Method and apparatus for testing hearing ability

Abstract

The present invention concerns a method for the detection of distortion products of otoacoustic emissions (DPOAE), comprising the steps: (a) output of at least one first pair of primary tones each consisting of a first primary tone with frequency f.sub.1,1 and sound pressure level L.sub.1,1 and a second primary tone with frequency f.sub.2,1 and sound pressure level L.sub.2,1 with f.sub.2,1>f.sub.1,1, and (b) detecting evoked distortion products of otoacoustic emissions (DPOAE), characterized in that the first primary tone {f.sub.1,1, L.sub.1,1} is output with a time delay t.sub.lag after the second primary tone {f.sub.2,1, L.sub.2,1}.

Claims

1. A method for detecting distortion products of otoacoustic emissions (DPOAE) in a hearing organ comprising the steps of: (a) output of a first primary tone pair {f.sub.1,1, L.sub.1,1, f.sub.2,1, L.sub.2,1} each comprising a first primary tone with frequency f.sub.1,1 and sound pressure level L.sub.1,1 and a second primary tone with frequency f.sub.2,1 and sound pressure level L.sub.2,1 with f.sub.2,1>f.sub.1,1, and (b) detection of evoked distortion products of otoacoustic emissions (DPOAE), wherein the first primary tone {f.sub.1,1, L.sub.1,1} is output with a time delay tag after the second primary tone {f.sub.2,1, L.sub.2,1}, wherein at least one further primary-tone pair is presented, consisting of a first primary tone with frequency f.sub.1,n and sound pressure level L.sub.1,n and a second primary tone with frequency f.sub.2,n and sound pressure level L.sub.2,n where f.sub.2,n>f.sub.1,n, wherein the second primary tone {f.sub.2,n, L.sub.2,n} of the at least one further n-th primary-tone pair has a time delay tag after the first primary tone {f.sub.1,n, L.sub.1,n} of this primary-tone pair, wherein {f.sub.1,1, L.sub.1,1}={f.sub.1,n, L.sub.1,n} and {f.sub.2,1, L.sub.2,1}={f.sub.2,n, L.sub.2,n}, and wherein the method further comprises a step of comparing the DPOAE evoked by output of the first primary-tone pair {f.sub.1,1, L.sub.1,1, f.sub.2,1, L.sub.2,1} with the DPOAE evoked by the output of each n-th further primary-tone pair {f.sub.1,n, L.sub.1,n, f.sub.2,n, L.sub.2,n}.

2. The method according to claim 1, wherein the output of the at least one further n-th primary-tone pair {f.sub.1,n, L.sub.1,n, f.sub.2,n, L.sub.2,n} takes place before or after the output of the first primary-tone pair {f.sub.1,1, L.sub.1,1, f.sub.2,1, L.sub.2,1}.

3. The method according to claim 1, where n=2.

4. The method according to claim 1, wherein the first primary tone {f.sub.1,1, L.sub.1,1} and/or {f.sub.1,n, L.sub.1,n} (“f.sub.1-pulse”) and optionally the second primary tone {f.sub.2,1, L.sub.2,1} and/or {f.sub.2,n, L.sub.2,n} (“f.sub.2-pulse”) can be presented pulsed.

5. The method according to claim 4, wherein a pulse length of the f.sub.1-pulse of the first primary-tone pair {f.sub.1,1, L.sub.1,1} is shorter than a pulse length of the f.sub.2-pulse of the first primary-tone pair {f.sub.2,1, L.sub.2,1}, and/or a pulse length of the f.sub.2-pulse of the n-th further primary-tone pair {f.sub.2,n, L.sub.2,n} is shorter than a pulse length of the f.sub.1-pulse of the n-th further primary tone pair {f.sub.1,n, L.sub.1,n}.

6. The method according to claim 4, wherein the f.sub.1-pulse of the first primary-tone pair {f.sub.1,1, L.sub.1,1; f.sub.2,1, L.sub.2,1} is switched off before or after the end of the f.sub.2-pulse of the first primary-tone pair.

7. The method according to claim 1, wherein the time delay tag is between 10 ms and 0.1 ms or between 5 ms and 0.5 ms.

8. The method according claim 1, wherein a duration of the f.sub.1-pulse of the first and optionally each further n-th primary-tone pair and/or a duration of the f.sub.2-pulse of the first and optionally each further n-th primary-tone pair is selected to be greater than a latency of the evoked DPOAE, or at least twice, or at least three times or at least five times as long.

9. The method according to claim 1, wherein a duration of the first primary tone {f.sub.1,1, L.sub.1,1} of the first primary tone pair and/or of the second primary tone {f.sub.2,n, L.sub.2,n} of the n-th further primary-tone pair is 200 ms or less, 100 ms or less, 50 ms or less, between 40 ms to 1 ms, between 30 ms and 2 ms or between 25 ms and 5 ms.

10. The method according to claim 1, wherein a set of the first primary-tone pair {f.sub.1,1, L.sub.1,1; f.sub.2,1, L.sub.2,1} and the at least one further primary tone pair {f.sub.1,n, L.sub.1,n; f.sub.2,n, L.sub.2,n} is output in a block which is repeated several times during a measurement period.

11. The method according to claim 10, wherein in a block a beginning of a primary-tone pair follows a beginning of the primary-tone pair immediately preceding in the block with a time interval T.sub.a, where T.sub.a is >10 ms.

12. The method according to claim 10, wherein in a block second excitation frequencies f.sub.2 of two immediately successive pairs of primary tones are at least one octave apart.

13. The method according to claim 10, wherein during the measurement period measured sound pressure levels of the DPOAE are averaged for primary-tone pairs of same second excitation frequencies f.sub.2.

14. The method according to claim 10, wherein the or each block of primary-tone pairs is presented during a block time selected such that there is a time interval of 30 ms to 100 ms, or at least 70 ms, between a beginning of a first and a subsequent primary-tone pair with same excitation frequencies f.sub.2.

15. The method according to claim 10, wherein sound pressure levels of the DPOAE are measured and averaged for all second excitation frequencies f.sub.2 contained in a or each set at a second sound pressure level L.sub.2 respectively associated with an excitation frequency f.sub.2, and measurements of sound pressure levels of the DPOAE are performed at least once for new sound pressure levels L.sub.2.

16. The method according to claim 1, wherein a duration of the first and second primary tones of each pair of primary tones is between 2 ms and 20 ms.

17. The method according to claim 1, wherein at a beginning of measurements it is checked whether a frequency f.sub.dp of one of the DPOAEs interferes with a spontaneous emission (SOAE).

18. The method according to claim 1, further comprising automatically determining an individual function of a DPOAE level map having p.sub.dp,I=f(L.sub.1, L.sub.2) to determine an optimal DPOAE excitation level: Reading a model function p.sub.dp,M=f(L.sub.1, L.sub.2) with model parameters of a DPOAE level chart, based on a number of N DPOAE measurements of an excitation frequency pair f.sub.1, f.sub.2 each with different level pairs {L.sub.1.sup.(1 . . . N), L.sub.2.sup.(1 . . . N)} in a population (p) of subjects with normal hearing, into a working memory of a computer unit, where N≥40 and p≥2, automatic presentation of n different level pairs {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . n)} of an excitation frequency pair f.sub.1, f.sub.2 via sound output means to an individual and detecting a corresponding DPOAE of the individual via sound recording means, wherein at least a first level pair {L.sub.1.sup.(1), L.sub.2.sup.(1)} is predefined and where n<<N, iterative fitting of a model level-map function p.sub.dp,M=f(L.sub.1, L.sub.2) to measured n DPOAE until an individual function is obtained p.sub.dp,I=f(L.sub.1, L.sub.2) with individual parameters of a DPOAE level map of the individual by the computer unit, output of the individual function p.sub.dp,I=f(L.sub.1, L.sub.2) and/or their individual parameters on an output device of the computer unit.

19. The method according to claim 18, wherein the model function has an approximately linearly increasing ridge (73) which is approximately linearly related to {L.sub.1.sup.(G), L.sub.2.sup.(G)} level pairs, at least half of measured level pairs {L.sub.1, L.sub.2} being at least 5 dB on both sides away from the group of signals evoked by the level pairs {L.sub.1.sup.(G), L.sub.2.sup.(G)} associated with the ridge (73).

