Methods and systems for determining the integrity of auditory nerve fibers and synapses

Abstract

The present invention belongs to the field of hearing testing. In particular, the present invention relates to systems and methods for determining the integrity of a subject's auditory nerve fibers (ANFs) and / or afferent auditory nerve synapses (ANS) and / or inner hair cells (IHCs).

Description

Technical Field

[0001] The present invention belongs to the field of hearing testing. In particular, the present invention relates to systems and methods for determining the integrity of a subject's auditory nerve fibers (ANFs) and / or afferent auditory nerve synapses (ANS) and / or inner hair cells (IHCs). Background Art

[0002] Hearing loss can be attributed to several measurable factors, including damage or loss of outer hair cells (OHC) and inner hair cells (IHC), and auditory nerve (AN) dysfunction or loss. AN dysfunction can occur when auditory nerve synapses (ANS) are damaged without damaging the fibers themselves; AN loss can occur due to damage to auditory nerve fibers (ANFs). A decrease in the number of ANFs and ANSs compared to the total number present in a healthy ear (30,000 to 45,000, depending on the species) is called cochlear synaptopathy.

[0003] Cochlear synaptopathy usually develops as a result of aging, exposure to loud noise, or ototoxic drugs. Figure 1 As shown, the cochlea is deeply embedded in the temporal bone, making it impossible to directly assess the functionality of ANFs and their synapses in living humans. Currently, damage to ANFs and ANS can only be quantified using post-mortem temporal bone histology (human or animal), where ANFs can be immunostained, labeled, and counted using imaging techniques.

[0004] Inner hair cell (IHC) deafferentation, i.e., a reduction in the number of afferent nerves contacting the IHC, can be associated with degeneration of the ANS and ANF, usually resulting in 100% synaptopathy. However, under certain conditions, the ANF and ANS may affect the IHC separately, for example, as may be caused by certain medical treatments, such as chemotherapy. This technique cannot reliably distinguish IHC damage from OHC damage.

[0005] Current clinical practice in assessing hearing loss is primarily through threshold testing, such as the pure tone audiogram, a sound perception test that aims to assess hearing sensitivity, i.e., the minimum pressure level required to detect a pure tone or sound. The test measures the level of sound that an individual needs to detect (in dB) compared to a reference standard normal hearing group to detect the same sound (i.e., the test is repeated for different sound frequencies). The measurement provides the hearing sensitivity loss in dB for frequencies between 0.125 and a maximum of 16 kHz. Since the audiogram measures hearing sensitivity, it is used as a measure of the hearing sensitivity of the inner ear (i.e., the cochlear cells that amplify sound in the inner ear, see Figure 1 ) are useful markers of outer hair cell (OHC) damage. However, hearing sensitivity tests are not suitable for predicting cochlear synaptopathy (i.e., Figure 1The light-colored boxes in the figure show that the ANS and ANF synapse onto inner hair cells in the cochlea).

[0006] In addition, there are non-invasive physiological tests to assess hearing status: these tests include, for example, otoacoustic emissions (OAE), or brainstem EEG, ABR / BERA tests, or auditory steady-state response (ASSR) tests. However, the listed tests mainly reflect outer hair cell (OHC) integrity (OAE), or are affected by the state of OHC (EEG, ABR / BERA, and ASSR). The latter test therefore fails to distinguish the difference between the damage of OHC and the degeneration of ANF / ANS, and in addition, they fail to provide an objective way to quantify the degeneration of ANF and / or ANS and the complete deafferentation or IHC loss in extreme cases. Currently, the listed clinical tests only determine whether sounds are detected, but the response level of the auditory system cannot be measured in an accurate quantitative manner. In addition, some of these tests are cumbersome to perform and may require a specific system to collect and process the measured data.

[0007] Prior art describes the use of sine amplitude modulation (SAM) auditory stimulation to mimic the rapidly changing timing characteristics in speech in order to study human auditory responses. For example, EP 1 658 810 describes the generation of various auditory quasi-sine stimulations to induce auditory steady-state responses. Alternatively, EP 3 281 585 describes the use of composite quasi-speech stimulations generated by applying automatic algorithms to induce auditory steady-state responses. However, this technology fails to provide a diagnostic test that reliably quantifies the degree of hearing loss associated with ANF and / or ANS in a subject.

[0008] Providing a non-invasive diagnostic test for determining the extent of a person's cochlear synaptopathy - and not just whether or not they have the disease - could provide new objective markers of hearing loss that are not currently considered clinically. The test would make hearing diagnosis more precise and could therefore improve treatment. Furthermore, it could allow hearing aids to be calibrated using user-specific parameters that better reflect an individual's hearing loss and could also provide improvements in augmented hearing tools.

[0009] Accordingly, there is a need for systems and methods that can assist in determining cochlear synaptopathy, preferably wherein the integrity of the ANF and ANS is independently determined or quantified (i.e., unaffected by the degree of OHC damage / loss). In particular, there is a need for sensitive, non-invasive diagnostic tests and systems that detect synaptopathy, particularly in living humans (not autopsy cochlear material). Similarly, there is a need for sensitive, non-invasive diagnostic tests and systems that detect IHC loss. In addition, there is a need for tests and systems that quantify ANF and ANS degeneration in a wide range of populations. In particular, people over the age of 50 often complain of self-reported hearing difficulties, which may be a consequence of cochlear synaptopathy. But listeners in that age group also often show OHC damage. Diagnostic ANS and ANF integrity tests should therefore also be sensitive and applicable to the elderly population with mixed hearing loss pathology. Summary of the invention

[0010] The present invention overcomes one or more of these problems. Preferred embodiments of the present invention overcome one or more of these problems, for example by:

[0011] (I) Enhanced EEG response strength beyond that of existing EEG-based auditory tests (e.g., in the ASSR), thus being able to explain significant individual differences in response strength;

[0012] (II) showed that the EEG response intensity is more sensitive to the integrity of the ANF and ANS than to OHC-related aspects of sensorineural hearing loss, thus providing a differential test for cochlear synaptopathy in listeners with or without mixed OHC-AN injury patterns;

[0013] (III) employing an EEG analysis method that extracts all signal-relevant EEG response components to enable interpretation in terms of the extent of AN damage; thereby providing a more accurate and sensitive method for determining ANF and ANS integrity than other existing ASSR analysis methods that can only extract single EEG response features;

[0014] (IV) using EEG analysis methods that reduce confounding effects associated with individual electrophysiological noise floors to generate relative markers of ANF and ANS integrity;

[0015] (V) Employs an EEG analysis method that is robust to recording artifacts of EEG responses at non-relevant frequencies such as eye movement artifacts. This method is more robust than other EEG-based audiometric tests that analyze in the time domain (e.g., BERA / ABR amplitudes);

[0016] (VI) using an EEG-based measurement procedure that has a relatively fast (approximately 10 min) and easy-to-implement protocol that can be integrated with existing EEG recording equipment;

[0017] (VII) using an improved sensitivity stimulation paradigm combined with the proposed EEG analysis method to produce noise-resistant measurements that enhance the diagnostic capability in quantifying ANF and ANS integrity in a real clinical setting (without requiring special equipment, such as electronic shielding and soundproof rooms, high-density EEG); and / or,

[0018] (VIII) Optimizing EEG methods and procedures requires fewer signal repetitions to achieve significant response levels above the noise floor (compared to the number of repetitions required by other methods at the same signal-to-noise quality), which reduces the overall required testing time of our new method.

[0019] The present invention relates to a computer-implemented method for determining the integrity of auditory nerve fibers and / or synapses of a subject, comprising the following steps:

[0020] a) receiving brain activity data, preferably EEG or MEG signals, more preferably EEG signals, of a subject exposed to an auditory stimulus s(t), wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); and,

[0021] b) The degree of integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses as determined from brain activity data.

[0022] The present invention also relates to a computer-implemented method for determining the integrity of auditory nerve fibers and / or synapses of a subject, comprising the steps of:

[0023] a) generating an auditory stimulus s(t), wherein the auditory stimulus comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); exposing the subject to the auditory stimulus s(t); and receiving brain activity data of the subject exposed to the auditory stimulus s(t), preferably an EEG or MEG signal, more preferably an EEG signal; and,

[0024] b) The degree of integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses as determined from brain activity data.

[0025] The present invention also relates to a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS) and / or an auditory nerve fiber (ANF) of a subject, the computer-implemented method comprising the following steps:

[0026] a) receiving neural activity data, preferably EEG or MEG signals, more preferably EEG signals, of a subject exposed to an auditory stimulus s(t); and,

[0027] b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANF as determined by neural activity data;

[0028] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%.

[0029] The present invention also relates to a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS), an auditory nerve fiber (ANF) and / or an inner hair cell (IHC) of a subject, the computer-implemented method comprising the steps of:

[0030] a) receiving neural activity data of a subject exposed to an auditory stimulus s(t), preferably an EEG or MEG signal, more preferably an EEG signal; and,

[0031] b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANFs and / or IHCs as determined by the neural activity data;

[0032] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%.

[0033] In some preferred embodiments, the subject is a human subject. In some preferred embodiments, the non-sinusoidal amplitude modulator m(t) has a rectangular waveform with a duty cycle d, wherein d=W / T, wherein the duty cycle is at least 10% and at most 40%, preferably at least 15% and at most 35%, preferably at least 20% and at most 30%, for example about 25%.

[0034] In some preferred embodiments, the non-sinusoidal amplitude regulator m(t) has a rectangular waveform with a slope deviation window δ between 0 and δ max In the range of max =0.5*d, where d is the duty cycle.

[0035] In some preferred embodiments, the subject's brain activity data is received in the time domain as recorded electrical potentials between at least two EEG electrodes.

[0036] In some preferred embodiments, step b) comprises the following steps:

[0037] b2) transforming said brain activity data, preferably received EEG signals, into the frequency domain (FD), preferably by applying a (discrete) Fourier transform (dFT) to the time segments of said EEG signals evoked by the auditory stimulus s(t).

[0038] In some preferred embodiments, step b) comprises the following steps:

[0039] b4') determines the frequency corresponding to the stimulation modulation frequency f m The response amplitude at the frequency F0 and at one or more frequencies F1 to Fn which are harmonics of the modulation frequency.

[0040] In some preferred embodiments, step b) comprises the following steps:

[0041] - Identify at least 2 peaks in the frequency domain, preferably all significant peaks above the noise floor, including at least:

[0042] - the modulation frequency F0 of said stimulation; and,

[0043] - one or more harmonics (F1 to Fn) of said modulation frequency;

[0044] - reconstructing a time domain waveform from said identified peaks, preferably by inverse Fourier transform (iFT); and,

[0045] - measuring the amplitude of the reconstructed signal from the reconstructed time domain waveform;

[0046] Therein, the amplitude is a measure of the level of integrity of the subject's AN fibers and / or synapses.

[0047] The invention also relates to a method for generating an individualized mathematical auditory peripheral model of a subject for processing sounds for a hearing device, the method comprising the following steps:

[0048] - determining the integrity of AN synapses and / or fibers of the subject according to the method and embodiments thereof as described herein;

[0049] -optionally, determining the fraction of outer hair cells that are damaged; and,

[0050] - Generating an individualized mathematical model for processing sound for a hearing device based on the integrity of the subject's AN synapses and / or fibers, and optionally the outer hair cell damage fraction.

[0051] The present invention also relates to a method for modifying a subject's desired auditory response to a sound, the method comprising the steps of:

[0052] - performing a method as described herein, or an embodiment thereof, to determine a specific AN fiber and / or synapse (and optionally OHC) hearing impairment profile in said subject;

[0053] -Include hearing loss profiles in the auditory peripheral model to simulate auditory responses to any acoustic stimulus;

[0054] - comparing the independently simulated auditory responses to the expected auditory responses; and,

[0055] - comprising a signal processing algorithm to adjust the sound stimulus so that the simulated auditory response matches the desired auditory response.

[0056] The present invention also relates to the use of the method as described herein as a screening method for the presence of a synaptopathology in a subject.

[0057] The present invention also relates to the use of a method as described herein as a screening method for quantifying the extent of synaptopathy in a subject, preferably wherein the extent of synaptopathy is defined as % of AN synapse and / or fiber loss.

[0058] The present invention also relates to a method for comparing the extent of hearing loss caused by synaptopathy and by outer hair cell (OHC) loss in a subject, comprising the steps of:

[0059] - performing a method as described herein, or an embodiment thereof, to quantify the extent of auditory nerve fiber and synaptic damage, preferably the extent of synaptopathy, in a subject;

[0060] - obtaining markers, such as EEG markers, for subjects that are only sensitive to outer hair cell damage; to quantify the extent of outer hair cell damage in the subject; and,

[0061] - Plot the extent of auditory nerve fiber and synapse damage and the extent of outer hair cell damage on a 2-dimensional graph.

[0062] The present invention also relates to a system for determining the integrity of an auditory nerve synapse (ANS) and / or an auditory nerve fiber (ANF) of a subject, the system comprising:

[0063] - sound generation and stimulation means for generating auditory stimulation s(t);

[0064] - neural activity screening hardware configured to record subject neural responses as neural activity data when said subject is exposed to auditory stimuli s(t) generated by said sound generating and stimulating device; and,

[0065] - A processing unit connected to the recording hardware, configured to:

[0066] a) receiving neural activity data of said subject exposed to said auditory stimulus s(t), and,

[0067] b) determining from the neural activity data the degree of integrity of all or a subset of the subject-specific ANS and / or ANF;

[0068] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%.

[0069] The present invention also relates to a system for determining the integrity of an auditory nerve synapse (ANS), an auditory nerve fiber (ANF), and / or an inner hair cell (IHC) of a subject, the system comprising:

[0070] - sound generation and stimulation means for generating auditory stimulation s(t);

[0071] - neural activity screening hardware configured to record subject neural responses as neural activity data when said subject is exposed to auditory stimuli s(t) generated by said sound generating and stimulating device; and,

[0072] - A processing unit connected to the recording hardware, configured to:

[0073] a) receiving neural activity data of said subject exposed to said auditory stimulus s(t), and,

[0074] b) determining from the neural activity data the degree of integrity of all or a subset of the subject-specific ANS and / or ANFs and / or IHCs;

[0075] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%.

