Topic 17 – Coding Frequency and Intensity
We’ll spend most of our time on coding of frequency
There were two competing theories of the coding of frequency reminiscent of the competition between trichromatic and opponent processes theories of hue perception.
The basic question is: How does the auditory mechanism distinguish high frequency from low frequency sounds?
Temporal theory, aka Frequency Theory, aka Telephone Theory
In this theory it was assumed the basilar membrane vibrated as a whole, like the membrane of a telephone microphone
Called the telephone theory
Assumed membrane vibrated in unison with sound – higher the sound frequency, faster the membrane vibrated.
Assumed that somehow, the membrane vibration was transmitted to higher neural centers. For example, neurons that fired each time the membrane moved.
Main problem with this theory: We can perceive sounds whose frequencies are as high as 20,000 Hz, but neurons cannot respond at rates higher than 1000 action potentials per second, if that high. So the theory, unaltered, cannot account for our ability to hear sounds above 1000 Hz.
One attempt to salvage temporal theory: Volley principle. Proposed that no single neuron responded with the membrane, but that neurons “took turns” responding, so that each individual neuron responded, say, every 10th vibration or every 20th. This would allow the collection of neurons to signal frequency while not requiring any one to fire at a rate greater than 1000 APs/sec.
Note the premise: Based on idea that the neural activity must be a mirror of the external stimulus.
But there is no evidence of neural activity forming a model of the external world in other areas with the exception that location of activity in the cortex roughly corresponds to location of stimulation in the visual field. So why require it here?
Plus, volley theory moves the problem of having something respond at the rate of the sound stimulus up the neural chain, still begging the question of what neural structure could respond at the rate of the sound stimulus.
Helmholtz’s Place Theory.
Helmholtz believed that the basilar membrane is composed of fibers running at right angles to the length of the membrane.
He believed that these fibers are strung taut, like the strings of a harp.
Sound caused them to vibrate, just as the strings of a harp vibrate in the presence of sounds.
Short fibers vibrate most to high frequency sounds. Long fibers vibrate most to low frequency sounds. So place of vibration is the signal for frequency.
Note that in this theory the fibers don’t have to vibrate in unison with the stimulus. The brain will know what the frequency of sound is by knowing where the vibration is occurring.
Fatal flaw in Helmholtz’s theory: The basilar membrane is not made of taut fibers. It’s more like a sheet hanging between two clotheslines.
So there was no way that the ear could code frequency in the specific fashion proposed by Helmholtz.
Traveling Wave theory – von Bekesy
Von Bekesy carefully examined inner ears of cadavers and built a model of the basilar membrane based on his examinations. (As G8 notes, because his research involved cadavers, he missed some important information about basilar membrane responses to sound.)
His examinations and models convinced him that the membrane is not taut, but fairly loosely slung.
He proposed that the response of the membrane to sound is a wave that travels from the base of the cochlea to the apex.
Like a sheet or rug being snapped to shake off dirt.
Shape of the membrane during its “shaking” is illustrate by this figure
Play VL 11-11 (Traveling Waves) here
The point of maximum “movement” grows in size from the base to the apex – reaching a maximum amplitude at some point - then diminishing in size.
Base Apex
As the basilar membrane “flaps” the hair cells on the organ of corti, located on top of the membrane are bent by the “flap” as it travels down the length of the membrane.
When bent the cilia on the hair cells bend causing release neurotransmitter substance that causes auditory nerves to emit action potentials. Those at the point of maximum movement are bent the most and thus release the most neurotransmitter substance.
Key: The point at which membrane movement is greatest depends on frequency of the sound.
High frequency sounds: Movement is greatest near the base – near the oval window end. Hair cells at the base release the most neurotransmitter substance.
Low frequency sounds: Movement is greatest near the apex. Hair cells near the apex release the most neurotransmitter.
Implications
Each hair cell is a frequency-specific receptor
If a sound is composed of several frequencies, it will activate several receptors – one for each of the frequencies comprising the sound.
Analogy: 100s of different cone types in the eye - a different type of cone for each wavelength of light.
Since each auditory nerve synapses only with hair cells at a specific place on the basilar membrane, this means that each auditory nerve responds to a specific frequency in the sound stimulus. Each auditory nerve is “tuned” to a different frequency. The collection of responses of the several thousand auditory nerves is like a spectrum.
Another way of thinking about this is that the receptors on the membrane perform a rough Fourier analysis of the sound stimulus, reporting the various frequency components of each complex sound.
Spectrum of a complex sound
The resurrection of frequency theory
Apparently, low frequency sounds cause movement of the whole membrane, in unison with the sound. This movement is indicated by responses of some auditory nerve neurons, much as supposed by the frequency theory proposed in the early 1900s.
Coding of intensity.
A given frequency activates auditory nerve neurons on the place on the basilar membrane corresponding to that frequency.
Each auditory nerve neuron responds at a level corresponding to the amount of activity on the basilar membrane.
So the neurons at the place of maximum amplitude respond the most.
Neurons next to those respond less, and so forth as the distance of an auditory nerve neuron from the place of maximum amplitude increases.
Graphically
If the sound is made more intense, ALL neurons respond a higher levels.
Graphically
The solid line shows that the neurons at all places on the basilar member are more active in response to the more intense sound. The place of maximum amplitude is the same, so the sound is heard as the same frequency as the low intensity sound, but since more neurons are active, it is perceived as being louder.
If the sound has low intensity, only a few neurons, those closest to the “place” on the membrane corresponding to the frequency will be activated.
Note that as intensity increases, the activity spreads away from the “place”. This amount of this spread probably represents intensity.
Auditory Pathway – G8 p. 281
Note that the only structures that are monaural are the cochlear nuclei. After that, all structures receive input from both ears – they are binaural.
Detail of the primary auditory cortex. Note that much of it is in a brain sulcus.
Tonotopic maps G8 p. 283
The neurons in the primary auditory receiving areas respond primarily to single frequencies. The frequency with which different neurons respond form tonotopic maps. This means that each neuron responds best to a particular frequency and that adjacent neurons respond to similar frequencies.
As you move away from the primary auditory receiving area, neurons in the auditory cortex surrounding the primary area respond to more complex sounds.
The auditory cortex may be partitioned into areas primarily involved in sound identification and other areas primarily involved in sound localization.
This suggests that the two key aspects of external stimulation – what is out there? where is it? – found in the processing of visual stimuli are also important in the processing of auditory stimuli. That makes sense.
Note that the “what” system originates in the general vicinity of the visual “what” system and that the “where” system may originate in the general vicinity of the visual “where” system.
This is shown nicely in Figure 11.40 in the text, shown above flipped horizontally. The green areas were activated when the observer had to determine the pitch of a sound (“what” sound it was) and the red areas were more active when the observer had to determine the location (“where”) of a sound.
Topic 17: Coding Frequency and Intensity - 8 3/21/5