CSD 4417 Final Notes.docx

33 Pages
138 Views

Department
Communication Sciences and Disorders
Course Code
Communication Sciences and Disorders 4417A/B
Professor
Ewan Macpherson

This preview shows pages 1,2,3,4. Sign up to view the full 33 pages of the document.
Description
CSD 4417 Final Notes Inner Ear Hair cell types: • In humans and most mammals there’s one row of inner hair cells and three of outer • The BM goes up and down in response to pressure signals varying across it • The tectorial membrane is gelatinous and its hairs bend when the BM moves • Nerve fibers come off outer and inner hair cells Two kinds of hair cells: • One row of IHC – 3500 cells o Each of the IHC has about 10 auditory nerve fibers attached to it o These have different diameters, different spontaneous rates and different threshold so they are able to cover a wide range of amplitudes of sounds then if they all had the same threshold • Up to 5 rows of OHC – 12000 cells • Both have stereocilia • The up and down motion of the BM causes shearing motion of tectorial membrane • For the most part the IHC are sticking out into the endolymph • Row of IHC along entire length of BM • Inner tunnel of corti is filled with cortilymph • The tips of the stereocilia of the OHC are embedded in the tectorial membrane • OHC change the mechanical properties of the TM to sharpen frequency responses • IHC transduce mechanical vibration into electrical activity in the auditory nerve • The BM analyses sound into spectral components IHC vs OHC • Same parts, different shapes o Inner are more like a light bulb o OHC have more rectangular, elongated shape with three rows of chevron pattern stereocilia 1 • At the top the part exposed to endolymph is stereocilia (similar to cilia, but have no internal structure and are non- motile) • Synapse between nerve fibers and hair cells to carry info • IHC are responsible for turning the movement of the BM into changes in firing rate of the auditory nerve • OHC are anatomically and physiologically quite different from IHC o Act as tiny motors that amplify the mechanical movement of the BM o Responsible for the sensitivity, wide dynamic range and sharp tuning of normal hearing o Their damage is the most usual because of sensory-neural hearing loss • Both cells have tiplinks, filaments on the tip of stereocilia to convert motion to electrical signal Depolarization of IHC activates ascending afferent neurons • BM moves = stereocilia bend • Endolymph is full of positive potassium and calcium ions • Near the top of all stereocilia are ion channels that let in positive ions when open • They are opened when the stereocilia bend and springs pull on the ion channels causing them to reconfigure and open allowing ions to flow into the cell • Influx of positive ions changes the voltage of cell = depolarization • This converts it to an electrical signal mapping the motion of the BM The BM vibrates due to the motion of the oval window • Pressure difference between the scala vestibule and tympani (top to bottom) causes the BM to go up and down • Depending on the frequency, different parts of the BM vibrates • Any spot on the BM vibrates with the same frequency as ear drum and ossicles, but the amount of vibration is different depending on the area • Closer to the apex there’s lower frequencies o If we lower the frequency of the signal putting in, it will move towards the apex Traveling Waves • Characteristic frequency – the frequency to which a particular place on the BM is tuned o Yost defined this as critical or central frequency • Mapping of frequency to a place of vibration on the BM • If we look at one of these spots over time with one frequency we can trace the envelope of the traveling wave 2 o The peak of the traveling wave traces an outline or envelope o The wave is not propagating down the BM • Note: the traveling wave doesn’t correspond to the movement of any material, it’s a consequence of each place on the BM moving up and down in response to stimulation • The frequency of vibration at each place on the BM is equal to the frequency of the tone • The higher the intensity of the sound = higher amplitude of wave • If you put in two tones you get a traveling wave that is high in amplitude in two places Traveling Wave envelope • The envelope indicates the max displacement of the BM at each point along the BM • The up and down vibration of each point is sinusoidal at the stimulus frequency, amplitude and phase vary along the length of the BM • The apparent traveling wave motion along the BM is not sinusoidal • The wave is asymmetrical – it builds up slowly then dies away quickly • Low frequencies – get more spread out wave than higher frequencies (takes up more of the BM) Traveling wave envelopes and frequency • Lower frequencies can cause a larger region of the BM to vibrate substantially o It uses more length