[0076] In some preferred embodiments, the duty cycle is at least 15% and at most 35%, preferably at least 20% and at most 30%, such as about 25%.

[0077] In some preferred embodiments, the rectangular waveform has a slope deviation window δ between 0 and δ max In the range of max =0.5*d, where d is the duty cycle.

[0078] In some preferred embodiments, the screening hardware comprises at least two EEG electrodes and an amplifier; and is configured to record the subject's neural response as an EEG signal in the time domain (TD).

[0079] In some preferred embodiments, the subject is a human subject, and the neural activity screening hardware is configured to record human neural responses.

[0080] In some preferred embodiments, the processing unit is configured to transform the neural activity data, preferably the EEG signal, into the frequency domain (FD), preferably by applying a Fourier transform to the time segment of the EEG signal induced by the auditory stimulus s(t).

[0081] In some preferred embodiments, the processing unit is configured to determine the frequency corresponding to the stimulation modulation frequency f m The response amplitude at the frequency F0 and at one or more frequencies F1 to Fn which are harmonics of the modulation frequency.

[0082] In some preferred embodiments, the processing unit is configured as follows:

[0083] - Identify at least 2 peaks in the frequency domain, preferably all significant peaks above the noise floor, including at least:

[0084] - the modulation frequency F0 of said stimulation; and,

[0085] - one or more harmonics F1 to Fn of said modulation frequency;

[0086] - reconstructing a time domain waveform from said identified peaks, preferably by inverse Fourier transformation; and,

[0087] - measuring the amplitude of the reconstructed signal from the reconstructed time domain waveform; wherein the amplitude is a measure of the level of integrity of the subject's ANS and / or ANF and / or IHC.

[0088] The present invention also relates to a system for determining the integrity of auditory nerve synapses and / or fibers of a subject, the system comprising:

[0089] -EEG-based hearing screening hardware, including electrodes and amplifiers;

[0090] - sound generation and stimulation means; and,

[0091] - A processing unit configured to perform the method according to any one of the methods described herein and their embodiments.

[0092] The present invention also relates to a system for generating an individualized mathematical auditory peripheral model of a subject for processing sounds for a hearing device, wherein the processing unit is additionally configured to:

[0093] -optionally, determining the fraction of outer hair cells that are damaged; and,

[0094] - generating an individualized mathematical model for processing sound for a hearing device based on the integrity of the subject's ANS and / or ANF and / or IHC and optionally the damaged fraction of outer hair cells.

[0095] The present invention also relates to a system for modifying a subject's desired auditory response to a sound, wherein the processing unit is further configured to:

[0096] - Determine subject-specific ANS and / or ANF hearing impairment profile;

[0097] - including said auditory impairment profile in a personalized mathematical auditory peripheral model to simulate the auditory response to any acoustic stimulus;

[0098] - comparing the independently simulated auditory responses to the expected auditory responses; and,

[0099] - comprising a signal processing algorithm to adjust the sound stimulus so that the simulated auditory response matches the desired auditory response.

[0100] The present invention also relates to a system for comparing the extent of hearing loss caused by ANS and / or ANF damage (synaptopathy) and / or IHC damage and / or OHC (OHC) damage in a subject, wherein the processing unit is additionally configured to:

[0101] - quantifying the extent of ANS and / or ANF damage, preferably synaptopathy and / or IHC loss in a subject, for example as % of ANS and / or ANF and / or IHC loss;

[0102] - obtaining markers, such as EEG markers, for subjects that are only sensitive to outer hair cell damage; to quantify the extent of outer hair cell damage in the subject; and,

[0103] - The extent of ANS and / or ANF and / or IHC damage and the extent of OHC damage are plotted, for example on a 2-dimensional graph.

[0104] The present invention also relates to a computer program, or a computer program product which can be directly loaded into the internal memory of a computer, or a computer program product which is stored on a computer-readable medium, or a combination of said computer program or computer program products, configured to perform the method described herein, or an embodiment thereof.

[0105] The present invention also relates to a computer program, or a computer program product directly loadable into the internal memory of a computer, or a computer program product stored on a computer readable medium, or a combination of said computer program or computer program products, configured to perform a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS) and / or an auditory nerve fiber (ANF) of a subject,

[0106] The computer-implemented method comprises the following steps:

[0107] a) receiving neural activity data of a subject exposed to an auditory stimulus s(t), preferably an EEG or MEG signal, more preferably an EEG signal; and,

[0108] b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANF as determined by neural activity data;

[0109] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%.

[0110] The present invention also relates to a computer program, or a computer program product directly loadable into the internal memory of a computer, or a computer program product stored on a computer readable medium, or a combination of said computer program or computer program products, configured to perform a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS), an auditory nerve fiber (ANF) and / or an inner hair cell (IHC) of a subject,

[0111] The computer-implemented method comprises the following steps:

[0112] a) receiving neural activity data of a subject exposed to an auditory stimulus s(t), preferably an EEG or MEG signal, more preferably an EEG signal; and,

[0113] b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANFs and / or IHCs as determined by the neural activity data;

[0114] Preferably, wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t) having a rectangular waveform with a duty cycle d, wherein the duty cycle is at least 10% and at most 40%. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] The following description of the drawings of the present invention is given by way of example only and is not intended to limit the present description, its application or uses.In the drawings, the same reference numerals refer to the same or similar parts or features.

[0116] Figure 1 A cross-sectional view of the human auditory periphery is shown, illustrating the location of the auditory nerve fibers and their synapses within the human cochlea (the box at the base of the IHC emphasizes the location of the AN structure at the base of the IHC). Degeneration of AN fibers and synapses has been associated with cochlear synaptopathy. Figure from Kujawa and Liberman in J. Neurosci. (2009); 29(45): 14077–14085.

[0117] Figure 2 shows an auditory stimulus s(t) created by modulating a pure tone carrier signal c(t) with a sinusoidal modulator m(t). Figure 2A shows the auditory stimulus s(t) in the time domain, and Figure 2B Auditory stimuli in the frequency domain are presented.

[0118] FIG3 illustrates an exemplary method for generating auditory stimulation according to a preferred embodiment of the present invention. In particular, Figure 3A is an exemplary periodic modulator m(t) with duty cycle d=W / T and δ=0. T represents the period of the modulator and 1 / T represents the modulation frequency f m . Figure 3B Shown in δ max Examples of possible variations of envelope shapes within the range. Figure 3C is through modulation with Figure 3A The rectangular modulator m(t) is used to generate the auditory stimulus s(t) using the pure tone carrier signal c(t). Figure 3D is through modulation with Figure 3A The rectangular modulator m(t) is used to generate the auditory stimulus s(t) by applying a broadband noise carrier signal c(t).

[0119] Figure 4 An exemplary electroencephalogram (EEG) apparatus suitable for performing aspects of the present invention is shown. A subject is exposed to an auditory stimulus according to a preferred embodiment of the present invention. The recorded EEG subject data (in the time domain) of the response envelope following response (EFR) can be transformed using a (discrete) Fourier transform (dFT) to identify the frequency of the modulator (in the frequency domain) and optionally the frequencies of its harmonics.

[0120] Figure 5 shows the results of a study based on young normal hearing (yNH) subjects. Figure 3B Schematic diagram of the calculation flow of the average EEG data for the rectangular stimulus in . Figure 5A The amplitudes of the extracted modulation frequency and its first four harmonics in the frequency domain are shown (PtN, vertical arrows). Figure 5B A reconstructed time domain waveform based on five frequency components (h0-h4) and their respective phases is shown.

[0121] FIG. 6 presents a rectangular (non-sinusoidal) amplitude modulation waveform according to a preferred embodiment of the present invention ( Figure 3B , Figure 6B ) compared to the use of a sinusoidal amplitude modulated waveform ( Figure 2A , Fig. 6A ) The results of auditory response tests performed with auditory stimulation. These results include:

[0122] Fig. 6A An exemplary auditory stimulus (two cycles) not according to the present invention is shown, the stimulus comprising a carrier signal amplitude modulated by a periodic modulator having a sinusoidal waveform.

[0123] Figure 6B An exemplary auditory stimulus (two cycles) according to a preferred embodiment of the present invention is shown, the stimulus comprising a carrier signal amplitude modulated by a periodic modulator having a non-sinusoidal (rectangular) waveform.

[0124] Figure 6C and Fig.6D Depicts simulated total ANF firing frequency from an ANF synapsed onto a single inner hair cell with a 4-kHz characteristic frequency, i.e., the cochlear location corresponding to the carrier frequency of the stimulus. Shown are responses to stimulation not according to the present invention ( Figure 6C ) and for stimulation according to the present invention ( Fig.6D ) of the three auditory profiles: no hearing deficit (NH), hearing loss with simulated synaptopathy (HI SYN ) and hearing impaired with 35dB hearing loss at all frequencies (HI FLAT ).

[0125] Fig. 6E It is demonstrated that the auditory EEG responses (i.e., representative of responses measured in humans using EEG electrodes) of the simulated source generator to stimulation according to the invention (squares) are greater than the responses to stimulation not according to the invention (circles), and show greater sensitivity to the same degree of simulated synaptopathy.

[0126] Fig. 6F The results are shown for the case where a sinusoidal amplitude modulated waveform ( Figure 2A , Fig. 6A Box plots and individual data points of response amplitudes measured from human EEG recordings to auditory stimuli of ) (i.e. not according to the invention) for three auditory profiles: young adults with normal hearing (yNH), older adults with normal hearing thresholds (oNH) and older adults with hearing impairment (oHI).

[0127] Figure 6G The waveforms resulting from the use of (non-sinusoidal) rectangular amplitude modulation according to the invention are shown ( Figure 3B , Figure 6BFigure 2 Box plots and individual data points of the measured response amplitudes from human EEG recordings to auditory stimuli with α ( ) for three auditory profiles: young adults with normal hearing (yNH), older adults with normal audiometric thresholds (oNH), and older adults with hearing impairment (oHI).

[0128] FIG7 presents regression plots of the auditory response test for two auditory profiles: no and OHC lesions (NH DP )-related hearing loss and OHC injury (HI DP ) hearing-impaired subjects. According to a preferred embodiment of the present invention, with a rectangular (non-sinusoidal) amplitude modulated waveform ( Figure 3B , Figure 6B ) compared to the sinusoidal amplitude modulated waveform ( Figure 2A , Fig. 6A ) The test was performed with auditory stimulation. The y-axis shows the estimated DP threshold (higher values ​​reflect greater OHC damage), and the x-axis shows the EEG response amplitude (EFR PTN ). These results include:

[0129] Fig. 7A , which shows the responses of a subgroup of NH participants to stimulation using a sinusoidal amplitude modulated waveform.

[0130] Figure 7B , which shows the responses of the subgroup of HI participants to stimulation using a sinusoidal amplitude modulated waveform.

[0131] Figure 7C , which shows the responses of a subgroup of NH participants to stimulation with a rectangular (non-sinusoidal) amplitude modulated waveform.

[0132] Fig.7D , which shows the responses of the subgroup of HI participants to stimulation with a rectangular (non-sinusoidal) amplitude modulated waveform.

[0133] FIG8 presents a regression plot between EEG response amplitude (EFR amplitude) and speech intelligibility in noise score (SRT) for three subgroups of listeners, namely normal hearing young adults (yNH), elderly adults with normal audiometric thresholds (oNH), and elderly adults with hearing impairment (oHI). Figure 3B , Figure 6B , Figure 8B ) compared to the sinusoidal amplitude modulated waveform ( Figure 2A , Fig. 6A , Fig. 8A ) Record EEG responses to auditory stimulation. Figure 8C and Fig.8DThe analysis method according to the preferred embodiment of the present invention is used to compare the relationship between speech intelligibility and the extracted EEG amplitude, that is, to extract the amplitude corresponding to the stimulus envelope and its harmonics (f0-f N, Fig.8D ), corresponding to the EEG analysis method that only considers f0 ( Figure 8C ) frequency. Note that Figure 8C and Fig.8D The speech material was not passed through a high-pass filter.

[0134] Fig. 9 A flow chart is presented of preferred steps for determining AN fiber and synapse profiles, and optionally using reference data to determine a subject-specific hearing profile.

[0135] Fig.10 A flow chart presents preferred steps for determining ANS / ANF and OHC profiles, and optionally using reference data to determine a subject-specific hearing profile.

[0136] Fig.11 A flow chart presenting preferred steps for determining a desired auditory response.The determined auditory response may be used to configure a hearing device, such as a cochlear implant or a hearing aid.

[0137] FIG12 presents the results for normal hearing (NH) and hearing impaired (HI) with outer hair cell damage. OHC ), model simulation of neural responses to rectangular amplitude modulation (RAM) waveforms as a function of duty cycle (%) ranging from 0.2% duty cycle (RAM0.2) to 25% duty cycle (RAM25). RAM25 stimulation produces improved response amplitude compared to lower RAM stimulation %.

[0138] Fig. 12A Model simulations for a modulation frequency of 10 Hz are presented.

[0139] Fig. 12B Model simulations for a modulation frequency of 120 Hz are presented.

[0140] Fig.13 Presentation of normal hearing (NH), hearing loss with outer hair cell damage (HI OHC ) and hearing loss with ANS / ANF damage (HI CS ), model simulation of the relationship between neural responses to rectangular amplitude modulation (RAM) waveforms and modulation frequency (Hz). A 120-Hz modulation rate produced the highest response amplitude, and RAM stimulation was more sensitive than SAM or TT stimulation for detecting CS-related deficits.

[0141] Fig.14 Presents a model simulation of the neural responses to three auditory stimuli:

[0142] The waveform shown in the upper thumbnail is: Sine amplitude modulated waveform (SAM- Fig.14 a), rectangular amplitude modulation waveform with 25% duty cycle (RAM25- Fig.14 b) and a rectangular AM waveform with a 50% duty cycle (RAM50- Fig.14 c) in.

