WO2019146809A1 - Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur - Google Patents

Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur Download PDF

Info

Publication number
WO2019146809A1
WO2019146809A1 PCT/KR2018/001048 KR2018001048W WO2019146809A1 WO 2019146809 A1 WO2019146809 A1 WO 2019146809A1 KR 2018001048 W KR2018001048 W KR 2018001048W WO 2019146809 A1 WO2019146809 A1 WO 2019146809A1
Authority
WO
WIPO (PCT)
Prior art keywords
gpi
tinnitus
stimulus
amplitude
ratio
Prior art date
Application number
PCT/KR2018/001048
Other languages
English (en)
Korean (ko)
Inventor
구윤서
서명환
김희찬
안중우
오승하
이준호
Original Assignee
서울대학교산학협력단
서울대학교병원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 서울대학교산학협력단, 서울대학교병원 filed Critical 서울대학교산학협력단
Priority to PCT/KR2018/001048 priority Critical patent/WO2019146809A1/fr
Publication of WO2019146809A1 publication Critical patent/WO2019146809A1/fr

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/12—Audiometering

Definitions

  • the present invention relates to a tinnitus test apparatus and method using a classifier, and more particularly, to a tinnitus test apparatus and method for testing whether a subject is aliased by inputting auditory evoked potential (AEP) measured by a subject into a classifier .
  • AEP auditory evoked potential
  • Tinnitus is a subjective feeling of noise heard in the ear. In other words, it feels that sound is heard in a situation where there is no auditory stimulation from the outside. In a totally insulated quiet room, about 95% of ordinary people feel less than 20 dB of tinnitus, but this is not clinically a tinnitus, and it is called tinnitus that you may notice noise or noise that annoys you.
  • Symptoms of this tinnitus include moderate to severe tinnitus (8%) and severe tinnitus (1%). Patients who complain of tinnitus feel various noises from their ears or head, and if they are severe, they will suffer from everyday life. Although the frequency of external noise exposure increases in modern society, tinnitus patients are increasingly recognized as an incurable disease. For these reasons, there is a growing demand for accurate diagnosis and treatment of tinnitus.
  • the conventional tinnitogram used in humans has a merit that it is relatively easy and simple to diagnose the presence or absence of tinnitus as a subjective test method.
  • tinnitus there is a problem in accurately diagnosing the elderly elderly or infants with tinnitus who are not able to understand the test procedure or expressiveness because it is a subjective test that relies on the cooperation of the subject with the stimulus sound.
  • the victim's economic compensation is related to external noise exposure or accident, it is necessary to make an objective judgment about the presence of tinnitus.
  • a problem to be solved in the embodiment of the present invention is to provide a technique for checking whether or not a person has tinnitus by using a clearing evoked potential and a classifier of an examinee so as to objectively examine the person's tinnitus.
  • the tinnitus test apparatus includes a first stimulus sound including background noise and pulse noise, a second stimulus sound including a background sound with a silent gap and a pulse sound, A first waveform of an auditory evoked potential (AEP) of the subject according to the first stimulus sound and a first waveform of an auditory evoked potential (AEP) of the subject according to the second stimulus sound, A GPI calculating unit for measuring a GPI (Gap Pre-pulse Inhibition) ratio, which is a ratio between the amplitude between the pole points of the first waveform and the pole points of the second waveform, And a classifier for storing a reference GPI ratio measured in a plurality of tinnitus patients and a normal group and receiving a GPI ratio measured by the subject and examining whether the subject is aliased based on the reference GPI ratio do .
  • GPI Gap Pre-pulse Inhibition
  • the stimulus sound generator generates a plurality of second stimulus sounds having different silence intervals
  • the amplitude and GPI calculator calculate a maximum and a minimum, respectively, extracted from a plurality of second waveforms corresponding to the plurality of second stimulus sounds
  • a plurality of GPI ratios may be measured based on a plurality of amplitudes between the poles, or a GPI ratio may be measured based on the largest amplitude among the plurality of amplitudes.
  • the silence interval of each second stimulus note may be 10 ms or more and 100 ms or less.
  • the amplitude between the pole points may be an amplitude between a maximum pole point and a minimum pole point of the first waveform and an amplitude between a maximum pole point and a minimum pole point of the second waveform.
  • the stimulus sound generator generates a plurality of first stimulus notes and second stimulus notes having different amplitudes of the background sounds or frequencies of the background sounds
  • the amplitude and GPI calculator generate the plurality of first and second stimulus notes
  • the classifier measures a plurality of GPI ratios measured from the subject and calculates an amplitude of the background sound or a frequency of the background sound with respect to a plurality of tinnitus patients and a normal group, And the GPI ratio to determine the tinnitus degree and tinnitus frequency of the subject.
  • the classifier receives the measured GPI ratio, and determines whether or not the subject's tinnitus according to the input GPI ratio from a classification model derived from a plurality of reference GPI ratios measured in a plurality of tinnitus patients and a normal group Can be inspected.
  • the classifier receives the amplitudes of the first waveform or the second waveform and the measured GPI ratio, and derives a result of the tinnitus patient or normal group using the determined reference amplitude and the reference GPI ratio It is possible to check whether the subject is aliased according to the inputted amplitude and the GPI ratio.
  • the classifier further stores a reference spectral power density of a reference GPIAS (Gap Pre-pulse Inhibition of Acoustic Start) and a clearing-induced potential measured in the plurality of tinnitus patients and the normal group,
  • the spectral power density of the clearing-induced potential is further input, and it is possible to check whether the subject is aliased based on the classification model derived using the reference GPI ratio, the reference GPIAS, and the reference spectral power density.
  • a tinnitus checking method includes a first stimulus sound including background noise and pulse noise, a second stimulus sound including a background sound with a silent gap and a pulse sound, A second waveform of the auditory evoked potential (AEP) of the subject according to the first stimulus sound and a second waveform of the auditory evoked potential of the subject according to the second stimulus sound; A GPI calculating step of measuring a GPI (Gap Pre-pulse Inhibition) ratio, which is a ratio between the amplitude between the pole points of the first waveform and the pole points of the second waveform, A diagnostic step of receiving a measured GPI ratio of the subject based on a reference GPI ratio measured in a plurality of tinnitus patients and a normal group and examining whether the subject is aliased based on the reference GPI ratio; Can.
