance by the identification module (1) within the system. The algorithms used to measure and evaluate the hearing health used in the invention are pure sound and pure sound masking test module, speech and speech masking test module, SISI test module, ABLB test module, AMBL test module, TEN test module, Weber test module, Tone Decay test module, Stenger test module, Tinnitus therapy test module it includes the MLD test module. These test modules include starting, ending and processing steps of the relevant tests autonomously in accordance with stimulation and patient (H) responses (T) of the relevant tests until the interpretation of the results. The MLD test module and the Tinnitus test module are intended only for the interpretation of the obtained test results.

In the system subject to the invention, the machine learning and/or artificial intelligence module (2) in the autonomous audiometer software performs these operations according to the defined rules as a result of defining the relevant tests in the database via the identification module (1). The processing steps performed here are described as follows.

1. Defining the rules: It is the first processing step of the machine learning and/or artificial intelligence module (2). Since the artificial intelligence system will replace the expert in the autonomous audiometer, the subject of the invention, it is defined which action the system will take as a result of the current answers (C) it gives depending on the patient's (H) consciousness and perception for each current situation (M). For each current situation, the patient's (H) reaction (T) are defined as input data to the software. In the algorithm, an action is defined to be taken for each input sent to the software by patient (H) or not sent while the software is waiting. Then this algorithm is converted into software through programming languages.

2, Information to be evaluated by the artificial intelligence system: In the autonomous invention system, the second processing step of the machine learning and/or artificial intelligence module (2) is to receive input data to the artificial intelligence system. In the system subject to the invention, even if the actual data is to be taken in accordance with the existing answers (C) given by the patient (H) depending on his consciousness and perception, a number of input data are also defined by the expert, even if the main data are to be taken in accordance with the data that will be evaluated both by the patient (H) and by the specialist as input to the system is realized by triggering a method in the software (trigger) by pressing a button.

The data that experts can define as input to the system are the commands to start the test and stop the test. In the system subject to the invention, the machine learning and/or artificial intelligence module (2) receives the start command of the test from the expert and immediately applies the ‘stop test’ command that may come from the expert in any case, even if the test process is not over.

In the system subject to the invention, the input data received from patient (H) to the machine learning and/or artificial intelligence module (2) and the main thing is that the patient (H) presses the button, presses the button and then releases it, and never presses the button. Three buttons with separate functions defined for each test to the artificial intelligence system are evaluated autonomously by the system for each test to be performed. Because the reaction (T) to which the buttons correspond for each test may differ. In addition, the frequency of pressing the button and the timing of pressing the button and the timing of the test flow are controlled in the algorithm here. Accordingly, the patient evaluates the cooperation and if there are inappropriate responses, the system stops the test.

The button-pressing actions in the system subject to the invention trigger the methods defined in the software. In the system subject to the invention, the function of releasing the button after pressing the patient's (H) button is also evaluated separately. The button release actions trigger the methods defined in the software. If the machine learning and/or artificial intelligence module in the system subject to the invention cannot receive any input data about the patient (H) pressing the button, it takes action by evaluating this situation according to the defined rules. 3, Evaluation of inputs to the artificial intelligence system: The third and final step of the machine learning and/or artificial intelligence module (2) in the system subject to the invention is to evaluate the input data of the artificial intelligence system.

In the system subject to the invention, before starting the testing process, the patient (H) is asked various questions, and then the medical history (A) information is entered by the expert into the machine learning and/or artificial intelligence module (2) (agent), from which the patient's (H) history is obtained to assist in diagnosis. After the medical history (A) information entered into the system about patient (H) is evaluated according to the defined rules (K), the testing process begins. The test process begins with the transmission of the first warning (U) to the patient (H), and a reaction (T) is expected from the patient (H) to this warning (U). At this stage, the necessary changes in the type, intensity, frequency, duration, modulation, continuity, orientation or any other aspect of the warning are made according to the defined rules.