[0143] The small figure shows SAM ( Fig.14 d in), RAM25( Fig.14 e) and RAM50( Fig.14 f) The sum of the corresponding single-unit firing rates of all ANFs connected to a single IHC (spikes / s).

[0144] The lower panel shows SAM ( Fig.14 h in), RAM25( Fig.14 g) and RAM50( Fig.14 i) The corresponding simulated evoked potential generator and recorded human evoked potential (μV).

[0145] Fig.15 Neural responses to auditory evoked potentials (AEPs) for three subgroups of listeners, young adults with normal hearing (yNH), elderly adults with normal audiometric thresholds (oNH), and elderly adults with hearing impairment (oHI), are presented, along with model simulations (triangles) for different sensorineural hearing loss profiles and stimulus types, namely a sinusoidal amplitude modulated waveform (SAM), a rectangular amplitude modulated waveform with a 50% duty cycle (RAM50), and a rectangular amplitude modulated waveform with a 25% duty cycle (RAM25). Comparison of the experimental data and model simulations shows that RAM25 (25% duty cycle) produces improved responses in terms of EFR / ASSR response amplitudes over RAM50 (50% duty cycle) and SAM. DETAILED DESCRIPTION

[0146] As used herein hereinafter, the singular form "a", "an" or "the" includes both the singular and the plural, unless the context clearly indicates otherwise.

[0147] As used below, the terms "comprise" and "comprises" are synonymous with "including", "include" or "contain", and are inclusive and open-ended and do not exclude additional unmentioned parts, elements or method steps. Where this specification indicates that a product or process "includes" certain features, components or steps, it refers to the possibility that other features, components or steps may also be present, but it may also refer to an embodiment that only includes the listed features, components or steps.

[0148] Numerical values ​​recited in numerical ranges include all values ​​or fractions within those ranges, as well as the recited endpoints.

[0149] When referring to a measurable value, such as a parameter, an amount, a time period, etc., the term "about" is used to include variations of + / -10% or less, preferably + / -5% or less, more preferably + / -1% or less, and still more preferably + / -0.1% or less of or from that particular value, which variations are currently applied to the invention disclosed herein. It should be understood that the value itself to which the term "about" refers has also been disclosed.

[0150] All references cited in this specification are deemed to be incorporated by reference in their entirety.

[0151] As used herein, percentages may also be indicated as dimensionless fractions or vice versa. A value of 50% may also be written, for example, as 0.5 or 1 / 2.

[0152] Unless otherwise defined, all terms disclosed in the present invention, including technical and scientific terms, have the meanings commonly given to these terms by those skilled in the art. For further guidance, definitions are included to further explain the terms used in the present specification.

[0153] Auditory nerve fibers (ANF) are located within the cochlea (i.e., the hearing area of ​​the inner ear) and transmit auditory information from the cochlea to the brain. Afferent auditory nerve synapses (ANS) activate the auditory nerve fibers; each ANF usually has a corresponding ANS. There are three types of ANFs with different levels of specific characteristics: low (LSR), medium (MSR), and high spontaneous rate (HSR) ANFs. Degeneration of the number of ANSs and the resulting ANF dysfunction or loss have been associated with cochlear synaptopathy, a pathology that leads to partial or even complete hearing loss. The ANF population may be damaged by hearing impairment (e.g., due to aging, excessive noise exposure, and / or ototoxic agents). Prior art methods focus on testing the hearing sensitivity of a subject, which is a suitable marker for determining outer hair cell (OHC) damage but cannot reliably evaluate and quantify ANF or ANS damage.

[0154] Inner hair cell (IHC) deafferentation, i.e., a reduction in the number of afferent nerves contacting the IHC, can be associated with degeneration of the ANS and ANF, usually resulting in 100% synaptopathy. However, under certain conditions, the ANF and ANS may affect the IHC separately, for example, as may be caused by certain medical treatments, such as chemotherapy. This technique cannot reliably distinguish IHC damage from OHC damage.

[0155] The present invention relates to systems and methods for determining (such as measuring and / or quantifying) the integrity of ANF and / or ANS and / or IHC of a subject. The term "subject" can refer to an animal. In one aspect, the animal can be a mammal. In another aspect, the mammal can be a human. The human can be a healthy person, such as a person with normal hearing, or can also be a patient with an abnormal health condition, such as partial or complete hearing loss.

[0156] The term "integrity" can refer to the functioning of the remaining ANFs, or to both the functioning of the remaining ANFs and their innervation by afferent cochlear synapses, and / or the functioning of the remaining IHCs. The term "integrity" can also refer to the quantification of the number and / or type of ANFs and / or ANS and / or IHCs that are damaged. A reduction in the integrity of the auditory nerve can result in dysfunction of the auditory nerve (AN), whereby synapses are damaged alone rather than the fibers themselves, or in the loss of the AN, whereby the fibers are damaged. As used herein, the terms "measuring integrity" or "determining integrity" can interchangeably refer to a qualitative measurement or a quantitative measurement.

[0157] It has been found that ANF shows strong neural responses to specific auditory stimuli (audio or multi-stimuli), that is, auditory stimuli can induce highly synchronized ANF responses across the cochlea and ANS groups. ANF responses can be recorded by brain neural activity measurements performed by neural activity recording hardware known in the art.

[0158] It is understood that 100% ANF damage can induce the same or similar neural response (or its absence) as 100% IHC damage. Accordingly, if the present method determines 100% damage or loss, this can be interpreted as (i) at least part of the IHC still remains but the ANF completely disappears (synaptopathy) or (ii) the IHC completely disappears, including the ANF attached thereto. Accordingly, if the present method determines less than 100% damage or loss, such as 90%, this indicates that at least part of the IHC still remains and only the ANF and / or ANC are damaged. When all ANS / ANF are completely damaged (i.e., complete cochlear synaptopathy, or complete IHC afferent block) or when the IHCs themselves are lost (e.g., due to ototoxic drugs), reduced integrity of the IHC may occur. One of ordinary skill in the art can therefore understand that the determination of the integrity of the ANS and / or ANF of a subject can be associated with the determination of the integrity of the IHC of the same subject. Neural responses to auditory stimulation can be recorded by measuring the electrical activity of the brain. This activity can be mapped via invasive recording electrodes (for animals) or by electroencephalography (EEG) for humans or animals, etc. For EEG, a number of electrodes, also called an electrode array, attached to the subject's scalp record the overall (neural) activity as a waveform pattern. The EEG data can be processed to determine the integrity of the subject's ANF and / or ANS. The integrity of all or a subset of the ANF / ANS population can be determined.

[0159] Other functional neuroimaging techniques may also be used within the purposes of the present invention to record neural responses. For example, the brain activity of the subject may be plotted by magnetoencephalography (MEG) or electrocardiography (EcochG). The skilled person will appreciate that EcochG / MEG data may be processed in an equivalent manner to the embodiments describing EEG data, and that the application of the present auditory stimulation is not limited to any particular neuroimaging technique. Different neuroimaging and / or hearing test data may also be combined to obtain more accurate or alternative results, such as determining damage to other parts of the hearing, such as outer hair cell (OHC) damage.

[0160] The following provides an initial overview of various aspects of the present invention and then further describes specific embodiments in detail. This initial overview is intended to help readers understand the technical concept more quickly, but is not intended to identify its key or primary features, nor is it intended to limit the scope of the claimed subject matter.

[0161] In addition, one of ordinary skill in the art understands, and upon reading this detailed description will recognize, that the various aspects may be combined unless otherwise stated. Thus, any particular implementation of a particular aspect may be understood to constitute a particular implementation of another aspect without detailed discussion herein. For example, an implementation of a system for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse and / or an inner hair cell of a subject also forms an implementation of a computer-implemented method for determining the same, or an implementation of a computer program configured for performing the computer-implemented method.

[0162] One aspect of the invention relates to a computer-implemented method for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse of a subject, e.g. as described above, comprising the steps of:

[0163] a) receiving brain activity data, preferably EEG or MEG signals, more preferably EEG signals, of a subject exposed to an auditory stimulus s(t), wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); and,

[0164] b) the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses as determined by brain activity data;

[0165] Preferably, the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t), which preferably has a rectangular waveform with a duty cycle d, wherein the duty cycle is preferably at least 10% and at most 40%.

[0166] One aspect of the invention relates to a computer-implemented method for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse and / or an inner hair cell of a subject, e.g., as described above, comprising the steps of:

[0167] a) generating an auditory stimulus s(t), wherein the auditory stimulus comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); exposing a subject to the auditory stimulus s(t); and receiving brain activity data of the subject exposed to the auditory stimulus s(t), preferably an EEG or MEG signal, more preferably an EEG signal; and,

[0168] b) the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses and / or inner hair cells as determined by brain activity data;

[0169] Preferably, the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t), which preferably has a rectangular waveform with a duty cycle d, wherein the duty cycle is preferably at least 10% and at most 40%.

[0170] One aspect of the invention relates to a system for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse and / or an inner hair cell of a subject, for example, as described above, the system comprising:

[0171] - sound generation and stimulation means for generating auditory stimulation s(t);

[0172] - neural activity screening hardware configured to record subject neural responses as neural activity data when said subject is exposed to auditory stimuli s(t) generated by said sound generating and stimulating device; and,

[0173] - A processing unit connected to the recording hardware, configured to:

[0174] a) receiving neural activity data of a subject exposed to an auditory stimulus s(t), and,

[0175] b) determining from the neural activity data the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses and / or inner hair cells;

[0176] Preferably, the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t), which preferably has a rectangular waveform with a duty cycle d, wherein the duty cycle is preferably at least 10% and at most 40%.

[0177] One aspect of the invention relates to a system for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse of a subject, for example, as described above, the system comprising:

[0178] - sound generation and stimulation means for generating auditory stimulation s(t);

[0179] - neural activity screening hardware configured to record a subject's neural responses as neural activity data when said subject is exposed to an auditory stimulus s(t) generated by a sound generating and stimulating device; and,

[0180] - A processing unit connected to the recording hardware, configured to:

[0181] a) receiving neural activity data of a subject exposed to an auditory stimulus s(t), and,

[0182] b) determining from the neural activity data the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses;

[0183] Preferably, the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t), which preferably has a rectangular waveform with a duty cycle d, wherein the duty cycle is preferably at least 10% and at most 40%.

[0184] One aspect of the invention relates to a computer program, or a computer program product directly loadable into the internal memory of a computer, or a computer program product stored on a computer readable medium, or a combination of said computer program or computer program products, configured to perform a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS), an auditory nerve fiber (ANF) and / or an inner hair cell (IHC) of a subject, said computer-implemented method comprising the following steps:

[0185] a) receiving brain activity data, preferably EEG or MEG signals, more preferably EEG signals, of a subject exposed to an auditory stimulus s(t), wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); and,

[0186] b) the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses and / or inner hair cells (IHCs), as determined by brain activity data;

[0187] One aspect of the invention relates to a computer program, or a computer program product directly loadable into the internal memory of a computer, or a computer program product stored on a computer readable medium, or a combination of said computer program or computer program products, configured to perform a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS), an auditory nerve fiber (ANF) of a subject, said computer-implemented method comprising the following steps:

[0188] a) receiving brain activity data, preferably EEG or MEG signals, more preferably EEG signals, of a subject exposed to an auditory stimulus s(t), wherein the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t); and,

[0189] b) the integrity of all or a subset of the subject's specific auditory nerve fibers and / or synapses as determined by brain activity data;

[0190] Preferably, the auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t), which preferably has a rectangular waveform with a duty cycle d, wherein the duty cycle is preferably at least 10% and at most 40%.

[0191] It has been found that when a non-sinusoidal amplitude modulator, and preferably a rectangular waveform, is used, the EEG response to sound stimulation, (also referred to herein as the envelope following response (EFR)) is informative about the extent of damaged ANF fibers and / or damaged IHCs. It has been found that when a non-sinusoidal amplitude modulator, and more preferably a rectangular waveform, is used, the overall range of EEG response amplitudes is much greater.

[0192] The present method provides a robust, non-invasive and potentially real-time EEG-based test for differential diagnosis of ANF and / or ANS and / or IHC damage, which can be applied to the diagnosis of synaptopathy and / or deafferentation, particularly in humans, and the method provides an objective test of speech intelligibility in noise. The method provides a quantifiable explanation for the reduced speech intelligibility in noise, a common complaint of the elderly and noise-exposed listeners.

[0193] The present method has been found to test speech intelligibility in noise better than any existing audiometric test and is insensitive to outer hair cell (OHC) damage that commonly occurs with aging and noise exposure. The present invention is therefore ideally suited for subjects with an unknown mix of synaptopathy and OHC damage. OHC damage and synaptopathy are present in elderly hearing-impaired people. For speech in noise testing, the EEG response amplitude (EFR) from the method is PTN The good correlation between the ) metric and intelligibility suggests a synaptopathy-dominated performance.

[0194] Due to the enhanced response amplitude and the proposed method of extracting relevant parts of the signal from EEG recordings, the present method may provide more robust diagnostic results (compared to other methods) for use in real clinical settings (e.g., without the need for electronic shielding rooms) and / or with portable EEG recording solutions.

[0195] The present method can be used with a wider range of listeners, even in older listeners, to better quantify individual response differences, and thus allow for the diagnosis of synaptopathy in a larger population than would otherwise be possible.

[0196] In some preferred embodiments, the subject is a human or animal subject, preferably a human subject. In some embodiments, the human subject suffers from both synaptic lesions and outer hair cells (OHC) loss, such as synaptic lesions and outer hair cells (OHC) loss due to aging or noise exposure. In some embodiments, the human subject suffers from inner hair cells (IHC) loss, such as inner hair cells (IHC) loss through medical treatment. In some embodiments, the human subject suffers from afferent nerve blockade of IHC caused by ANS / ANF damage. In some embodiments, the human subject suffers from both synaptic lesions and inner hair cells (IHC) loss, such as synaptic lesions and inner hair cells (IHC) loss due to aging or noise exposure. In some embodiments, the human subject suffers from IHC damage of non-ANS / ANF damage; for example, due to administration of specific medical treatments such as chemotherapy.

