Why is speech audiometry importnat?
-speech = most common stimulus we hear
-gives an estimate of how the pure-tone audiogram is related to the perception of speech
-gives an estimate of speech recognition and comprehension
SDT stands for...
Speech Detection Threshold aka Speech Awareness Threshold (SAT).
indicates the level of input signal
What does the SDT indicate?
Just that you HEARD the sound, not understood.
The threshold is when the patient can detect 50% of the target words/sentences.
Patient responds with a hand raise when she/he hears the speech signal.
Any speech signal can be the stimulus (names, familiar objects)
Used mostly with children
Speech Recognition Threshold.
-Aka ST (spondee threshold)
-Use spondees and ask the patient to repeat back (@ various input levels).
-If the words are repeated correctly, the level is decreased.
-If the words are repeated incorrectly, the level is increased.
*Goal: To find the level at dB HL that the patient can repeat 50% of the target words.
-More frequently used than the SDT. Can use for anyone with vocab level.
-Use spondees b/c they are easily recognized by listeners.
SRT: As level of signal is increased, performance ____ rapidly.
-As the level of the signal is increased, performance increases rapidly.
Performance Level (% correct) on Y axis and Intensity of signal on X axis
Guidelines for SRT that are recommended for measurement
-familiarize patients w/ spondee words- avoid vocabulary effects!
-ensure that each word can be recognized auditorily
-ensure that the patient's responses can be understood by the clinician
-set the start level at 30 dB HL (up 5, down 10 rule)
-change the word for each trial.
-threshold in dB HL = lowest level with at least 3 correct responses at that intensity
-does NOT use a carrier phrase
SRT types of presentation
monitored live voice (MLV) by looking at the VU monitor
recorded (CD or tape) *preferred because each client/patient receives the same set of stimuli and each word is presented in a consistent fashion for intensity and speech pattern
when would monitored live voice be used for SRT?
when greater flexibility in response time is required
when words have to be chosen carefully due to vocabulary restrictions (ex: young)
Word Recognition Testing
-NOT a threshold measure
-Words are presented at a FIXED level
-use monosyllabic words
-uses carrier phrase: "say the word..."
-use either recorded stimuli or MLV
What is the Word Recognition Score?
% correct at a given level. Unlike the SRT where you use the 50% performance on the psychometric function to determine threshold in dB HL, the WRT score is the % correct at a given level (that WE choose).
How to determine SRT?
Goal: To find the level at dB HL that the patient can repeat 50% of the target words.
Use the Performance-Intensity function to go to 50% on the Y axis and look at what the Intensity level is.
Explain the diff between word recognition score and speech recognition threshold.
In the word recognition score, we choose an intensity level (that is above the SRT so it is audible to lsitener) and the score is simply the % of words that they got right at that level.
In speech recognition, we increase and decrease the intensity level while testing. Using the psychometric function, we find the threshold in dB HL where they perform at 50%.
WRT = set the level, maintain. SRT= set level and based on response, inc or dec
The level above or relative to threshold.
ex: 40 dB SL is 40 dB above a patient's threshold for pure tones or speech. SL = presentation level - SRT level
WRT words are presented at :
30 to 40 dB SL re the SRT.
how does the monosyllabic word psychometric function differ from the spondee one?
mono are harder to recognize than spondees, so the mono curve is more to the right (higher Sound Pressure Level) and is more shallow. this is b/c they can miss the first letter (cat v pat)
phonetically balanced words- they are used in WRT.
phonemes appear approximately in proportion to that observed in running speech.
use familiar words w/ equal difficulty
four 50-item lists
most widely used words today
open set word recognition with NU-6 words is the most common practice
Performance Intensity functions for word recognition
the shape gives diagnostic info to the tester
the PI function of a sensorineural loss
PB max does not achieve 100% even when signal is raised to suprathreshold level (40 dB HL)
the PI function for conductive loss
PB max can go to 100% but starts at higher levels of dB HL
maximum performance (when using phonetically balanced word list) aka word recognition testing
neural hearing loss (retrocochlear hearing loss)
the PI function is unusually shaped
a rollover function: once it reaches the PB max, as the level inc the performance goes back down.
HINT Hearing in noise test
tests understanding of sentences in varying amounts of noise
evaluates level of noise where person can repeat sentences correctly
valid estimate of performance in noise, benefit from amplification
Speech tests for comfort level
MCL: most comfortable loudness level (40-55 dB SL for normal)
UCL or LDL :Uncomfortable loudness level or loudness discomfort level (100-110 dHLB for normal)
use running speech as the test signal
Relation of speech audiometry to pure tone audiogram
SRT can be predicted based on PTA (w/i 5-10 dB)
WRT cannot be predicted easily from audiogram (since scores change w/ level)