Calibrating Bone Conductors using OAE’s

The most common method of calibrating an audiometer for bone conduction testing requires an artificial mastoid where the bone transducer is positioned on the mastoid and an electrical input signal is measured and equated into a certain sound pressure level.

The problem with using an artificial mastoid is that it does not represent the anatomical uniqueness of each person who will be tested.  That is, a bone conductor calibrated using an artificial mastoid is calibrated to replicate the mechanical impedance of the head.  But not all heads are the same.  A static force of approximately 5.4 newtons (N) (ANSI, 1996) is the current standard, but this varies depending on the size of a person’s head. 

In this article, an objective method of calibrating a bone conductor using otoacoustic emissions (OAE’s) will be discussed.  Distortion product OAE’s are elicited using two pure tones (f1 and f2).  Most normal ears will produce a recordable emission at 2f1 – f2, where the magnitude of the emission varies as the magnitude of f1 (L2) and f2 (L2) are varied.      

The theory underlying this calibration technique is that OAE’s are not restricted to pure tone air conducted stimuli, but can also be elicited using pure tone bone conduction stimulus.  Barany (1938), Lowy (1942), Tonndorf (1966) have shown that the cochlea responds equally to both air and bone conducted stimuli.  That is, a traveling wave will be established along the basilar membrane regardless of whether it is elicited by an air or bone conducted stimuli (Zwislocki, 1953).

DPOAE with stimulus magnitude can be represented using a contour plot that is illustrated below.  In the graph below, L1 and L2 (level/magnitude of f₁ and f₂ tone in dB SPL) are represented on the Y and X-axis respectively.  The lighter shaded region represents the space where the highest magnitude DPOAE’s were recorded (L_dp = 5-10 dB SPL).  The darkest regions represent the area where the DPOAE’s were weakest (L_dp = -20-15 dB SPL).  At these levels, it is difficult to distinguish whether the response is a DPOAE or whether it is simply part of the noise floor.        

figure 1 from Purcell et 1999

If we were to hold either L₁ or L₂ constant in magnitude, then a graph with a single curve can be obtained.  This graph is called an IOgram and is illustrated below with L1 held constant.

figure 2 from Purcell et al 1999

The IOgram is used to indicate nonlinear emission growth with changes in L₂.  That is, with L₁ held constant, we can see how the magnitude of the emission changes as the magnitude of the f₂ tone (L₂) is varied.  A general pattern with DPOAE’s as seen on an IOgram is that emissions tend to increase up to a certain level, and then level off.  This very pattern is seen in the above IOgram where the growth decay is quite obvious.  According to Purcell et al.  1999, it is advisable to hold L₁ constant and vary L₂ rather than the reverse because, “emissions tend to be recordable above the NF (noise floor) for a larger range of L₂ and of L₁”. 

So what do these graphs mean towards calibration you may ask?  Well, the IOgrams are used to compare emissions elicited through BC with those from AC.  As I mentioned before, a critical point about DPOAE’s is that they can be elicited by either BC or AC because both methods set the basilar membrane into motion. 

Using IOgrams with a fix L₁ value ensures that emissions are compared in the same region in stimulus space.

Once again, let’s review these important steps.

1.    f1 and f2 tones are fixed at a certain frequency

2.   The L₁ magnitude is fixed and L2 is free to vary in order to find the maximum emission level

3.   Two IOgrams will then be generated, one by AC and the other by BC

It is critically important that L1 remains constant during this procedure, therefore it is recommended that the same stimulus transducer be used during all measurements.  Purcell et al. 1999 recommends that f₁ is always presented using an air speaker in a probe inserted in the ear canal.  This same probe contains the microphone for the response measurement as well as a second speaker for the f₂ tone when performing an AC measurement.  When a BC measurement is being performed, everything remains the same, except that instead of using the probe to generate the f₂ tone, a bone conductor at f₂ is used.   

According to Purcell et al. 1999, if the IOgrams are similar under bone and air conduction stimulation, it follows then that the f₂ tone must be reaching the cochlea with similar magnitudes in both cases. 

For example, if the f2 tone by air conduction produces an IOgram that looks like figure 2 above where the maximum emission occurs at 50 dB SPL, we can expect that the same f2 tone by bone conduction will produce a similarly looking IOgram.  If the maximum emission by bone conduction does not occur at 50 dB SPL or close, then it is safe to assume that the bone conductor is not properly calibrated.  When this happens, the maximum emission point by BC must be located and marked as being 50 dB SPL because this is the point of maximum emission by AC (assuming that AC is properly calibrated).  By repeating the IOgram correlations through the frequency range of interest, the bone conductor can be objectively calibrated against the known air stimulus (Purcell et al. 1999 pg 378).

Note:  This site is best viewed using IE 4.X or higher under 800x600 resolution