Speech Science M12-M13

categorical perception experiment

demonstrated that listeners classify acoustic stimuli, such as Voice Onset Time (VOT), into discrete phonetic categories (e.g., /b/ vs. /p/) rather than perceiving gradual changes, even when acoustic differences are equal.

These findings, using synthetic speech and identification/discrimination tasks, shaped theories that speech perception is special, innate, and maps acoustic signals to articulatory gestures rather than just auditory features

Shaped theories because it showed that speech perception is not always direct, but can depend on acoustics and different articulations and phonetics, which leads into future theories. 

pattern playback device and categorical perception

Using the pattern playback device, the finding that a burstless pattern elicited the perception of stop consonants, with the correct place or articulation dependent on the pattern of transitions. 

Listeners identified the correct place of articulation based on these frequency transitions, even when the consonant release burst was removed, proving transitions are crucial, context-dependent

continuous perception

perceived responses gradually change

no sharp breaks, just a smooth linear relationship b/w signal and perceived response

ex. brightness or loudness, where small changes are perceived as small differences

categorical perception

a cognitive phenomenon where continuously varying sensory stimuli (like sounds or colors) are perceived as belonging to distinct, discrete categories.

It enhances perception of differences between categories while making items within the same category seem more similar, reducing the cognitive load of processing continuous, analog input.

categorical perception for stop place

refers to the tendency to hear distinct consonants rather than a smooth continuum, despite continuous changes in acoustic cue

duplex perception

can be elicited for both speech and no speech signals

demonstration of simultaneous activation of the “speech mode” and general auditory processing mechanisms

sound into both ears and brain combines them

brain can process same acoustic info in multiple ways at the same time

trading relations and revised motor theory

trading relations- brain doesn’t rely on just 1 acoustic cue to identify a speech sound but it uses multiple cues and these cues can compensate for each other

ex. say word pay vs. stay, 2 important acoustic cues are 1 closure duration, a longer closure duration makes us more likely to hear stay (stop)

connection to revised motor theory- we don’t just perceive sounds, we perceive the underlying articulatory gesture; brain is trying to figure out how the sun was produced, not just what it sounds like; when acoustic cues vary, we still perceive same intended speech gesture

categorical perception of stop voicing distinction: VOT

VOT- time b/w release of stop consonant and start of vocal fold vibration or voicing

on graph, even though VOT time changes gradually, perception changes suddenly

category boundary- sharp shift, this is where distinguishing sounds right is easiest

takeaway- voice-down time is a powerful example of how small timing differences lead to big perceptual changes and acoustic signal is continuous

motor theory

proposes that we perceive speech by identifying the articulatory gestures used to produce it

articulatory position/gestures

auditory theory

speech perception is based on acoustic signal itself

brain analyzes frequencies, timing, and intensity using general auditory mechanisms, not something unique to speech

involves cognitive processing

direct realism

articulatory gestures from the acoustic signal. direct perception of gestures

it does not rely on higher level cognitive processing

info is present in signal and brain picks it up directly

perception is more immediate

word intelligibility

context plays a major role in how well we understand speech

words in sentences are more intelligible than same words presented alone

brain integrates info across sounds, words, and context

speech intelligibility tests; why are they used

understand why communication breaks down and where the problem is occurring

speaker, medium, and listener

to evaluate a listening medium; medium

to evaluate a persons’ ability to perceive speech; listener

to evaluate a speaker’s intelligibility, speaker

word identification tests (% intelligibility)

percentage of intelligibility

measure percent of words correctly understood by listener

often used in clinical evaluation of hearing disorders, the effectiveness of hearing aids and/or cochlear implants, and as a metric of severity in speech disorders

can take the form of qualitative estimates, or (preferably) formal tests in which carefully chosen words in isolation or words in sentences are presented to listeners for orthographic or multiple choice responses; word lists may be structured to obtain estimates of % phonetic features correct

scaling tests

measure perceived intelligibility more subjectively

visual, excuse me, analog scales, rating scales, magnitude estimates

allow listeners to rate how understandable speech is, mot just whether its correct

capture more than just phonetic accuracy

Equal-appearing interval scales, visual analog scales, direct magnitude estimation (and others)

