Detecting the existence of a sound that has been lost in thermal noise

[tonyjeffs]

ultra sensitive microphone or technique for listening to inner ear

A friend has very loud tinnitus. He says it is constantly louder than anything else he can hear.
Nothing can be heard by an independant observer such as myself, but it is possible that a sound with a small power level is so very close to the stereocillia in his inner ear that it is very loud to him, but is lost in thermal noise by the time it reaches the outside world.

Is there any non-invasive way of listening to see if a real, extremely low power, sound exists inside his ears? Perhaps it could be described as the auditory equivalent of an MRI?

(I have a feeling that this is a dumb question - & can't see how it could be possible to achieve, but thought I'd ask just in case someone has any ideas)

...
The current fashion with tinnitus is to say that it is an entirely neurological phenomenon, and there is definitely no sound wave involved. I think this is bad science as in the vast majority of cases there is absolutely no evidence to identify the process involved
...

Grateful for any thoughts.

Tony

 

[berkeman]

<< Merged thread from Medical Sciences with this one in General Physics. See Post #7 for a clarification of the OP's question >>

What physical mechanism are you postulating for creating this sound? You would need something vibrating inside, near the inner ear, right? Are you thinking maybe some kind of an aneurysm or something that causes oscillations in portions of the nearby blood vessels? But then the sound would be pulsing with the beats of the heart, right. Is the noise reported to be a steady ringing?

 

[Evo]

A friend has very loud tinnitus. He says it is constantly louder than anything else he can hear.
Nothing can be heard by an independant observer such as myself, but it is possible that a sound with a small power level is so very close to the stereocillia in his inner ear that it is very loud to him, but is lost in thermal noise by the time it reaches the outside world.

Is there any non-invasive way of listening to see if a real, extremely low power, sound exists inside his ears? Perhaps it could be described as the auditory equivalent of an MRI?

(I have a feeling that this is a dumb question - & can't see how it could be possible to achieve, but thought I'd ask just in case someone has any ideas)

...
The current fashion with tinnitus is to say that it is an entirely neurological phenomenon, and there is definitely no sound wave involved. I think this is bad science as in the vast majority of cases there is absolutely no evidence to identify the process involved
...

Grateful for any thoughts.

Tony

There are two types of tinnitus, one can actually be heard by others.

I will look up the articles later.

 

[tonyjeffs]

Berkeman,
I am thinking along the lines of a noisy stray capilliary, that has grown in the wrong place.
Something much smaller than (what I'd think of as) an anneurysm. Maybe it could even be the normal surface noise of a tiny blood vessel ,ordinarily inaudable, but in this case so close to the inner ear receptors that the person can hear it.
It might, I wonder, alternatively be something to do with a resonance created by damage to the inner ear, or one of the surface structures such as the oval window, so that the noise arises from a localised sea-shell effect.


Evo, agreed.
Typically, pulsatile tinnitus, which is connected with heartbeat, can be audible to others. This can sometimes be cured by moving a blood vessel.
Cat's ears can apparently be heard whistling away sometimes.

My argument, though, is that the fact that external observers can not hear the sound in most cases does not mean or even imply that the perceived noise is purely neurological in origin.

 

[Evo]

There are many possible causes of tinnitus, I've never heard of it being believed to be mostly neurological.

http://www.aafp.org/afp/20040101/120.html

 

[tonyjeffs]

Hi Evo,
I have been heavily involved with tinnitus charities and personally with some of the researchers in the field. The people involved largely overlap and understandably interact and know each other.
Generally the medical-science approach is to describe it as neurological, and that flows into the associated research. There's a lot of what could be called hype or propaganda. If you tell a patient it is neurological, you can discuss things like neuronal plasticity, dazzle him with big concepts, give them something to think about, and devise psychological treatments. Pavel Jastreboff is the main researcher who started that way of looking at it.

The article you cite to me is along the same lines:

The authors say "Subjective tinnitus, which is more common, is heard only by the patient. " Although they don't sepcifically do so here, this sort of phrase is generally used as evidence that tinnitus is neurological, and not a real sound.