of the BM per Hz change in characteristic frequency • The amplitude of the traveling wave = the envelope • Red bars show a region where the vibration is emphasized at each frequency Different ways of drawing the traveling wave • Just representing the positive part of the envelope • Amplitude of vibration on different spots on the BM 3 • Lower frequency = closer to the apex • Higher frequencies have more narrow peak • Waves are asymmetrical • If we play 100 Hz tone the peak will vibrate the most and closer to the tails will vibrate less o We can make the 400 spot vibrate at the 100 Hz level pretty well • But if we play the 400 Hz tone it doesn’t vibrate as much at the 100 Hz spot where the overlap • It’s easer to get low frequency to leak into high frequency channels then the reverse o Low f sounds can interfere with your ability to hear high f sounds  It has to travel all the way down the BM before it peaks and dies as opposed to high that dies off quickly Evidence for auditory filterbank theory: von Bekesy’s experiments • Observed BM motion in human cadavers • Observations near the cochlear apex • Used intense sounds to elicit responses big enough to see under the light microscope Traveling wave characteristics • Always starts at the base and moves towards the apex • For a tone, all of the BM vibrates at the same frequency, but amplitude and phase of the vibration change along the length o Sound waves travel quickly in fluid so all places on the BM are stimulated simultaneously, it’s the consequence of the filtering properties of the BM • The position along the BM at which its amplitude is highest depends on frequency of the stimulus Characteristic frequencies on the basilar membrane are logarithmically spaced • Tonotopy: frequency specific tuning • Low tones relatively spread out – high tones squashed together • Physical properties changed from base to apex • High frequencies are more squished together • Low tones are relatively spread out, high ones are squished together Low frequency coding by place Cochlear Mechanics 4 The BM response looks like band pass filtering • CF = characteristic frequency = lowest input level required • Measuring what level of sound you have to put in to get a certain spot to vibrate a specific amount at different frequencies • Stimulus must be high level to achieve the desired BM displacement at frequencies about the CF • The frequency response of this place on the BM looks like an asymmetrical band-pass filter o Extremely steep high frequency roll off and more gradual low frequency roll off Tonotopy: frequency selectivity of the BM • Each sound frequency causes maximal displacement/vibration in only a small region of BM • Each region of the BM has its own characteristic frequency • Map is preserved along the whole auditory system Why does the BM show frequency selectivity • From the base to apex BM: o Width increases, mass increases o Stiffness decreases, k decreases • f =1/2pi * sqroot(stiffness/mass) o Both mass and stiffness change to decrease resonant frequency • Heavier mass on a spring = slower vibration • Stiffness of the BM really drives it o Heavy and floppy at the apex and stiff at the base o Mass only changes by a small ratio Cochlea as a series of filters • Cochlea acts as a series of filters that are distributed along the length of the BM who resonant frequencies drop from base to apex • Resonant frequency changes along BM due to changes in stiffness • IHC at a particular place are more stimulated when the BM motion is greater at that place Passive and Active BM • Active cochlea dramatically increases response for a given input • Active – amplitude is much bigger, amount of BM vibrating is much smaller on passive and more spread out • Still have asymmetry Cochlear amplifier • Von Bekesy measured passive response of the BM (dead = passive), frequency tuning that results from the mechanical properties of the BM 5 • With functional OHC, this passive mechanism is supplemented by active mechanism OHCs actively modify vibrations • Change in length amplifies movement of the BM at a particular place for its own characteristic frequency • Stereocilia are bent and change their length with the hair cells o This exerts a force on the rest of the structure • OHCs respond to vibration by stretching and contracting at a rate equal to stimulating frequency o This serves to sharpen the tuning of the BM • OHCs contract  arch of corti tilts downwards bending the stereocilia in the opposite (inhibitory) direction BM response to a 1 kHz tone demonstrating level dependence • Change the level of the tone and measure vibration on a log scale • At higher tone levels there’s no more sharp peak = return to passive pattern • This spot likes 1 kHz o Measuring away from this area  increasing the level of the tone by 20 dB = amplitude of vibration increases about the same