Scaling techniques allow factors other than phonetic characteristics (such as voice quality) into the estimates of speech intelligibility; for the same group of participants, they tend to provide a wider range of intelligibility than word tests (which are primarily affected by phonetic factors)

Scaling tests are moderately-to-strongly correlated with word tests

phonetic transcription tests

listener transcribes exactly what they hear using phonetics symbols; not so easy

very detailed info about speech specific errors and patterns

difficult to obtain acceptable reliability; requires substantial expertise in transcription

phonetic transcription data moderately correlated with word tests

pinna

outer ear

conductive mechanism

contribute to localization of sound in the up-down dimension

collect movement of sound waves, refraction, and compression and channel it to the external auditory meatus

external auditory meatus

outer ear

conductive mechanism

primarily functions to channel sound waves from the outer ear to the tympanic membrane while protecting the middle and inner ear from foreign objects and pathogens.

•External auditory meatus (EAM) is slanted up from the opening in the concha of the pinna to the termination of the tube at the tympanic membrane

•EAM is a tube closed at one end; the walls of the initial section of the tube are cartilage, the walls of the remaining length of the tube are bony

•At the boundary between cartilage and bone, the EAM turns slightly toward the back of the head; tube is slightly crooked

•EAM resonates like a tube closed at one end; first resonant frequency is ~ 3000-3500 Hz (depending on length of the tube)

tympanic membrane

outer ear\

conductive mechanism

It separates the outer ear from the middle ear and plays a crucial role in hearing by receiving sound vibrations from the air.

The membrane transfers sound waves to the auditory ossicles in the middle ear, which then conduct the vibrations to the inner ear.

Additionally, it acts as a protective barrier, safeguarding the middle ear from dirt, bacteria, and debris.

Overall, the tympanic membrane is essential for both sound transmission and protection of the ear structures

•Tympanic membrane (TM) is composed of three layers of tissue (except for the pars tensa, which is composed of two sheets); when viewed through an otoscope the TM is translucent

•The tissue layer that faces the EAM is part of the outer ear; the tissue layer that faces the middle ear is part of the middle ear

•The middle tissue layer of the TM is composed of densely interwoven, stiff tissue that gives the TM its high sensitivity to air molecule-size vibration; the outer layers are not so sensitive to these vibrations

•Otoscopic image is from the right ear, based on the angle of the cone of light

middle ear

conductive mechanism

transformed from an acoustic signal now to a mechanical

The primary function of the middle ear is to transmit and amplify sound vibrations from the eardrum to the inner ear while equalizing air pressure.

•Air-filled cavity encased in temporal bone

•Cavity contains muscles, ligaments, nerves

•Cavity has two openings connecting it to other air-filled spaces: 1) the aditus, connecting the upper part of the middle ear cavity to the antrum, and 2) the auditory (Eustachian) tube, connecting the middle ear cavity to the nasopharynx

•The ossicles (tiny bones in the middle ear cavity) are the malleus, incus, and stapes. The malleus is attached to the tympanic membrane, the stapes is attached to the bony labyrinth, and the incus connects the malleus to the stapes

•Ligaments run between the ossicles and walls of the middle ear cavity, providing stability; one of these ligaments holds the footplate of the stapes in place within the oval window

•Muscles (2) stiffen the ossicular chain when they contract

•A portion of cranial nerve VII runs through the walls of the cavity

ossicles

conductive mechanism

•Top: disarticulated ossicles

•Bottom: articulated ossicles

•Ossicles articulate by synovial joints—incudo-malleolar joint, and incudo-stapedial joint

•Handle of malleus is attached to the tympanic membrane; footplate of stapes is attached to the oval window of bony labyrinth