Description

FIGURES

(1) Further advantages result from the following description of the attached figures.

(2) FIG. 1 Comparison of the amplitudes of the envelopes of the 2f.sub.1-f.sub.2-DPOAE after excitation by PTA-I (grey line) and PTA-II (black line).

(3) FIG. 2 Transduction curve of the OHC and its effect on the receptor potential for DPOAE excitation sounds of different strengths.

(4) FIG. 3 Measurement of the SPDPOAE or the resting potential shift with an excitation with extended f.sub.1-pulse (pulse length 20 ms).

(5) FIG. 4 State-of-the-art PTA-I excitation scheme in which the f.sub.2-pulse (black, solid line) is presented during the continuous f.sub.1-pulse (dotted line).

(6) FIG. 5 A system for automatically determining an individual function of a DPOAE level map.

(7) FIG. 6 A model function whose three-dimensional graph corresponds to a model level map.

(8) FIG. 7 Process steps for the automatic determination of an individual function of a DPOAE level map.

(9) FIG. 8 Detailing of the procedural step according to the marking IV in FIG. 7.

(10) FIG. 9 Comparison of the amplitudes of the envelopes of the 2f.sub.1-f.sub.2-DPOAE after excitation by PTA-II (left) and PTA-I (right).

(11) FIG. 10 Curves superimposed from FIG. 9.

(12) It goes without saying that the features mentioned above and the features to be explained below can be used not only in the combination indicated, but also in other combinations or in a unique position, without leaving the scope of this invention.

(13) Examples of how the invention may be executed are explained in more detail in the following description, with reference to the attached drawings.

(14) FIG. 1 shows the envelopes of the 2f.sub.1-f.sub.2-DPOAE after excitation by PTA-I (grey line) and PTA-II (black line). PTA-I and PTA-II lead to different amplitudes. If the f.sub.1-tone is switched on later than the f.sub.2-tone (here: 5 ms later) (PTA-II), a higher amplitude of the envelopes of the 2f.sub.1-f.sub.2-DPOAE (black line) is obtained than if the f.sub.2-tone is switched on 5 ms later than the f.sub.1-tone (grey line) (PTA-I).

(15) In this example, the response curves of the DPOAE show only the response of the nonlinear source, while the response of the coherent reflection source in this subject and at this frequency is too low to be detected. The response of the nonlinear source is relevant because its amplitude is normally greater and, unlike the coherent reflection source, it is not dependent on any additional process (roughness of the impedance function).

(16) In this example, the response of the nonlinear source (i.e., the envelope of the 2f.sub.1-f.sub.2-DPOAE time signal) is 3 dB greater for PTA-II than for PTA-I. This difference is large enough to be detected with certainty.

(17) FIG. 2 shows the transduction curve of the OHC (201) and its effect on the receptor potential for DPOAE excitation sounds of different levels. Input is the deflection of the stereocilia, output is the receptor potential, which controls the force coupling of the cell into the vibrations of the organ and thus the amplification process. Since the ion channels can only be completely closed or completely opened, the transduction curve at high input values reaches a maximum or minimum value, in the normalization selected here the value 1 or 0. Furthermore, two input signals ((2a) and (2b)) are shown, which correspond to an excitation with two primary tones. The resulting beating signal is clearly visible. Firstly, it is excited with a relatively low amplitude ((2a), solid line), which responds to a still somewhat linear part of the transduction curve. Secondly, an excitation with a significantly higher amplitude is shown ((2b), dotted line), which is clearly processed non-linearly. The resulting output signals (3a, 3b) are also shown. In the case of small-signal excitation ((3a) solid, strong line), the system response cannot be visually distinguished from the input signal. (204) represents the mean value, i.e. the DC component of the signal. With stronger excitation (5) the influence of non-linearity is clearly visible. Due to the asymmetry of the transduction curve, the negative half-waves of the potential fluctuations are strongly cut off, while the positive half-waves still show a somewhat sinusoidal course. As a result, the mean value, i.e. the DC value of the potential fluctuations, shifts.

(18) In a healthy case, the operating point shifts from the steepest point to a less steep one. This must lead to some reduction in the gain of the OHC. Preferably, the shift of the resting or DC potential takes place largely frequency-independently within 1-2 ms.

(19) FIG. 3 shows the measurement of the SPDPOAE or the resting potential shift with an excitation with extended f.sub.1 short pulse. If the second primary tone {f.sub.2, L.sub.2} is switched on first and the extended f.sub.1 short pulse delayed by 5 ms, in this example the DPOAE reaches a maximum of 113 μPa at the beginning, then drops by more than 3 dB within 2 ms, and then rises again. This subsequent increase can in principle be caused by the influence of the second source, but also by slower control processes of the difference between intra- and extracellular potentials. (301) Excitation by PTA-II, (302) Excitation by PTA-I.

(20) Pulse Interlacing Method

(21) The inventive method can also be combined with the pulse interlacing method known from WO 2015/192969 A1 (“time-limited multifrequency method”).

(22) Pulsed DPOAE have the disadvantage compared to continuous DPOAE that the measurement with one frequency sound-pressure level combination generally has a low duty factor and thus a correspondingly lower signal-to-noise ratio within the same measurement time. In the procedure described in WO 2015/192969 A1, however, this disadvantage is considerably reduced by interlacing several measurements in time-frequency space, i.e. presenting them alternately with a time delay. Within a block, for example, seven frequencies can be stimulated and analyzed with a time delay. The measurement procedure is thus accelerated by parallel and adaptive stimulation and analysis steps, and allows the measurement of growth functions for 5-7 frequencies f.sub.2 in typically 2-2.5 min.

(23) In the method described in WO 2015/192969 A1, at least two different pulse pairs (each with a pulsed first primary tone {f.sub.1, L.sub.1}) and a pulsed second primary tone {f.sub.2, L.sub.2}) with different excitation frequencies f.sub.2 (and consequently different excitation frequencies f.sub.1) are presented in a block which is repeated several times during a measurement period. In a block, a first primary tone pair presented in pulsed mode is followed by a first primary tone pair with {f.sub.2,1, f.sub.1,1} a second pair of primary tones with different frequencies {f.sub.2,2, f.sub.1,2} and if necessary further ones with {f.sub.2,n, f.sub.1,n} wherein the frequency ratio is preferably always close to f.sub.2,n/f.sub.1,n=1.2 is held.

(24) WO 2015/192969 A1 provides that the pulsed primary tones of each pair are presented according to the PTA-I excitation scheme (with f.sub.2-short pulse). The f.sub.1-pulse of a pair of pulses starts before the f.sub.2-pulse and ends after the f.sub.2-pulse has ended, i.e. the f.sub.1-pulse is longer than the f.sub.2-pulse of a pair of pulses.

(25) The pulse interlacing method according to the invention provides for the presentation of at least one primary-tone pair according to the PTA-II excitation scheme (with f.sub.1-short pulse). The other primary-tone pairs in a block can be presented either according to PTA-II (pure PTA-II pulse interlacing method) or according to PTA-I (combined PTA-I/PTA-II pulse interlacing method).

(26) In the pulse interlacing method according to the invention, the beginning of a (PTA-I or PTA-II) pulse pair follows in a block, preferably with a time interval T (T.sub.SLOT or T.sub.S), the beginning of the (PTA-I or PTA-II) pulse pair immediately preceding in the block, where T.sub.S generally corresponds at least to the length of the preceding pulse. Preferably T.sub.S is >10 ms. This measurement time reserved for a (PTA-I or PTA-II) pulse pair in a block is referred to below as the slot. It should be noted that slots do not overlap, but one slot follows the other when the previous slot is finished.