[0197] This method has the following advantages, it can be applied to most people of all ages and various types of sensorineural hearing disorders and can be used in different conditions: watching movies, sleeping, subconscious, non-verbal (e.g., newborns). People undergoing cancer treatment can also be considered. In some embodiments, the method is used for people with severe IHC-ANF damage (e.g., genetic) who need cochlear implants or for people with some brainstem lesions. Several types of drugs may affect EEG responses and should preferably be considered before testing.

[0198] In some embodiments, the human subject is at least 40 years old, such as at least 50 years old, such as at least 60 years old. Elderly people often exhibit some degree of hearing loss and are therefore particularly suitable for use in the present invention.

[0199] In some embodiments, brain activity data is obtained from a signal, preferably an EEG (electroencephalogram) or MEG (magnetoencephalogram) signal, preferably an EEG signal. Preferred and exemplary embodiments of EEG signals are also preferred and exemplary embodiments of MEG signals. The proposed stimulation paradigm enhances the auditory evoked responses generated by the ANF and ANS groups, which can be reflected from different tests. The present EEG and MEG methods can provide a non-invasive method for hearing screening with high temporal accuracy. As used herein, the term "EEG" also includes EcochG (electrocardiogram), because its setting is basically an EEG recording from the ear canal (tiptrode) or through the middle ear cavity through the tympanic membrane (clinical setting required).

[0200] The auditory stimulus s(t) comprises a carrier signal c(t) amplitude modulated by a non-sinusoidal periodic modulator m(t). Preferably, the audio stimulus s(t) is an external audio stimulus. The external audio stimulus may be generated by sound generation and stimulation. Additionally or alternatively, the auditory stimulus s(t) may be provided via a bone conduction device (vibration on the temporal bone). The audiogram may also be recorded via bone conduction headphones.

[0201] It has been found that ANF shows enhanced evoked responses, especially for rectangular waves. Therefore, auditory stimulation with rectangular waves can induce highly synchronized ANF responses across the cochlear ANF and ANS groups, and thus stimulate increased brain activity, which can be recorded as stronger EEG or MEG signals. When a specific ANF type, or some ANF and / or ANS are vacant or non-functional, brain activity is reduced, resulting in weaker EEG or MEG signals. When the present data analysis method is applied, the percentage of AN fiber and synapse loss can be quantified.

[0202] In some preferred embodiments, the non-sinusoidal amplitude modulator m(t) has a rectangular waveform.

[0203] The envelope of a rectangular wave may have an approximately rectangular shape. As used herein, such an audio stimulus is referred to as a "rectangular wave". The smooth curve outlining the waveform extremes is also referred to as the envelope of the waveform. max The envelope can cover a range of varying non-sinusoidal waveforms, such as square waves (high slope), triangles (low slope), trapezoids, exponentials, etc. In Example 1 and with reference to Figure 3A and Figure 3B The parameters of the rectangular wave and its envelope shape are further detailed.

[0204] As used herein, the term sinusoidal waveform refers to a waveform having a substantially "sinusoidal shape" as understood in the art. Thus, the term non-sinusoidal wave refers to any waveform that does not fall within the former definition of a sinusoidal waveform. In addition, the auditory stimulus includes a carrier signal c(t) that is amplitude modulated by a sinusoidal periodic modulator, referred to as a sinusoidal amplitude modulated (SAM) stimulus. SAM stimuli have been described in the art to mimic the rapidly changing timing characteristics of speech and category-specific vocalizations.

[0205] As used herein, the term rectangular waveform refers to a waveform having a substantially rectangular shape but also including a small amount of deformation (defined by the slope deviation window δ), such as Figure 3A and Figure 3B For example, in Figure 3B The shapes S1, S2, and S3 in are all considered to have deviation δ in the range 0≤δ≤δ max The "rectangular waveform" within. Figure 3A shows a rectangular envelope shape with δ=0 and Figure 3BThe trapezoidal shape S3 in FIG. 1 shows another example with δ not equal to 0. A rectangular waveform is a non-sinusoidal waveform. In addition, the auditory stimulus includes a carrier signal c(t) that is amplitude modulated by a rectangular period modulator, referred to as a rectangular amplitude modulated (RAM) stimulus. The advantages of using a non-sinusoidal waveform over a sinusoidal waveform are discussed below.

[0206] The inventors have found that the speed of the auditory stimulus envelope rise is advantageous in terms of differential sensitivity of markers used for auditory nerve (AN) defects. A rectangular waveform as defined above has a fast rising envelope that can bring the AN directly into its sustained state and produce insufficient time for outer hair cell (OHC)-mediated cochlear responses to affect the firing of the AN. Therefore, the degree of OHC damage will not affect the AN response and the integrity of the AN can be determined differentially, whether or not OHC damage is present.

[0207] In contrast, the slowly rising (sinusoidal or "sinusoidal-like") envelope allows time for the OHC to modulate the AN response and affect the strength of the marker used. For those stimuli, it is not possible to obtain the % reduction in the ANF population from the strength of the neural response amplitude when OHC damage is present. Thus, the sinusoidal waveform at least cannot induce the cycling of neural responses that are not affected by the OHC damage and is therefore not suitable for determining AN integrity, For example AN integrity in older adults who typically present with some degree of OHC damage.

[0208] In some preferred embodiments, the non-sinusoidal amplitude modulator m(t) has a rectangular waveform with a duty cycle d, defined as d=W / T, where d is the duty cycle, W is the pulse width (pulse active time), and T is the total period of the signal, where the duty cycle can range from 1% (single pulse) to 100% (continuous pulses).

[0209] In some preferred embodiments, the duty cycle d is at least 10%, preferably at least 15%, preferably at least 20%, even more preferably about 25%, for example, 24% or 26%. The inventors have found that a stimulation envelope with a duty cycle of at least 10% can allow sufficient time for AN exocytosis to establish and release neurotransmitter material to produce a stable sustained response. For "click" stimulation (duty cycle below 10%), the plateau duration is too short, and this impairs a strong sustained AN response for the stimulation rate.

[0210] In some preferred embodiments, the duty cycle d is at most 40%, preferably at most 35%, preferably at most 30%, even more preferably about 25%, such as 24% or 26%. The inventors have found that for long duty cycles exceeding 40, the AN adaptation time constant may take effect and reduce the overall sustainable AN firing efficiency. At the same time, the neural response may start to show multiple extremes during the stimulation plateau duration (in Fig.14h comparison in Fig.14 yNH traces in i), and that this may impair the strength of the marker to the original stimulus modulation frequency, thus rendering the neural response less sensitive to individual differences in the integrity of the AN synapse / fiber population (also compare Fig.15 RAM25 and RAM50EFR strength).

[0211] In some preferred embodiments, the duty cycle is at least 10% and at most 40%, preferably at least 15% and at most 35%, preferably at least 20% and at most 30%, even more preferably about 25%, for example, 24% or 26%. Further to the duty cycle described above, the inventors have found that a stimulus envelope with a duty cycle between 20% and 30%, for example, about 25%, can induce improved clear single-unit auditory nerve responses and ASSR / EFR waveforms compared to auditory stimulation in the art. Fig.14 Comparative model simulations are found in , which are further discussed in Example 6.

[0212] The deviation δ max The preferred upper limit of is defined as half of the duty cycle d (d = W / T, see Figure 3A ). In some preferred embodiments, δ is at most δ max = d / 2 (or 50%*d), and can have a value between 0 and δ max In some preferred embodiments, the non-sinusoidal amplitude regulator m(t) has a rectangular waveform with a slope deviation window δ between 0 and δ max In the range; max =0.5*d.

[0213] The slope deviation window δ is preferably 0 and can be a maximum of δ max , different modulator shapes can be defined in the delta window ( Figure 3B Several possibilities are shown).

[0214] In some embodiments, δ max Preferably at most 0.4*d, preferably at most 0.3*d, preferably at most 0.2*d, preferably at most 0.1*d, preferably at most 0.05*d, preferably at most 0.02*d, for example at most 0.01*d, most preferably 0.00.

[0215] In some more preferred embodiments, the non-sinusoidal amplitude modulator m(t) has a rectangular waveform with δ = 0. A rectangular waveform may refer to a range of stimulus envelope shapes that fall within a parameter space defined by:

[0216]

[0217] Among them, fm is the modulation frequency, refers to the modulation phase shift, and d refers to the duty cycle (in %). Figure 3A and Figure 3B The parameters of the rectangular wave and its variations are further detailed.

[0218] In some embodiments, the non-sinusoidal periodic modulator m(t) has a period of 2π. A period of 2π has the advantage that the response can be easily analyzed using Fourier transforms.

[0219] In some embodiments, the modulation frequency f of the non-sinusoidal amplitude modulator m(t) is m (defined as 1 / T) is at least 50 Hz, preferably at least 60 Hz, preferably at least 70 Hz, preferably at least 80 Hz, preferably at least 90 Hz, preferably at least 100 Hz, preferably at least 110 Hz, preferably about 120 Hz.

[0220] In some embodiments, the modulation frequency f m It is at most 1000 Hz, preferably at most 800 Hz, preferably at most 600 Hz, preferably at least 400 Hz, preferably at least 300 Hz, preferably at most 200 Hz, preferably at most 150 Hz, preferably about 120 Hz.

[0221] In some embodiments, the modulation frequency f m It is at least 50Hz and at most 1000Hz, preferably at least 60Hz and at most 800Hz, preferably at least 70Hz and at most 600Hz, preferably at least 80Hz and at most 400Hz, preferably at least 90Hz and at most 300Hz, preferably at least 100Hz and at most 200Hz, preferably at least 110Hz and at most 150Hz, preferably about 120Hz. Lower modulation frequencies (<50Hz, preferably <80Hz) induce both cortical and subcortical EEG responses and are therefore less sensitive to synaptic lesions. Higher modulation frequencies (>80Hz) generate subcortical responses and are therefore more sensitive to ANF and ANS degeneration. It is understood that when modulation is too high and carrier tones, resolved harmonics may occur in the cochlea; therefore, the optimal modulation frequency parameters depend on the carrier frequency.

[0222] In some embodiments, the modulation phase shift of the non-sinusoidal amplitude modulator m(t) is at least 0 and at most 2π, e.g. 3π / 2. If there is no slope change (δ=0), then there is no benefit in using a different phase, but if there is a slope change (δ non-0), there is a benefit in starting with a phase shift so that the instantaneous amplitude at the beginning of the envelope waveform is equal to 0.

[0223] In some embodiments, the duty cycle d of the non-sinusoidal amplitude modulator m(t), defined as d=W / T, where d is the duty cycle, W is the pulse width (pulse active time), and T is the total period of the signal, is at least 10%, preferably at least 15%, preferably at least 20%, for example, about 25%. In some embodiments, the duty cycle is at most 40%, preferably at most 35%, preferably at most 30%, for example, about 25%. In some embodiments, the duty cycle is at least 10% and at most 40%, preferably at least 15% and at most 35%, preferably at least 20% and at most 30%, for example, about 25%. Including a duty cycle can ensure the presence of silent gaps between consecutive auditory stimuli. This can allow the stimulated ANFs to return to their resting state and recover, so that the evoked response can stay at the "maximum" for subsequent stimulations. In other words, if silent gaps are not included, subsequent stimulations may result in a reduction in the "maximum" response from the ANF and ANS due to the ANF adaptation characteristics. The shorter the duty cycle, the more synchronized the ANF response. For short duty cycle values, such as 25%, there are longer inter-peak intervals for the same modulation frequency. If the duty cycle becomes too short (i.e. <10%), the ANFs do not receive long enough sustained stimulation to fire according to their "maximal" state (a property related to the kinetics of neurotransmitter release across synapses).

[0224] In some embodiments, the carrier signal c(t) comprises one or more pure tones. In some embodiments, the carrier signal c(t) comprises noise or speech. Pure tone carriers can induce frequency-specific responses to identify the pitch topology configuration of sensorineural hearing loss profiles. Pure tone carrier signals can provide more frequency-specific evaluations, for example, the integrity of ANF, ANS, and / or IHC can be determined for different octaves. Using a speech-like carrier, the frequency range of ANF, ANS, and / or IHC associated with speech intelligibility can be targeted.

[0225] In some embodiments, the carrier frequency is broadband / band-limited noise. If there is no need for frequency-specific assessment, a broadband noise carrier signal can stimulate the entire AN fiber and / or synaptic population. Simultaneous stimulation of the entire hearing range is beneficial for rapid screening tests. In addition, band-limited noise can be used as a carrier, which allows assessment of synaptopathy in frequency regions limited by the cutoff frequency of the noise. Broadband carriers generally provide more robust EEG responses than narrowband noise responses.

[0226] In some embodiments, the carrier frequency f of the carrier signal c(t) is c It is at least 500 Hz, preferably at least 1000 Hz, preferably at least 1500 Hz, preferably at least 2000 Hz, preferably at least 2500 Hz, preferably at least 3000 Hz, preferably at least 3500 Hz, preferably about 4000 Hz.

[0227] In some embodiments, the carrier frequency f c It is at most 32000 Hz, preferably at most 24000 Hz, preferably at most 16000 Hz, preferably at least 12000 Hz, preferably at least 8000 Hz, preferably at most 6000 Hz, preferably at most 5000 Hz, preferably about 4000 Hz.

[0228] In some embodiments, the carrier frequency f c It is at least 500 Hz and at most 32000 Hz, preferably at least 1000 Hz and at most 24000 Hz, preferably at least 1500 Hz and at most 16000 Hz, preferably at least 2000 Hz and at most 1200 Hz, preferably at least 2500 Hz and at most 8000 Hz, preferably at least 3000 Hz and at most 6000 Hz, preferably at least 3500 Hz and at most 5000 Hz, preferably about 4000 Hz. It can be understood that the optimal carrier frequency f c It also depends on the species of the subject, for example rodents have a higher high frequency hearing range than humans.

[0229] In some embodiments, the carrier phase shift of the carrier signal c(t) is is at least 0 and at most 2π, for example 0.