  • GPI Gap Pre-pulse Inhibition
  • the stimulus sound generation step may include generating a plurality of second stimulus sounds having different silent intervals
  • the step of calculating the amplitude and GPI may include the steps of generating a plurality of second stimulus sounds corresponding to the plurality of second stimulus sounds Using the amplitudes between the extracted maximum and minimum pole points or measuring the GPI ratio based on the largest amplitude.
  • the silence interval of each second stimulus note may be 10 ms or more and 100 ms or less.
  • the amplitude between the pole points may be an amplitude between a maximum pole point and a minimum pole point of the first waveform and an amplitude between a maximum pole point and a minimum pole point of the second waveform.
  • the diagnosing step may include receiving the measured GPI ratio and determining whether the subject's tinnitus according to the inputted GPI ratio from a classification model derived from a plurality of reference GPI ratios measured in a plurality of tinnitus patients and a normal group . ≪ / RTI >
  • the generating of the stimulus sound may include generating a plurality of first stimulus notes and a second stimulus note having different amplitudes of the background sounds or different frequencies of the background sounds, 1 < / RTI > and a second stimulus sound, wherein the diagnostic step comprises receiving a plurality of measured GPI ratios from the subject and determining an amplitude of the background sound for a plurality of tinnitus patients and a normal group Or determining the tinnitus degree and tinnitus frequency of the subject from the classification model derived using the amplitude and the GPI ratio measured at different frequency of the background sound.
  • the diagnostic step further stores the amplitude between the poles of the first waveform measured from the plurality of tinnitus patients and the normal group as a reference amplitude, and further comprises, in addition to the measured GPI ratio of the subject, And comparing the reference GPI ratio with the reference amplitude to check whether the subject is aliased.
  • the diagnosis step further stores a reference spectral power density of a reference GPIAS (Gap Pre-pulse Inhibition of Acoustic Start) and a clearing evoked potential measured in the plurality of tinnitus patients and the normal group, and the measured GPIAS Or the spectral power density of the clearing-induced potential, and examining whether the subject is aliased based on the classification model derived using the reference GPI ratio, the reference GPIAS, and the reference spectral power density. have.
  • a reference GPIAS Gap Pre-pulse Inhibition of Acoustic Start
  • the auditory evoked potential of the subject is measured through the stimulus sound which can confirm the degree of tinnitus, and the information analyzed through the auditory evoked potential is input to the classifier to objectively examine the subject's tinnitus .
  • FIG. 1 is a block diagram showing a configuration of a tinnitus test apparatus according to an embodiment of the present invention.
  • FIG. 2 is a diagram illustrating a configuration of a first stimulus note and a second stimulus note according to an embodiment.
  • FIG. 3 is a graph showing the results of comparing the clearing evoked potentials between the tinnitus patient group and the normal group when the silent interval of the second stimulation sound was changed to 20 ms and 100 ms.
  • FIG. 4 is a graph showing the results of comparing amplitudes between a maximum pole point and a minimum pole point extracted from the clearing evoked potentials of the tinnitus patient group and the normal group according to a second stimulus sound having a different silence interval.
  • FIG. 5 is a graph showing the results of comparing the GPI ratios of the tinnitus patient group and the normal group according to the results of FIG. 4.
  • FIG. 5 is a graph showing the results of comparing the GPI ratios of the tinnitus patient group and the normal group according to the results of FIG. 4.
  • FIG. 6 is a diagram for explaining an algorithm for diagnosing a tinnitus patient using a feature vector according to an embodiment.
  • FIG. 7 is a flowchart showing a procedure of a tinnitus checking method according to an embodiment.
  • FIG. 8 is a flowchart illustrating a procedure of a method for selecting a GPI ratio or a GPI ratio according to a plurality of second stimulus tones according to an embodiment to check the tinnitus.
  • FIG. 9 is a flowchart showing a procedure of a method for checking the degree of tinnitality or tinnitus frequency of a subject according to an embodiment.
  • a component when referred to as being connected or connected to another component, it may be directly connected or connected to the other component, but it should be understood that there may be other components in between.
  • FIG. 1 is a block diagram showing the configuration of a tinnitus test apparatus 100 according to an embodiment.
  • the tinnitus test apparatus 100 includes a stimulus tone generator 110, a clarification induced potential measurement unit 120, an amplitude and GPI calculator 130, and a classifier 140.
  • the stimulus sound generator 110 generates a stimulus sound having a silent gap having a predetermined duration while the first stimulus note including the first background noise and the pulse noise and the second background sound continue And then generates a second stimulus sound including a pulse sound generated during the third background sound. At this time, the stimulus sound generator 110 may generate a plurality of second stimulus sounds with different durations of silent intervals.
  • the background sound refers to a sound that lasts for a predetermined time at a predetermined frequency and a predetermined amplitude
  • the pulse sound refers to a sound having a strength greater than a background sound for a time shorter than a predetermined time duration.
  • FIG. 2 is a diagram illustrating a configuration of a first stimulus note and a second stimulus note according to an embodiment.
  • the first stimulus sound may include a pulse sound having a duration of 20 ms in the middle of the first background sound for about 1000 to 3000 ms
  • the second stimulus sound may include a pulse sound of 1000 to 3000 ms
  • a silence interval having a duration of 20 ms starting from 200 to 120 ms before the pulse sound.
  • a pulse having a duration of 20 ms after the third background sound is continued for 100 ms at the same amplitude and same frequency again Sound.
  • the graph of (A) represents a first stimulus sound including a first background sound that lasts for a predetermined time at a predetermined frequency and a predetermined amplitude, and a pulse sound that has a duration shorter than a background sound in the middle of a continuous sound
  • the second background sound continues for a predetermined time at a predetermined frequency and a predetermined amplitude, and after a silence interval for a predetermined time, the third background sound continues at the same frequency and the same amplitude again,