(27) It is advantageous to present the second (PTA-I or PTA-II) pair of pulses in this way only after the DPOAE evoked by the first (PTA-I or PTA-II) pair of pulses has been sufficiently decayed (to approx. 1 to 10% of the output value) so that there are no noticeable interferences when measuring the sound pressure levels (L.sub.dp) of the individual DPOAE, which have only a very low sound pressure level compared to the sound pressure levels L.sub.1 and L.sub.2 of the f.sub.1 and f.sub.2-pulses. In addition, the increased time interval between the presentation of (PTA-I or PTA-II) pulse pairs with the same excitation frequencies f.sub.1 and f.sub.2 allows sufficient recovery time for the complete decay of the DPOAE triggered in the previous measurement block.

(28) Preferably, the or each block of (PTA-I or PTA-II) pulse pairs can be presented during a block time T.sub.B. The block time T.sub.B is defined as the sum of the slot lengths T.sub.S of a block. Preferably, T.sub.B is chosen so that a (PTA-I or PTA-II) pulse has sufficiently approached its steady state when repeated. A time interval of 30 ms to 100 ms, preferably of at least 70 ms, preferably lies between the beginning of a first and a subsequent pulse pair with the same excitation frequency f.sub.2.

(29) Preferably, the f.sub.2-excitation frequencies of two consecutive (PTA-I or PTA-II) pulse pairs in a block are at least one octave apart. The choice of such a frequency spacing of at least one octave between the f.sub.2-excitation frequencies advantageously ensures that the frequencies f.sub.dp of the evoked DPOAE are sufficiently far apart so that there is preferably no noticeable interference when measuring the individual DPOAE.

(30) For the excitation frequencies f.sub.1 and f.sub.2, the above-mentioned applies, for example a preferred frequency ratio of f.sub.1/f.sub.2=1.2. An example of two different pulse pairs of a block are a first pulse pair with an excitation frequency f.sub.2 of 1.5 kHz and an excitation frequency f.sub.1 of 1.25 kHz and a second pulse pair with an excitation frequency f.sub.2 of 4 kHz and an excitation frequency f.sub.1 of 3.33 kHz.

(31) A preferred set (i.e. panel or frequency time pattern of excitation frequencies in a block.) of f.sub.2 excitation frequencies in a block consists of the excitation frequencies f.sub.2=1 kHz, f.sub.2=3 kHz, f.sub.2=1.5 kHz, f.sub.2=6 kHz. These f.sub.2-excitation frequencies are repeatedly presented in a block in this order. Another preferred set (panel) of f.sub.2-excitation frequencies in a block consists of the excitation frequencies f.sub.2=2 kHz, f.sub.2=4 kHz, f.sub.2=1.5 kHz, f.sub.2=3 kHz.

(32) Depending on the frequency, the (PTA-I) f.sub.1-pulse can preferably be switched on 3-10 ms earlier and switched off 3-10 ms later than the (PTA-I) f.sub.2-pulse, so that the (PTA-I) f.sub.1-pulse briefly reaches a steady state during the presentation of the (PTA-I) f.sub.2-pulse.

(33) However, it is also possible to work with two equally or similarly short (PTA-I) f.sub.1 and f.sub.2-pulses, which are so time-shifted that both excitations occur simultaneously at the diagnostically most valuable characteristic place of the (PTA-I) f.sub.2-pulse in the cochlea, in order to save as much time as possible. Then the (PTA-I) f.sub.1-pulse is switched on about 0.1-3 ms later, since its propagation time to the more basal (direction of the foot plate) place of the f.sub.2-pulse is shorter than for the (PTA-I) f.sub.2-pulse. If this setting is selected optimally, there is no effect from the afferent-efferent feedback loop of the medial olivocochlear reflex.

(34) The duration (length) of the (PTA-I) f.sub.1 and f.sub.2-pulse in a pulse pair can preferably be 2 to 20 ms. With regard to the pulse length, the above applies with regard to PTA-II.

(35) In the pulse interlacing procedure of the present invention, the sequence of the pulse pairs and the time interval between two successive (PTA-I and/or PTA-II) pulse pairs, i.e. the slot time (T.sub.SLOT), can be constant in a block. With this block rigid procedure, as many blocks are measured and averaged until the desired SNR is reached for each excitation frequency in the set (panel). Alternatively, if the desired SNR for one excitation frequency f.sub.2 is reached, the remaining pulse pairs for this excitation frequency f.sub.2 and consequently their averages can be skipped. In addition, the remaining pulse pairs can be used to continue measuring in shortened blocks, i.e. with fewer slots, which further shortens the measuring time.

(36) Furthermore, two sets with at least partially different (PTA-I and/or PTA-II) pulse pairs with respect to the second excitation frequency f.sub.2 can be selected in the pulse interlacing method of the present invention according to the invention, wherein the blocks of the individual sets are presented one after the other in time and the DPOAE are measured and averaged. The blocks are therefore processed one after the other. In this block-flexible method with fixed pulse arrangement, for example, seven (PTA-I and/or PTA-II) pulse pairs with different excitation frequencies f.sub.2 are arranged in such a way that, according to general experience, they are all assigned approximately the averaging time required to achieve a certain signal-to-noise ratio.

(37) Preferably, according to the invention, several sets are presented one after the other, to which (PTA-I and/or PTA-II) pulse pairs are distributed in such a way that (PTA-I and/or PTA-II) pulse pairs with low excitation frequencies f.sub.2 (and thus also low excitation frequencies f.sub.1) occur in several sets.

(38) Using the pulse interlacing method according to the invention, it is also possible to continuously check for each (PTA-I and/or PTA-II) pulse pair whether a desired SNR is achieved. For the further measurement the (PTA-I and/or PTA-II) pulse pairs can be discarded for this excitation frequency f.sub.2 and the remaining (PTA-I and/or PTA-II) pulse pairs can be redistributed to the blocks if necessary.

(39) In this block-flexible method with free (PTA-I and/or PTA-II) pulse pair arrangement, the length of the measurements for the individual (PTA-I and/or PTA-II) pulse pairs is no longer fixed relative to each other. Instead, for each (PTA-I and/or PTA-II) pulse pair, the SNR is continuously checked to see if it has been reached; as soon as this is the case, the SNR is checked to see if another (PTA-I and/or PTA-II) pulse-pair measurement is incomplete. In this way, the completed (PTA-I and/or PTA-II) pulse-pair measurements are successively eliminated from the measurement, and only the remaining (PTA-I and/or PTA-II) pulse pairs are further presented. On the one hand, it is checked whether the octave spacing between two consecutive (PTA-I and/or PTA-II) pulse pairs is maintained. If this is no longer the case, (PTA-I/PTA-II) pulse pairs may no longer be processed in the blocks to which they were originally assigned, but in other (newly defined) blocks. In addition, it is checked whether the required time interval T (corresponding to the time for a slot (T.sub.SLOT) plus the required decay time (T.sub.decay)) between (PTA-I and/or PTA-II) pulse pairs with the same excitation frequency f.sub.2 is observed.

(40) As described above with regard to PTA-II, the sound pressure levels of the (PTA-I and/or PTA-II) pulse pairs within a block are preferably similar.

(41) The DPOAE is preferably measured and averaged for all excitation frequencies f.sub.2 contained in the set or sets at a sound pressure level L.sub.2 assigned in each case to the excitation frequency, and then at least one new measurement is carried out at new sound pressure levels L.sub.2, the new sound pressure level L.sub.2 for the respective new measurement being determined preferably in a threshold value approximation method from the measured DPOAE for each excitation frequency f.sub.2. This procedure is repeated until for each excitation frequency f.sub.2 a growth curve can be determined from measured values of the sound pressure levels of the DPOAE for 3 to 4 different sound pressure levels L.sub.2, from which the respective threshold values are then determined. A more detailed description of this procedure can be found in WO 2015/192969 A1.

(42) Level Map Method

(43) The procedure according to the invention can be combined with the level map method known from PCT/EP2017/000334. A “DPOAE level map” (DPOAE level map, see Shera and Guinan, J Acoust Soc Am. 2007 February; 121(2):1003-16) and Martin et al. J. Acoust. Soc. Am. 127 5, p. 2955-2972) denotes the amplitude of a level (here the 2f.sub.1-f.sub.2 distortion product) as a function of the primary tone levels.