[0230] In some embodiments, the modulation depth md is at least 80%, preferably at least 85%, preferably at least 90%, for example, about 95% or about 100%. In some embodiments, the modulation depth md is at most 100%, preferably at most 99%, preferably at most 97%, for example, about 95%. In some embodiments, the modulation depth md is at least 80% and at most 100%, preferably at least 85% and at most 99%, preferably at least 90% and at most 97%, for example, about 95%. Such a modulation depth avoids silent gaps in each modulation cycle. Generally, higher md% (up to 100%) can induce higher ANF responses. Below 80%, the ANF rate response begins to deteriorate, but is still measurable.

[0231] In some embodiments, the received brain activity data, such as the subject's EEG signal, is received in the time domain. In some preferred embodiments, the subject's EEG signal is received in the time domain as a recorded electrical potential between two EEG electrodes.

[0232] Preferably the time domain signal is sampled via an EEG amplifier. The sampling frequency can be in the range of 500 Hz to 96 kHz depending on the amplifier and the type of test. It is preferred to maintain a sufficient sampling frequency for the analysis so that the highest target frequency, e.g. the minimum 2*f in the EEG recording m and maximum 4*f mLess than the sampling frequency / 2 (i.e., the Nyquist–Shannon sampling theorem).

[0233] In some embodiments, EEG signals may be obtained by placing EEG electrodes, preferably on the subject's scalp, near one or both ears, and / or within one or both ear canals.

[0234] The position of the electrode is preferably suitable for capturing the dipole source generated by the auditory nerve fiber group or the auditory brainstem. In some embodiments, the electrode is placed to measure the AN fibers and synaptic groups of a single ear. In some embodiments, the electrode is placed to measure the AN fibers and synaptic groups of both ears simultaneously. Preferably, the electrodes include a recording electrode placed on the top of the head (Cz, the top of the head) or in the front (Fz), and / or a reference electrode placed on the mastoid or earlobe. The EEG electrodes can be interchangeable and can include any type of EEG electrode, such as a wet-attach or dry EEG electrode or a lead connector placed in the ear canal.

[0235] In some embodiments, step b) comprises the following steps:

[0236] b1) Preprocessing brain activity data, preferably EEG or MEG signals, preferably EEG signals.

[0237] In some preferred embodiments, step b1) includes one or more steps from the following list, including: averaging the EEG recording into multiple stimulation repetitions, noise reduction, filtering (low pass, band pass and / or notch filter), artifact suppression, baseline calibration, epoching, and / or windowing.

[0238] In a preferred embodiment, step b) comprises the following steps:

[0239] b2) Transforming the brain activity data, preferably received EEG or MEG, preferably EEG signals, into the frequency domain (FD).

[0240] In some embodiments, step b2) transforms brain activity data, such as the EEG signal that has been pre-processed in step b1).

[0241] In some preferred embodiments, step b2) comprises applying a (discrete) Fourier transform (dFT) to the time segments of the brain activity data, such as the EEG signal evoked by the auditory stimulation.

[0242] In some embodiments, step b) comprises the following steps:

[0243] b3) Determine brain activity data, such as subject-specific noise floor on EEG recordings.

[0244] In some preferred embodiments, the subject specific noise floor is obtained from the frequency domain transformation performed in step b2) or via bootstrapping / resampling.

[0245] In some preferred embodiments, step b3) includes averaging the noise floor at at least two, preferably at least three or more, e.g. ten frequency points around the stimulation modulation frequency and harmonics of the modulation frequency, and subtracting the average noise floor estimate from the frequency domain (FD) transformed signal at the stimulation modulation frequency and its harmonics. Such a step can produce a relative and noise-corrected response amplitude.

[0246] In some embodiments, step b) comprises the following steps:

[0247] b4) Determine the frequency corresponding to the stimulation modulation frequency f m The response amplitude at frequency F0.

[0248] In some preferred embodiments, step b) comprises the following steps:

[0249] b4') determines the frequency corresponding to the stimulation modulation frequency f m The response amplitude at the frequency F0 and at one or more frequencies F1 to Fn which are harmonics of the modulation frequency.

[0250] As used herein, the term "F0" refers to the modulation frequency of the envelope. As used herein, the terms "F1, F2, ..., Fn" refer to the first harmonic, the second harmonic, and the nth harmonic, respectively. In some preferred embodiments, n is at least 1, preferably at least 2, such as at least 3, 4, 5, 6, 7, 8, 9, 10, etc.

[0251] In some preferred embodiments, the one or more frequencies that are harmonics of the modulation frequency include at least 3 and more preferably at least 4. The human subject harmonics F1-F4 can be considered as the most significant harmonics above the noise floor (NF) estimate, while F5-Fn may not be too obvious in the EEG record. However, the number of significant harmonics may depend on the type. As used herein, a "significant" harmonic or peak can refer to the number of harmonics that produce a significant record and can be used to determine the amplitude of the response after processing.

[0252] In some embodiments, a significant harmonic is a harmonic that is visible in the frequency domain. In some preferred embodiments, a significant harmonic is a harmonic that is 2 standard deviations (p=0.05) above the noise floor.

[0253] In some preferred embodiments, step b) comprises the following steps:

[0254] b4″) determining the response amplitude at the frequency F0 corresponding to the stimulation modulation frequency fm, as well as all frequencies that are harmonics of the modulation frequency, preferably all significant frequencies that are harmonics as defined above.

[0255] Generally, the more harmonics are included, the better the signal reconstruction and thus the accuracy of the method.In this way, the response amplitude can be determined from all harmonics of the modulation frequency Fn, preferably above the noise floor (NF).

[0256] It has been found that in the above method and its embodiments, the EEG response amplitude (which is taken as a measure of the integrity of the subject's AN fibers and / or synapses) shows a better correlation with psychoacoustic measures of hearing performance, such as speech reception threshold in noise. The more harmonics that can be included in the EEG analysis, the more accurately the speech reception threshold in noise of the corresponding individual can be determined.

[0257] In some preferred embodiments, steps b4), b4') or b4") are performed in the frequency domain (FD) obtained in step b2).

[0258] In some embodiments, the absolute amplitude is determined in step b4) or b4'), preferably at a frequency corresponding to the modulation frequency of the envelope and at one or more frequencies which are harmonics of the modulation frequency.

[0259] In some embodiments, the amplitude and phase are determined in steps b4), b4') or b4"), preferably at a frequency corresponding to the stimulation modulation frequency and at one or more frequencies that are harmonics of the modulation frequency. Subtracting the noise floor (NF) from the spectral peaks at the modulation frequency and harmonics can allow direct manipulation of signal values ​​without confounding factors associated with individual NF fluctuations. Typical NF fluctuations may be related to general biophysical noise, head size, age, etc. Since all available harmonics are summed according to the phase information of each harmonic, interpreting the phase information at the modulation frequency and harmonics can allow more accurate recovery of the response amplitude from the original EEG recording.

[0260] The reaction amplitude can be obtained, for example, by summing the signal energies of the previous steps in the frequency domain.

[0261] The response amplitude can alternatively be obtained by estimating the amplitude of a reconstructed signal (e.g., via an inverse (discrete) Fourier transform (iFT) with or without the original phase information), the reconstructed signal comprising only the FD energy from the fundamental frequency and / or harmonics in the reconstruction (with or without noise floor correction).

[0262] In some preferred embodiments, step b), preferably in step b4), b4') or b4"), comprises the following steps:

[0263] - Identification of one or more peaks corresponding to the modulation frequency and its harmonics (F1 to Fn) in the frequency domain, preferably at least 2 peaks, preferably all significant peaks above the noise floor;

[0264] - reconstructing a time domain waveform from said identified peaks, preferably by inverse Fourier transform (iFT); and,

[0265] - measuring the amplitude of the reconstructed signal from the reconstructed time domain waveform;

[0266] Among other things, the amplitude is a measure of the integrity of the subject's AN fibers and / or synapses.

[0267] In some preferred embodiments, the amplitude is half of the peak-to-peak amplitude of the reconstructed time domain signal. The skilled person will appreciate that the amplitude may be quantified as the root mean square (rms) of the time domain signal, or other known data processing techniques.

[0268] In some preferred embodiments, step b), preferably in step b4'), comprises the following steps:

[0269] - Identify at least 2 peaks in the frequency domain, preferably all significant peaks above the noise floor, including at least:

[0270] - the modulation frequency F0 of the stimulation; and,

[0271] - one or more harmonics (F1 to Fn) of said modulation frequency;

[0272] - reconstructing a time domain waveform from said identified peaks, preferably by inverse Fourier transform (iFT); and,

[0273] - measuring the amplitude of the reconstructed signal from the reconstructed time domain waveform;

[0274] Therein, the amplitude is a measure of the level of integrity of the subject's AN fibers and / or synapses.

[0275] It has been found that this method produces larger response amplitudes, which capture more aspects of the recorded signal. After the discrete Fourier transform (FT) of the pre-processed, averaged and time-segmented EEG recordings, both the individual noise floor and the spectral peaks at the frequency of the modulator and its harmonics can be identified. The energy of these peaks on the noise floor can be obtained in the frequency domain but is preferably part of the reconstructed signal calculated by the inverse dFT. The amplitude of the time domain signal is defined as the waveform of the evoked response amplitude and is an EEG marker of ANF and ANS integrity developed by this method.

[0276] Depending on the quality of the EEG amplifier and the EEG electrodes, the stimulation is preferably repeated anywhere between at least 10 and 4000 times to achieve a reliable signal that significantly exceeds the biological / electronic noise floor. The more repetitions are needed, the longer the recording lasts. For animals, about 30 times are sufficient due to better electrode contact, but for humans, at least 400 repetitions are usually required to improve the average value. The same signal quality can be achieved by recording fewer average values ​​but using different electrode types (e.g., lead connectors inserted in the ear canal). Alternatively, more recording electrodes can be used to achieve a better signal-to-noise ratio (as the average signal of several electrodes) with fewer stimulation repetitions. In a preferred embodiment, the stimulation is repeated between at least 400 and at most 2000 repetitions. This sub-range can be applied to human recordings to obtain a balance between recording time and signal quality for long stimulation periods (epochs). It can be understood that due to stronger evoked responses, this method requires fewer repetitions than other methods.

[0277] The term "determining integrity" can be both a qualitative measurement and a quantitative measurement. In some embodiments, the "degree of subject-specific ANF / ANS integrity" provides a qualitative result; for example, whether synaptic pathology is present or not, or whether synaptic pathology is more (or less) present (e.g., than another subject such as a "healthy" subject, or than a previous measurement of the same subject). In some embodiments, the degree of subject-specific ANF / ANS integrity provides a quantitative result; for example, defining the degree of synaptic pathology as the % of AN synapse or fiber loss.

[0278] Depending on the stimulation specificity, the method can target different subclasses of AN fibers (low, medium or high spontaneous rate) or the method can target all AN fiber types and ANS in a specific frequency region (i.e. between 2-4kHz). This allows the method to measure the integrity of the ANF / ANS population within the targeted frequency region of interest.

[0279] In some embodiments, the method comprises the following steps:

[0280] b') determining the degree of integrity of subject-specific AN fiber and synaptic populations from brain activity data, preferably EEG or MEG signals; preferably quantifying the degree of integrity of subject-specific AN fiber and synaptic populations from brain activity data, preferably EEG or MEG signals.

[0281] In some embodiments, the method comprises the following steps:

[0282] b”) determining the degree of integrity of a subset of subject-specific AN fiber and synaptic populations from brain activity data, preferably EEG or MEG signals; preferably quantifying the degree of integrity of a subset of subject-specific AN fiber and synaptic populations from brain activity data, preferably EEG or MEG signals.

[0283] In some embodiments, the subset is frequency specific, i.e., the carrier may be tonal, targeted to a specific frequency within the hearing range, or broadband to cover all or a wider portion of the hearing range with one stimulus (e.g., using a band cutoff). In such embodiments, the method targets the health of only a subset of the ANF / ANS population in the target frequency region. This has the advantage of not only providing a yes-no answer, but the % of loss can provide a more sensitive diagnostic tool within the target frequency range.

[0284] In some embodiments, the subsets are subclass specific, i.e., the carrier may be tonal, targeting specific ANF subclasses within the ANF population, such as low (LSR), medium (MSR), and high spontaneous rate (HSR) ANFs, by varying the modulation depth md of the auditory stimulus. In such embodiments, the method targets the health of only a subset of the ANF population in the targeted ANF subclass.

[0285] In some embodiments, methods described herein are used to quantify the extent of percentage loss when ANF and / or ANS clusters are suspected to be involved. ANF and / or ANS clusters may be impaired due to hearing loss (e.g., due to aging, excessive noise exposure, genetic loss, and / or ototoxic agents).

[0286] The present invention also relates to the use of the method as described herein, and embodiments thereof, as a screening method for the presence of a synaptopathology in a subject.

[0287] In some embodiments, the method can be used to identify whether an amplitude above the noise floor is observed. For example, if the response is found to have a higher signal-to-noise ratio than a predetermined value, and / or if the response is found to have a higher amplitude than a predetermined value, it may be found that the subject does not have a synaptopathy. Conversely, if the response is found to have a lower signal-to-noise ratio than a predetermined value, and / or if the response is found to have a lower amplitude than a predetermined value, it may be found that a synaptopathy is present.

[0288] The present invention also relates to the use of the method as described herein, and embodiments thereof, as a screening method for quantifying the extent of synaptopathy (i.e., ANF / ANS loss) in a subject. Preferably, the extent of AN fiber and / or synapse loss is defined as % of healthy subjects (i.e., subjects with normal hearing) in the total population. AN fiber and synapse loss can also be compared to specific subgroups of subjects, such as by age, gender, occupation, etc.

[0289] In some embodiments, the parameters of the modification stimulation are targeted to one or another subgroup. There are three types of neural fibers with different levels of specific properties. Low (LSR), medium (MSR) and high spontaneous rate (HSR) ANF fibers. It has been found that the method as described herein is particularly sensitive to all spontaneous rate fiber types. For narrowband stimulation, subgroups can be targeted by changing the modulation depth of the stimulation paradigm. (For example, reducing the modulation depth of the signal s(t) may enhance the selectivity for the loss of LSR fibers).

