Speech/hearing Science: Feedback
Closed loop feedback
- systems utilize feedback - they return the results of output back to the original input source for the purposes of comparison and correction (motor plan for a sentence being executed by the nervous system -> the ears picking up the acoustic result of the output -> and feeding it back to the nervous system to compare to the original program)
Lombard effect
- the natural tendency of speakers to increase their vocal intensity when the air conducted signal is masked (speakers cannot hear themselves), such as in a noisy room - Conversely, if the air conducted signal is amplified, speakers tend to decrease their vocal intensity
Some target models are spatial in nature:
- we learn specific targets (spacial coordinates) for specific sounds, we make memories for these targets, and we replicate them - In our brain we have a map of our mouth that tells our brain how to produce a sound
Delayed Auditory Feedback (DAF)
- when a speech signal from a speaker is fed back to their ears but there is a delay in time (like hearing an echo) - effect: slows down speech - sometimes a treatment for stuttering - HOWEVER, people can accommodate DAF
In speech production, we use both Open and Closed loop systems...
-Closed loop is more important early in life -Later in life becomes more for open loop
On argument against the essential need for auditory feedback in skilled speakers is that...
...it takes too long for the auditory signal to get back to the brain for it to influence the next sound or even word - the speaker has already produced these before the auditory feedback can have a corrective effect. As you all know, when you speak, you can often produce an error before you have a chance to correct it.
Tendon Reflex
Golgi tendon organs are involved in a reflex arc which prevents muscles from being ripped away from the bone when they contract
Stretch reflex
Muscle spindles are involved in a reflex arc which prevents muscles from being over stretched
alpha motor neurons
Nerves that cause extrafusal muscle fibers to activate or contract
Intrafusal fibers are oriented in _____ to the extrafusal fibers, so that when the extrafusal fibers stretch, such as the bicep muscle fibers do when you extend your arm from a by contracting your tricep, the intrafusal fibers will ____ ______.
parallel; also stretch
When muscles contract/stretch...
sensory receptors send messages to the brain
Muscle spindles
slender inner muscle fibers located within the larger outer muscle fibers
Model
some simplified representation of a system; the system is responsible for the phenomenon; allow us to test our theories
Serial Order
sounds of speech are in serial order and that order is important for meaning; one sound blends into the sound that comes after it
When proprioceptive info is blocked...
speakers are able to compensate and maintain a high degree of speech intelligibility.
Production of speech sounds requires...
synchronized and instantaneous use of systems
Source-filter theory
the observations we make when we acoustically analyze speech are the result of a vocal fold vibrating sound source being modified by the vocal tract above; changes to the filter will cause changes to the resulting sound
Speech =
the oral expression of language
Feedback occurs when...
the output of a system is returned to input source - ex: acoustic signal output from mouth to ears
If there is a breakdown or error in IF...
the speaker can recognize it via external feedback, and make corrections, although the errors will always be detected after the production has occurred
Because muscle spindles deliver sensory information to the CNS quickly,
they may be good candidates for control of rapid movements, such as speech gestures.
Goal of Speech production =
to produce acoustic patterns in the air
Associative Chain Theory
when a movement is initiated, it stimulates what you are going to do next (stimulus response)
gamma motor neurons
neurons which cause the intrafusal muscle fibers to contract
TF is probably (essential/not essential) at least in learned speakers
not essential
Bone conduction -
other listeners don't hear it (you hear your own, that's why you sound differently in a recording)
When demands call for a specific sound to be produced...
it's motor plan is retrieved from memory and refined based on the context of the utterance it is produced in
PF is located in the...
joints , tendons, and muscles of the body
Feedback is essential to ________ speech patterns.