•Acoustic vibrations are transmitted across the ossicles from the tympanic membrane to the footplate of the stapes; movement of the footplate into the cochlear fluid starts the process of sound sensation

tensor tympani

middle ear muscle

muscle arises from a bony canal in the front wall of the middle ear cavity, just above the cartilaginous portion of the auditory tube

As the muscle emerges from the tube into the middle ear cavity, it gives off a slender tendon that turns toward the lateral wall (i.e., in the direction of the tympanic membrane); the tendon attaches to the handle of the manubrium

When the tensor tympani m. contracts it pulls the handle of the manubrium medially which pulls the tympanic membrane in the same direction. Contraction of the tensor tympani m. stiffens the ossicular chain

Contraction of the tensor tympani m. may reflect a response to very intense sound energy, and even chewing; the muscle is innervated by motor branch of cranial nerve V (trigeminal)

stapedius

middle ear

arises within the pyramidal eminence, a tiny, cone-like “bump” in the posterior wall of the middle ear cavity

As the muscle emerges from the pyramidal eminence if gives off a tendon that runs upward and toward the anterior wall of the middle ear cavity; the tendon attaches to the crus and neck of the stapes

When the stapedius m. contracts it pulls the footplate of the stapes away from the “fit” in the oval window, preventing the footplate from excessive displacement into the cochlear fluids; action of the muscles stiffens the entire ossicular chain

Contraction of the stapedius m. is the muscular component of the acoustic reflex, an “automatic” response of the auditory system to intense sound; the reflex is thought to protect cochlear structures from excessive displacement of the cochlear fluids and possible damage to the hair cells. The stapedius m. is innervated by a motor branch of cranial nerve VII (facial)

conduction of sound energy from outer ear to cochlea

•For the same applied force, air molecules move a greater distance than fluid molecules: fluid offers greater impedance to displacement compared with air

•In the peripheral auditory system, there is an impedance mismatch as sound energy is transmitted from the conductive to sensorineural mechanism (vibratory force of air applied to tympanic membrane; same force applied to cochlear fluid at the footplate of the stapes)

•Pressure = force per unit area (F/A). The area of the tympanic membrane is 20 times that of the area of the footplate of the stapes, so the same force applied to both surfaces (tympanic membrane; footplate of stapes) results in a much higher pressure at the footplate of the stapes compared with the tympanic membrane

•The impedance mismatch is overcome by the area difference; sound pressure at the footplate of the stapes is ~ 20-25 dB greater than sound pressure at the tympanic membrane. That is, the area difference “solves” the impedance mismatch

•Other possible mechanisms to overcome some of the impedance mismatch: lever action of the ossicles; cone effect of tympanic membrane

bony labrynth

inner ear

sensorineural mechanism

•Bony labyrinth encases fluid-filled ducts and chambers of the semicircular canals, ampullae, vestibule, and cochlea

•Motion of the head is coded by displacement of fluid within the semicircular canals, which stimulates hair cells in the ampullae

•Position of head relative to the body is coded by hair cells in the saccule and utricle, located within the vestibule of the bony labyrinth

•Audition is coded by displacement of fluid within the cochlear ducts, which stimulates hair cells

cochlea

sensorineural mechanism

a spiral-shaped part of the inner ear that helps us hear by turning sound waves into signals for the brain.

•Horizontal cut: note the bony casing, the ducts, the hollow, center core containing nerve fibers and the spiral ganglion

•Cochlea has a basal, middle, and apical turn; each turn shows three ducts

•All ducts are fluid-filled: scala vestibuli and scala tympani with perilymph, scala media (also called the cochlear duct) with endolymph

•The scala vestibuli and scala tympani are connected at the tip of the cochlea by the heliocotrema

inner hair cells

•cells provide the direct frequency analysis of incoming sound; the nerves issuing from the base of the inner cells are gathered into the afferent auditory nerve

basilar membrane

The basilar membrane converts sound waves into electrical signals and separates frequencies along its length, enabling pitch perception

organ of corti

converts mechanical sound vibrations into electrical signals that are transmitted to the brain, enabling hearing.