(44) The procedure described in PCT/EP/2017/000334 is used to automatically determine an individual function of a DPOAE level map. This is preferably done to avoid errors in the extrapolation of growth functions, which may occur in other state-of-the-art methods for measuring the distortion product threshold L.sub.edpt due to their principle. The procedure described in PCT/EP2017/000334 can be used for conventional (i.e. quasi continuous) measured (PTA-II excited) DPOAE. It can also be combined with a pulsed (PTA-II excited) DPOAE process as described in DE 102014108663 A1. In particular, the procedure described in PCT/EP2017/000334 can be combined with the combined pulse interlacing described above.

(45) In particular, the procedures of the present invention can include the following steps for automatically determining an individual function of a DPOAE level chart with p.sub.dp,I=f(L.sub.1, L.sub.2) of a human or animal hearing: Reading a model function p.sub.dp,M=f(L.sub.1, L.sub.2) with model parameters of a DPOAE level map based on a number of N DPOAE measurements of an excitation frequency pair {f.sub.1, f.sub.2} each with different level pairs {L.sub.1.sup.(1 . . . N), L.sub.2.sup.(1 . . . N)} in a population (p) of normal hearing persons into a main memory of a computer unit, where N is ≥40 and p≥2, automatic presentation of n different level pairs {L.sub.1.sup.(1 . . . N), L.sub.2.sup.(1 . . . N)} of an excitation frequency pair {f.sub.1, f.sub.2} via sound output means to an individual and detecting the corresponding DPOAE of the individual via sound recording means, wherein at least the first level pair {L.sub.1.sup.(1), L.sub.2.sup.(1)} is predefined and where n is <<N, iterative adjustment of the model function p.sub.dp,M=f(L.sub.1, L.sub.2) to the measured n DPOAE until an individual function is obtained p.sub.dp,I=f(L.sub.1, L.sub.2) with individual parameters of a DPOAE level chart of the individual by the computer unit, and Output of the individual function p.sub.dp,I=f(L.sub.1, L.sub.2) and/or their individual parameters at an output unit of the computer unit.

(46) With iterative fitting (curve fitting), the model function is fitted to experimentally determined measured values. For this purpose, parameters of this function are changed with a suitable algorithm until the deviation between the measured values and the stepwise changed function according to an optimality criterion is minimal (e.g. minimization of the quadratic error). Algorithms for such an iterative adaptation are known to the expert from the state-of-the-art (e.g. Isqcurvefit or Isqnonlin).

(47) The first level pair {L.sub.1.sup.(1), L.sub.2.sup.(1)} can set a level L.sub.1 of 67±10 dB and a level of L.sub.2 of 57±10 dB. These levels of L.sub.1 and from L.sub.2 have proven to be particularly favourable initial levels. For normal hearing people, these excitation levels are still in the range up to which the level map increases approximately linearly, and even with hearing losses of up to approx. 40 dB, a DPOAE can still be measured at these levels. Thus, in most cases, values are obtained that are valid for the recording of the level map.

(48) The model function defines a linearly rising ridge, to which {L.sub.1.sup.(G), L.sub.2.sup.(G)} level pairs are assigned with linearly interrelation (where “G” is the index for “assigned to the ridge”). Preferably, at least half of the measured level pairs {L.sub.1.sup.(i), L.sub.2.sup.(i)} can be positioned by at least 5 dB to both sides off the ridge assigned to the {L.sub.1.sup.(G), L.sub.2.sup.(G)} level pairs (where “i” is the index of the measurement from 1 to n).

(49) The different level pairs {L.sub.1.sup.(i), L.sub.2.sup.(i)} are preferably presented in a sequence that is identical for each individual. This greatly simplified and standardized (rigid) procedure makes the approximation to the individual function of a level map somewhat less accurate, but this procedure is very fast in execution.

(50) It can also be advantageous if the pre-defined, different level pairs (L.sub.1, L.sub.2) are presented in a sequence comprising a number of k subsequences whose level pairs are {L.sub.1, L.sub.2} substantially transverse to the level pairs linearly linked to each other and associated with the ridge {L.sub.1.sup.(G), L.sub.2.sup.(G)} lie. By switching on the subsequences, the ridge can be scanned in several places, which increases the accuracy of determining the individual function of the DPOAE level chart.

(51) Preferably, n≥5 and ≤12, preferably 6≤n≤8. Due to the small number of planned measurements, a short measurement duration is achieved while the individual function of the DPOAE level map is recorded with high quality.

(52) The number of subsequences k≥2 and ≤5 is advantageous, whereby a good sampling of the ridge of the DPOAE level map is achieved.

(53) It is also advantageous if the level pairs {L.sub.1.sup.(2 . . . n.sup.l.sup.), L.sub.2.sup.(2 . . . n.sup.k.sup.)} of the subsequence with n.sub.k measurements following the first predefined level pair {L.sub.1.sup.(1) L.sub.2.sup.(1)} are determined via a function {L.sub.1.sup.(i), L.sub.2.sup.(i)}={L.sub.1.sup.(i−1)+μ.Math.ΔL.sub.1, L.sub.2.sup.(i−1)+μ.Math.ΔL.sub.2} from the respective previous level pair L.sub.1.sup.(i−1), L.sub.2.sup.(i−1), where μ=±1, in particular +1, and ΔL.sub.1, ΔL.sub.2 denotes a level distance of two successive level pairs and values of ΔL.sub.1=4 to 14 dB, preferably from 6 to 10 dB, and ΔL.sub.2=0 up to −2.78 dB, preferably ΔL.sub.2=−1.52 to −2.78 dB. The factor μ defines the search direction (towards smaller or larger L.sub.1 measurement levels) across the ridge.

(54) Preferably, if the first pair of levels is {L.sub.1.sup.(1), l.sub.2.sup.(1)} and the second pair of levels {L.sub.1.sup.(2), L.sub.2.sup.(2)} generates two DPOAE with p.sub.dp,I.sup.(1 . . . 2) which each have a signal-to-noise ratio of >=4 dB, preferably >=10 dB, the level of a subsequent third pair of levels {L.sub.1.sup.(3), L.sub.2.sup.(3)} is set at least by ΔL.sub.1≥4 dB differently than the level of the previous level pair {L.sub.1.sup.(2), L.sub.2.sup.(2)} if p.sub.dp,I.sup.(2)−p.sub.dp,I.sup.(1)>0, and, on the other hand, the level of a subsequent pair of levels {L.sub.1.sup.(3), L.sub.2.sup.(3)} at least by ΔL.sub.1≤−4 dB set differently than the level of the first level pair {L.sub.1.sup.(1), L.sub.2.sup.(1)} if p.sub.dp,I.sup.(2)−p.sub.dp,I.sup.(1)≤0. This procedure ensures that at least one point is measured to the left and one point to the right of the ridge and, in between, a point near the ridge.

(55) Preferably, if the first pair of levels {L.sub.1.sup.(1), L.sub.2.sup.(1)} generates no DPOAE p.sub.dp,I.sup.(1) which have a signal-to-noise ratio of >=4 dB, preferably >=10 dB, the search is continued in the same direction until either the maximum or minimum excitation level L.sub.1.sup.(i) or a group of three valid DPOAE with p.sub.dp,I.sup.(i . . . i+2) which each have a signal-to-noise ratio of >=4 dB, preferably >=10 dB. In contrast to a rigid procedure, this ensures that the ridge is also found if it is positioned clearly apart from the position to be expected for normal hearing, as may be the case with sound-conduction loss, for example.

(56) Preferably, if in the first subsequence after measurement at i excitation level pairs no group of three valid DPOAE is produced which each have a signal-to-noise ratio of >=4 dB, preferably >=10 dB, a further subsequence with a higher level pair {L.sub.1.sup.(i+1), L.sub.2.sup.(i+1)} is started, whereby the start level pair for the new subsequence is set to L.sub.2.sup.(i+3)=L.sub.2.sup.(1)+20±10 dB, L.sub.1.sup.(i+3)=L.sub.1.sup.(1)+20±10 dB is set. The level is preferably limited to the maximum technically achievable or reasonable level. This maximum level can be e.g. 75-85 dB SPL sound pressure. By this procedure, individual level maps and/or their function can still be determined, even if they are strongly deviating from the average.