[0290] The present invention also relates to the use of methods as described herein, and embodiments thereof, as screening methods for quantifying the degree of hearing loss in a subject. Preferably, the degree of AN fiber and synaptic loss is defined as the % of damage. Preferably, the % of damage is combined with another diagnostic marker that can determine the degree of OHC damage. It has been found that the methods described herein allow for the distinction between OHC and ANF damage, therefore, allowing for a more accurate determination of sensorineural hearing loss and separating it into sensory (OHC) and neural (ANF / ANS) loss portions.

[0291] The extent of AN fiber and synaptic integrity and / or IHC integrity can be quantified using various methods described below.

[0292] In some embodiments, the degree of synaptopathy can be quantified using a first quantification method. In such a first quantification method, the method as described herein is combined with a computational model (ie, simulation), wherein ANF parameters can be modified and wherein auditory brainstem EEG sources can be simulated.

[0293] In some embodiments, a second quantification method can be used to quantify the degree of synaptic pathology. In such a second quantification method, the method described herein is combined with data from a large reference data set (e.g., EEG data) of listeners with or without hearing pathology.

[0294] In some embodiments, a third quantification method can be used to quantify the degree of synaptic pathology. In such a third quantification method, the method described herein is combined with data from a computational model and data from a large reference data set.

[0295] In some embodiments, as described herein, the EEG response of the subject to the present auditory stimulus is experimentally measured using an EEG device and the EEG response is inserted into a biophysical model of the auditory periphery. For example, individual parameters of ANF / ANS damage derived from EEG responses can be inserted into a biophysical model of the auditory periphery to simulate individual auditory processing. The model may include data from, for example, simulations and / or experimental frequency and / or type-specific ANF / ANS damage profiles. In some embodiments, an ANF / ANS damage profile is determined based on experimental data (e.g., EEG recordings). ANF / ANS data can be subdivided based on a subset of the ANF / ANS group; this can include high spontaneous rate fibers (HSR), medium spontaneous rate fibers (HSR), and low spontaneous rate fibers (LSR).

[0296] In some embodiments, possible EEG responses to all or a subset of ANF / ANS groups can be simulated to obtain predicted auditory responses to auditory stimuli (e.g., auditory EEG responses, auditory sound perception, cochlear, ANF, and brainstem processing). Calculating the response amplitudes of all possible EEG responses (from simulations) can allow the creation of various auditory profiles corresponding to different ANF profiles or other input parameters. Using category-based parameters, auditory profiles can be further subdivided based on, for example, age, gender, etc. or other parameters. The calculated profiles can be stored or retrieved through a database.

[0297] In some embodiments, the EEG response data of the subject processed by predictive simulation data interpretation can be used to assign an auditory profile to the subject. The assignment can be performed automatically by a matching algorithm. The assigned profile is preferably based on the most likely match between the simulated and recorded EEG response amplitudes. The integrity of the subject's ANF / ANS and / or IHC can be determined based on the assigned auditory profile. It has been found that LSR and MSR fibers and, in particular, their afferent synapses are the most vulnerable parts during noise-induced hearing loss. Histological studies have also shown that LSR fibers and synapses disappear first.

[0298] In some embodiments, the methods for determining the integrity of ANF / ANS and or IHC in a subject may also be extended to determining the extent of damage to outer hair cells (OHC) in said subject.

[0299] In some embodiments, the biophysical model of the auditory periphery can be extended to also include, for example, simulated and / or experimental frequency-specific OHC lesion profiles. The OHC lesion profile can be determined based on experimental data, i.e., frequency-specific OHC lesions (e.g., derived from audiogram testing, otoacoustic emissions). Alternatively, the OHC lesion profile can remain variable so that the matching algorithm for both the ANF and OHC profiles can be optimized simultaneously.

[0300] In some embodiments, methods for determining the integrity of AN fibers and / or the integrity of synapses and / or the integrity of IHCs and / or damage to OHCs in a subject can be used to modify the subject's desired auditory response to sound. The difference between the desired auditory response and the auditory response corresponding to an individual's AN / OHC damage profile can be used to optimize the signal processing algorithms of a hearing device such as a cochlear implant or a hearing aid.

[0301] In some preferred embodiments, the (individual and / or simulated) auditory response to sound (e.g., auditory EEG response, sound understanding, cochlear, ANF, and brainstem processing) is used to adjust specific aspects of the sound stimulation in the time domain or frequency domain, preferably adjusting the intensity and / or temporal envelope shape (e.g., initial sharpening / envelope depth enhancement). The expected auditory response to the sound (e.g., normal hearing or auditory feature enhancement response) can be simulated or recorded. The difference between the expected auditory response and the auditory response corresponding to the subject's AN fiber and synaptic integrity and / or IHC integrity and / or OHC damage profile can then form a feedback loop to the processing unit of the hearing device. For example, the feedback loop can be used to optimize the signal processing algorithm to adjust the sound stimulation in these devices.

[0302] The present invention also relates to a method for modifying a subject's desired auditory response to a sound. The method preferably comprises the following steps:

[0303] - performing a method as described herein to determine the subject's auditory nerve fiber and / or synapse (and optionally OHC) hearing damage profile; and,

[0304] - Hearing profiles included in the auditory peripheral model to simulate auditory responses to any acoustic stimulus.

[0305] In some embodiments, the method comprises the following steps:

[0306] - Comparing the individual simulated auditory response to an expected auditory response, for example, wherein the expected response can be experimental or simulated, such as a response from a normal hearing subject or a response with enhanced characteristics.

[0307] In some embodiments, the method comprises the following steps:

[0308] - Including a signal processing algorithm to adjust the sound stimulus so that the simulated auditory response matches the expected auditory response. The matching algorithm may, for example, ultimately filter, onset sharpening, compress and / or expand the audio stimulus.

[0309] The invention also relates to a method for generating an individualized mathematical auditory peripheral model of a subject for processing sounds for a hearing device, the method comprising the following steps:

[0310] - determining the integrity of AN synapses and / or fibers and / or IHCs in a subject according to the methods, and embodiments thereof, as described herein;

[0311] -optionally, determining the fraction of outer hair cells that are damaged; and,

[0312] - Generating an individualized mathematical model for processing sound for a hearing device based on the integrity of the subject's AN synapses and / or fibers and / or IHCs, and optionally the fraction of outer hair cell damage.

[0313] In some embodiments, the method includes configuring a hearing device, such as a cochlear implant or a wearable hearing aid. Preferably, the method includes adjusting the sound stimulus to minimize the difference between the desired auditory response (e.g., a normal hearing response or an auditory feature enhanced response) and the auditory response simulated from an individualized mathematical model for processing sound.

[0314] The present invention also relates to a computer-implemented method for comparing the extent of hearing loss in a subject caused by a combination of ANF / ANS damage and by outer hair cell (OHC) damage or loss.

[0315] The method can be based on the degree of AN fiber and synaptic and / or IHC integrity obtained by the methods described herein. The method can be combined with another EEG marker to form a sensorineural hearing loss map with differential sensitivity for the two aspects of sensorineural hearing loss (AN and OHC).

[0316] In some preferred embodiments, a method for comparing the extent of hearing loss caused by synaptopathy and by outer hair cell (OHC) loss in a subject comprises the following steps:

[0317] - performing a method as described herein, or an embodiment thereof, to quantify the extent of ANF / ANS damage, preferably the extent of synaptopathy, in a subject;

[0318] - obtaining a marker sensitive only to outer hair cell damage (OHC) of the subject, preferably an EGG marker; to quantify the extent of outer hair cell damage in the subject; and,

[0319] - The extent of ANF / ANS damage and the extent of outer hair cell damage were plotted on a 2-dimensional graph.

[0320] The OHC damage marker may be an EEG marker. Alternatively, the marker is obtained by performing another test (e.g., audiogram, otoacoustic emission threshold) that is sensitive only to outer hair cell (OHC) damage on the subject. The graph preferably forms a sensorineural hearing loss graph. For example, the graph may include an EEG marker of synaptopathy obtained on the X-axis (or Y-axis) and an EEG marker that is uniquely sensitive to OHC damage on the Y-axis (or X-axis).

[0321] Using the above methods, different individuals can be placed on a sensorineural hearing loss map to drive decisions about treatment options.

[0322] In some embodiments, the method is frequency specific. In such embodiments, the method targets only the health of a subset of the ANF / ANS population in the target frequency hearing range. This has the advantage of being able to obtain a frequency specific sensorineural hearing loss map.

[0323] The present invention also relates to a computer program, or a computer program product which can be directly loaded into the internal memory of a computer, or a computer program product which is stored on a computer-readable medium, or a combination of said computer program or computer program products, configured to perform the method described herein, or an embodiment thereof.

[0324] The present invention also relates to a system for determining the integrity of an auditory nerve fiber and / or an auditory nerve synapse and / or an IHC of a subject, the system comprising:

[0325] -EEG-based hearing screening hardware, including electrodes and amplifiers;

[0326] - sound stimulation generation and delivery means; and,

[0327] - a processing unit configured to perform a method as described herein, or an embodiment thereof.

[0328] EEG hardware, also referred to herein as an EEG recording device, may include clinical or research EEG amplifiers and electrode caps (eg, Neurospec, Biosemi, Interacoustics, BrainScan, and Smart Hearing Systems).

[0329] In some embodiments, the system includes at least one reference electrode, a signal electrode, and a ground electrode.

[0330] In some embodiments, the system allows recordings with sampling frequencies above 4 kHz and 4 electrodes (2 reference, 2 signal) and provides pre-amplification of electrode signals with a minimum gain of 40 dB, i.e., this feature is preferably more stringent than for brain-computer-interface EEG, portable EEG, or cortical EEG.

[0331] Preferably, the system has a sampling frequency that is at least twice the highest modulation frequency of interest. In some embodiments, this corresponds to a sampling frequency of at least 1000 Hz, preferably at least 2 kHz, preferably at least 4 kHz, preferably at least 8 kHz, preferably at least 12 kHz, such as about 16 kHz. This allows for the generation of sufficient frequency resolution, such as short duration stimulation to extract spectral peaks in analysis.

[0332] Preferably, the recorded EEG signal is amplified using more stringent gain parameters than the brain-computer interface, giving a small amplitude of the target signal (of the order of 0.5 microV), which is much smaller than the cortical EEG signal. Amplification has the advantage of avoiding the need for more signal repetitions in order to obtain a sufficient signal-to-noise ratio in the rating recording.

[0333] The EEG electrode configuration may include any of conventional multi-channel cap EEG, peri-auricular electrodes (eg, cEEGrid), or adhesive electrodes placed on Cz and / or Fp (eg, dual-channel IHS recording amplifier).

[0334] The reference and signal electrodes are preferably positioned in such a way that they best capture the dipole generated by the AN fibers and synapses in a particular ear. This dipole is usually oriented from the bottom to the top of the head, so the placement of mastoid-Cz or mastoid-Fz or earlobe-Cz or earlobe-Fz has been found to be the best position to obtain this response. In some cases, you can also place the recording electrode on one mastoid and the reference electrode on the other mastoid.

[0335] The present invention also relates to a kit for performing the method as described herein, and embodiments thereof, the kit comprising:

[0336] - a system as described herein, and embodiments thereof; and,

[0337] - a computer program as described herein, and embodiments thereof, and / or a computer program configured to perform a method as described herein, and embodiments thereof.

[0338] The kit may comprise a computing device preconfigured with a computer program for performing the method as described herein, and embodiments thereof.

[0339] Example

[0340] Example 1: Generating auditory stimulation

[0341] An auditory simulation not according to the invention is illustrated in Figure 2. The figure illustrates an auditory stimulus s(t) created by modulating a pure tone carrier signal c(t) with a sinusoidal modulator m(t). Figure 2A shows the auditory stimulus s(t) in the time domain, and Figure 2B Auditory stimuli in the frequency domain are presented.

[0342] An exemplary method for generating auditory stimulation according to a preferred embodiment of the present invention is further described in detail in FIG3. In particular, Figure 3A is an example periodic modulator m(t) generated by the following square() function in MATLAB R2014b:

[0343]

[0344] Use the following modulation parameters:

[0345] Modulation frequency (f m )=120Hz

[0346]

[0347] Duty cycle (d) = 25%

[0348] Envelope slope change and / or envelope distortion window (δ has 0≤δ≤δ max , where δ max =d / 2)( Figure 3A ).

[0349] The duty cycle d is the ratio of the pulse duration or width (W) of the waveform to the total duration (T), that is, d = W / T*100%, such as Figure 3A As shown in . Figure 3A The exemplary stimulus has an optimal δ = 0, but in δ max Envelope slope / shape deviations from the rectangular shape are allowed within a margin. Examples of possible slope / shape deviations are provided in Figure 3B middle.

[0350] The periodic modulator m(t) described above is used to modulate the carrier signal c(t) to produce the auditory stimulus s(t). Figure 3C is an exemplary implementation of an auditory stimulus s(t) generated by the following sin() function in MATLAB R2014b.

[0351]

[0352] Use the following carrier parameters:

[0353] Carrier frequency (f m )=4000Hz

[0354]

[0355] Modulation depth (md) = 95%, also Figure 3C Displayed in.

[0356] The generated auditory stimulus s(t) has a rectangular waveform with δ=0 and period 2π. It presents a stimulus s(t) generated by modulating a pure tone carrier signal c(t) with the modulator m(t) described above.

[0357] In practice, pure tone carrier signals can induce frequency-specific responses to identify the tone topology configuration of sensorineural hearing loss profiles. For rapid hearing screening tests, broadband noise can be used as the carrier c(t) to produce a single response that stimulates all cochlear tone topology frequencies at once. Figure 3D is an exemplary implementation of an auditory stimulus s(t) generated by the rand() function in MATLAB R2014b. It presents a stimulus s(t) generated by modulating a broadband noise carrier c(t) with the modulator m(t) described above.