inner hair cells

•cells provide the direct frequency analysis of incoming sound; the nerves issuing from the base of the inner cells are gathered into the afferent auditory nerve

outer hair cells

cells serve to make frequency analysis of the inner hair cells more precise, and to enhance their sensitivity to very low-energy sound waves

hair cells and sterocilia

sensorineural mechanism

Hair cells and stereocilia are essential components of the inner ear, responsible for converting physical stimuli into electrical signals that the brain can interpret. They play a crucial role in the senses of hearing and balance. Here's how they function:

Hearing: Stereocilia on inner hair cells respond to sound waves, causing them to bend and generate electrical signals that are sent to the brain via the auditory nerve. This process is called mechanotransduction.

Balance: Stereocilia on vestibular hair cells respond to head movements, providing information about the orientation of the head and body. This helps maintain balance and spatial orientation.

•Stereocilia are connected at their ”tips”—they are “tip-linked” and move as a unit

•Displacement of the linked stereocilia in the direction of the longest one opens + ion gates in stereocilia membranes, which in turn causes depolarization of the hair cells, resulting in an action potential in the nerve attached at the base of the hair cell; displacement of the stereocilia toward the shortest one may inhibit the nerve fiber from generating an action potential

This hair cell/stereocilia model is generally applicable to the role of hair cells in the sensorineural function of both the vestibular and cochlear mechanisms, but with some differences across systems

traveling waves in cochlea

•Transmission of sound energy across the ossicular chain causes the footplate of the stapes to be displaced into the oval window, which displaces the perilymph in the scala vestibuli

•The fluid displacement takes the form of a traveling wave, which causes motion of the basilar membrane with differential amplitudes along the length of the membrane; the location of maximum displacement of the basilar membrane is determined by the frequency (ies) of the sound energy

•High frequencies result in greatest displacement of the basilar membrane toward the base, low frequencies cause greatest displacement of the membrane toward the apex: the displacement of the basilar membrane depends on the frequency (ies) of the signal, and therefore stimulates the hair cells at specific locations (i.e., specific locations along the tonotopically-arranged hair cells)

tuning curves for sinusoidal inputs

•Neural response of a single nerve fiber to basilar membrane displacement and therefore displacement of hair cells

•Note that the greatest sensitivity of a nerve fiber (lowest point on the curve, called the “characteristic frequency”) is at the frequency of the input signal, which corresponds to a specific place along the basilar membrane

•Even when monitoring the “characteristic frequency” nerve fiber, the fiber responds to other sinusoidal frequencies, although with less sensitivity compared with that of the characteristic frequency

•The plot of nerve fiber sensitivity for a single nerve fiber, across frequencies, is called a tuning curve

central auditory pathways

•dedicated to processing auditory information, from brainstem to cortex

•Roughly 75% of auditory fibers in the auditory part of cranial nerve VIII entering the brainstem on one side of the head cross over to the other side in the pons, ascending to the cortex on the side opposite their point of entry

•As the auditory fibers (tracts) ascend in the central nervous system they make synapses in several nuclei

•Tonotopic arrangement of the auditory nerve, and fibers tracts and nuclei along the central auditory pathways, is the rule

brainstem circuitry for acoustic reflex

•Acoustic reflex is the forceful, bilateral contraction of the stapedius muscle in response to very high-intensity sound waves transmitted across the ossicular chain

•Bilateral because very intense sound waves in one ear results in contraction of the stapedius muscle in both middle ears

•Circuitry for bilateral reflex: 1) high-intensity sound introduced in right ear (as if listener is facing you) is carried to the brainstem by the auditory part of cranial nerve VIII, which makes synapses in two auditory brainstem nuclei along the central auditory pathway, which in turn send output to the facial motor nucleus; facial motor nucleus issues tract that exits the brainstem as cranial nerve VII, and delivers impulses to stapedius m. on the same side as the sound (red circuit). 2) fibers from the brainstem on the same side the sound cross over to the facial motor nucleus whose tracts exit the brainstem as cranial nerve VII on the side opposite the sound (green circuitry)