amount (linear) o Near 1 kHz  increase by 20 dB = less than 20 dB change o At the peak the input/output cure is kind of flat o At 1 kHz, high level sounds lose sharp tuning when there’s more vibration, goes back to linear state • This sharpening is due to action of the OHCs, for some reason they can’t keep up with the input BM response at a given place with CF 18 kHz • At low levels we have this sharp tuning, but at higher levels it broadens out • Everything looks like a sharp bandpass filter • Region at CF where hair cells are firing we get compression o This gives our auditory system more dynamic range Active vs Passive Response of BM • Gain increases frequency selectivity • More gain at low levels results in compression • OHCs firing = extra vibration, gain and sharper frequency tuning • If frequency is fixed and the input level is changed: o Play a low level tune we get a linear line at a passive state o BM just vibrating a bit and cells can push and give extra vibration o Works for low levels, but at higher levels they can only move so much so their contribution gets smaller and doesn’t make a difference if you have OHC or not • Passive cochlea without functioning OHC = broad passive tuning (top figure) 6 • When there is OHC we get a lot of gain and sharper tuning seen by peak (top figure) o Don’t get this amplitude gain at frequencies away from CF o When vibration is big, OHC contribution becomes negligible • Bottom figure: o Fixed frequency o Low intensity levels = OHC gives a certain gain o Increase input level and at some point OHC are maxed out and amplification they provide decreases = flatter slope o Eventually they don’t make a difference • Take away: to get the BM to vibrate x amount, we can have a smaller input if OHC are working o One form of hearing impairment – damage to OHC causes less or no gain 9 kHz tone vs 1 kHz @ CF = 9 kHz • 9kHz tone = compression o Steep at low levels – the OHC provide the same amount of gain o Higher levels the OHC start maxing out and contribute less, so curve becomes shallower than 1:1 – region of compression  If we put in 10 dB we get less than 10 dB out o When the slope gets steep again it doesn’t matter having OHCs • 1 kHz – less vibration and less gain o Don’t see any compression at frequency far from proffered o OHC aren’t doing anything because it’s far from preferred frequency Changes in BM compression after temporary OHC deactivation with furosemide • The motion drops down at 11-19 minutes and becomes more linear • As the drug washes out at 40 minutes we see some gain and compression again • At 1 kHz, it doesn’t effect it if the OHCs are on or off The active mechanism • OHCs are relatively more active for low level sounds than higher • The effect is restricted to a small region of BM around the active OHCs who respond only to stimuli near the CF • Increase sensitivity (lowers thresholds) o Because they amplify sound from BM • Leads to compressive loudness growth • Increases frequency selectivity • Produces otoacoustic emissions 7 o Activity of OHCs produce vibration recordable in canal with sensitive microphone o Vibration of the OHCs can travel outwards Otoacoustic emissions • Byproduct of the activity of OHCs in healthy ears • Nonlinear phenomenon of the ear • Used to: o Screen for hearing loss in infants, newborns and young children o Determine whether a hearing loss is cochlear or more central o Monitor medications that are toxic to the ear – abnormal OAEs reveal cochlear dysfunction before actual hearing loss is present o Assess the effects of noise/music exposure in occupational settings • OHCs are typically the first to suffer from noise exposure Transduction Neuroanatomy Auditory biological transducer • Transduction = transformation of energy from one form to another • Sound pressure  peripheral auditory system  neural response • Peripheral auditory system: hydromechanical vibration  IHC  neural response • Pressure oscillates at the ear drum causing acoustic amplification, ossicles move oval window, pressure waves in fluid, basilar membrane moves up and down  stereocilia bend = IHCs depolarize when ion channels open Resting Potentials: measurable regardless of activity in cochlea • Endolymph has a higher voltage than perilymph o Maintained by stria vascularis pumping positive ions (mostly K+) into this region and boosting electrical potential in endolymph  Important that the stria vascularis is healthy to provide concentration of K+ • Inside the IHCs, just like in the OHCs, they try to maintain the voltage inside as negative o Gives a strong driving force to make positive ions go into the IHCs Motion- to-voltage transduction by IHCs • Vertical movement of the BM causes the IHC stereocilia to move from side to side • Movement towards the OHC causes the tip links to stretch  opening of ion channels = depolarizes • When hairs are bent