  • a second stimulus sound including pulse noise.
  • the lengths of the first, second and third background sounds described above, the length of the silent intervals, and the length between the end of the silent interval and the start of the pulse sound are exemplary values, but the invention is not limited thereto.
  • the first stimulus note and the second stimulus note generated by the stimulus note generation unit 110 may be heard by the subject through an amplifier, a headphone, or the like.
  • a sound level meter such as a probe microphone for measuring the sound pressure in front of the eardrum in the ear canal is connected to the stimulus sound generator 110 so that a stimulus sound of a strength appropriate for the subject is applied, can do.
  • the clarification induced potential measurement unit 120 measures a first waveform of the auditory evoked potential (AEP) of the subject according to the first stimulus sound and a second waveform of the stimulation induced potential of the subject according to the second stimulation sound do.
  • AEP auditory evoked potential
  • the celestial evoked potential is the response measured by measuring the electrophysiological activity of the auditory nerve. More specifically, the electric nerve stimulation generated in the cochlear cells of the cochlea after the transmission of the auditory organ by acoustic stimulation is measured on the scalp.
  • the auditory evoked potential is the potential generated from the auditory pathway from the cochlea of the inner ear to the auditory cortex of the cerebral cortex after the stimulation of the sound, do.
  • the auditory brainstem response is called the auditory brainstem response in about 10 ms after the stimulus and seven positive poles generally up to I - VII in the auditory brainstem response appear.
  • the response that occurs in the thalamo-cortical pathway between the brainstem and the cerebral auditory cortex is called the auditory middle-latency response and is called the N0, P0, Na, Pa, Nb, .
  • the responses that occur in the cerebral auditory cortex are called the auditory late response, and include the poles typically referred to as P1, N1, P2, and N2.
  • the amplitude of the pole referred to in the specification of the present invention means the amplitude of at least one pole among a plurality of poles appearing at the clearing induced potential, and is not limited to any pole. Also, the magnitude of the amplitude may be the magnitude of the baseline-to-peak itself and may be the magnitude of the peak-to-peak magnitude.
  • FIG. 3 is a graph showing the results of comparing the clearing evoked potentials between the tinnitus patient group and the normal group when the silent interval of the second stimulation sound was changed to 20 ms and 100 ms.
  • the auditory evoked potentials for the first stimulus sound and the second stimulus stimulus are compared with each other.
  • the difference between the maximum pole point and the minimum pole point in the clarification induced potential according to the second stimulus sound is smaller than the difference between the pole points in the clarification induced potential according to the first stimulus sound.
  • the central nervous system contrasts with the pulse sound that occurs after the silence interval, so that the change of the cognitive evoked potential is small, It is similar to the response to the first stimulus tone because it does not recognize the silence interval due to the influence of the tinnitus.
  • the embodiment of the present invention is based on the above principle and is based on the first waveform of the auditory evoked potential of the subject according to the first stimulus sound which does not include the silence interval and the first waveform of the auditory evoked potential of the subject in accordance with the second stimulation sound including the silence interval
  • An index for judging whether or not the patient is tinnitized is measured by measuring a second waveform of the induced potential and measuring a GPI (Gap Pre-pulse Inhibition) ratio, which is a ratio between the amplitude between the pole points of the first waveform and the pole points of the second waveform .
  • GPI Gap Pre-pulse Inhibition
  • the amplitude and GPI calculator 130 measures the amplitude between the pole points of the first waveform and the pole points of the second waveform from the first waveform and the second waveform measured by the clarification induced potential measurement unit 120 , The ratio of the amplitude of the first waveform to the amplitude of the second waveform is measured.
  • the tinnitus patient may be referred to as the first stimulus note and the second stimulus note in terms of the ratio between the amplitude of the first waveform and the amplitude of the second waveform.
  • the GPI ratio is close to 1, and the GPI ratio of the normal group is smaller than the GPI ratio of the tinnitus patients because the difference between the stimulation - induced potentials is small.
  • the auditory evoked potential is the potential generated from the auditory pathway from the cochlea of the inner ear to the auditory cortex of the cerebral cortex after the stimulation of the sound, do.
  • the auditory brainstem response is called the auditory brainstem response in about 10 ms after the stimulus and seven positive poles generally up to I - VII in the auditory brainstem response appear.
  • the response that occurs in the thalamo-cortical pathway between the brainstem and the cerebral auditory cortex is called the auditory middle-latency response and is called the N0, P0, Na, Pa, Nb, .
  • the responses that occur in the cerebral auditory cortex are called the auditory late response, and include the poles typically referred to as P1, N1, P2, and N2.
  • the amplitude and the GPI calculating unit 130 can measure the GPI ratio by measuring the amplitudes of the N1 and P2 reactions and the amplitudes between N1 and P2, which are the extreme points of the clearing-induced potential in the late phase reaction.
  • the amplitude of the pole referred to in the present invention means the amplitude of one or more pole points among a plurality of pole points appearing at the clearing-induced potential, but as long as various types of EEG show similar patterns to N1 and P2 responses, You can check for the presence of tinnitus.
  • the magnitude of the amplitude may be the magnitude of the baseline-to-peak itself and may be the magnitude of the peak-to-peak magnitude.
  • FIG. 4 is a graph showing the results of comparing amplitudes between a maximum pole point and a minimum pole point extracted from the clarification evoked potentials of the tinnitus patient group and the normal group according to the second stimulus sound having a different silence interval, and FIG. And GPI ratios of the tinnitus patient group and the normal group.
  • the GPI ratio average and standard deviation of the tinnitus patient group were 0.82 0.30, The standard deviation was 0.79 ⁇ 0.20, indicating that there was no significant difference between the tinnitus and normal groups.
