Chapter 10: The Ear and Perception

According to what we have learned so far, the fundamental frequency of vibrations in the air is the biggest factor in determining the perceived pitch of a note and the energy per second per square meter (intensity) determines how loud the sound is. We also know that our perceptions do not perfectly match frequencies and intensities measured by laboratory instruments. In this chapter we look at how the ear turns vibrations into the perception of sound. Some of the exact details of this process are still not completely understood but the general picture of how we hear is fairly well established. Once again we will see that the human hearing mechanism gives us experiences that do not correspond exactly to laboratory measurements.

Key Terms:

Pinna, middle ear, inner ear, timpanic membrane, ossicles, cochlea, semicircular canals, vestibular nerve, auditory nerve, inner hair cells, outer hair cells, place theory of hearing, timing theory of hearing, missing fundamental, auditory illusions, attack frequencies, critical bands, conductive hearing loss, otoacoustic emission, sensorineural hearing loss, presbycusis.

10a: Structure of the Ear

The image below shows the main parts of the ear. The outer ear or pinna captures sound and funnels it into the auditory canal where it reaches the timpanic membrane or eardrum. The eardrum changes the vibrations in the air into mechanical vibrations which travel through as small set of bones called ossicles (the malleus, incus and stapes) in the middle ear. These bones are connected to the cochlea which is in the inner ear. The cochlea changes the mechanical vibrations into nerve impulses that travel to the brain.

The outer ear or pinna, the part of the ear you can see, funnels sound into the auditory canal which travels to the eardrum. It is thought that the curvature of the outer ear allows us to judge the direction from which sound is coming. Experiments have been done with plastic models of a head with the ears attached backwards or with different shapes and microphones where the auditory canal is located. When recorded sounds from these microphones are played back to a real person using headphones, their sense of where sounds are coming from change drastically, indicating that the shape of the ear does have an effect on the perception of sound direction. A few more comments on how we perceive the direction to a sound source are found in Chapter 16: Acoustics.

When sound reaches the eardrum or tympanic membrane it starts to vibrate. A tiny bone called the mallus is attached to the back of the timpanic membrane and carries the vibrations to the incus and then to the stapes. The stapes connects to the cochlea so that vibrations in the three bones cause fluid in the cochlea to move.

Air vibrates more easily than fluid in the cochlea so there is a problem getting energy to move from the air into the liquid inside the cochlea. The problem of getting vibrational energy to move from a medium to one with different transmission properties is called impedance mismatch and we will talk about this in more detail when we talk about musical instruments. The structure of the middle ear has two mechanisms to overcome this problem. The three bones (together called ossicles) act like levers so that a small movement of the timpanic membrane becomes a larger movement of the stapes. The ratio is about 1 to 1.3, in other words, if the eardrum moves 1 mm the end of the stapes connected to the cochlea moves 1.3 mm. Recall from chapter one that pressure is proportional to area. The eardrum has 17 times the area of the base of the stapes. Because of the difference in size between the eardrum and the base of the stapes and the extra leverage in the ossicles the pressure at the cochlea is 17×1.3 = 22 times the pressure on the eardrum. These two mechanisms overcomes the problem of impedance mismatch between the air at the eardrum and the fluid in the cochlea.

The three bones in the middle ear also have muscles attached to them. If a very loud sound comes into the ear the muscles tense up in an automatic reflex to dampen the vibrations of the three bones. This helps protect the inner ear from damage from sudden intense sounds.

The cochlea has two functions. The three semicircular canals are oriented at 90 degrees to each other and give us our sense of balance. The canals are filled with a fluid. When we move our head the inertia of the fluid causes it to remain stationary while the body moves. This relative motion causes small hairs at locations called ampulla inside the canals to be bent, sending signals to the brain via the vestibular nerve. Two other locations, the utricle and the saccule have similar arrangements of small hairs that detect motion. There are three canals to account for rotational motion along three axes; left to right rotational head motion, up down head motion and tilting head motion to the left or right. The bending of the hair cells in the five ampulla due to motion of the head is shown in the picture on the right.

The second function of the cochlea is to convert mechanical vibrations of the fluid inside into nerve impulses that we perceive as sound. The cochlea is basically one long tube, folded over so that sounds go up one side (the scala visibuli) and back down the other (the scala tympani). Vibrations start at the oval window where the stapes is attached and travel through the tube to the apex, then go down the other side until they reach the round window. The tube is coiled and embedded in the bones of the head. The round window is a thin, flexible membrane that keeps pressure from getting too high inside the cochlea. The section between the two halves is called the basilar membrane. The basilar membrane is thickest at the apex where the tube doubles back in the other direction and thinnest at the oval and round window end.

Below is a cross section of the basilar membrane which divides the two parts of the cochlea. If the right frequency is present this part of the membrane will undergo resonance and vibrate. When the membrane vibrates the part called the tectorial membrane tries to remain stationary because of inertia. The flapping together of the basilar membrane and the tectorial membrane bend the tiny hairs between them.