towards the tallest stereocilium = cell voltage is increased = increased firing rates 8 • Stereocilia are direction sensitive, when they are bent away from the tallest  cell’s voltage is decreased IHC receptor potential is low pass filtered • Below 1 kHz, stimulus fine structure is well represented • Above 1 kHz, fine structure is attenuated • Above 5 kHz, only stimulus envelope is transduced o At high frequencies the hair cells can’t tell the frequency of the oscillations and all the carrier Info is thrown away and it just shows the envelope Voltage to chemical to electrical transduction by IHCs and auditory neurons • Depolarization of IHC causes release of NT • Sufficient NT release can create action potentials in ascending afferent neurons • Threshold is generally about 15 mV above the cell’s resting membrane potential, occurring when the inward sodium current exceeds outward potassium • Net influx of positive charges carried by sodium ions depolarizes the membrane potential leading to further opening of voltage gated sodium channels o Channels support greater inward current causing further depolarization = positive feedback cycle that drives membrane to very depolarized level Action Potential • Electrical potential exists between inside and outside of neuron • Cell membrane permeability changes dramatically in the axon when positive feedback cascade is initiated by adequate stimulus • Ions cross membrane and briefly depolarize the cell at one place on the axon • That place is then in refractory period – steady state potential is restored • Depolarization propagates along the axon in one direction • Figure: o Na rushed in brings us to B o Na channels close and K open bringing us t C Two kinds of auditory nerve fibers • Tallest tips of stereocilia are embedded in tectorial membrane • OHC change mechanical properties of the tectorial membrane to sharpen frequency response • IHCs transduce mechanical vibration into electrical activity in the auditory nerve • Each of the IHCs has about 10 auditory nerve fibers attached to it 9 o These fibers have different diameters, different spontaneous rates and different thresholds so they are able to cover a wider range of amplitudes of sound than if they all had the same threshold • Afferent = to the brain • Efferent = from the brain • OHCs amplify vibration of BM • IHC generate AP in nerve fibers connected to them and send it to the brain • Efferent connect directly to OHC, connected to IHC in a more indirect way • Efferent pathways control amount of amplification • Efferent fibers can decrease output from IHCs or decrease motility of OHCs Afferent neurons: ascending • Two kinds of afferent: • Type I radial: o Only synapse with one or two IHCs o About 20-30 fibers per IHC o Thicker (faster) than type II fibers o Shoot straight in through spiral lamina and into the center of the cochlea  Cell bodies in the auditory neurons (bipolar cells) is in the madialis of the cochlea o 80-95% are type I • Type II Outer spiral: o On average synapse with 10 OHCs Summary of innervation • IHC: o Radial afferent o Lateral olivocochlear efferents • OHC: o Spiral afferents o Medial efferents Neural Code – how are stimulus parameters like frequency and amplitude represented using AP? Encoding frequency and amplitude (and time) 10 • All sound info is carried to the brainstem by the auditory nerve, which must encode a sound’s characteristics into electrical impulses • If we can distinguish two sounds then the pattern of impulses in the nerve must be different • The signals are decoded and further processed by higher centers of the auditory nervous system Single unit recordings – spontaneous rates • Electrical activity is recorded near or within single auditory neurons • In the absence of an acoustic stimulus, the unit will have a spontaneous rate at which it fires • Auditory fiber spontaneous rates can be from 0-100 discharges/sec but each neuron has its own average rate Neural thresholds • Spike rate is proportional to basilar membrane velocity (<200 Hz) or displacement o Higher levels cause bigger displacements and higher velocities are needed to travel those displacements within the stimulus cycle period T • A single neuron can only fire so fast due to its refractory period • When strongly stimulated most auditory neurons fire at rates <500 spikes/sec o Firing + refractory period is about 1 ms giving a theoretical max of 1000 spikes/sec • Neural threshold – minimum stimulus level causing a specified increase in discharge rate • IHC respond to BM velocity for frequencies below 100-200 Hz, Above that BM displacement is the effective stimulus for IHC response • For an individual neuron, high discharge rate probably means high stimulus level • But, dynamic range (threshold to saturation) is typically < 35 dB, once saturated, the single neuron can’t tell us more about stimulus