An later, they say:
"Most cases of tinnitus result from the same conditions that cause hearing loss. "
and
"NIHL is the most common type of acquired hearing loss. It is irreversible,"
i.e. they're effectively saying that most cases of tinnitus are caused by sensaurial hair cell damage - i.e. neurological.

And later on they say:
"Objective tinnitus is rare."
Objective tinnitus, caused by a real sound, is rare.


You might disagree with me and say that I'm reading too much between the lines into this particular article, by associating subjective with neuronal, and objective with real vibration, for example, but that is pretty much the standard association made by people involved in the vast majority of research and treatment related to the condition.

Having said all that, I am not supporting the 'predominantly neurological' position at all, and think your and my views are very much more aligned.

Here are some abstracts :

http://www.neurology.org/cgi/content/abstract/50/1/114
(imo These researchers confuse the perception of tinnitus with actual tinnitus. They don't have adequate controls in their research. They don't use enough subjects for the result to be statistically significant):

http://www.ncbi.nlm.nih.gov/pubmed/2175858
This is the paper that got the whole "tinnitus is neurological" trend started. It is a heavy read, and gives the reader impression that part one has somehow proved part two, but they've not quite understood it. - But in my opinion this is a kind of trick, though I'm sure the authors didn't do it intentionally. The first half of the paper considers the different types of neurons and receptors in the auditory neuronal pathwaay. The second half considers the possibility that tinnitus in somehow neurological in origin. The conclusions in the second half don't really follow on from the first, and there is a barely noticeable leap of faith in the middle.

I didn't mean to get so much into discussing tinnitus, though it is interesting and very refreshing to hear your straight impartial scientific view, which is refreshingly independant from that of those who are active within the field.

...

Any thoughts on the microphone question? I guess that really it was a dumb question, and there isn't any way of detecting a sound that is so far away as to be lost in thermal noise.

Tony

 

[tonyjeffs]

Suppose there is a quit sound inside the centre of a spherical box. the sound is so quiet that it cannot carry beyond the surface of the box. - at that distance from the source it is much quieter than the natural thermal noise of the box. It cannot be heard with even a perfect microphone.

Does anyone have any ideas how the sound might be detected? Theoretical will do- it doesn't matter if the idea not practically possible.

~One possibility, for example, might be to supercool the box, thereby reducing thermal noise, and making the sound level significant relative to it, so that it can be heard with a microphone


Tbanks

Tony

 

[Borek]

What do you know about the sound? Any cyclic pattern?

 

[berkeman]

Suppose there is a quit sound inside the centre of a spherical box. the sound is so quiet that it cannot carry beyond the surface of the box. - at that distance from the source it is much quieter than the natural thermal noise of the box. It cannot be heard with even a perfect microphone.

Does anyone have any ideas how the sound might be detected? Theoretical will do- it doesn't matter if the idea not practically possible.

~One possibility, for example, might be to supercool the box, thereby reducing thermal noise, and making the sound level significant relative to it, so that it can be heard with a microphone


Tbanks

Tony

The traditional way is to perform narrow-band detection. Since the noise is broadband in nature, you put a sharp bandpass around the desired signal frequency to improve your signal to noise ratio (SNR). You can also sweep your bandpass to look for different signal frequencies (this is basically how a spectrum analyzer works).

 

[rbj]

i might suggest posting this to the USENET newsgroup comp.dsp. it's possible that a sound that is inaudible to your friend, but audible to the those of us with good hearing, can have its frequency components identified, isolated, and amplified, without amplifying the ambient broadbanded noise. it's far less likely that a sound that "cannot be heard with even a perfect microphone" can be isolated and selectively amplified.

it's maybe just a little be tragic (for a signal processing engineer who works in audio and music) but i also have tinnitus and had something like it for at least a couple years. but i really noticed it getting bad in 2006. 30 dB loss at 4 kHz and I'm 52. i think my case might be less bad than your friend's because for me it is not deafeningly loud. late at night, when i lie quietly awake in bed, it sounds like a constant drone of a few hundred crickets or locusts in the distance (not quite like white or pink noise). i listened to too much loud rock 'n roll in the 70s and 80s, which is probably the cause of it (plus the inherited thing, my dad and his brothers were, sometimes comically, hard of hearing).