The machine learning and/or artificial intelligence module (2) in the system subject to the invention evaluates the issues contained in the following articles at every step according to predefined rules. a) Medical history (A) information entered about patient (H) before starting the test, b) From the very beginning of the test to the current state (M), the patient (H) said,

“I heard, I didn't hear, I heard from the left, I heard from the right, I heard 30 seconds, I also heard from the middle, what I heard was different, what I heard was the same." button presses corresponding to many different types of diversifiable answers (C) in the form of, c) The current state (M) in the test (the type of sound, its intensity, frequency, how many ms the sound travels to which ear, etc.) d) The patient's (H) answer (C) to the current situation (M)

The machine learning and/or artificial intelligence module (2) in the system subject to the invention takes action against these situations after evaluating the issues contained in the above articles according to the predefined rules at each step and during the test period. This action in question determines the type, intensity, frequency, duration, orientation of the stimulus (U), etc. one of the options is to include a new warning (U) in the test by changing it, or to continue the test by removing a warning (U) from the test, or to finish the test directly.

Claims

CLAIMS - The invention relates to an algorithm that automates and interprets hearing tests, the feature of which is; each of the current situation (M), the patient (H), depending on perception, consciousness and that gave the existing answers (C) or the patient (H) in case the software by software that expects to sent or unsent for each entry, the system will take action rules already defined to the system in which data in the form of the information that was where the entrance to the base, the identification module (1), for the patient's (H) pure voice and pure voice masking test module, speech and speech masking test module for measuring and evaluating hearing health, SISI test module, ABLB test module, AMBL test module, SKIN test module, Weber test module, Tone Decay test module, Stenger test module, Tinnitus therapy and Tinnitus test module, The MLD test module it includes machine learning and/or artificial intelligence module (2), which contains and was developed as a rule-based system (rule-based system). - The algorithm mentioned in Claim 1, which automates and interprets hearing tests, has a feature; it is characterized by the fact that the defined rules are entered into the system database by the definition module (1) as a result of which these data are interpreted and the results are obtained by including rule-based artificial intelligence module (2) and machine learning methods. - As mentioned in Claim 1 it is an algorithm that automates and interprets hearing tests, and its feature is that the test modules include a machine learning and/or rule-based artificial intelligence module (2) that performs the operations of autonomously starting, terminating, and interpreting the results of tests in accordance with the stimulation and patient (H) reaction (T) of the relevant tests. - As mentioned in Claim 1 it is an algorithm that automates and interprets hearing tests, and its feature is that it is characterized by containing a rule-based artificial intelligence module (2) that interprets test results by the MLD test module and the Tinnitus test module that it contains. - As mentioned in Claim 1 it is an algorithm that automatically performs and interprets hearing tests, its feature is; the patient (H), along with the data received in accordance with the existing answers (C), which he gives depending on his consciousness and perception by the specialist it is characterized by the fact that the system includes a machine learning and/or rule-based artificial intelligence module (2), which interprets the data in a combined way as a result of saving the predefined rules into the definition module (1). - As mentioned in Claim 1 it is an algorithm that automates and interprets hearing tests, and its feature is that it is characterized by the fact that it includes a machine learning and/or artificial intelligence module (2), where the medical history (A) information is entered by a specialist, from which the patient's (H) history is obtained to help diagnose after various questions have been asked to the patient (H) before starting the testing process. - In accordance with Claim 1 or Claim 6 automatic hearing tests, which makes the algorithm interprets and property; sick (H) entered into the system about the medical history (a) the information with the identification module (1) in the defined rules (K) according to the hearing status of the patient engaged in the process of evaluation of the machine learning and/or rule-based artificial intelligence module (2). - As mentioned in Claim 1 automatic hearing tests, and interprets the algorithm and feature; the first warning (U) the patient (H) transmitting by means of the start of the test process and the patient (H) the warning (U) reaction (T), expected phase of alert type, intensity, frequency, duration, modulation, continuity, orientation, or identification of necessary changes in regard to a different module (1) in the defined rules (K) that makes it according to a machine learning and/or rule-based artificial intelligence module (2). - As mentioned in Claim 1 it is an algorithm that automates and interprets hearing tests, and its feature is; medical history (A) information entered about patient (H) before starting the test (may not be received), from the very beginning of the test to the current state (M), the button press corresponding to the patient's (H) answers (C), the current state (M) in the test (sound type, intensity, frequency, how many ms the sound goes to which ear, etc.) and is characterized by the fact that the patient's (H) answers (C) to the current situation (M) include the machine learning and/or artificial intelligence module (2), which performs the operations of checking.