(57) Preferably, after the acquisition of the DPOAE of at least 3 level pairs {L.sub.1.sup.(1 . . . 3), L.sub.2.sup.(1 . . . 3)} which are preferably belonging to a subsequence, from these 3 level pairs {L.sub.1.sup.(1 . . . 3), L.sub.2.sup.(1 . . . 3)} the position of the ridge {L.sub.1.sup.(G), L.sub.2.sup.(G)} is determined along the line formed by the 3 level pairs, and then a fourth level pair {L.sub.1.sup.(4), L.sub.2.sup.(4)} is acquired lying at a given distance downhill near the ridge, where the group mean of the ridge direction, φ is used to estimate its position and wherein, on the basis of the level pairs presented from the four {L.sub.1.sup.(1 . . . 4), L.sub.2.sup.(1 . . . 4)} a gradient m of the linear ridge of the level map is determined for a certain DPOAE.

(58) Preferably, if in the first or second subsequence a group of three valid DPOAE is recorded with p.sub.dp,I.sup.(i−2 . . . i) having a signal-to-noise ratio of >=4 dB each, preferably >=10 dB, by automatically adapting a suitable calculation function to the corresponding DPOAE p.sub.dp,I.sup.(i . . . i−2) the level pair below the ridge {L.sub.1.sup.(G), L.sub.2.sup.(G)}={L.sub.1.sup.(i−2)+ε.Math.ΔL.sub.1, L.sub.2.sup.(i−2)+ε.Math.ΔL.sub.2} is determined where ε is calculated so that p.sub.dp,I(L.sub.1.sup.(G), L.sub.2.sup.(G)) forms a maximum, and from there a fourth pair of levels {L.sub.1.sup.(i+1), L.sub.2.sup.(i+1)} is presented, with a function {L.sub.1.sup.(i+1), L.sub.2.sup.(i+1)}={L.sub.1.sup.(i)+ΔL.sub.1, L.sub.2.sup.(i)+ΔL.sub.2} where ΔL.sub.2=−15±10 dB, and the level pair is preferably set on the projection of the expected ridge on the L.sub.1, L.sub.2-plane, i.e. with ΔL.sub.1/ΔL.sub.2≈0.51±0.15 and where, based on the level pairs presented from the four L.sub.1.sup.(i−2 . . . i+1), L.sub.2.sup.(i−2 . . . i+1) the slope m of the approximately linear ridge of the level map is determined.

(59) On the basis of the determined slope m of the linear ridge of the level map, at least two, preferably three, further level pairs L.sub.1.sup.(i+1 . . . i+3), L.sub.2.sup.(i+1 . . . i+3) are automatically defined, whose excitation levels are grouped in a subsequence, and which are determined on the basis of the already known position and slope of the ridge in such a way that valid DPOAE can be expected within a measurement time of t.sub.m≤40 s by adapting a model function to the four preferably already valid measured DPOAE, and then determining the last two or three pairs of levels in the model function in such a way that the expected DPOAE levels are preferably measured at p.sub.DP,I.sup.(i+1 . . . i+3), p.sub.DP,I.sup.(i+1 . . . i+3)≥10 μPa.

(60) Preferably the level pairs {L.sub.1.sup.(1−n), L.sub.2.sup.(1−n)} are presented pulsed with a duration T.sub.D of 2 to 40 ms. By using such a pulsed presentation, the influence of the two source contributions of a DPOAE can be suppressed or separated.

(61) Preferably, the level pairs {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . n)} are presented in blocks of several subsequent pulsed level pairs {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . n)}, where level pairs {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . m)} following directly each other are presented with different excitation frequencies {f.sub.2, f.sub.1}. In one block, a first pulsed level pair with {f.sub.2,1, f.sub.1,1} is followed by a second pair of levels with different frequencies {f.sub.2,2, f.sub.1,2} and if necessary further ones with {f.sub.2,m, f.sub.1,m}, wherein the frequency ratio is always close to f.sub.2,m/f.sub.1,m=1.2. Several blocks with time-frequency interlaced pulse pairs can be averaged before evaluation takes place. This measure makes it possible to use the time in which the pulse response to a presentation decays at one frequency pair to measure at another frequency, thus reducing the measurement time compared to a purely sequential approach with regard to the desired measurement frequencies.

(62) The individual function of a DPOAE level map and its parameters determined by the computer unit are stored in a non-volatile memory in one processing step. The raw data determined by the computer unit can also be stored in the non-volatile memory. The stored data can be used by the computer unit for the continuous extension of the data set underlying the model function of a level map.

(63) FIG. 5 shows a system for the automatic determination of an individual function of a DPOAE level map of a human or animal hearing in a possible configuration. System 501 includes a probe unit 20 that can be positioned on one ear, in particular an OAE probe, and a computer unit 10. The probe unit has a probe tip 24 that can be inserted into the ear canal of one ear. The probe unit 20 contains a sound recording device 23, e.g. a microphone, which is set up to record sounds coming from the auditory canal. In the probe unit 20 a first and a second sound output means 21 and 22 are further provided which function as f.sub.1 sound emitters (sound output means 21) and as f.sub.2-sound emitters (sound output means 22). The sound output devices 21, 22 can be designed as loudspeakers, for example. It is also possible to provide only one sound output medium or only one loudspeaker, which is set up for simultaneous playback of two tones f.sub.1, f.sub.2 and in particular has a highly linear characteristic. The probe unit 20 is, for example, connected via a cable connection 502 to the control unit, which contains the computer unit 10. Cable connection 2 preferably contains shielded cables 503, 504, 505 via which the sound output devices 21, 22 and the sound recording device 23 are connected to an AD/DA converter unit 12 of the control unit. The AD/DA converter unit 12 is in turn connected to the computer unit 10 via at least one line 6 for bi-directional data exchange. As an alternative to cable connection 2, the probe unit 20 could also communicate wirelessly with the control unit or with the computer unit 10. The wireless connection could be, for example, a Bluetooth® radio link or another suitable radio connection that preferably has a short range.

(64) The computer unit 10 has a working memory of 15 and a non-volatile memory of 16 in which a model function p.sub.dp,M=f(L.sub.1, L.sub.2) for a model level map of a human or animal auditory system and the parameters associated with that model function. The non-volatile memory 16 also contains the instructions for carrying out a procedure described here. System 1 also has an output device 11 or a display unit, such as a display, monitor or the like, through which a determined individual function of a DPOAE level map of a human or animal hearing and its parameters can be output by System 1 and made accessible to a user. The output device 11 can also be implemented in the form of an interface via which an external output device, such as a printer or monitor, can be connected to the system.

(65) To perform an automatic measurement operation to create an individual function of a DPOAE level map of a human or animal ear, the probe unit 23 is inserted into the ear canal 31 of an ear 30 (indicated in FIG. 5) in the direction of arrow 40. The procedure is explained below with reference to FIGS. 7 and 8.

(66) First, however, FIG. 6 shows an example of a model function whose three-dimensional graph 70 corresponds to a model level map. The model function presented as an example is based on the measured data from p≥2, available p=6, normal hearing individuals with N≥40, available N=47, measured different level pairs {L.sub.1.sup.(1 . . . N), L.sub.2.sup.(1 . . . N)} at excitation frequencies f.sub.2=2 kHz and f.sub.1=1.67 kHz. The excitation frequencies f.sub.2 and f.sub.1 of a level pair {L.sub.1, L.sub.2} are preferably linked via a frequency ratio f.sub.2/f.sub.1=1.2. A certain distortion product was evaluated, which preferably lies at the frequency f.sub.dp=2f.sub.1−f.sub.2. The superscript indices in brackets indicate the 1st to N-th measuring point.