 

[tonyjeffs]

Sorry, I think I messed up by mentioning my friend's tinnitus as it has taken us off topic, and detracted from the physics question I wanted to consider.
The question is really "Is there any way of detecting a sound that cannot be heard".
An example of an answer might be "Yes, if you can see the needle moving in the groove of a gramaphone record that is insulated with a vacuum, then you can detect that particular sound even though it cannot be heard"



All we know about the sound is: if it exists at all, it no longer exists at the obervation point as a detectable pressure wave.



RBJ -Common story -similar to my own - but I just bought a specialised sound card to get on with some amateur home recording anyway.

Tony

 

[GTrax]

Hi tonyjeffs

First - about the tinnitus. This is something I have, and know something about.
Hearing a noise is not at all the same as there actually being a noise there. Victims of injury or disease where the whole hearing mechanism is totally destroyed can still "hear" tinnitus.

In my case, monitoring electrodes on my head while playing pulsed sounds into my ears established the auditory nerve system was functioning, yet I knew the approx 8kHz whistle was not real because it did not produce beat frequency 400Hz notes when I played 7600Hz tones into my ear. When I added in a separate phone playing a real 7600Hz note, along with an 8000Hz note I was using to compare, then the 400Hz appeared. I concluded the 8kHz squeal was "in my head".

There is an explanation that what we sufferers perceive as sound is actually the normal way the brain generates sound perception, but gone a bit out of control in trying to compensate for loss of hearing function (e.g. rock gig damage). Its as if the gain has been turned up high. Tinnitus is only there when one is awake, and we hear full scale sounds and conversations in our dreams, not corresponding to the sounds really around us as we sleep. Possibly the ear mechanism is a controlling thing, modulating and moderating the tinnitus-like scheme in our brains by which we have sound awareness.

As for very sensitive detection of real sound pressures, the human ear is very hard to beat. The fine hairs in the cochlea are small enough to be deflected by very tiny air molecule movements. The human average is a sound pressure of about 20uP (micropascals). There may be instruments that can do better.

 

[tonyjeffs]

Gtrax
That's a good, controlled, experiment you did.
In virtually all the related professional research I've read on the topic, control is woefully inadequate imo.
Are you fairly confident that your 8k tinnitus is a fairly pure tone? You would of course still get a beat frequency if it is not - clearly a piano note falls into this category although it is not a pure sine wave, for example.

But I think it would be possible to produce a complex tone that sounds like and centres on 8khz, yet masks any beat frequency with its harmonics, and comparing my case to yours, I'm inclined to think that might be the real explanation. It'd be interesting to experiment with that and see if I'm right or wrong.

On the subject of tinnitus - I guess I can't put that darn genie back in the bottle on this thread - a fairly obvious comparison is pain. A person without an arm can feel pain in the missing hand - phantom limb pain. This means that in their circumstance the pain signal is not due to a hand injury but is sourced in the brain. The important factor to remember though is that this cannot be extrapolated to 'most people'. In most cases the pain in the hand IS due to damage in the hand.
It is therefore wrong to jump to any conclusion or generalisation. The fact that something is true for a group of people does not mean, or even imply, it is true for all.
That's why I'd like to research it.

I guess if the noise is there to detect, we can detect it though the equipment would be bulkier than the cochlea.

The question I'm trying to solve is not about tinnitus though; It is "How can we detect a pressure wave that exists at a distance, but not at the point of observation"

You made some interesting points, and I'm impressed that you actually did the experiment.


