Re: Wasn't v. Helmholtz right? (Andrew Bell )

Subject: Re: Wasn't v. Helmholtz right?
From:    Andrew Bell  <bellring(at)SMARTCHAT.NET.AU>
Date:    Tue, 20 Jun 2000 10:31:40 +1000

-----Original Message----- From: Ben Hornsby [mailto:ben.hornsby(at)] Sent: Friday, 16 June 2000 12:17 To: Andrew Bell Subject: Re: Re: Wasn't v. Helmholtz right? Andrew, I am a doctoral student in Audiology and have been following your discussion to some degree. One comment you made caught my eye: " Of course, at higher SPLs (above about 60 dB), vertical movement of the partition does begin, but only as a means of damping excessive motion." I noticed that other respondents to the discussion either agreed with this or didn't comment on it. I was under the impression (based on BM tuning curves) that a response on the BM could be observed at levels much lower than this. I believe I am missing something fundamental in the discussion. What do you base this observation on? Any comments are appreciated. Ben Hornsby -------------------------------- Dear Ben: Thankyou for your interest and your perceptive question. Yes, you're right: BM responses can be observed right down to near zero. What I meant to convey by my statement was that whole-scale up and down movement of the cochlear partition (like that which is supposed to happen under the traveling wave theory) does not begin until about 60 dB SPL. Below this level, there is insufficient differential pressure developed across the partition for it to be pushed up and down. However, there is still sufficient common-mode pressure, a factor usually neglected in mathematical models, for the OHCs to detect and respond. Another way of saying this is that the acoustic impedance of the helicotrema is much lower than usually assumed, so that this hole effectively short-circuits the pressure difference across the partition. The common-mode pressure is simply the pressure built up uniformly throughout the cochlea by the inward movement of the oval window (like pushing in the loudspeaker cone on a sealed-box speaker enclosure). Both the low- and high-level processes involve the tectorial membrane, and only incidentally the basilar membrane (where doppler velocimeter measurements are made with reflective beads). Thus, at low SPLs, the OHCs are creating ripples on the underside of the tectorial membrane. There is effectively no movement of the top layer, which is covered with a net (indeed, the covering net is designed to damp movement). The OHC resonant cavity is a regenerative receiver of oscillating sound pressure (it's like a radio receiver in the cochlea picking up a broadcast from the transmitter at the oval window). It amplifies the signal and sends the output to the IHC via ripples on the underside of the TM. Accompanying this activity, some energy will be communicated to the BM; however, because the wavelength is small (the distance between OHC1 and OHC3, about 30 um), and because the reflective beads dropped on the BM are about that size, the measured movement is much less than the movement experienced by the IHCs when they sense the radially propagating wave on the TM. Hence, the measurements are sensing only a residual fraction of what is really going on. The result is that we end up with unimaginably small figures for the threshold of detection: figures claiming that at 0 dB the IHC can detect BM movements of hundredths of nanometres and stereocilia deflections of hundredths of a degree (see Dallos, in The Cochlea, ed. P. Dallos et al, Springer, 1996). Of course, the IHC is detecting far bigger movements because it is closer to the output of the regenerative receiver. Above about 60 dB, however, the regenerative receiver saturates, and a second process begins. (Saturation occurs when the membrane potential of all three OHCs falls to the same resting level: there is now no amplification because there is no difference in the response of OHC2 compared to its neighbouring OHC1 and 3.) The cochlear partition begins to move because the differential pressure across it finally becomes sufficient to exert enough force. That is, there is some acoustic impedance at the helicotrema to allow a degree of differential pressure build-up. Now the traveling wave picture can be introduced, but with a difference - the pressure difference is across the tectorial membrane, not the basilar membrane. This is because acoustic pressure builds up across a material only when the acoustic resistance of that material is appreciable. Think of a thin rubber sheet immersed in a swimming pool: that sheet is invisible to passing acoustic waves and the waves pass straight through. By contrast, a sheet of metal suspended in a pool is detectable because waves bounce off it. The difference in the two cases is that the speed of sound in the water differs substantially from that in steel, but not the speed in rubber. In the same way, the speed of sound in the basilar membrane differs little from that in water. However, the tectorial membrane is a gel with tensile properties (and corresponding compressional wave speed) appreciably different to those in water. The result is that acoustic pressure builds up across the TM, not the BM. What this means in practice is that the TM is vibrated backwards and forwards against the vestibular lip, once more generating ripples which are launched from the lip (that's what it's there for), and propagate along the underside of the TM towards the IHC. In this respect, we can understand why the lip is so remarkably sharp and well defined; the lip launches ripples simultaneously along its length, and because of graded ripple speed along the partition (based on TM surface tension and thickness), we get a graded delay of response at the IHCs - otherwise interpreted as a 'traveling wave' delay. In summary, there are two ripple-generating processes going on: one in the OHC cavity at low SPLs, and the other from the vestibular lip (at high SPLs). Note that such an arrangment leads to cancelation of the IHC response at some intermediate sound level (observed as 'peak splitting' in cochlear nerve studies); it also leads to dynamic range compression. Compression occurs because the amplitude of the ripples is a non-linear function of SPL. It also happens because the OHC detectors saturate, and can become active absorbers at higher SPLs (simply by changing relative membrane potential between OHC2 and OHC1&3). Movement of the basilar membrane, which is usually thought of as the result of resonant mass/compliance elements in the partition, is simply spreading of agitation from OHC activity. It is also a process which tends to absorb energy (Martin Braun [Hearing Research, 97 (1996), 1-10] suggested that this is what the BM is for). While this is happening, remember that the tuning of the OHC cavity is about half an octave higher than the mass/compliance tuning of the TM (and probably an octave or more higher than the strongly damped movement of the whole partition). Here is an explanation for the curious half-octave shift of McFadden (In: Basic and Applied Aspects of Noise-Induced Hearing Loss, Salvi et al (eds), Plenum, 295-312). Clearly, dropping reflective beads on the BM, or even the top of the TM, is not going to give us a clear picture of what's going on in the OHC cavities, particularly at low SPLs. Another misleading result from invasive experimentation is that drilling a hole in the cochlea is going to disrupt the pressure response of the OHCs. Andrew Bell. ___________________________________________

This message came from the mail archive
maintained by:
DAn Ellis <>
Electrical Engineering Dept., Columbia University