(67) The model function defines an approximately linearly increasing ridge 73, which is linked to approximately linearly related level pairs {L.sub.1.sup.(G), L.sub.2.sup.(G)}. Lines across the ridge can be defined by the relationship L.sub.2+aL.sub.1=C where C is any constant, and where a is the slope parameter of the projection of the ridge onto the {L.sub.1, L.sub.2}-plane. In a mathematical sense, the position of the ridge is defined by a series of gradient vectors of the scalar field formed by the DPOAE, with all other field lines formed by gradient vectors running towards this ridge and swiveling in. Into the {L.sub.1, L.sub.2}-plane 71 depicted below and being shifted by p.sub.dp=100 μPa for the sake of clarity, the transformed {L.sub.1′, L.sub.2′}-coordinate system 72 as well as contour lines of the level map at 20 μPa intervals are drawn. The coordinate system 72 is generated by shifting the origin to the {L.sub.1,edpt, L.sub.2,edpt} and by rotation by arc tan (a). The L.sub.2′-axis corresponds to the projection of the ridge of the level map on the {L.sub.1, L.sub.2}-plane. The L.sub.1′-axis orthogonally cuts the model hill fitted to the level map. This section through the hill transverse to the ridge is approximated by a 2-nd order parabola, the spread of which is given by a parameter c, and which enters the following equation:
L.sub.dp′=−c(L.sub.1′).sup.2+L.sub.dp′.sup.(G)
with
L.sub.dp′.sup.(G)=30 log.sub.10(m(L.sub.2′))
L.sub.dp′ and L.sub.dp′.sup.(G) is the level of any DPOAE or a DPOAE on the ridge and m is the slope of the ridge along the L.sub.2′-axis.

(68) The {L.sub.1′, L.sub.2′}-coordinate system is located in the coordinate system defined by the known coordinate system {L.sub.1, L.sub.2} of the primary tone level. The previously mentioned coordinate transformation can be expressed as follows:
L.sub.1′=(L.sub.1−L.sub.1,edpt)cos(φ)−(L.sub.2−L.sub.2,ept)sin(φ)
L.sub.2′=(L.sub.1−L.sub.1,edpt)sin(φ)+(L.sub.2−L.sub.2,ept)cos(φ)

(69) Here, the projection of the ridge of the L.sub.dp-hill onto the {L.sub.1, L.sub.2}-plane corresponds to the L.sub.2′-axis. Furthermore, at the point {L.sub.2,edpt, L.sub.1,edpt} the ridge of the L.sub.dp-hill intersects the {L.sub.1, L.sub.2}-plane, and φ is the angle between the L.sub.2-axis and the projection of the ridge of the L.sub.dp-hill onto the {L.sub.1, L.sub.2}-plane, given by the already mentioned L.sub.2′-axis. The angle φ is therefore the angle by which the L.sub.2′-axis is rotated with respect to the L.sub.2-axis. The base of the ridge can be interpreted in a broader sense as equivalent but not identical to the estimated distorsion product level (edpt) as known from [P. Boege and T. Janssen., J. Acoust. Soc. Am., 111(4): 1810-1818, 2002].

(70) For positive L.sub.dp values, the model function for the level map is valid and can be described by five free parameters: a; b; c; L.sub.2,edpt′; m. To be able to calculate this function from measured values, at least 5 DPOAE are required.

(71) The method is based on the adaptation of the three-dimensional model function to a coarsely sampled three-dimensional DPOAE level map with preferably at least 5 measurements. In a first execution example of the procedure for the automatic determination of an individual function of a DPOAE level map with p.sub.dp=f(L.sub.1, L.sub.2) of a human or animal hearing, the ear of an individual is presented with n predefined, e.g. n=6 predefined, excitation level pairs from the system shown in FIG. 5 {L.sub.1, L.sub.2} 51, 52, 53, 54, 55, 56. These six predefined excitation level pairs {L.sub.1, L.sub.2} 51, 52, 53, 54, 55, 56 are shown as examples in FIG. 6 of the {L.sub.1, L.sub.2}-level being selected. From these 6 predefined excitation level pairs {L.sub.1, L.sub.2} 51, 52, 53, 54, 55, 56, presented in two subsequences 57, 58, DPOAE are then determined, which are used via the procedure for determining the individual function of a DPOAE level map.

(72) According to FIG. 7, in a first step 110 of the procedure, the model function already described is first read from the non-volatile memory 16 into the computer unit 10 of system 1 or the working memory 15 of the computer unit 10. After the model function has been read in, system 1 in the second step 120 generates a number of different level pairs. {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . n)}, of an excitation frequency pair {f.sub.1, f.sub.2} is output or presented to an individual via the sound output means 21, 22 of the probe unit 20 and the corresponding DPOAE of the individual is detected via the sound recording means 23, wherein at least the first pair of levels {L.sub.1.sup.(1), L.sub.2.sup.(1)} is predefined and where n is <<N. The corresponding DPOAE are sent by the AD/DA converter unit 12 to the computer unit 10 for further processing. The superscripts in brackets indicate the 1st to n-th measuring points.

(73) The second step 120 contains a series of substeps 121 to 127 in a first variant of the procedure, which can be described as an adaptive variant, which are explained in more detail below with reference to FIG. 8.

(74) According to FIG. 8, in a first substep 121 of the second step 120, first a start level {L.sub.1.sup.(1), L.sub.2.sup.(1)} (preferably a level of L.sub.1 of 67±10 dB and a level of L.sub.2 of 57±10 dB) for a first level pair is read from the non-volatile memory 16 into the computer unit 10. Furthermore, in this substep 121, the step sizes ΔL.sub.1, ΔL.sub.2 for the other level pairs {L.sub.1.sup.(2 . . . n), L.sub.2.sup.(2 . . . n)} as well as thresholds for search direction decisions L.sub.dp,min.sup.(G, 1) SNRmin) are read into the computer unit 10. SNRmin denotes the desired SNR (signal-to-noise ratio) and L.sub.dp,min.sup.(G, 1) denotes the DPOAE level on the ridge, which must be present at least along a first subsequence of k subsequences, so that a next subsequence below the first can still be expected with sufficient SNR. If this value is not reached, the next subsequence is sampled above the first, i.e. up hill, in order to avoid too long a measurement time to reach a sufficient SNR. The step sizes ΔL.sub.1, ΔL.sub.2 denote the level spacing of two successive level pairs where ΔL.sub.1 in particular has a value of ΔL.sub.1=4 to 14 dB, preferably from 6 to 10 dB, and wherein ΔL.sub.2=0 is up to −2.78 dB, preferably ΔL.sub.2=−1.52 to −2.78 dB.

(75) In a second substep 122 of the second step, the measurements of the first subsequence of k subsequences are now performed across the assumed degree of individual function of a level map. The DPOAE are performed at the excitation frequencies f.sub.2=2 kHz and f.sub.1=1.67 kHz already described above. The excitation frequencies f.sub.2 and f.sub.1 of a level pair {L.sub.1, L.sub.2} are preferably linked via a frequency ratio of about f.sub.2/f.sub.1=1.2. The subsequence belongs to a number of k subsequences, where k is ≥2 and ≤5. In each subsequence a number of n.sub.k level pairs {L.sub.1, L.sub.2} are measured.

(76) The start level {L.sub.1.sup.(1), L.sub.2.sup.(1)} is changed according to the specified step sizes ΔL.sub.1, ΔL.sub.2 according to the formula L.sub.1.sup.(n+1)=L.sub.1.sup.(n)+ΔL.sub.1′ cos(φ) and the further formula L.sub.2.sup.(n+1)=L.sub.2.sup.(n)−ΔL.sub.1′ sin(φ). If a descending flank or no ascending flank is measured in the measured subsequence, the search direction is reversed and ΔL.sub.1′=−ΔL.sub.1′.