Tony


Tony

 

[GTrax]

Clearly there are two aspects to this thread folk are interested in.
re: Detecting sound
Detecting a sound pressure wave "at a distance" requires that some of the energy, however minute, makes it to the detector. There are serious attempts mixing research from biology, nanotechnology, and electronics to make ultra-sensitive microphones and artificial cochleas. Microelectronic etched lever structures on piezoelectric substrates are one type. You cannot probe a sound pressure wave "at a distance" unless maybe the medium is locally modified and probed by some (say) laser light. Search for "mechanical cochlea" keyed with "MEMS devices"

re: The tinnitus experiment.
What I tried wasn't that controlled, and started out as an attempt to check if the relatively clean pitch tone "heard" had frequency at or near the known fall-off in the frequency response of my ear. They do coincide, but that does not actually prove much. I was out to find a way to establish whether the "sound" was a brain-generated fault condition tricking the mind into delivering the perception, or whether any real mechanical vibrations of any part of the living bits were involved. The little vibrating cilia might be accompanied by types that can self-vibrate (flagella) as part of the built-in mechanisms for processing arriving sound pressure waves - but I don't know

Is it a "pure" tone? No is the answer. Definitely not ! By clamping my teeth together, I can alter the character of the note without altering its frequency.

When you hear a squeak as high as 8kHz, its much harder to tell its not pure as compared to say 440Hz. Almost anyone can spot as little as 2% clipping distortion in a 440Hz tone. It takes more distortion of a triangular wave before we become aware of it. Third harmonic content is very obvious, even in small amounts like 1.5%. Square wave also. Even harmonic distortion is harder to tell. It just makes the note "richer" and "nicer". I do not know why it is that even non-musical folk can instantly tell if frequencies are in simple number ratio, and can instantly spot a discord.

Is it purely "in the brain" - no real sounds involved?
I think so yes. That is what I was trying to prove to myself.

When you live with it day in,day out, you get to know stuff about it. I knew that mixing signals linearly using two sine tones (generated from a music editing suite program) just let's you hear two tones together, provided the playback earphone does not have serious non-linearities. Making a perfect sum waveform was not going to work for what I wanted to test, so I used both Sony (L & R) earpieces coupled into plastic pressure tubing and a T-piece, similar to those washable hygiene types used in hospitals. The PC provided one sine tone, and an old audio signal generator supplied the other. Then I simply stuck the tube up to my ear.

Consider when you hear "beats", say from two guitars near each other, both twanged, one slightly mis-tuned relative to the other. Are the "beats" caused by the sound pressure waves from one source messing with the vibrations of the other source in a non-linear way? .. or do the waves arrive and mix and interfere with each other producing "beats", as opposed to simply adding in pure fashion?

When I tried the single tone, it was easy to match the frequency, but I did not hear "beats". Even at high levels of single tones, especially exploring downwards from 8kHz, I could hear the tinnitus, and the tone, as separate. When I tried the 8Khz and the 7600Hz generated tones together, I could actually hear the 400Hz component in there. The perception of beat products might be something built into the brain, or it might have been the waves reaching each other's transducer and causing non-linear mixing. I could even produce low frequency throbbing. In all of this, the one thing clear was that to hear a beat note required two real generated pressure waves. The tinnitus bit was always separate. No amount of single tone would mix with it to cause "beats"

I accept that this was a rough and ready homebrew thing, and that I could be wrong about it all, but I concluded that hearing beat notes required real sound pressure waves interfering to produce envelope products to land at the ear, and that the failure to make beat notes with tinnitus was evidence that the sound heard was "only in the brain".