level o Dynamic range = range that it can tell us exactly the level of the sound • Difference between threshold and saturation is usually pretty small • Fig. stimulus threshold at 25 dB SPL, saturation above 50 dB SPL • Fibers of different spontaneous rates will saturate at different stimulus levels • Different fibers have different thresholds, dynamic ranges and spontaneous rates – if firing is observed across many neurons for a given part of the BM – level encoding is across a larger range 11 • This is one way that stimulus level may be encoded = rate encoding • Fig. Second fiber (dashed line) has a different threshold, rate and range. Looking at the activity from both of these you can find out more info about the sound because the dynamic range is doubled • We encode intensity of sound by how fast all the individual fibers are firing Spread of excitation on the BM • At higher stimulus levels, more of the BM moves significantly so more neurons will be stimulated • This leads to greater activation in the auditory nerve • Low level sound = small range of nerve fibers firing about threshold, turning up the level = more fibers firing Place theory • One way the ear encodes frequency • Which fibers fire tell which BM place and therefore what stimulus frequency • Stimulus frequency is mapped to vibration in different place • Fibers in nerve are tonotopic • If we put in 4 kHz, then we get activity in fibers around that range etc. Phase locking • Neurons with CF < 5 kHz fire in a phase-locked manner in response to low frequency stimuli o The neuron may fire at the peal of most stimulus cycles • If the stimulus is < 500 Hz, the neuron may fire at the peak of most cycles o Due to the fact that most fibers’ max firing rates are <500 Hz so it makes sense they cannot fire on every cycle for frequencies above 500 Hz o If it does not fire on every cycle there is imperfect entrainement • Due to the refractory period, at higher frequencies the neuron will skip stimulus cycles • Rectification: only one peak of the stimulus waveform causes depolarization and neuron firing • If we get a spike it is always at the peak of the oscillation, but we don’t always get a spike at every peak o This means the timing of spikes has something to do with the phase of the oscillation, these are phase locked • Relationship between timing of spikes and waveform because we are more likely to get them around the peak, can see this when looking at temporal info of phase locked histogram 12 o Time between histogram peaks is the stimulus period T Interval or Interspike Interval (ISI) histograms • ISI histograms show the time between successive spikes for a long duration stimulus • fig the frequency of the stimulus must have been 1 kHz. But there were many intervals where the spikes skipped one or more cycles, this is expected since 1 kHz is the theoretical limit of a neuron’s ability to fire every cycle • Hair cells act like low pass filters, they can’t follow the individual cycles o By 400-500 Hz they can’t follow o This kind of coding doesn’t work at high frequencies • Our ability to discriminate different frequencies is much better at < 2 kHz • At higher frequencies, spikes are more likely to skip multiple cycles of the stimulus o Shows the loss of ability to track the individual cycles o At high frequencies we lose the phase locking Temporal theory • Another way the ear can encode frequency • Phase locking in auditory neurons up to 5kHz • Periodicity of the neuron discharges could be used to determine stimulus frequency • If response is observed across many fibers then skipped cycles are filled in – Volley Theory o A population of nerve fibers does not require an entrainment of every fiber to convey a temporal code • If the brain can time the intervals of phase locking it can estimate the frequency of the tone o Timing theory of pitch perception of pure tones • If phase locking is not occurring, the time intervals will be random and give no info about the frequency of the tone • Phase locking also provides interaural time differences in localization Auditory nerve activity • Info about frequency carried by: o Place coding: fibers servicing a particular place increase firing o Temporal coding: produced by phase locking, only works at low frequencies • Info about stimulus level is carried by: o Firing rate: extent to which low and high spontaneous rate fibers fire o Spread of excitation: more neurons will be stimulated at higher levels Auditory sensitivity 13 Physical Psychological Sound Auditory sensation Intensity Loudness Frequency Pitch Spectral distribution and temporal Timbre characteristics Example of detection • The ear canal amplifies resonance from 3 kHz and up – that is why you hear more at 4 kHz • Hard for low frequencies to get into the middle ear • Threshold of audibility is dependent on the room o Noise background