Dead Hair Cells?

I was searching the literature for experimental evidence of “calcification” of hair cells as a possible explanation for decruitment hearing. By “calcification” I mean simply some mechanism which could make hair cells less compliant to mechanical forces – making them stiffer in some sense, and thereby requiring greater sound power to have them produce the same output as non-stiffened hair cells.

I did find some papers describing chemical actions in the hair cells, based on exposing lab rats to damaging levels of sound, and then performing autopsies on them. They showed some graphs indicating as much as 80% damage to hair cells after exposure to sounds louder than 106 dBSPL.

Obviously, these researchers, and perhaps the whole field itself, still believe in the damage theory of hair cells as causing hearing impairment.

So let’s look at what follows from this belief.


  • The still-functioning hair cells produce a collective signal to the brain.
  • There is some function that describes how the signal from a hair cell grows with increasing sound levels.
  • There is some thresholding action that prevents us hearing below some small level of sound.

Choose a frequency, say 1 kHz, where we have defined the sonic effects so that 40 dBSPL = 40 Phon = 1 Sone, and where an increase of 10 dBSPL corresponds to an increase of 10 Phon, and a factor of 2 multiple in Sones. And we know that threshold level sounds correspond to 20 microPascals of RMS sound pressure.

Now consider that the function describing Sones in relation to Phons is S(p), for signal level p Phons. And consider that there are N functioning hair cells at that frequency. Then the threshold level has an aggregate signal of N S(0), where 0 Phons describes our hearing threshold, for normal hearing. From the 20 microPascal RMS threshold level, we know that S(0) = 0.0021 Sones.

The thresholding effect of our hearing is normally described as being that level of sound such that listeners can identify the signal as present for 2 out of 3 trials.

Now suppose a hearing impairment is found to require 20 dBHL to reach threshold conditions. This would be a very mild impairment, essentially considered at the upper bound of normal hearing.

If that hearing impairment were due to a death of hair cells, where some fraction, f, of hair cells are dead, then we can describe the threshold of impaired hearing as:

(1-f) N S(20) = N S(0)

It follows then, that the absolute number of hair cells, N, falls out, leaving f = 1 - S(0)/S(20).

We haven’t specified what function S(p) actually looks like. But we have experimental evidence that it shows cube root behavior above 40 Phons, and linear behavior near 0 Phons.

I have a model, called EarSpring, which fits these boundary conditions. It may, or may not, be a correct model. My model says that S(20) = 0.15 Sones. A linear model would indicate 0.21 Sones, and a cube root model would predict 0.01 Sones. So my model falls between these two extreme models, but closer to linear at this level. At 20 dBHL we are still very near threshold sound levels, so it is likely to be closer to linear behavior than cube root behavior.

So the dead hair-cell model would have us believe that f = 0.99 (!!). A whopping 99% of hair cells need to be damaged to produce a mere 20 dBHL threshold elevation !? This suspends belief.

Furthermore, at 120 dBSPL (480 Sones) – about as loud as a running jet engine at 1 meter distance, well into the range of cube root compression above 40 dBSPL – it would have this mildly impaired listener hearing something about as loud as 60 dBSPL (4.7 Sones) to everyone else – about as loud as a loud conversation between two adults at 1m separation!

A normal, but loud, adult conversation (60 dBSPL = 4.7 Sones) would be heard as a mere faint whisper at 0.047 Sones (about 14 dBSPL to the rest of us)!?

Remember, someone with only 20 dBHL of impairment is considered within the range of normal hearing.

Beyond this, there is no way to produce recruitment hearing on the basis of reduced numbers of participating hair cells. With recruitment hearing we have experience of loud sounds becoming essentially normal, even for much more severe levels of impairment. That’s why we instinctively shout at people with hearing loss – it helps but mostly irritating. But the simple model here would have a constant Sones fractional multiple everywhere.

So, on the basis of these crazy predictions, we know that typical sensioneural hearing impairment cannot be caused by the lack of numbers of hair cells.

Note that I have not imposed my world view of the EarSpring Sones model here. Just a simple line of reasoning about numbers of participating hair cells is all that is needed to show how foolish is this model of hair cell death.

Back to the initial paragraph about calcification – it strikes me that this would be impossible to determine. I am not talking about cell ossification, which might show up someday on a high-precision CAT scan. And unless you kill the victim to perform a precision autopsy, we simply can’t see what is happening in hair cells.

So, I wonder if some (most?) of the damage seen in autopsies might have been caused by handling of the cells in preparation for microscopy?

  • DM

I also have anecdotal evidence (my own) that supposedly dead hair cells can come alive again, at least temporarily.

It is well known that injecting chemical agents into our bodies can cause short term remissions of many conditions. I had that happen to me with my hearing about 10 years ago. A foreign substance (a blood serum) was injected into my neck, and within minutes my superb acute hearing returned. I was able to listen to and comprehend a whisper conversation across the room at more than 20 feet separation. My hearing remained acute for about 2 weeks after that. But my impairment did eventually return.

So there is evidence that hearing impairment is a chemical interaction with hair cells. That hair cells don’t need to die to produce hearing impairment.