Let us not stress the word "only"! When the hearing is damaged, the brain tries to compensate. I do not know what pilot signal tricks the mind into tinnitus, but I do think we can explore its properties. The ear mechanism should be seen as an extension of the brain. What in our minds gives us awareness of sounds is unknown, but it does not require real pressure waves (dreams, tinnitus). It does interact with the signals from real pressure waves at our cochlea transducers to give us hearing. Tinnitus is tragic to live with, but I am maybe lucky that the tone is so high, and not so loud, and that my "in the mind" sounds have not yet taken the form of voices telling me things. :|

 

[tonyjeffs]

GTrax,
Are the "beats" caused by the sound pressure waves from one source messing with the vibrations of the other source in a non-linear way? .. or do the waves arrive and mix and interfere with each other producing "beats", as opposed to simply adding in pure fashion?
Here's a fairly obvious experiment that I haven't tried or heard of anyone doing, though I imagine it must've been done...
What happens if you put an 8000 hz pure tone into the left ear say 8400hz into the right ear. Does the listener hear a beat frequency?

...
Going off on a tangent that you might have some thoughts on, I don't believe Von Bekesy's theory/observations on frequency detection in the cochlea. The idea of a traveling wave, whereby sound frequency is somehow converted to distance along a membrane doesn't seem to me to have any basis in the physics i know. I've never heard of such a thing happenning in a shallow stream, a beach, a guitar string, a home made flute.
So I think Bekesy is wrong.
If he was right, there would be cases where the frequency of neuronal hearing loss shifetd a little as the Basilar membrane changed over time, but I don't think that ever happens.

Tony

I can't imagine ho

 

[GTrax]

Hi tonyjeffs

Better to try 8000 along with 7600. Only young children hear up there. Also, the contribution above 8000Hz more modifies the quality of the sound by enriched harmonic content allowing faster rising edges to waveforms. Hi-Hat cymbals and percussive sharp sounds are different. There are not many sounds that use 8kHz fundamental. The majority of most sound power is between 150Hz and 2000Hz.

There is evidence that sound processing occurs, even at the cochlea. The "hairs" are in groups of various lengths, which are joined at the tips by a minute protein strand. Lack of connective strands is accompanied by deafness, and sound filtering of some sort is implied. The movements are in picometres - apparently almost of atomic radius order. There is lots I still have to read about it.

Now that my PC is no longer using a Microsoft system, the music edit program is no longer available. I need to load a new sound tool to generate the tones when I get opportunity. Then I will try the two-ear experiment. :)

 

[mdshaver]

I have only briefly read some of these posts but in response to the OP with regards to the tinnitus and concerns of a non-neurological origin, I would like to comment on that. In the vast majority of cases, the origin of tinnitus is in the cochlea or located more centrally and does not warrant any medical concern. In cases where the tinnitus is unilateral, it should be taken more seriously especially if there is not a known reason for its unilateral occurrence (e.g., noise trauma). In these cases, it could be due to a vestibular schwannoma, Meniere's disease, or a variety of less common pathologies. It also may be vascular in origin. Most tinnitus is subjective (only patient can hear it); however; it some cases it may be objective. In the unlikely case that it is an objective case of tinnitus, I would recommend looking for spontaneous otoacoustic emissions (SOAEs). These are very low-level, narrowband sounds that are normal in ~60% of people; however, if the amplitude is much higher than expected and the frequency is related to the user's perception of the tinnitus pitch then it may be an objective case of tinnitus. The vast majority of tinnitus may be due to damage to the auditory efferent fibers which innervate the hair cells and auditory nerve fibers allowing for enhanced frequency selectivity and amplitude compression. The activity of the auditory efferent system is thought to explain many of the phenomena that could not be accounted for by Bekesy. The auditory efferent system plays an inhibitory role and with damage to this system it is possible that the discharge rates of neurons increases causing the perception of tinnitus. I do not do research in the area of tinnitus so my facts may not be entirely accurate but I think they are pretty close. It is interesting to note that in a study where patients with intractable tinnitus underwent auditory nerve section, only ~50% report reduced tinnitus perception. This, along with brain imaging studies, suggests that there may be cortical representation of tinnitus in many persons.

 

[mdshaver]

Oh btw, if anyone wants any pdf articles on tinnitus research or tutorials on tinnitus assessment, management, or mechanisms, I'm pretty sure I've got some. My email is mdshaver@wichita.edu.