  • the above detailed values may vary according to the stimulus sound condition such as the pulse sound magnitude applied by the stimulus sound generator 110, but the effect of the presence or absence of tinnitus is influenced by the length of the silence interval included in the stimulus sound Can be confirmed.
  • the silence interval is less than 10 ms
  • the effect of the silence interval is not obtained even in the GPI ratio of the normal group.
  • the difference in the cognitive evoked potential due to the first stimulus sound and the second stimulus sound is not large.
  • the patients with tinnitus had a length enough to recognize the interval of silence.
  • the GPI ratio did not differ between the tinnitus patient and the normal group, the second stimulus sound with a silence interval of less than 10 ms or greater than 100 ms, It is not appropriate to distinguish between the two.
  • the stimulus sound generator 100 may generate a second stimulus sound including a silence interval of 10 ms or more and 100 ms or less.
  • the stimulus tone generator 110 generates a plurality of second stimulus sounds by adjusting at least two values having different silence intervals
  • the amplitude and GPI calculator 130 generates a plurality
  • a plurality of GPI ratios are measured based on a plurality of amplitudes between a maximum pole point and a minimum pole point extracted from a plurality of second waveforms corresponding to a second stimulus sound of the plurality of amplitudes or a GPI ratio is measured based on the largest amplitude among a plurality of amplitudes can do.
  • a plurality of GPI ratios are measured based on a plurality of amplitudes between a maximum pole point and a minimum pole point extracted from a plurality of second waveforms corresponding to a plurality of second magnetic pole sounds or a plurality of GPI ratios measured based on a plurality of amplitudes
  • the reason for use is as follows. According to the results of FIG. 5, it is easy to distinguish the GPI ratio between the tinnitus patient and the normal group when the silent interval is 10 to 100 ms. However, in measuring the tinnitus per individual, It is because.
  • the second waveform having the largest amplitude among the amplitudes between the maximum pole point and the minimum pole point extracted from the plurality of second waveforms is selected by dividing the silence interval of the second stimulation sound used when the amplitude of the second waveform is the largest, This means that it is easy to compare with the GPI ratio of the normal group.
  • the present embodiment it is possible to measure a plurality of GPI ratios based on a plurality of amplitudes between a maximum pole point and a minimum pole point extracted from a plurality of second waveforms corresponding to at least two second pole notes having different silence intervals,
  • the GPI ratio can be measured based on the largest amplitude of the amplitudes of the tinnitus patients.
  • the classifier 140 performs the function of assigning an input value to one of a plurality of classes.
  • the input value may be the cognitive evoked potential itself, or it may mean a set of one or more feature values extracted from the cognitive evoked potential, i. E. A feature vector.
  • the classifier 140 classifies the class into a class when new data that need to be classified are input using a plurality of raw data, which are feature values extracted from the cognitive evoked potential itself or the cognitive evoked potential, A classification model can be generated.
  • a classification model is generated by using the input of the classifier 140 as the input of the clearing-induced potential measured from a plurality of tinnitus patients and the normal group, or a classification model
  • the classification model can be generated from the pole-point amplitudes and the GPI ratios extracted from the clearing-induced potentials according to the first waveform of the induced potential or the second waveform.
  • the process of creating a classification model is called training or learning.
  • the classifier 140 classifies a feature space into a plurality of classes by using a linear classifier that linearly classifies using a decision hyperplane and a nonlinear classifier that classifies by using a decision hypersurface (nonlinear classifier).
  • the model may be classified according to whether the model is selected based on probability based on the model to be studied.
  • a naive Bayesian classifier is used as the classifier 140 using a representative probability distribution, and a non-probabilistic classification model
  • the classifier 140 to be used is a support vector machine.
  • the classifier 140 described herein is exemplary only and is not intended to limit the invention.
  • a classifier 140 may include a stimulus generator 110, a clarification induced potential measurer 120, and a reference measured through an amplitude and GPI calculator 130 in a plurality of tinnitus patients and a normal group, GPI ratio, and receives the amplitude and the GPI ratio of the subject measured by the GPI calculating unit 130, and can check whether the subject is aliased based on the reference GPI ratio stored in the database.
  • the reference GPI ratio means a raw data for calculating training data constituting a classification model through a classification algorithm.
  • the reference GPI ratio can further store reference GPI ratio values for the tinnitus patient group or the normal group calculated through the plurality of second stimulus tones generated by adjusting to at least two values having different silent intervals.
  • the classifier 140 samples a plurality of reference GPI ratios that are determined to be tinnitus patients or normal patients, and determines whether the numerical value of the inputted GPI ratio is a numerical value corresponding to the tinnitus patient or a normal group You can decide.
  • the classifier 140 receives the amplitude and the GPI ratio measured from the GPI calculator 130, and extracts training data from the classification model derived from the plurality of reference GPI ratios measured from the plurality of tinnitus patients and the normal group It is possible to check whether the subject is aliased by analyzing the input GPI ratio.
  • the classifier 140 receives the amplitudes of the first waveform or the second waveform and the measured GPI ratios from the amplitude and GPI calculator 130, and calculates a plurality of tinnitus patients or normal group From the training data by the classification model derived by using the amplitude and the reference GPI ratio, it is possible to check the subject's tinnitus according to the inputted amplitude and the GPI ratio.
  • the classifier 140 calculates the amplitude between the poles of the first waveform of the subject's auditory evoked potential according to the first stimulus sound, which does not include the amplitude and the silence interval measured through the GPI calculator 130, and the silence interval And one or more values among the amplitudes and GPI ratios of the second waveforms of the subject's clearing-induced potential according to the second stimulus sound included in the second stimulation waveform may be selected as a feature vector.
