Method and apparatus for measuring the Acoustic Reflex with artifact management by using multiple probe tones

Abstract

A method and apparatus is disclosed that performs a measurement of the acoustic reflex. In contrast to prior art, more than one probe tone is used simultaneously to detect the acoustic reflex and separate the reflex response from artifacts. In a preferred configuration, at least one probe tone frequency is selected to be higher than the middle ear resonance frequency.

Claims

1. A method and apparatus to measure the acoustic reflex that presents and records at least two probe tones of different frequencies simultaneously, comprising a. a probe with at least one speaker and at least one microphone, to be placed in the human ear canal b. means for signal generation of said probe tones and presentation via said loudspeaker(s) c. means for recording back said probe tones via said microphone(s) d. evaluation of said recorded probe tones for amplitudes and/or phases to detect reflex and artifacts

2. A method and apparatus according to claim 1 which provides means for presentation of a stimulus, either ipsilateral or contralateral or both, to trigger the acoustic reflex

3. A method and apparatus according to claim 1, that provides an interface for an external trigger to allow detection of the acoustic reflex from external stimulation

4. A method and apparatus according to claim 1, where signal generation and analysis are implemented digitally

5. A method and apparatus according to claim 1, which displays the compliance as derived from the recorded probe tones or other features of those signals graphically for visual inspection

6. A method and apparatus according to claim 1, which automatically detects the acoustic reflex by using two or more of the recorded probe tone signals simultaneously

7. A method and apparatus according to claim 1, which automatically detects or suppresses artifacts by using two or more of the recorded probe tone signals

8. A method and apparatus according to claim 1, which automatically adapts artifact suppression parameters during no-stimulus phases

9. A method and apparatus according to claim 1, that also evaluates the phase of the recorded tones to separate reflex from artifact response

10. A method and apparatus according to claim 1, that includes a tympanometric pump to allow reflex testing under static pressure

11. A method and apparatus according to claim 6, that automatically selects stimulus frequencies and/or levels to automatically find the subject's reflex threshold

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

[0028] FIG. 1: illustrates the admittance change with the acoustic reflex: The resonance frequency, which corresponds to a maximum in admittance, moves to higher frequencies. The change in admittance that can be recorded at 3 different frequencies differs. Real, measured data can be found in Peake W T, Rosowski J J. (1997), pp 1343, FIG. 5a.

[0029] FIG. 2 illustrates the admittance change with probe movement (typical artifact). The residual volume is modulated, resulting in a similar change in admittance for most frequencies.

[0030] FIG. 3 illustrates a 3-channel recording of a reflex. Since the three traces correspond to different probe frequencies, one of which is above the middle ear resonance frequency, they respond differently to the reflex.

[0031] FIG. 4 illustrates a 3-channel recording of volume-changing artifacts. Since a change of the residual volume changes the measured compliance of all frequencies to the same direction, the response of all three channels is to the same sign. Both directions (volume smaller, volume bigger) can occur.

[0032] FIG. 5 illustrates a simplified principle schematic of the method. 3 different probe frequencies are generated (1), amplified (2) and presented to the ear probe's (3) speaker (4). The ear probe's microphone (5) records back the signal which is amplified and separated by narrow band filters (6) which are tuned to the probe frequencies. The amplitude of the component frequencies is measured (7), offset removed and presented (8). means to trigger the reflex is not shown for simplicity. It can be acoustically on either ear or electrical (e.g. via a cochlea implant).

[0033] FIG. 6 illustrates a more realistic implementation of the invention, where all signal processing is implemented digitally. A processor (11) is connected to a D/A converter (12) and a A/D converter (16) which provide probe tone and stimulus output and recording of the microphone signal. The processor can be connected to various interfaces (17) which can be a direct user interface, such as display, keyboard, touch-screen, or a connection to a computer for operation. Again, means to trigger the reflex is not shown.

DETAILED DESCRIPTION OF THE INVENTION

[0034] The (human) middle ear transforms the external sound field that enters the ear canal to a sound in the inner ear. Since the inner ear is filled with a fluid, its acoustic impedance is much different, basically towards higher sound pressure and lower medium velocity. The middle ear handles this by connecting a relatively big ear drum membrane to a small stapes footplate, acting as a piston into the inner ear. The middle ear efficiency peaks at a frequency of around 1 kHz. This frequency corresponds to the main resonance of the middle ear, where mobility is highest. The mobility is usually described as compliance, which is in fact the imaginary part of the so-called acoustic admittance. Acoustic impedance is defined as p/q, acoustic admittance as q/p, where p is the sound pressure and q is the sound volume flow (volume/time).

[0035] A muscle, called musculus stapedius, is connected to the stapes bone and, if activated, drags the stapes bone sidewards. This temporarily reduces the mobility of the stapes bone and in turn the ear drum. However, since the muscle stiffens the overall middle ear mechanics, it also moves its resonance frequency up.

[0036] The musculus stapedius is activated when loud sound is received by either ear, and is thought to be a protective mechanism. Other purposes are under discussion. Since the muscle acts non-voluntary, its activation is called stapedius reflex or acoustic reflex. The sound level that is needed to excite the stapedius reflex is called acoustic reflex threshold.

[0037] In audiometry, the acoustic reflex can be used for objective measurements, since the subject under test does not have to voluntary respond to the examiner. This makes this test possible in sleeping subjects, young children, or during surgery. It can also be used in context with implantable hearing devices, such as cochlea implants, which stimulate the inner ear electrically.