(77) In a third sub-step 123, it is checked whether at least three valid DPOAE were measured. If this check is positive, i.e. three valid DPOAE were measured, the procedure continues to the next sub-step 124. If no three valid DPOAE were measured, then the measurement is repeated, where from the original excitation level L.sub.1.sup.(old), L.sub.2.sup.(old) a new excitation level L.sub.1.sup.(1), L.sub.2.sup.(1) is computed according to
L.sub.2.sup.(1)=L.sub.2.sup.(old)+L.sub.2′ cos(φ),
L.sub.1.sup.(1)=L.sub.1.sup.(old)+L.sub.2′ sin(φ),

(78) In the following fourth sub-step 124, the position of the ridge of the individual function is determined on the basis of the three valid measured values, e.g. by solving a parabolic equation and finding the individual maximum according to the function:
L.sub.dp′.sup.(G,1)=f(L.sub.1.sup.(1 . . . 3),L.sub.2.sup.(1 . . . 3))

(79) Here means L.sub.dp′.sup.(G, 1) is the point on the estimated ridge of the individual model function whose associated excitation level pair is located on line formed by the level pairs {L.sub.1.sup.(1 . . . 3), L.sub.2.sup.(1 . . . 3)}. The position of the ridge of the individual function at the higher excitation levels L.sup.(1 . . . 3) is now already known from substep 124, but the slope of the ridge, i.e. the parameter m, is not.

(80) In the following fifth substep 125 of the second step 120, a second subsequence is measured along the assumed ridge, whereby only one measured value is determined. The measurement is carried out according to the formula L.sub.1.sup.(4)=L.sub.dp′.sup.(G, 1)−ΔL.sub.2′ sin(φ).

(81) Undershoots L.sub.dp′.sup.(G, 1) a preset limit (L.sub.dp,min′.sup.(G, 1)), this step is executed at higher levels (ΔL.sub.2=−L.sub.2).

(82) The value of the DPOAE (L.sub.dp.sup.(4)) is subsequently used in the sixth substep 126 to determine the individual slope of the ridge, m.

(83) In the sixth substep 126 of the second step 120, the individual slope of the ridge is calculated according to the formula m=f(L.sub.dp.sup.(G, 1), L.sub.1.sup.(G, 1), L.sub.2.sup.(G, 1), L.sub.dp.sup.(4), L.sub.1.sup.(4), L.sub.2.sup.(4)). Based on the determined gradient m of the ridge, a start level is now determined L.sub.1(5), L.sub.2(5) for the third subsequence.

(84) In the seventh substep 127 of the second step 120, the measurements of the third subsequence are now performed across the assumed ridge of the function: A variation of ΔL.sub.1′ in L.sub.1.sup.(n+1)=L.sub.1.sup.(n)+ΔL.sub.1′ cos(φ).

(85) Preferably, at least half of the level pairs {L.sub.1, L.sub.2} at which the measurement is performed, are positioned by at least 5 dB on both sides away from the group of the level pairs {L.sub.1.sup.(G), L.sub.2.sup.(G)} associated with the ridge (of the model function).

(86) With the measured values determined in accordance with the second step 120 and its substeps 121 to 127, the model function already presented is now fitted to the measured values obtained in a third step 130 in the computer unit 10. Here, the three-dimensional model function p.sub.dp,M=f(L.sub.1, L.sub.2) is fitted to the measured DPOAE values. The fitting is carried out with the mathematical methods of the regression calculation, e.g. with the least squares method, i.e. with the iterative minimization of the difference between the n measured values and the values of the model function p.sub.dp,M=f(L.sub.1, L.sub.2) to the measured n DPOAE (corresponding to the associated L.sub.1, L.sub.2-coordinates) until an individual function p.sub.dp,I=f(L.sub.1, L.sub.2) is obtained with individual parameters of a DPOAE function and level map of the individual by the computer unit 10. Thus an individual function/level map of the hearing of an individual can be easily obtained with greatly reduced measuring effort in a short time.

(87) In a fourth step 140, the output of the individually fitted function and its function parameters takes place in an output device 11 of system 1, such as a display, monitor, printer, etc. The output function parameters contain in particular the parameters already described above: a; b; c; L.sub.2,edpt′ and the slope of the ridge m. The output medium 11 can, as already mentioned, also be implemented in the form of an interface, via which an external output device, such as a printer or a monitor, can be connected to System 1.

(88) In a possible further process step, the determined individual function of a DPOAE level map and its parameters can be stored by the computer unit 10 in the non-volatile memory 16. The measured raw data from the computer unit 10 can also be stored in the non-volatile memory 16. The stored data can be used by the computer unit 10, e.g. for the continuous extension of the data set underlying the model function of a level map.

(89) The following findings can be gained from the individually adapted function obtained and the associated function parameters: An approximate distortion product threshold can be calculated which provides information about the threshold of the input signal for the inner hair cells of the measured ear. The corresponding parameter is L.sub.2,edpt. The width of the ridge, in the function defined by the parameter c is a measure for the compression and thus for the frequency resolution of the underlying travelling waves in the measured ear. The position and angle, expressed in function by the parameters a; b contains information about the nature of a hearing loss: In a pure conductive loss, the angle (expressed in function by the parameter a) does not change, instead the hill shifts in the first approximation to the same extent in the direction of higher L.sub.1- and L.sub.2-Level. If, for example, the displacement of the hill (relative to the standard values, or relative to a reference measurement of the individual at an earlier point in time) coincides with the deterioration of the distortion product threshold, i.e. ΔL.sub.2≈ΔL.sub.1≈ΔL.sub.2,edpt can be inferred from a pure sound conduction loss. The slope of the ridge, expressed by the parameter m allows conclusions to be drawn about a possible sound-conduction loss. As long as the hearing loss is below 30 dB, it can be expected that the slope of a pure conductive loss corresponds to the standard values, while a deviation from the standard value indicates in a proportional reduction of the retrograde middle-ear transmission at f.sub.dp.

(90) In an alternative variant of the procedure, instead of the sub-steps 121 to 127, a number of n fixed or predefined but different level pairs {L.sub.1, L.sub.2} (where n preferably ≥5 and ≤12, in particular ≥5 and ≤8) is output by the system and the response of an individual's hearing to these level pairs is {L.sub.1, L.sub.2} is recorded. This variant can be described as a rigid procedure. The level pairs {L.sub.1, L.sub.2} can in turn be used in a number of k subsequences (57, 58; see FIG. 6) are measured, where k≥2 and ≤12 is. The n Level pairs are then largely static and the second and possibly third subsequence are not adapted to the results of the measurements of the first subsequence, as is the case with the procedure described above. The number n of the level pairs {L.sub.1, L.sub.2} is designed after weighing the measuring time (as few measuring points as possible) against the achievable accuracy (as many points as possible).

(91) In this rigid procedure with fixed excitation levels, for example with L.sub.2′=40 dB for the three higher excitation levels, and L.sub.2′=25 dB for the three lower excitation levels, and within a group of three excitation levels, respectively L.sub.1′=0±6 dB can be measured. In the {L.sub.1, L.sub.2}-coordinate system, this corresponds to the excitation levels L.sub.2=68.1; 65.6; 63.1; 42.8; 45.3; 40.3 and L.sub.1=68.0; 73.5; 79.0; 63.3; 57.8; 68.7 (see FIG. 6). The choice of L.sub.1′=0±6 dB aims at recording three points perpendicular to the presumed position of the ridge, to reliably measure its position: at f.sub.2=2 kHz the DPOAE falls with ΔL.sub.1′±6 dB typically drops to about 50% of the maximum value. If normal hearing individuals are to be measured first and foremost, as is usually the case with screening tests, the rigid arrangement will give good results. But outliers must be detected. This is possible via the quadratic error of the model adjustment. If the error is too high, i.e. the rms error (rms: root-mean-square) is greater than 5 μPa, for example, further level pairs must be measured until the error is low enough. Here, more ΔL.sub.1′-steps might be suitable. The same procedure must be followed if individual measuring points cannot be registered because the signal-to-noise ratio is too low.

(92) Finally, it should be noted that in deviation from the used and previously described frequency ratio f.sub.2/f.sub.1 a different frequency ratio can also be selected instead of 1.2. So the frequency ratio f.sub.2/f.sub.1 can be e.g. also another suitable value between 1.15 and 1.35. In addition, the frequency ratio f.sub.2/f.sub.1 could be a function of f.sub.2.