o Standing waves • Proper experiment in anechoic room Absolute threshold • MAF – minimum audible field (yellow) o Loudspeakers o Level measured with mic where listener was o Little dip in the graph is due to ear canal resonance around 3-4 kHz   Ear canal is now amplifying the sound • MAP – minimum audible pressure (blue) o Headphone/insert o SPL at ear drum • The middle ear is like a high pass filter • At 10 kHz, if the normal threshold is 12 dB and someone has a threshold of 25 their hearing would be 13 dB HL worse o Subtract threshold o HL = hearing level; always referencing it to what the normal population is like dB hearing level (dB HL) • Since MAP and MAF are known for normal hearing population this scale is defined with respect to the normal • Across frequency, the normal threshold of hearing is 0 dB HL • Ex. At 1 kHz, an individual’s threshold for detecting a tone is 30 dB HL so their threshold is 30 dB above average normal • Hearing is normal for threshold < 20 dB HL due to wide inter-individual differences • HL can be converted to SPL for MAF or MAP conditions o Ex. Using earphones at:  1 kHz  0 dB HL = 0 dB SPL  200 Hz  0 dB HL = 18 dB SPL 14 Threshold summary • Middle ear, canal and pinna help shape threshold curves • Below 1 kHz, sensitivity decreases at 6 dB/octave • Above 4 kHz, sensitivity decreases at about 24 dB/octave • We lose sensitivity because we don’t have portions of the BM devoted to those functions Weighted Sound pressure (dBA SPL) • “A” weighting is a standard type of filtering that can be applied to a microphone signal • commonly used for measurement of environmental and industrial noise • Used by sound level meters to approximate how humans perceive loudness at modest levels • Eg. A sound level meter may say the room was at 70 dB SPL o This does not fit your perception because you are not hearing low frequency ventilation noises dominating the measurement o In a quiet room there’s lots of sounds at frequencies we can’t hear – we just aren’t sensitive to them Thresholds of hearing as a function of duration • Same amplitude sound played at different durations • As you move to the lower duration you don’t hear the sound’s beeps anymore • 300-500ms is the range where duration doesn’t matter for hearing Problem of criterion • Perceptual decisions are influenced by the person’s sensory abilities and their motivations and expectations • Eg. Willingness to say yes or no in threshold experiments depends on… o Consequences of those responses (rewards and costs) o Likelihood of an event • Performance differences among people or groups of people could be attributed in part to criterion differences Two-alternative forced choice 2AFC • Before we talked about yes-no experiments o Tendency to say yes or no will affect performance (ie. Bias not sensitivity) • Instead, on each trial present two stimuli: one with a tone and one without, subjects report which internal contained the signal o Bias applies equally to the two observations and is cancelled out • When the stimulus is too weak to be detected or the stimulus difference is too small to be noticed, performance will be 50% • Features: o Minimizes bias effects, more confidently estimate sensitivity o Can calculate % of trials correctly identified as containing signal then threshold may be taken as level that gives performance 75% correct 15 o Curve given is a psychometric function • 2AFC tasks can be applied to measuring: o Detection thresholds (is a stimulus present) o Discrimination thresholds (are two stimuli different) • 2AFC must have an assessment of order o Don’t confuse with same-different tasks which are harder to interpret with detection theory and produce worse sensitivity than YES-NO tasks Differential sensitivity • About detecting changes in level or frequency o The just-noticeable difference (JND) o Or difference limen (DL) • Discrimination Weber Fraction • ΔS/S = constant (Weber’s law) • ΔS = the JND and S is the smaller of two values being discriminated • For limited stimulus ranges, the Weber fraction holds for frequency and intensity discrimination Frequency discrimination • A 2AFC type task might be used to determine the JND with a question like: which of the last two intervals was lower in frequency? • Threshold might be chosen as the Δf where participants could correctly answer questions for 75% of trials • At every frequency there’s a place code but only at low frequencies is time coding o This hurts your frequency discrimination Level discrimination • Weber’s rat
More Less
Unlock Document

Only pages 1,2,3,4 are available for preview. Some parts have been intentionally blurred.

Unlock Document
You're Reading a Preview

Unlock to view full version

Unlock Document

Log In


OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


OR

By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.


Submit