learning
Fairbanks
- Closed loop model - suggested that control of speech production operated in a closed-loop self-regulating manner, where external (auditory and tactile) feedback and proprioceptive feedback play a significant role in ongoing speech production
The 4 types of feedback:
- External feedback: Auditory feedback (AF)―slow Tactile feedback (TF)―slow - Proprioceptive feedback (PF)―relatively fast - Internal feedback (IF)―theoretically most rapid
Kozhevnikov and Christovich
- Open loop model - suggested that the speech production was controlled at the level of the syllable, not the individual sound, and that different sounds within a syllable can be initiated simultaneously, such as what happens during coarticulation
Lashley 1951
- Sequential organization is central to much of animal and human behavior, including speech production - we can co-articulate sounds where they are almost produced at the same time... no stimulation/ response going on - something in the brain organizes speech in time before we utter it = Schema for action - must be organized by something other than direct associative connections
Degrees of freedom
- When we produce speech, we have to control a very large number of muscles, associated with three different systems and over at least a dozen different structures - Many of these structures can move in different ways, at different speeds, and in different relationships or combinations with each other. Each potential muscular contraction of an individual muscle is referred to as a degree of freedom
Proprioceptive feedback
- a relatively quick form of feedback which sends information related joint angle, muscle tension, and muscle length to the CNS ***the sense of your body in space***
2 elements to a motor plan:
- a spacial element (place) - a temporal element (time) *** you're going to move articulators to a certain place, in a certain amount of time***
Some suggest Acoustic in nature:
- acoustic pattern of air that the listener will recognize (doesn't matter how you place articulators in the mouth) ***Some say that we start our using Spatial to learn and then end up using Acoustic when we get older; others say we use both***
Golgi tendon organ
- another type of proprioceptor that provides information about changes in muscle tension - located in series between a muscle and a tendon, which connects the muscle ton bone - When a muscle contracts, the Golgi tendon organ detects the amount of tension and sends this information back to the CNS
Target Models
- attempt to explain speech as a process where speakers attempt to achieve a sequence of speech sound targets, or goals - spacial or acoustic
Because IF is learned and stored in memory, ___ ___ for ongoing speech is not required.
- external feedback (although auditory, tactile, and proprioceptive information is always available for the speaker to monitor ongoing speech) ***when learning speech sounds, we learn the spacial and temporal goals for different speech sound, and when they're learned they are stored in brain so.... Any time you want to say a speech sound, just REACLL THE MOTOR PLAN - NO NEED FOR FEEDBACK!***
Linguistically oriented models (LOM)
- focused on the sound structure of speech; often a phonological perspective - tries to explain how we plan the production of a word ***IPA is LOM***
Internal feedback
- often referred to as feed-forward information. - this type of information can occur much faster than other feedback because it does not require the perception and relaying of results; open loop control - delivery of information from brain about motor commands prior to motor response itself
Open loop feedback
- systems do not use feedback to modify the ongoing output (think of this as a pre-programmed motor plan which is executed and does not need feedback, because it is highly learned and automated)
TF is measured by:
2-point discrimination and oral stereognosis
Tactile Feedback (TF)
Sensations of touch (light touch, deeper pressure) from: - Articulators in contact; as our articulators make contact with each other, or as the pressurized air travels along the tissues of the vocal tract (imparting a degree of pressure on the tissue), these sensory receptors are stimulated and send feedback into the central nervous system - Air pressure and flow changes at glottis and in supraglottal cavities; The oral cavity is rich in mechanoreceptors - special sensory receptors which respond to touch and pressure
All theories of speech production have to deal w/ 3 important issues:
Serial-order issue, Degrees of freedom, Context-sensitivity
Context-sensitivity
The specific way/location a speech sound is produced can vary with the phonetic context in which it is uttered
What is the FILTER in source-filter theory? SOURCE? LOCATION?
The vibrating vocal fold; Vocal tract; everything above the vocal folds
A muscle cannot be activated unless...
a nerve stimulates it
Language =
a shared set of symbols used to represent ideas
Theory
a statement that describes a phenomenon that helps us understand things that we experience
When we speak, our auditory nerve is stimulated by activation in the cochlea of the inner ear. It is stimulated by two sources:
an air conducted signal and a bone conducted signal
sensory receptors
detect changes in length of the intrafusal fibers, such when it is stretched
The outer muscle fibers, which is what you can see contracting underneath your skin, such as the bicep when you flex your arm, are called _________. The fibers of the muscle spindle are called _________.
extrafusal muscle fibers; intrafusal muscle fibers
The muscle spindles are thus innervated by this ___ ___ ___, and also by ___ ___which wrap around the intrafusal fibers.
gamma motor neuron; sensory receptors
Two-point discrimination
has shown that the mechanical resolution of the anterior tongue (especially the tongue tip) is much greater (they can discriminate between 2 points of stimulation that are very close together) than that of the posterior tongue
People who become post-lingually deaf...
have hearing into adulthood, then lose hearing but still mostly maintain normal speech
How do we learn? =
have to be exposed, repetition (child has to practice), feedback (correcting)