[0358] Example 2: Neural Response to Auditory Stimuli

[0359] Compared with the auditory stimulation not according to the present invention (i.e. Figure 2A , Fig. 6A The auditory stimuli generated according to the present method (i.e., the sinusoidal amplitude modulation shown in FIG. 2 ) were compared to those generated by the present method. Figure 3C , Figure 6B Using the same carrier frequency (f = 4 kHz) and ratio (f m =120Hz, starting phase = 3π / 2) to generate two amplitude modulated pure tone stimuli. The only difference between the two stimuli is the shape of their modulator waveforms: in particular, the stimulus s1(t) not according to the present invention uses a sinusoidal amplitude modulator with the following equation:

[0360]

[0361] The second stimulus s2(t), generated by a non-sinusoidal (rectangular) amplitude modulator as described in Example 1 above, has a duty cycle d of 25%. A modulation depth md of 0.95 (95%, -0.45dB re. 100%) is used to avoid silent gaps in each modulation cycle. Both stimuli are windowed using a 2.5% tapered cosine window, have a duration of 0.4 seconds, and each stimulus is repeated 1000 times (500 times for each polarity).

[0362] More repetitions allow to improve the overall signal quality (signal quality improves proportional to the square root (without averaging)). The longer the duration, the better the resolution of the spectral part in the frequency domain will be. The minimum duration of the signal is preferably at least 100 ms to ensure that the AN fibers reach their steady state condition of firing. Before that time, initiation effects may occur.

[0363] The interstimulus interval (ISI) consisted of uniformly distributed random silent jitters (mean ISI: 100 ms ± 10 ms). The stimuli not according to the invention were presented at 70 dB SPL, and the present stimuli were presented at the same peak-to-peak amplitude corresponding to 68 dB SPL.

[0364] ISI refers to the time between two periods of 0.4 duration stimulation, not the time between cycles of stimulation. The ISI is preferably jittered to avoid time locking of the EEG signal to, for example, a constant frequency heartbeat or line noise. The ISI preferably has a minimum duration of at least 60 ms, preferably at least 80 ms, for example about 100 ms, so that the AN fibers "restart" after stimulation.

[0365] Test Group

[0366] Three participant groups were recruited, using age (20≤age≤30 or 60≤age≤70) and audiometric pure tone thresholds (≤ or >25dB HL for frequencies up to 4kHz) as classification criteria. This grouping criterion does not take into account the individual degree of synaptopathy, which is a variable that can be expressed at an individual level within each participant group. 15 young normal hearing (yNH: 24.5±2.3y / o, 8 females), 16 elderly normal hearing (oNH: 64.3±1.9y / o, 8 females), and 13 elderly hearing impaired (oHI: 65.2±1.8y / o, 8 females) participants participated in the study. There was no significant age difference between the oNH and oHI participant groups (t(27)=-1.41, p=0.17). Otoscopic examination was performed before data collection to ensure that participants had no obstruction or other visible pathology.

[0367] The suitability of the participants was verified by pure tone audiometry thresholds (PT) collected for frequencies between 0.125-8kHz using, for example, a Sennheiser HDA200 headset using a clinical audiometer AT900 (Auritec). Normal hearing participants had 4-kHz pure tone thresholds of 3.3±3.6 (yNH) and 11.6±4.0dB HL (oNH), respectively. The oHI group had a high-frequency drop-type audiogram with a 4-kHz threshold of 37.7±6.7dBHL. Distortion product otoacoustic emissions (DPOAE) were recorded at 4kHz as a direct estimate of OHC integrity. Stimulation was presented on an ER-2 speaker using foam earplugs and recorded using an ER10B+OAE microphone system (Etymotic Research) and a customized MATLAB script (Mauermann, 2013). Two pure tones (f1, f2) were presented simultaneously using the main frequency sweep method with a fixed f2 / f1 ratio of 1.2. The L1 levels followed the Scissors paradigm, assuming a primary L2 level of 30–60 dB SPL in 6 dB steps (in the oHI listeners, additional L2s were collected at 66 and 72 dB SPL). The DPOAE threshold was determined as the median of the bootstrapped L2 levels, at which the extrapolated fit curve reached a level of −25 dB SPL (1.12 mPa).

[0368] Individual speech intelligibility

[0369] The speech reception threshold (SRT) determines the speech-to-noise ratio (in dB) at which 50% of the words presented in a 5-word sentence are correctly perceived using an adaptive tracking program. The German matrix test is used to determine the SRT in the presence of spectral noise. The SRT for 50% correct word recognition is determined using a 1-up / 1-down adaptive program with a variable step size based on the word score. Speech and noise signals are generated by applying a 1024th order FIR high-pass filter with a 1.65kHz cutoff frequency. The noise level is kept fixed at 70dB SPL while the speech level is varied to determine the SRT.

[0370] EEG responses to auditory stimulation

[0371] Electroencephalogram (EEG) recordings were performed in a double-walled electrically shielded measurement room (IAC Acoustics), and participants watched silent movies while sitting in a reclining chair. EEG was measured using a 64-channel cap (Easycap) with equidistant electrode spacing, and active Biosemi Ag / AgCl electrodes were connected to a Biosemi amplifier. A 16384 Hz sampling rate (fs) and 24-bit AD conversion were used to store raw data traces.

[0372] The common mode sensing (CMS) active electrode was placed on the midfrontal line and the right leg drive (DRL) powered electrode was placed on the tip of the nose. A reference electrode was placed on each earlobe. The electrode offset voltage (DC value of the common mode signal) was kept less than 25 mV. Stimuli were generated in MATLAB (R2015b) at a sampling rate of 48 kHz and the same ear simulator type 4157 (Bruel & )) The same equipment was calibrated. A Fireface UCX sound card (RME) connected to a TDT-HB7 headphone driver (Tucker-Davis) was used to convert the digital signal to drive 330ER-2 in-ear headphones (Etymotic Research) using the open source portaudioplayrec ASIO codec (Humphrey, 2008). Stimuli were presented in mono based on the better ear of the audio metric. An exemplary EEG setup is shown in Figure 4 middle.

[0373] EEG processing - predicted results

[0374] Raw EEG recordings were extracted in Python (version 2.7.10 | Anaconda 2.3.0 (64-bit), (www.python.org) and MNE-Python (version; 0.9.0 Gramfort et al., 2013, 2014), and all channels were rereferenced to the offline averaged earlobe electrodes. Data were segmented into 0.4 s windows (epochs) starting from stimulus onset and baseline corrected by the mean amplitude of each epoch.

[0375] The preprocessing was performed in MATLAB (R2014b). Based on the modulation frequency and its first four harmonics (h0-h4 = k×f m , k = [1..5]) are estimated for each stimulation condition and participant to account for all envelope-related energies in the EEG signal. This process can be observed in Figure 5, which is a schematic diagram of the calculation process for each bootstrap run based on the average yNH EEG response of a single participant to the present rectangular modulation stimulus.

[0376] A) The amplitudes of the modulation frequency and its first four harmonics (h0-h4) are extracted from the average spectral estimate of the neural frequency response (Fn) (PtN, vertical arrows). The horizontal lines at the base of the frequency peaks represent the respective noise floor level (NF) for each considered frequency.

[0377] B) Reconstructed time domain waveform based on five frequency components (h0-h4) and their respective phases. The amplitude was extracted from the averaged, bootstrapped and reconstructed time domain waveforms (ie, the peak-to-peak amplitude divided by 2).

[0378] Model prediction results

[0379] Figure 6C and Fig.6D Describes the summed ANF firing frequency from simulations of ANFs synapsed to a single IHC with a 4-kHz CF, i.e., without stimulation according to the present invention ( Figure 6C ) and the stimulation according to the invention ( Fig.6D ) carrier frequency. The simulation was performed using a computational model of the human auditory periphery (Verhulst et al., Hearing Research, 2018). The responses to the two cycles of stimulation described above are shown in three SNHL profiles:

[0380] (i) Total NH reactions on 19ANF(NH);

[0381] (ii) a synaptopathy response (HIsyn), in which the number of ANFs is reduced to seven high-SR fibers; and,

[0382] (iii) OHC injury response (HI flat) with 19 ANFs across all frequencies and 35 dB hearing sensitivity loss.

[0383] In summary, the present stimulation elicits larger and more synchronous ANF responses with sharper switch settings compared to stimulation not according to the invention. When the number of available ANFs (HIsyn) is low, the ANF firing frequency decreases for both stimulation types. The introduction of OHC damage (HI flat) affects stimulation not according to the invention and stimulation according to the invention differently: when the ANF rate to tones not according to the invention is increased and a bell-shaped pattern begins to appear, the response to the present stimulation is almost unaffected. The results show that the on-CF ANF response to the present stimulation is sensitive to the synaptopathy aspect of sensorineural hearing loss but is insensitive to the OHC damage aspect.

[0384] Single-unit ANF simulations predict this stimulus:

[0385] (i) produce overall larger response amplitudes, which can enhance the sensitivity of the method to characterize individual differences; and,

[0386] (ii) Induce responses that are primarily sensitive to synaptopathies, even when co-occurring OHC lesions are also present.

[0387] Fig. 6EThe simulated responses to stimulation according to the invention (squares) are shown to be larger than the responses to Sinusoidal Amplitude Modulated (SAM) stimulation not according to the invention (circles). Together with the overall increased amplitude, the present stimulation also produces a larger sensitivity range. Compared to the effect of synaptopathy, the predicted effect of OHC damage (HI flat) is smaller and becomes smaller as the degree of synaptopathy increases.

[0388] Experimental Results

[0389] Fig. 6F The amplitudes of the experimental EEG-recordings described in the Figure 1 were consistent with model predictions and confirmed to be larger than all native response amplitudes. Second, the predicted sensitivity to different aspects of SNHL may be reflected in the different behaviors of the oNH group under different stimulation conditions. Whereas the oNH response results to SAM stimulation overlapped to a large extent with those of the yNH group, there was a clear separation between the response amplitudes of the yNH and oNH groups to native stimulation (i.e., showing a higher sensitivity to synaptopathy).

[0390] In addition, FIG7 shows that in similar objective hearing threshold groups (distortion product OAE threshold> or <25dB SPL), rectangular envelope stimulation ( Figure 7C and Fig.7D ) than SAM stimulation ( Fig. 7A and Figure 7B ) can better be based on EEG response amplitude (EFR PtN ) separated the young and old listeners. This finding is consistent with how age-related synaptopathy would be expected to affect EEG responses. Age-related synaptopathy could be detected based on the EEG responses to rectangular stimuli, but not based on the DP threshold in the <25 dB SPL group (C). Meanwhile, for DP thresholds >25 dB SPL (HI DP ) listeners, the EEG responses to the rectangular stimulus (D) placed the elderly subjects of the same age in the same range (i.e., similar degrees of cochlear synaptopathy), whereas they were separated from each other when the SAM stimulus (B) was used. The latter observation is consistent with the observation that the SAM EEG response is affected by both OHC damage and synaptopathy, whereas the present rectangular stimulus is primarily sensitive to the synaptopathy aspect of age-related sensorineural hearing loss. Together with Figure 6C and Fig.6D These experimental findings support the prediction of the model in that the EEG response to this rectangular stimulus is primarily sensitive to the synaptopathy of sensorineural hearing loss.

[0391] The combined simulation and experimental approach enabled an assessment of the sensitivity of currently considered EEG markers of synaptopathy to threshold or suprathreshold TENV deficits, which may result from OHC and ANF / ANS losses, respectively.

[0392] Fig. 8A and Figure 8B Displays a stimulus with a sinusoidal envelope ( Fig. 8A , Figure 2A ) compared to the stimulus with a rectangular envelope ( Figure 8B , Figure 3C ) can better predict the performance of individual speech intelligibility in noise. These data show that the present method for determining ANF / ANS integrity provides improved sensitivity to predict the performance of individual speech intelligibility in background noise.

[0393] in addition, Figure 8C and Fig.8D A preferred method for obtaining amplitude from an EEG recording (a marker of ANF / ANS integrity) is shown, which includes the stimulus envelope and all available harmonics represented by the FD of the EEG recording (f0-f5) to extract the amplitude, compared to a method applied to the same EEG recording that only considers f0 to calculate the amplitude ( Figure 8C ), the preferred method can better predict the intelligibility of a single speech ( Fig.8D ). The preferred method produces an overall larger amplitude (EFR), which makes the preferred method more robust under background noise. In addition, the preferred method is able to predict individual speech intelligibility more robustly (and significantly more) than using EEG analysis methods based on f0.

[0394] Example 3: Method for determining the integrity of AN fibers and synapses in a subject

[0395] refer to Fig. 9 Discussion A model for determining the integrity of a subject's auditory nerve fibers and synapses according to a preferred embodiment of the present invention presents a flow chart of preferred steps for determining an ANF / ANS integrity profile and optionally using reference data to determine a subject-specific auditory profile. Recordings are compared to a normative dataset of a "normal" person with normal ANF / ANS. By comparing the reference to the subject, a subject-specific auditory profile can be obtained.

[0396] (100) is an auditory stimulus (e.g., sound) that induces an auditory response across AN fibers and synaptic populations along the cochlea. The stimulus can be used for EEG recordings to diagnose ANF / ANS damage. The stimulus characteristics can be designed to target limited or broadband ranges of hearing. In a preferred embodiment, the auditory stimulus can be a carrier signal c(t) (e.g., broadband noise or a pure tone) that is amplitude modulated by a periodic modulator having a non-sinusoidal (rectangular) waveform m(t).

[0397] (200) is a biophysical model of auditory peripheral signal processing, (which preferably includes a digital description of cochlear mechanisms, outer and inner hair cell function, and firing frequencies representing AN synapses and firing). The model can include data from (210) of, for example, simulated and / or experimental frequency and / or type specific ANF damage profiles. The damage profile of the ANF / ANS (210) can be determined based on experimental data (e.g., EEG recordings). The ANF / ANS data can be subdivided based on subsets of ANF populations; this can include high spontaneous rate fibers (HSR), medium spontaneous rate fibers (HSR), and low spontaneous rate fibers (LSR), and / or subgroups of these fibers in a selected hearing frequency range.