  • the GPI ratio values calculated through the plurality of second stimuli generated by adjusting the amplitude and the silence interval between the punctuation-induced intense pole points for the first stimulus to the first stimulus are adjusted to at least two values different from each other have.
  • the classifier 140 performs a learning process to generate a classification model that classifies the tinnitus patient and the normal group using the extracted feature vectors.
  • FIG. 6 is a diagram for explaining how the classifier 140 according to an embodiment diagnoses a tinnitus patient using a feature vector.
  • FIG. 6 illustrates a case where a second stimulus sound having a 20 ms silence interval is applied together with a first stimulus sound.
  • the GPI ratio value which is the ratio between the amplitudes of N1 and P2 among the amplitudes of the clearing-induced potential poles for the first stimulus sound and the amplitudes between the first stimulus sound and the poles of the clearing evoked potentials obtained from the second stimulus sound, And shows an embodiment when it is used.
  • the number of feature vectors used as the input values may increase, but since the dimension of the space increases according to the number of feature vectors to be used, an example of two-dimensional classification using only two feature vectors / RTI >
  • the x-axis is the GPI ratio value calculated from the subjects, and the y-axis represents the amplitude between N1 and P2 in the clearing-induced potential for the first stimulus.
  • the optimal crystal hyperplane refers to a straight line when the distance between the feature vectors of the closest class and the crystal hyperplane is the same for the two classes.
  • the feature vectors nearest to the decision hyperplane are called support vectors, and the margin of the decision hyperplane is defined as twice the distance between the nearest feature vectors and the decision hyperplane.
  • the principle of the support vector machine is to find w that maximizes the size of the margin, which is solved by solving the constrained optimization problem.
  • a soft margin technique may be used to increase classification accuracy while recognizing some feature vectors as errors.
  • a method of mapping the feature vector space to a higher dimension is used.
  • the mapping function used is referred to as a kernel function, Is the Radial Basis Function (RBF).
  • RBF Radial Basis Function
  • the classifier 140 further stores at least one of the amplitude between the poles of the first waveform or the poles of the second waveform measured from the plurality of tinnitus patients and the normal group, and in addition to the GPI ratio measured by the subject, An amplitude between the pole points of the first waveform measured by the subject and an amplitude between the poles of the second waveform may be further inputted and the subject may be checked for tinnitus based on the stored reference GPI ratio.
  • the classifier 140 may classify a plurality of tinnitus patients and a normal group as measured by a conventional GPIAS (Gap Pre-pulse Inhibition of Acoustic Start) method or a plurality of tinnitus patients, The spectral power density of the potential is further stored, and the spectral power density of the measured GPIAS or the celestial evoked potential from the subject is further inputted, so that it is possible to check whether the subject is aliased.
  • GPIAS Gap Pre-pulse Inhibition of Acoustic Start
  • the factors that can be used as the input values of the classifier 140 are not limited to this. If the factors that can increase the reliability in measuring the tinnitus of the patient, the measured values extracted from the tinnitus patient and the normal group, 140, and using a plurality of factors as an input value of the classifier 140, it is possible to increase the reliability of the tinnitus patient test.
  • the classifier 140 included in the above-described embodiment may be implemented by an arithmetic unit including a memory including instructions programmed to perform these functions, and a microprocessor that executes these instructions.
  • FIG. 7 is a flowchart showing a procedure of a tinnitus checking method according to an embodiment.
  • the tinnitus checking method according to FIG. 7 can be performed by the tinnitus testing apparatus 100 described with reference to FIG. 1, and each step will be described below.
  • the stimulus sound generator 110 generates a first stimulus sound including a background sound and a pulse sound, a second stimulus sound including a background sound with a silence interval, and a pulse sound (S710).
  • the clearing-off evoked potential measuring unit 120 measures a first waveform of the clearing-induced potential of the subject according to the first stimulation sound and a second waveform of the clearing-induced stimulation potential of the subject according to the second stimulation sound in step S720.
  • the amplitude and GPI calculating unit 130 measures the GPI ratio, which is a ratio between the amplitude between the pole points of the first waveform and the pole points of the second waveform (S730).
  • the classifier 140 receives the measured GPI ratio, and determines whether or not the subject's tinnitus according to the GPI ratio input from the classification model derived from the plurality of reference GPI ratios measured in the plurality of tinnitus patients and the normal group (S740).
  • FIG. 7 Each step of FIG. 7 is the same as that of the apparatus shown in FIG.
  • FIG. 8 is a flowchart illustrating a procedure of a method for selecting a GPI ratio or a GPI ratio according to a plurality of second stimulus tones according to an embodiment to check the tinnitus.
  • the tinnitus checking method according to FIG. 8 can be performed by the tinnitus testing apparatus 100 described with reference to FIG. 1, and each step will be described as follows.
  • the stimulus sound generator 110 generates a first stimulus sound including a background sound and a pulse sound, a second stimulus sound including a background sound with a silence interval and a pulse sound, To generate a plurality of second stimulus notes (S810).
  • the clearing-off evoked potential measurement unit 120 measures a first waveform of the clearing-induced potential of the subject according to the first stimulation sound and a plurality of second waveforms of the clearing-induced stimulation potential of the subject according to the plurality of second stimulation sounds (S820).
  • the amplitude and GPI calculating unit 130 measures a plurality of GPI ratios based on a plurality of amplitudes between a maximum pole point and a minimum pole point extracted from a plurality of second waveforms corresponding to a plurality of second magnetic pole sounds, The GPI ratio is measured based on the largest amplitude among the amplitudes of the first and second amplitudes (S830).
  • the second waveform having the largest amplitude among the amplitudes between the maximum pole point and the minimum pole point extracted from the plurality of second waveforms is selected when the silence interval of the second stimulation sound used when the amplitude of the second waveform is the largest This means that it is easy to compare with the GPI ratio of the normal group.
