Does our hearing generate neural signals for harmonics?

Question

First off, the (unrolled) length of our cochlea maps out frequencies such that high frequency sounds cause excitation (neural signals) from the front part of the cochlea, and low frequencies result in neural signals from the end part of the cochlea. Thus we can associate the neurons at each position along the cochlea with a specific frequency.

The aspect is that I'm unsure of is whether a sinusoidal wave will cause the neural signals from the regions that correspond to the harmonics of the wave's frequency. I.e. when a person listens to a 440Hz sine wave, the neurons corresponding 880,1320,1760... are activated. I seem to recall this idea from somewhere, and this claim is made in this video, but it is unsubstantiated.

Basically, I'd like to know how (and by whom) this phenomenon has been studied, in order to better understand the physiological basis for the sensations of harmony.

Answer 1

When you use a pure sinusoidal excitation (at reasonably low levels) you will only excite the basilar membrane at a relatively small area and only the neurons associated with that area will be active. At higher levels the middle ear can become non-linear by itself so you will see some harmonics and the according neural activity as well.

Establishing a one to one relationship between specific neurons and pitch perception is overly simplistic. The mechanical properties of the basilar membrane don't allow for as much frequency selectivity than humans actually have. There are a fair bit of other physiological, neurological and cognitive processes involved as well.

If you want to deep dive, I suggest starting with the work of Georg von Bekesy (http://en.wikipedia.org/wiki/Georg_von_B%C3%A9k%C3%A9sy) who pioneered this field of research and actually got a Nobel Price for it. A quick summary of his work can be found here http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3449028/

Answer 2

Well, the excitation of the receptors happens according to their sinoidal components, and as long as no significant of energy is being dissipated in the hearing organs and attenuation responses (like the coupling in the middle ear) happen on a larger time scale than the vibration itself, the response of the various ear mechanics is essentially linear. And that means that the sensory receptors for 880Hz... will not get triggered.

With regard to following neural path ways, signals are highly correlated with their harmonics, so the following processing pathways will likely get coexcited. If you are looking at a partially obstructed circle, your brain will complete the picture of the circle and actual neurons usually related with the obstructed circle path will also fire.

Now some stuff will obviously not fire. Locating the source of a sinoid signal is actually pretty hard since the wavefronts of the harmonics and the different behavior when turning the head and the auricle are used for location pointers. But without a good sense of location, neural signals will be inconclusive in some respect.

The respective neurons will be tuned to be "trigger-happy", but there is no concrete signal that would make them fire in an information-carrying pattern in connection with the neurons triggered by the fundamental.

Answer 3

Re. the "by whom this has been studied": As I recall the local deity is Sethares. Do a search on the Alternate Tuning Yahoo Mailing List (https://groups.yahoo.com/neo/groups/tuning/info) for his name, 'cochlea', 'ERB', or a related concept 'harmonic entropy', and you'll find more babbling than you'll ever want to read about it.

Answer 4

Yes for loud enough sounds. Fundamentals of Muscial Acoustics (A. Benade 1975) describes this in Chpt. 14. The bare bones study involves presenting a loud sinusoidal sound to the subject at a given frequency, and then adding in a (sinusoidal) search tone. When the search tone is close to one of the harmonics of the main tone, beating between the search tone and the induced harmonic are heard by the subject. More careful experiments by J. Goldstein in the late 60's early 70's allowed for the estimation of the effective amplitude of these components as a function of the strength of the original tone.

This book also cites (as a note) "A good survey of heterodyne phenomena in the ear can be found in four papers by Donald D. Greenwood: "Aural Combination Tones and Auditory Masking" J. Acoust. Soc. Am 50 (1971): 502-43; and three that appear together in J. Acoust. Soc. Am. 52 (1972): 1137-67: "Masking by Narrow Bands of Noise in Proximity to More Intense Pure Tones of Higher Frequency", "Masking by Combination Bands;" and "Combination Bands of Even Order".

Answer 5

It's as if you saw a treestump that looks like a human. Audio neural response is similar to image neural response. Your brain will definitely see a human where the tree is, and your brain will ALSO experience 880Hz and related subfrequencies of audio where the 440Hz is because they look similar.

Difference between similar shapes and similar sounds, the tree requires your full attention, and the 880Hz subfrequency scan/comparison happens more efficiently, briefly and unconsciously, at the same speed that you can differentiate an oak leaf and a fig leaf, nearly instantly.

Sound will cause a complex pathway of neurons to signal to previously connected neurons which will attribute conscious memories and feelings to the sound. The brain has to compare the tone with other known tones, in a tree of neurons through wich the sound travels...

So the first neurons can identify "static pitch / glissando", and "frequency spectrum of tone" ... measure loudness of "880 / 3rd / 4rth partial" = none. You would think "sounds like a perfect sine wave"

So your neurons project a variety of possible imaginary variations onto the nerve signal that your ears receive, and you imagine for a second that the sound is them, just like you can see a tree that looks like a cat, or a person walking and think "it's a person walking" ... "oh it's only a tree" your mind actively experiences other sounds and signals and compares them to incoming data...

You imagine other sounds consciously or subconsciously(it's felt as you becoming tuning awareness of something "what's that?"), so effectively, your brain is comparing the sound to all those other frequencies, but it is not activating full stimulus for them, only sensing if that frequency region is responding to the sound.

with a-lot of the human musical response being interpreted by the vocal interpretation structures of humans which are 10-100 times bigger than animal sound-interpretation regions (for voice communication).