[0038] The standard method to record the stapedius reflex is to apply a moderate level, low frequency tone of typically 226 Hz to the ear canal via a probe. The probe tone level needs to be well below the reflex threshold. The probe is inserted into the ear canal. The probe also contains a microphone that records back the 226 Hz signal. A stimulus is then applied to the same ear (ipsilateral) or the opposite ear (contralateral). A separate speaker may be provided for this. The 226 Hz component in the microphone signal will change slightly in level and/or phase while the stapedius reflex is present and temporarily stiffens the ear drum.

[0039] Usually, the instruments that record the acoustic reflex are combined in so-called tympanometers. Tympanometry measures the mobility of the ear drum as a function of external static air pressure, and is a diagnostic method for the middle ear. The method to detect the change in mobility via a probe tone is closely related to acoustic reflex test, which is why the methods are often combined. If the tympanometry pump is available in the reflex system, applying a static pressure to a value where the ear drum mobility is maximal, can improve reflex testing reliability. Therefore, before reflex testing, the tympanometry is often executed to determine this pressure of maximum compliance.

[0040] In tympanometry, alternate frequencies are often provided which can be used instead of the standard 226 Hz, such as 1 kHz. An alternate approach (wide band tympanometry, Liu et. al 2008) uses broad band clicks to calculate the admittance over a broad frequency band as a function of static pressure. A similar approach is proposed by Freeney M., Keefe D. H. (1999). However, the repetition rate of the click is somewhat limited, its sound level needs to be quite high and some averaging is needed to achieve a sufficient signal-to-noise-ratio (SNR). This means the detection signal is at risk to evoke the reflex by itself.

[0041] The standard frequency of 226 Hz in tympanometry and reflex testing is historically selected because of a relation between acoustic admittance and effective acoustic volume: At 226 Hz, the acoustic volume in cm.sup.3 and the acoustic compliance are numerically equivalent. The frequency of 226 Hz is below the middle ear resonance. The middle ear system acts as a mass-spring-system. At frequencies below the resonance, it mainly acts as a spring, and the reflex will stiffen this spring.

[0042] Since the stapedius reflex will shift the middle ear resonance towards higher frequencies, the compliance at a higher probe frequency, where the middle ear mainly acts as a mass, can also rise during the reflex. Therefore, using a probe tone above the middle ear resonance, results in an inverted behavior of the compliance change during the reflex. This is illustrated in FIG. 1.

[0043] Since the change in the microphone signal due to the acoustic reflex is small, the measurement can easily be disturbed by artifacts. One obvious artifact is external noise that can interfere with the probe tone recording. This can be handled quite well by using a narrow band filter for the microphone signal. In state-of-the-art implementations, this filtering is implemented digitally. This narrow-band recording allows very moderate probe tone levels, in contrast to wide band approaches.

[0044] Another type of artifact includes any movement of the probe position within the ear canal. Such a movement will change the residual volume between probe and ear drum. Since the residual volume contributes to the overall compliance that the probe measures, a change can easily be mistaken as a reflex. A similar artifact occurs if the subject under test swallows or yawns.

[0045] As a consequence, an experienced examiner (who is aware of the artifacts) may tend to increase stimulus levels well above the threshold until a clear reflex is recorded. The threshold would then be overestimated. It may also be necessary to repeat testing specific stimuli until artifact-free recordings are achieved, thus prolonging total examination time.

[0046] An artifact as described above will modulate the residual volume. The compliance change during the artifact can therefore be expected to be similar for various probe frequencies. In contrast to this, the acoustic reflex, since it shifts the middle ear resonance, will act differently for different probe frequencies.

[0047] The idea of the present invention is to use this different behavior for artifact suppression. It presents more than one probe tone frequency simultaneously, and records all of them back. The different probe frequencies can be separated via dedicated narrow band filters from the common microphone signal, so that more than one compliance values can be derived simultaneously at any time. The overall loudness of the tones can still be kept moderate enough to be safe against triggering the reflex.

[0048] If, for example, frequencies 226, 678, 1356 Hz (1, 3, 7 times 226 Hz) are used, compliance at 226 would drop, compliance at 678 would not change much, and compliance at 1356 would rise during the acoustic reflex. In contrast to this, the measured compliance at all frequencies would rise if the probe is moved outwards or drop if it is moved inwards the ear canal, due to a mere residual volume change.

[0049] The test system can therefore separate artifacts and reflex in the response data. This separation can either be done by the examiner, who observes, for example, 3 traces instead of just one, knowing how reflex response and artifact response look like. The data can also be evaluated automatically, and a derived, artifact-free reflex signal shown to the examiner. Such evaluation may contain calculation of differences of compliance traces, correlation methods, etc. It may also make use of the phases of the recorded signals. A preferred implementation weights the recorded traces with factors (some of which are negative), which are selected to result in an artifact-free reflex response trace.

[0050] An apparatus which implements the method would typically contain a processor and digital-analog and analog-digital conversion. All signal generation and recording is implemented digitially. The filters would typically be implemented using quadrature detection, which allows phase-locked recording. In such an implementation, the number of probe tones that are used simultaneously is only limited by the overall sound level they produce, which needs to be below the reflex threshold, and the processing power of the digital system. A typical implementation would use 3 probe tones.

[0051] The configuration of the probe tones (number and frequencies) can be configurable, to, for example, optimally handle middle ears of different age groups or middle ear disorders. Said factors for deriving an artifact-free signal could also be adjustable or even automatically learned by the system during no-stimulus phases.