[0398] (300) Responses to all or a subset of the ANF / ANS group can be simulated to obtain a predicted auditory response to an auditory stimulus. The auditory response can be a simulated auditory EEG response, simulated auditory sound perception, and / or simulated cochlear, ANF, and brainstem processing). Calculation of the response amplitude (from the simulation) for the EEG response to the present stimulus or a different stimulus can allow creation of various auditory responses corresponding to different ANF / ANS profiles or other input parameters. Using category-based parameters, the auditory response can be further subdivided based on, for example, age, gender, etc. or other parameters. The calculated auditory responses and corresponding ANF / ANS damage profiles can be stored in a database or made available through a database.

[0399] (400) The EEG response of a subject to the present auditory stimulus (100) can be experimentally measured using EEG equipment. Processing of the EEG data allows calculation of the amplitude of the subject-specific EEG response to the stimulus.

[0400] (500) The EEG response data of the processed subject can be interpreted using the predicted simulation data (300) to assign an auditory profile to the subject. The assignment can be performed automatically by a matching algorithm (500). The assigned profile is preferably based on the most likely match between the simulated and recorded EEG response amplitudes. Based on the assigned auditory profile, the integrity of the subject's AN fibers and synapses can be determined. For example, in this figure, the subject is assigned an ANF / ANS profile with 54% HSR, 0% MSR, and 0% LSR damage profile characteristics. Since the best matching ANF / ANS profile does not return 100% ANF type in all ANF categories, the subject has a certain degree of cochlear synaptic lesions.

[0401] Example 4: Method for determining outer hair cell (OHC) damage in a subject

[0402] With respect to Example 3 described above, the method for determining the integrity of AN fibers and synapses in a subject may also be extended to determining damage to outer hair cells (OHC) in the subject. Fig.10 Methods are described which present preferred steps for determining an individual's ANF / ANS and OHC impairment profile and optionally use subject data to determine a subject-specific auditory profile.

[0403] In particular, the biophysical model (200) of the auditory periphery can be extended to also include, for example, simulated and / or experimental frequency-specific OHC lesion profiles (220). The OHC lesion profile (220) can be determined based on experimental data of frequency-specific hearing loss (e.g., from audiogram testing, otoacoustic emissions). Alternatively, the OHC lesion profile (220) can remain variable so that the matching algorithm (500) for both the AN and OHC profiles can be optimized simultaneously to find the best subject match. For example, in this figure, based on the subject's experimental EEG recording and the best match to a specific simulated auditory response to the same stimulus in a database of simulated auditory responses of multiple auditory profiles (including ANF / ANS and OHC lesions), the subject is assigned an OHC profile characterized by 50% OHC lesions. It is determined that the subject in the description has a certain degree of OHC-related hearing loss.

[0404] Example 5: Method for modifying a subject's desired auditory response to a sound

[0405] With respect to Example 4 or Example 5 described above, one or more methods for determining the integrity of a subject's ANF / ANS and / or OHC lesions can be used to modify a desired auditory response to sound for the subject. Fig.11 The method is described, presenting a flow chart of preferred steps for determining a signal processing algorithm (600) for modifying an auditory stimulus that produces a desired auditory response. The determined signal processing algorithm can be used to configure a hearing device, such as a cochlear implant or a hearing aid.

[0406] The captured (individualized) auditory response to sound (e.g., auditory EEG response, sound perception, cochlear, ANF, and brainstem processing, 400) can be used to determine a subject-specific ANF and OHC impairment auditory profile (500). The auditory profile can be included in an auditory peripheral model to simulate the auditory response to any acoustic stimulus (600). The independently simulated auditory response can be compared with the expected auditory response (700). The expected response can be experimental or simulated and can be, for example, a response from a normal hearing subject or a response with enhanced characteristics. A signal processing algorithm (800) is then included to adjust the sound stimulus in a manner such that the simulated auditory response (600) matches the expected auditory response (700). For example, the matching algorithm (800) can ultimately end with filtering, initial sharpening, compression, and / or expansion of the audio stimulus (100).

[0407] Example 6: Neural Response to Auditory Stimuli

[0408] With respect to Example 1 described above, to illustrate how the stimulus envelope shape affects the EFR sensitivity to different SNHL aspects, an additional number of simulations are provided in which the duty cycle ( FIG. 12 ) and modulation rate ( Fig.13 ).

[0409] refer to Fig. 12A , which shows that a rectangular stimulus (RAM) with a duty cycle of 25% (RAM25) at a modulation frequency of 10 Hz produces improved response amplitude compared to lower RAM stimulation %. It can be appreciated that neural responses to very short duty cycles ("click-like") produce lower amplitudes (compare the leftmost to the rightmost RAM) and are more sensitive to OHC damage (compare the bottommost to the topmost triangles). Thus, for differential diagnosis of synaptopathy, lower duty cycles are less preferred.

[0410] Additional references Fig. 12B , which shows that the insensitivity of the response amplitude to OHC injury is more pronounced at the higher 120 Hz modulation frequency. Note that for the RAM25 condition (RAM25-120 Hz modulation rate), only a minimal effect of OHC injury can be seen on the response amplitude. Accordingly, outside of the conditions shown, RAM25 provides a preferred embodiment to differentially diagnose synaptopathy. Accordingly, outside of the conditions shown, RAM25 at a 120 Hz modulation rate provides an even more preferred embodiment to differentially diagnose synaptopathy.

[0411] refer to Fig.13 , which shows that a rectangular stimulus (RAM) with a 25% duty cycle (RAM25) and a 120 Hz modulation frequency (RAM100% MD or 120 Hz data points) produces higher responses (triangles after NH) than a sinusoidal (SAM) modulator or a pitch-shifted tone (TT; or exponential modulation). Because the responses are higher, the sensitivity of the responses is increased for the same degree of synaptopathy / deafferentation applied (compare % decrease from NH triangles to filled CS triangles).

[0412] In summary, the % EFR reduction (i.e., normalized to NH EFR) was -10, -4, and +4% for the 35dB HL OHC lesion pattern (NH vs OHC triangles), whereas it was 81% for the most severe, 0L, 0M, 3H synaptopathy pattern (NH to filled CS triangles). In absolute EFR values, the RAM EFR for the same degree of synaptopathy was reduced by 0.08, 0.10, and 0.12 μV, respectively, indicating superior sensitivity of the RAM EFR criteria for diagnosis compared to SAM or TT stimulation at the same level and modulation rate.

[0413] In addition Fig.14 The figure shows that the RAM25 condition (rectangular stimulation with 25% duty cycle-120Hz modulation) induces better single-unit auditory neural responses (middle panel-NH) and ASSR / EFR waveforms (bottom panel-NH) compared to both SAM (sinusoidal stimulation) and RAM50 modulation (rectangular stimulation with 50% duty cycle and longer platform duration). At the same time, the comparative experimental data and model simulations show that the RAM25 duty cycle provides better EFR / ASSR intensity than RAM50 and SAM ( Fig.14 , bottom panel). Meanwhile, the single peak response observed under RAM25 conditions ( Fig.14 g) in the figure, the RAM50EFR waveform ( Fig.14 i) in FIG. 5 shows an undesirable bimodal response (yNH data). The bimodal response in RAM50 produces an overall smaller EFR amplitude that is less sensitive to detecting synaptopathy than the RAM25 response in the same listener.

[0414] Fig.15 This is further demonstrated in , which shows comparative experimental data and model simulations for three subgroups of listeners. RAM25 (25% duty cycle) produced improved responses over RAM50 (50% duty cycle) and SAM in terms of EFR / ASSR response amplitude. As shown in the model simulations (triangles), this results in improved sensitivity for quantifying the extent of individual synaptic pathology.

Claims

1. A system for determining the integrity of an auditory nerve synapse (ANS) and / or an auditory nerve fiber (ANF) of a subject, the system comprising: - Sound generation and stimulation devices for generating auditory stimulation s (t); - neural activity screening hardware configured to detect when the subject is exposed to auditory stimuli generated by the sound generating and stimulating device s (t), recording the subject's neural response as neural activity data; and, - A processing unit connected to the recording hardware, configured to: a) receiving exposure to the auditory stimulus s(t) neural activity data of the subject, and b) determining from the neural activity data the degree of integrity of all or a subset of the subject-specific ANS and / or ANF; Characterized in that the auditory stimulation s (t) comprising a non-sinusoidal periodic modulator having a rectangular waveform with a duty cycle d m (t) AM carrier signal c (t), wherein the duty cycle is at least 10% and at most 40%.

2. The system according to claim 1, wherein: The duty cycle is at least 15% and at most 35%.

3. The system according to claim 2, wherein: The duty cycle is at least 20% and at most 30%.

4. The system according to any one of claims 1 to 3, wherein: The rectangular waveform has a slope deviation window δ between 0 and δ max In the range of max = 0.5* d ,in, d is the duty cycle.

5. The system according to claim 1, wherein: The screening hardware includes at least two EEG electrodes and an amplifier; and is configured to record the subject's neural responses as EEG signals in the time domain (TD).

6. The system according to claim 1, wherein: The subject is a human subject, and the neural activity screening hardware is configured to record human neural responses.

7. The system according to claim 5, wherein: The processing unit is configured to transform the neural activity data into a frequency domain (FD).

8. The system according to claim 7, wherein: The processing unit is configured to transform the EEG signal into a frequency domain (FD).

9. The system according to claim 8, wherein: The processing unit is configured to s(t) The time period of the EEG signal induced applies a Fourier transform to transform the EEG signal into the frequency domain (FD).

10. The system according to claim 1, wherein: The processing unit is configured to determine the f m The response amplitude at the frequency F0 and at one or more frequencies F1 to Fn which are harmonics of the modulation frequency.

11. A system according to any one of claims 7 to 9, wherein: The processing unit is configured to - identifying at least 2 peaks in the frequency domain, comprising at least: - the modulation frequency F0 of said stimulation; and, - one or more harmonics F1 to Fn of said modulation frequency; - reconstructing a time domain waveform from said identified peaks; and, - measuring the amplitude of the reconstructed signal from the reconstructed time domain waveform; wherein the amplitude is a measure of the level of integrity of the subject's ANS and / or ANF and / or IHC.

12. The system according to claim 11, wherein: The peaks are all significant peaks above the noise floor.

13. The system according to claim 11, wherein: The time domain waveform is reconstructed from the identified peaks by inverse Fourier transform.

14. A system for generating an individualized mathematical auditory peripheral model of a subject for processing sound for a hearing device, comprising the system according to any one of claims 1 to 13; wherein, The processing unit is additionally configured to - optionally, determining the fraction of outer hair cells (OHC) damaged; and, - generating an individualized mathematical model for processing sound for a hearing device based on the integrity of the subject's ANS and / or ANF and / or IHC and optionally the damaged portion of OHC.

15. A system for modifying a subject's desired auditory response, comprising the system according to any one of claims 1 to 14; wherein: The processing unit is additionally configured to - Determine subject-specific ANS and / or ANF hearing impairment profile; - including said auditory impairment profile in a personalized mathematical auditory peripheral model to simulate the auditory response to any acoustic stimulus; - Comparison of independent simulated auditory responses with expected auditory responses; and, - comprising a signal processing algorithm to adjust the sound stimulus so that the simulated auditory response matches the desired auditory response.

16. A system for comparing the degree of hearing loss caused by ANS and / or ANF damage and / or IHC loss and / or outer hair cell (OHC) loss in a subject, comprising the system according to any one of claims 1 to 15; wherein, The processing unit is additionally configured to - quantifying the extent of ANS and / or ANF damage and / or the extent of IHC damage in said subject; - obtaining markers for subjects that are sensitive only to outer hair cell damage; to quantify the extent of outer hair cell damage in the subject; and, - Plot the extent of ANS and / or ANF and / or IHC damage and the extent of OHC damage.

17. The system of claim 16, wherein: The extent of synaptopathology and / or IHC loss in the subject is quantified.

18. A computer program, or a computer program product directly loadable into the internal memory of a computer, or a computer program product stored on a computer readable medium, or a combination of said computer program or computer program products, configured to perform a computer-implemented method for determining the integrity of an auditory nerve synapse (ANS) and / or an auditory nerve fiber (ANF) of a subject, The computer-implemented method comprises the following steps: a) receiving exposure to auditory stimuli s(t) neural activity data of the subject; and, b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANF as determined by neural activity data; Characterized in that the auditory stimulation s (t) comprising a non-sinusoidal periodic modulator having a rectangular waveform with a duty cycle d m (t) AM carrier signal c (t), wherein the duty cycle is at least 10% and at most 40%.

19. The computer program according to claim 18, wherein: Receive exposure to auditory stimuli s(t) of the subject's EEG or MEG signal.

20. A computer program according to claim 18 or 19, wherein: The duty cycle is at least 15% and at most 35% and / or The rectangular waveform has a slope deviation window δ between 0 and δ max In the range of max = 0.5* d ,in, d is the duty cycle.

21. The computer program according to claim 20, wherein: The duty cycle is at least 20% and at most 30%.

22. A computer-implemented method for determining the integrity of an auditory nerve fiber (ANS) and / or an auditory nerve synapse (ANF) of a subject, the computer-implemented method comprising the steps of: a) receiving exposure to auditory stimuli s(t) neural activity data of the subject; and, b) the degree of integrity of all or a subset of the subject-specific ANS and / or ANF as determined by neural activity data; Characterized in that the auditory stimulation s (t) comprising a non-sinusoidal periodic modulator having a rectangular waveform with a duty cycle d m (t) AM carrier signal c (t), wherein the duty cycle is at least 10% and at most 40%.

23. The computer-implemented method of claim 22, wherein: Receive exposure to auditory stimuli s(t) of the subject's EEG or MEG signal.

24. The computer-implemented method of claim 22 or 23, wherein: The duty cycle is at least 15% and at most 35%; and / or, The rectangular waveform has a slope deviation window δ between 0 and δ max In the range of max = 0.5* d ,in, d is the duty cycle.

25. The computer-implemented method of claim 24, wherein: The duty cycle is at least 20% and at most 30%.

Images