  • the classifier 140 which stores the reference GPI ratios measured in the plurality of tinnitus patients and the normal group, receives the amplitudes and the measured GPI ratios of any one of the first waveform and the second waveform, (S840) whether the subject is alive or not according to the input amplitude and the GPI ratio from the classification model derived using the reference amplitude and the reference GPI ratio, the result of which is determined as the patient's normal group.
  • the reference amplitude is a sample for calculating the training data constituting the classification model through the classification algorithm), and may be either the amplitude between the pole points of the first waveform or the amplitudes between the pole points of the second waveform, or both can do.
  • the largest amplitude among the maximum and minimum poles of the second waveform of the subject measured through the plurality of second magnetic noises generated by adjusting at least two values having different silence intervals By measuring the GPI ratio based on the second waveform with amplitude, it is possible to more accurately check whether the patient is tinnitus.
  • FIG. 9 is a flowchart showing a procedure of a method for checking the amplitude or tinnitus frequency of a subject's tinnitus according to an embodiment.
  • the tinnitus checking method according to FIG. 9 can be performed by the tinnitus testing apparatus 100 described with reference to FIG. 1, and each step will be described below.
  • the stimulus sound generator 110 generates a first stimulus sound including a background sound and a pulse sound, a second stimulus sound including a background sound with a silence interval and a pulse sound, And generates a plurality of different first stimuli and second stimuli (S910).
  • the stimulus sound generator 110 can generate the background sound with the amplitude of ⁇ 1 dB SL and ⁇ 2 dB SL in units of 1 dB SL. At this time, the pulse sound is kept constant with about 65 dB SL amplitude.
  • the degree of tinnitus felt by the subject is determined by measuring the GPI ratio by varying the amplitude of the background sound.
  • the stimulus sound generator 110 may generate a stimulus sound by presetting the amplitude of the background sound based on the degree of tinnitus sensed by the subject through the tinnitogram loudness match, which is a conventional tinnitus test method.
  • the stimulus sound generator 110 may generate a stimulus sound having a frequency of 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 9-10, 11-12, 14-16 kHz Can be generated and converted.
  • the background sound may be pure tone and may be narrow band noise with a certain width.
  • it may be various sounds such as broadband noise (white noise) and the like.
  • the pulse sound has a frequency of about 1 KHz and is kept constant.
  • the tinnitus patient can not recognize the silence interval. Therefore, when the largest value of the GPI ratio measured when the frequency of the tinnitus is different is derived, The frequency of the tinnitus felt by the subject is determined by measuring the GPI ratio by varying the frequency of the background sound.
  • the stimulus sound generator 110 may generate a stimulus sound by presetting the frequency of the background sound based on the frequency of the tinnitus sensed by the subject through the tinnitus matching test (tinnitogram), which is a conventional tinnitus test method.
  • tinnitogram tinnitus matching test
  • the clearing-off evoked potential measuring unit 120 measures a plurality of first waveforms of the clearing-induced potential of the subject according to the plurality of first stimuli, and a plurality of second waveforms of the clearing-induced potential of the subject according to the plurality of second stimuli
  • the waveform is measured in step S920.
  • the amplitude and GPI calculator 130 measures a plurality of GPI ratios, which is a ratio between the amplitudes between the pole points of the plurality of first waveforms and the amplitudes of the plurality of second waveforms, in step S930.
  • the amplitude and GPI calculating unit 130 compares the measured plurality of GPI ratios, and selects the amplitude or frequency of the used background sound when the GPI ratio is the largest, and determines the amplitude or frequency of the background sound It can be judged based on the degree or frequency of the tinnitus and input to the classifier 140. However, if the classifier 140 is used, the amplitude and the plurality of GPI ratios measured by the GPI calculator 130 may all be input to the classifier 140, and the classifier 140 may make such a determination.
  • the classifier 140 receives a plurality of GPI ratios measured by the subject and classifies the plurality of tinnitus patients and the normal group using the amplitude and the GPI ratio measured by varying the frequency of the background sound or the background sound frequency
  • the tinnitus degree and tinnitus frequency of the subject are inspected from the model (S940).
  • a plurality of tinnitus patients and / or a plurality of tinnitus patients using a plurality of GPI ratios measured through a plurality of first stimulus notes and second stimulus notes having different frequencies of the background sound or the background sound The classifier 140 learned through the classification model using the amplitudes of the background sounds or the frequencies of the background sounds and the GPI ratios measured using the frequencies of the background sounds can determine the presence or absence of tinnitus according to a plurality of inputted GPI ratios , The degree of the tinnitus or the frequency of the tinnitus sensed by the examinee can be ranked and output for the plurality of inputted GPI ratios.
  • the auditory evoked potential of the subject is measured through the stimulus sound which can confirm the degree of tinnitus, and the information analyzing the measured auditory evoked potential is input to the classifier 140, Can be objectively inspected.
  • embodiments of the present invention can be implemented by various means.
  • embodiments of the present invention may be implemented by hardware, firmware, software, or a combination thereof.
  • the method according to embodiments of the present invention may be implemented in one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs) , FPGAs (Field Programmable Gate Arrays), processors, controllers, microcontrollers, microprocessors, and the like.
  • ASICs Application Specific Integrated Circuits
  • DSPs Digital Signal Processors
  • DSPDs Digital Signal Processing Devices
  • PLDs Programmable Logic Devices
  • FPGAs Field Programmable Gate Arrays
  • processors controllers, microcontrollers, microprocessors, and the like.
  • the method according to embodiments of the present invention may be implemented in the form of a module, a procedure or a function for performing the functions or operations described above.
  • the software code can be stored in a memory unit and driven by the processor.
  • the memory unit is located inside or outside the processor, and can exchange data with the processor by various known means.