REFERENCE CHARACTER LIST

(93) 1 System 2 Cable connection 3 First line 4 Second line 5 Third line 6 Fourth line 10 computer unit 11 output device 12 AD/DA converter unit 13 DA converter 14 A/D converter 15 Main memory 16 non-volatile memory with stored model function 20 Probe unit, OAE probe 21 First sound output medium, f.sub.1-sound generator 22 Second sound output device, f.sub.2-sound generator 23 Sound recording medium, microphone 24 Probe tip 30 Ear 31 Ear canal 40 Arrow 51 excitation level pair {L.sub.1, L.sub.2} 52 excitation level pair {L.sub.1, L.sub.2} 53 excitation level pair {L.sub.1, L.sub.2} 54 excitation level pair {L.sub.1, L.sub.2} 55 excitation level pair {L.sub.1, L.sub.2} 56 excitation level pair {L.sub.1, L.sub.2} 57 First subsequence 58 second/further subsequence 70 Graph/Model level map 71 {L.sub.1, L.sub.2}-map 72 transformed {L.sub.1′, L.sub.2′}-coordinate system 73 ridge (of the DPOAE model level map) 110 First process step 120 Second process step 121 First substep 122 Second substep 12 Third substep 124 Fourt substep 125 Fifth substep 126 sixth substep 127 seventh substep 130 Third process step 140 Fourth process step

(94) The procedure described here for the automatic determination of an individual function of a DPOAE level map can be used to avoid errors in the extrapolation of growth functions that may occur in the procedures for measuring the distortion product threshold according to the state-of-the-art described above. In addition, additional data can be obtained, which is then available for diagnosis. In addition to the distortion product threshold Leapt and the slope of the growth function, the procedure described in PCT/EP2017/000334 can also record data on the frequency resolution and compression of the underlying travelling waves and sound conduction loss. It is therefore advantageous to obtain four instead of two informations for the measuring points at the same or even lower expenditure of time, and to reduce estimation errors for the parameters obtained so far (the distortion product threshold Leapt and the slope of the growth function).

(95) In the automatic determination of an individual function of a DPOAE level map described here, the primary tones of each level pair are preferably selected in such a way that both travelling waves at theft-characteristic place have approximately equal amplitudes. Since the f.sub.1-travelling wave has not yet reached its maximum at the f.sub.2-characteristic place, it is preferably excited much more strongly, at least at moderate overall excitation levels at which the cochlear amplifier is active. However, if the excitation level combination is clearly next to the individually optimal path, the different strength of the travelling waves in the range of f.sub.2 may attenuate or intensify the phenomenon shown. Therefore, a combination with the procedure described in PCT/EP2017/000334 can be useful to define an individual function. p.sub.dp,I=f(L.sub.1, L.sub.2) with individual parameters of a DPOAE level map.

(96) For a detailed description of the “level map method” please refer to PCT/EP2017/000334.

(97) Objects

(98) The procedure according to the invention is characterized, among other things, by the following objects:

(99) 1. A method for detecting distortion products of otoacoustic emissions (DPOAE) in a hearing organ comprising the steps of:

(100) (a) output of a first primary tone pair {f.sub.1,1, L.sub.1,1, f.sub.2,1, L.sub.2,1} each comprising a first primary tone with frequency f.sub.1,1 and sound pressure level L.sub.1,1 and a second primary tone with frequency f.sub.2,1 and sound pressure level L.sub.2,1 with f.sub.2,1>f.sub.1,1, and

(101) (b) Detection of evoked distortion products of otoacoustic emissions (DPOAE),

(102) characterized in that the first primary tone {f.sub.1,1, L.sub.1,1} is output with a time delay t.sub.lag after the second primary tone {f.sub.2,1, L.sub.2,1}.

(103) 2. Method according to object 1, characterized in that at least one further primary-tone pair is presented, consisting of a first primary tone with frequency f.sub.1,n and sound pressure level L.sub.1,n and a second primary tone with frequency f.sub.2,n and sound pressure level L.sub.2,n, where f.sub.2,n>f.sub.1,n.
3. Method according to object 2, characterized in that the second primary tone {f.sub.2,n, L.sub.2n} of the at least one further n-th primary-tone pair is presented with a time delay t.sub.lag after the first primary tone {f.sub.1,n, L.sub.1,n} of this primary-tone pair, the output of the at least one further n-th primary-tone pair {f.sub.1,n, L.sub.1,n, f.sub.2,n, L.sub.2n} optionally being presented before or after the presentation of the first primary-tone pair {f.sub.1,1, L.sub.1,1, f.sub.2,1, L.sub.2,1}.
4. Method according to one of the preceding objects, characterized in that the first primary tone {f.sub.1,1, L.sub.1,1} and/or {f.sub.1,n, L.sub.1,n} (“f.sub.1-pulse”) and optionally the second primary tone {f.sub.2,1, L.sub.2,1} and/or {f.sub.2,n, L.sub.2n} (“f.sub.2-pulse”) is presented pulsed.
5. Method according to object 4, characterized in that the pulse length of the f.sub.1-pulse of the first primary-tone pair {f.sub.1,1, L.sub.1,1} is shorter than the pulse length of the f.sub.2-pulse of the first primary-tone pair {f.sub.2,1, L.sub.2,1}, and/or the pulse length of the f.sub.2-pulse of the n-th further primary-tone pair {f.sub.2,n, L.sub.2n} is shorter than the pulse length of the f.sub.1-pulse of the n-th further primary-tone pair {f.sub.1,n, L.sub.1,n}.
6. Method according to one of the preceding objects, characterized in that the time delay t.sub.lag is between 10 ms and 0.1 ms, preferably between 5 ms and 0.5 ms.
7. Method according to one of the preceding objects, characterized in that the duration of the f.sub.1-pulse {f.sub.1,1, L.sub.1,1} of the first primary tone pair and/or of the f.sub.2-pulse {f.sub.2,n, L.sub.2,n} of the n-th further primary-tone pair is 200 ms or less, 100 ms or less, 50 ms or less, between 40 ms to 1 ms, between 30 ms and 2 ms or between 25 ms and 5 ms.
8. Method according to one of the preceding objects, characterized in that the f.sub.1-pulse of the first primary-tone pair {f.sub.1,1, L.sub.1,1; f.sub.2,1, L.sub.2,1} is switched off after the end of the f.sub.2-pulse of the first primary-tone pair.
9. Method according to one of the objects 2 to 8, characterized in that a set consisting of the first primary-tone pair {f.sub.1,1, L.sub.1,1; f.sub.2,1, L.sub.2,1} and the at least one further primary-tone pair {f.sub.1,n, L.sub.1,n; f.sub.2,n, L.sub.2,n} is presented in a block which is repeated several times during the measuring period.
10. Method according to any of the foregoing objects, further comprising the automatic determination of an individual function of a DPOAE level map having p.sub.dp,I=f(L.sub.1, L.sub.2) to determine the optimal DPOAE excitation level:

(104) Reading a model function p.sub.dp,M=f(L.sub.1, L.sub.2) with model parameters of a DPOAE level map, based on a number of N DPOAE measurements of an excitation frequency pair f.sub.1, f.sub.2 each with different level pairs {L.sub.1.sup.(1 . . . N), L.sub.2.sup.(1 . . . N)} in a population (p) of subjects with normal hearing, into a working memory of a computer unit, where N≥40 and p≥2,

(105) automatic presentation of n different level pairs {L.sub.1.sup.(1 . . . n), L.sub.2.sup.(1 . . . n)} of an excitation frequency pair f.sub.1, f.sub.2 via sound output means to an individual and detecting the corresponding DPOAE of the individual via sound recording means, wherein at least the first level pair {L.sub.1.sup.(1), L.sub.2.sup.(1)} is predefined and where n<<N,

(106) iterative fitting of the model function p.sub.dp,M=f(L.sub.1, L.sub.2) to the measured n DPOAE until an individual level-map function is obtained p.sub.dp,I=f(L.sub.1, L.sub.2) with individual parameters of a DPOAE level map of the individual by the computer unit,

(107) Output of the individual function p.sub.dp,I=f(L.sub.1, L.sub.2) and/or their individual parameters on an output device of the computer unit.

(108) 11. Use the procedure according to one of the previous objects to adjust a hearing aid.