There are about 16,000 hair cells on the basilar membrane. Two sets of hairs are found at each location on the basilar membrane, both of which are connect to nerve cells. Bending of the hairs causes a chemical channel to open and the nerve sends a signal to the brain along the auditory nerve. The inner hair cells are mainly responsible for hearing a particular frequency. The outer hair cells also bend and cause nerve impulses to be formed when the basilar membrane vibrates. These impulses, however, act to modify the signals being sent by the inner hair nerve cells. The result of this is a single signal being sent to the brain with information about the frequency being heard.

One interesting property of the outer hair cells is that they react to incoming sound by contracting. This contraction can actually produce sound that can be detected outside the ear. It is possible that the produced sound acts to reinforce and sharpen the response of the inner hair cells. The process of the ear producing sound is called otoacoustic emission and can be used to test for hearing loss. Sharp clicking sounds can be presented to the ear (including the ears of babies and children too young to respond to a hearing test) and if the ear is working correctly, there will be a sound response coming from back from the ear that can be detected with a small microphone near the ear.

Video/Audio examples:

Overview YouTube video of how sound turns into nerve impulses.
Simulated graphics of how sound is transmitted from the outer ear to the cochlea.
Lectures from Interactive Biology YouTube series:
1. How sound is transmitted from the outer ear to the cochlea (12 min.);
2. How the cochlea works (9 min.);
3. The organ of corti (8 min.);
4. The function of the hair cells (5 min.).

10b: The Place Theory of Hearing

The ear-brain system is a complex instrument. Currently there are two overlapping theories of how we hear; the place theory of hearing and the temporal theory of hearing. Neither of these concepts alone appears to be sufficient to explain the richness of auditory phenomena that we experience.

We know from the structure of the cochlea that different parts resonate at different frequencies; the end closest to the stapes resonates at high frequencies and the end furthest from the ossicles resonates at low frequencies. Nerves are connected to hairs located along the cochlea which are stimulated when vibrations are present. A logical conclusion is that each place in the cochlea corresponds to the perception of a given frequency. This is called the place theory of hearing

How might the place theory work? Below is a screen shot from the driven springs simulation you experimented with in Chapter 4. There are five masses but only one (number two) has very much of an amplitude at a driving frequency of 1.5 Hz. If we drive the masses at a different frequency one of the other masses will have a large amplitude but not mass number two. In other words mass number two has a resonance frequency at 1.5 Hz. (Go back and experiment with the simulation if you don't remember what was going on in this simulation.).

We can imagine this as a model of the cochlea, straightened out with the stapes end on the right. The less massive parts of the cochlea (to the left- opposite of the masses in the simulation) have high resonance frequencies and the part furthest from the ossicles (to the right in the picture) have lower frequency resonance. A given frequency presented to the cochlea only causes motion in one part of the cochlea. This part of the cochlea sends a nerve impulse to the brain and we perceive that frequency.

A problem with the place theory is that the resonance curves turn out to be very broad and they overlap, as shown in the graph below (compare with the resonance graph of amplitude versus frequency in Chapter 4). In other words the sections of the cochlea are low Q-factor resonators. This would seem to make it very difficult for the ear to pick out frequencies which are close together but we know that the just noticeable difference in frequency is about 1 Hz for frequencies lower than 1000 Hz for most people. If the place theory of hearing was correct we would expect that changing the frequency from, say 250 Hz to 240 Hz would shift the region of the basilar membrane that vibrates and trigger different nerve cells. But what actually happens is that the region that vibrates at 250 Hz overlaps with the region that vibrates at 240 Hz to such an extent that pretty much the same nerves are firing. The figure below shows four nearly overlapping resonance curves for four different frequencies.

We also know from the uncertainty principle, discussed in Chapter 9, that a sharp resonance would mean less information about the duration of the sound. If sharp resonance peaks (high Q-factor for the cochlea) were the mechanism that enabled us to hear frequencies that are close together, we would not be able to hear sudden changes in frequency. But we are easily able to hear frequencies that change in less than a tenth of second so the resonance peaks cannot be very sharp which means the place theory cannot be the whole story of how we hear different frequencies.

A possible explanation to save the place theory might be that neighboring nerves are inhibited by the nerve firing at the center of the excited region. This is known to happen in touch and vision (for example our touch sensation is more localized than the actual number of skin cells that are activated) but has not been shown to happen in hearing.

10c: The Temporal Theory of Hearing

A second theory of hearing is called the periodicity or temporal theory of hearing. In this theory it is the timing of the firing of nerve impulses that carries information about the perceived pitch. A simple sine wave at 500 Hz repeats with a period of 1/500 Hz = 0.002 s. The simplest form of the theory says that the vibration causes a nerve to fire every 0.002 s, sending a signal to the brain that is interpreted as a 500 Hz sound. Different sections of a complex sound waveform also repeat periodically. We would expect that hair cells might also fire with the same periodicity of sections of the waveform. For complex waveforms there might be more nerves firing for larger amplitude parts of the wave so that information about the wave shape is also transmitted to the brain.