Abstract

Un appareil de diagnostic d'acouphène, selon un mode de réalisation de la présente invention, comprend : une unité de génération de tonalité de stimulation pour générer une première tonalité de stimulation, qui comprend un premier bruit de fond et un bruit de pulsation, et une pluralité de deuxièmes tonalités de stimulation, chacune des deuxièmes tonalités de stimulation ayant une durée différente d'un intervalle silencieux et comprenant un bruit de pulsation généré tandis qu'un troisième bruit de fond continue après qu'un deuxième bruit de fond a été terminé, puis un intervalle silencieux est maintenu pendant une durée prédéterminée ; une unité de mesure de potentiel évoqué auditif pour mesurer une première forme d'onde du potentiel évoqué auditif d'une personne examinée en fonction de la première tonalité de stimulation, et une pluralité de secondes formes d'onde du potentiel évoqué auditif de la personne examinée en fonction de la pluralité de deuxièmes tonalités de stimulation ; une unité de calcul d'inhibition d'amplitude et d'intervalle de pré-pulsation (GPI) pour mesurer une pluralité de rapports GPI, qui sont des rapports entre une amplitude crête à crête de la première forme d'onde et une amplitude crête à crête des secondes formes d'onde respectives ; et un classificateur, dans lesquels des rapports GPI de référence mesurés à partir d'une pluralité de patients souffrant d'acouphènes et d'un groupe contrôle sont pré-stockés, pour diagnostiquer si la personne examinée est touchée ou non d'un acouphène sur la base des rapports GPI de référence par réception d'une entrée d'une pluralité de rapports GPI mesurés à partir de la personne examinée.
PCT/KR2018/001048 2018-01-24 2018-01-24 Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur WO2019146809A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/KR2018/001048 WO2019146809A1 (fr) 2018-01-24 2018-01-24 Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/KR2018/001048 WO2019146809A1 (fr) 2018-01-24 2018-01-24 Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur

Publications (1)

Publication Number Publication Date
WO2019146809A1 true WO2019146809A1 (fr) 2019-08-01

Family

ID=67396146

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2018/001048 WO2019146809A1 (fr) 2018-01-24 2018-01-24 Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur

Country Status (1)

Country Link
WO (1) WO2019146809A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115024719A (zh) * 2022-08-11 2022-09-09 安徽声讯信息技术有限公司 一种基于听力仿生的耳鸣检测方法及装置
WO2024023676A1 (fr) * 2022-07-28 2024-02-01 Cochlear Limited Techniques d'administration d'un stimulus pour le traitement des acouphènes

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2011251058A (ja) * 2010-06-03 2011-12-15 Panasonic Corp 聴性定常反応測定方法および測定装置
KR20120100597A (ko) * 2011-03-04 2012-09-12 단국대학교 산학협력단 뇌파를 이용한 이명 유무 검사 장치 및 이를 이용한 이명 유무 검사 방법
KR20120131778A (ko) * 2011-05-26 2012-12-05 삼성전자주식회사 청력 검사를 수행하는 방법과 이것을 이용한 청력 보조 장치
US20130039517A1 (en) * 2010-04-16 2013-02-14 Widex A/S Hearing aid and a method for alleviating tinnitus
JP2017018420A (ja) * 2015-07-13 2017-01-26 株式会社電通サイエンスジャム 耳鳴患者判別システム、耳鳴患者判別方法および耳鳴患者判別プログラム

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130039517A1 (en) * 2010-04-16 2013-02-14 Widex A/S Hearing aid and a method for alleviating tinnitus
JP2011251058A (ja) * 2010-06-03 2011-12-15 Panasonic Corp 聴性定常反応測定方法および測定装置
KR20120100597A (ko) * 2011-03-04 2012-09-12 단국대학교 산학협력단 뇌파를 이용한 이명 유무 검사 장치 및 이를 이용한 이명 유무 검사 방법
KR20120131778A (ko) * 2011-05-26 2012-12-05 삼성전자주식회사 청력 검사를 수행하는 방법과 이것을 이용한 청력 보조 장치
JP2017018420A (ja) * 2015-07-13 2017-01-26 株式会社電通サイエンスジャム 耳鳴患者判別システム、耳鳴患者判別方法および耳鳴患者判別プログラム

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024023676A1 (fr) * 2022-07-28 2024-02-01 Cochlear Limited Techniques d'administration d'un stimulus pour le traitement des acouphènes
CN115024719A (zh) * 2022-08-11 2022-09-09 安徽声讯信息技术有限公司 一种基于听力仿生的耳鸣检测方法及装置

Similar Documents

Publication Publication Date Title
US8277390B2 (en) Method for automatic non-cooperative frequency specific assessment of hearing impairment and fitting of hearing aids
Boothroyd et al. Vibrotactile thresholds in pure tone audiometry
WO2012121450A1 (fr) Dispositif de test des acouphènes utilisant les ondes cérébrales et méthode de test des acouphènes l'utilisant
Whitehead et al. Measurement of otoacoustic emissions for hearing assessment
US9345419B2 (en) Method and apparatus for the objective detection of auditive disorders
CN103313653A (zh) 具有电极验证的个人eeg监测装置
US4201225A (en) Method and apparatus for measuring stimulated acoustic reflex latency time
DK2572640T3 (en) Method and device for performing a survey rentoneaudiometri
WO2013103247A1 (fr) Procédé et appareil de production d'un signal audio et procédé d'ajustement du volume du signal audio
KR20190114303A (ko) 인지 기능 검사 방법 및 시스템
WO2019146809A1 (fr) Appareil et procédé de diagnostic d'acouphène utilisant un cassificateur
Mertes et al. Concurrent measures of contralateral suppression of transient-evoked otoacoustic emissions and of auditory steady-state responses
Dommes et al. Auditory cortex stimulation by low-frequency tones—An fMRI study
Tikka et al. Video Head Impulse Test (vHIT) Findings in Patients With Superior Semicircular Canal Dehiscence: A Case–Control Study
Jarollahi et al. Sound-field speech evoked auditory brainstem response in cochlear-implant recipients
Gama et al. Recording the human brainstem frequency-following-response in the free-field
WO2019151810A1 (fr) Dispositif et procédé pour tester les acouphènes en tenant compte de l'âge
CN113288127B (zh) 一种基于时频调制察觉阈的听力检测装置及检测方法
Antunes et al. Multichannel search strategy for improving the detection of auditory steady-state response
KR101824116B1 (ko) 분류기를 이용한 이명 검사 장치 및 방법
RU2722875C1 (ru) Способ определения оптимальных параметров настройки слухового аппарата
Zhao et al. Pure tone audiometry and speech audiometry
Kochanek et al. A universal computer audiometer for objective hearing testing and screening
Kumar et al. Utility of acoustic change complex as an objective tool to evaluate difference limen for intensity in cochlear hearing loss and auditory neuropathy spectrum disorder
Garnham et al. Assessment of aided ABR thresholds before cochlear implantation

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18902362

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18902362

Country of ref document: EP

Kind code of ref document: A1