One difficulty with this theory is that the nerves attached to the hair cells in the cochlea don't seem to fire as often as the theory would predict (and can't fire at a rate of 20,000 Hz, at the high end of human hearing). For example instead of firing every 0.002 s for a 500 Hz signal the nerve might initially fire at a period of 0.002 s but then skip to 0.004 s intervals, 0.006 s intervals or some combination of these periods. This could still transmit the correct information, however. The brain could possibly interpret a sequence of 0.002 s, 0.004 s, 0.006 s firings etc. as multiples of 0.002 and conclude the real signal is 0.002 s. Another idea is that the nerves in the cochlea itself somehow filter and combine signals to send a message that is read by the brain as the signal for a 500 Hz signal. There are two locations of hair cells inside the cochlea. It is possible that impulses from the outer hair cells could affect signals from the inner hair cells in a feedback process to produce some combined signal that has information about frequency and waveform. Alternative schemes explaining how the temporal theory could still work with a nerve firing rate not equal to the period of the signal have been proposed but as yet there is not enough experimental evidence to settle the issue.

10d: Hearing Loss

Hearing loss can occur at several points along the path between outer ear and nerve impulse sent to the brain. If the outer ear fills up with wax or is damaged, fewer vibrations will reach the ear drum. Very loud sounds may damage or even burst the eardrum. Although a burst eardrum usually grows back, it may grow back with scar tissue that changes its ability to vibrate properly. The inner surfaces of the middle ear can also be damaged by loud sounds and some diseases (for example inner ear infections). This may cause scars to form in the middle ear region that block or modify the transmission of vibrations along the bone passageway.

Another type of hearing loss involves problems with the three bones in the middle ear. If they become disconnected (which could happen because of a very loud sound) or if they get calcium deposits due to aging they may not effectively transmit vibrations to the cochlea. The above types of hearing loss are called conductive losses and they sometimes can be repaired by surgery.

Damage to the cochlea or the auditory nerve generally cannot be fixed by surgery. This type of hearing loss is called sensorineural loss. If a very loud vibration causes the stapes to pierce the cochlea vibrations will no longer be efficiently transmitted to the hair nerve cells. Likewise a very large vibration inside the cochlea may break the hair cells so that they can no longer bend in response to vibrations of the basilar membrane. Aging also affects the hairs by making them gradually become stiffer. This means they do not react the same way to vibrations of the basilar membrane or may even break off.

Everyone gradually loses some of their hearing as they get older. Age related hearing loss is called presbycusis and is the most common form of hearing loss after damage done by loud sounds. Both exposure to loudness and aging generally affect higher frequency perception more than lower frequencies. The typical 25 year old male has lost about 15 dB for frequencies above 6,000 Hz and the 25 year old female has lost 5 dB in the same frequency range. Recall that doubling the loudness (two instruments instead of one) is a 3 dB increase so decreasing your hearing ability by 3 dB decreases the perceived intensity by about half. By the time you reach 55 years of age the average person has lost about 30 dB in loudness for frequencies above 4,000 Hz if you are male and 10 dB if female. On the other hand, frequencies under 1000 Hz are generally not affected much by aging.

Some types of hearing loss can be overcome by amplifying the sounds before they enter the middle ear. This is done by the use of battery operated hearing aids. Hearing aids generally only amplify sound and so may not be helpful if the loss of hearing is in a certain frequency range. For Sensorineural loss it may be possible to implant wires directly into the cochlea or the auditory nerve region and attaching an external microphone. This is called a cochlear implant.

Video/Audio Examples:

Mayo Clinic on hearing loss hearing loss.
YouTube discussion of tinnitus (permanent ringing or noise in the ear).
An online hearing test from the University of New South Wales.
Two tests of frequency range of hearing: one, two. Note: Your computer speakers won't be able to play the lower frequencies.

10e: The Missing Fundamental

One important auditory phenomena that cannot easily be explained by either theory of hearing is the missing fundamental or virtual pitch. Some sources refer to the effect as residue pitch. As we saw in the chapter on Fourier analysis, musical instruments have overtones which are harmonic, meaning they are multiples of the fundamental frequency. When you hear a guitar play a note with a fundamental frequency of 100 Hz there are harmonic frequencies of 200 Hz, 300 Hz, 400 Hz, etc. present in the sound. Your ear-brain system perceives this as a single pitch of 100 Hz instead of a collection of individual frequencies. An interesting auditory phenomena is that if you are listening to a series of frequencies of 200 Hz, 300 Hz, 400 Hz, etc. your ear-brain system will ALSO perceive a single note of 100 Hz even if the 100 Hz fundamental is missing.