Speech Audiometry

Dynamics of speech
Intensity
Whisper - 20 dB HL
Normal conversational speech - 50 to 60 dB
Loud speech - 70 dB
Shouting - 90 dB

What is speech audiometry?

Refers to the assessment of a clients hearing using speech

stimuli.
A clients performance on speech audiometric tasks, especially

those that are suprathreshold, can provide an audiologist with a

more accurate view of the clients communication performance in
the real world.
Materials

Nonsense syllables
Monosyllabic words
Spondaic words
Sentence tests

Contribution of Speech Audiometry

To determine the extent of speech-recognition difficulty.
To provide a cross-check to ensure reliable data.
It can be sensitive to auditory problems that pure tone

testing might miss (sensitivity)

To aid in diagnosis of retro-cochlear problems.

Assists in the selection of amplification systems.

Helps clinician educate patients about loss and make a

prognosis about treatment outcomes.

Test environment
Normally sound treated booth
Pt should not see the examiners face to avoid lip reading cues.

Setup
Input selection: to select the

desired source of speech material

tape, CD player for recorded

speech

microphone for live voice

Input level control: used with the

VU (volume unit) meter to

ensure that the speech signal are
of at the level necessary for them
to being properly calibrated
Attenuation: to control the level
of speech being presented to the
patient.
Output selector: through which
mode you want to present the
speech signal to the patient

Live Voice Testing:

Controlled Vocal Effort.
Adjust microphone sensitivity
To have the speech balanced at 0 dB onVU meter.

However, it is recommend to use CDs whenever possible.

VU

Presentation of Speech Tests

Monaural (one ear at a time, usual

method).
Binaural (both ears
simultaneously).
Can be presented with earphones.
Can be presented via bone
conduction.
Can be presented in the sound field
using speakers.

Monitored live-voice (MLV)

Pre-recorded lists on CD or cassette.
Lists of words that patient repeats.
Standardized picture tests that require

the patient to point to a picture that

matches the spoken word
Standardized speech-in-noise tests.

Speech audiometry tests

Speech Recognition Thresholds
Speech Detection Thresholds
Word Recognition Testing

Level of auditory ability assessed

Awareness: tests that require the patients to simply indicate

that a sounds was detected.

Discrimination: tests that require the patient to detect a
change in the acoustic stimulus.
Identification/recognition: tests that require the patient to
attach a label to the stimulus either by pointing to a
corresponding picture or object or repeating the stimulus
orally:
Speech recognition threshold (SRT)
Word recognition scores or sentence recognition scores.

Speech Recognition Thresholds

Abbreviated as SRT
Defined as
The minimum hearing level for speech at which an individual

can recognize 50% of the speech material (ASHA, 1988).

Uses spondees as speech material

SRT relation to pure tone audiogram

Clinically, PTA should be within 10 dB of the SRT.

SRTs can be predicted by finding the average of 500, 1000,

and 2000 Hz (Pure tone average, PTA).

In some cases, SRT may be higher (worse) than the three

frequency PTA.
Age, or disorders of the CANS.

In other cases, the SRT may be much lower (better) than the

PTA.
When the audiogram falls precipitously in the high frequencies.

Use of the SRT as an

average hearing level
can be problematic
because it is possible
for a clients SRT to
reflect only one of the
three frequencies often
associated with it.

Test administration
Instruction:
I am going to review a series of words with you. Please repeat

each word that you hear.After this review, I will begin to

present these words at softer and softer levels. Please continue
to repeat the words that you hear, even if they are quite soft and
even if you have to guess
Familiarization:
This is an important part of test administration and allows

audiologists to omit words that may not be familiar to the

client.
ASHA guideline strongly suggests that familiarization not be eliminated

from the test protocol.

Test administration.procedure
Although ASHA (1988) guidelines suggest use of a descending

approach to obtaining the SRT, various approaches (e.g.,

ascending, descending, or ascending/descending) can yield
very similar results when testing in 5-dB steps. The basic idea of
all of these approaches is to present a series of spondee words several
times, until 50 percent understanding is achieved at one level.

ASHA recommended procedure

Determination of threshold
Step one: determine initial starting level:
Present one spondaic word at a level 30 to 40 dB HL above the anticipated SRT. If a
correct response is received, drop the level in 10-dB steps until an incorrect response
occurs. Once an incorrect response is received, present a second spondaic word at
the same level. If the second word is repeated correctly, drop down by 10-dB steps
until two words are missed at the same level. Once you reach the level where two
spondees are missed, increase the level by 10 dB. This is your starting level.
Step two:Threshold estimation
Present five spondaic words at the starting level. An individual should get the first
five spondaic words correct at the starting level, if not increase starting level by 10.
Drop the level by 5 dB, and present five spondaic words. Continue dropping the level
by 5 dB until the individual misses all five spondaic words at the same level; test
is then terminated.

Calculate threshold as:

Starting level - # of correct responses + 2 dB (correction factor)

Speech Detection Thresholds

The SDT is not commonly part of routine audiological assessment,

probably because it provides less information than other speech

measures.
ASHA guidelines suggest that the methods used for pure tone threshold
testing be applied to SDT testing.
Speech material:
Stimuli for SDT testing are not standardized.According to ASHA, the exact

stimulus used is not critical because the SDT measures detection and not
recognition.
For purposes of consistency, use of bup-bup-bup or uh-oh is suggested.

Use of the child or clients name should be avoided because it will result in a highly

variable stimulus.

Other possible speech materials include speech babble, running speech, or familiar

words.

Regardless of the type of stimuli used, it is important to record the material type on

the audiogram for reporting and comparison purposes.

Word Recognition Testing

Word recognition testing may be defined as a measure of

speech audiometry that assesses a clients ability to identify

one-syllable words that are presented at hearing levels that
are above threshold to arrive at a word recognition score.
How is it different than SRT?
Uses monosyllabic words (words of one-syllable), not spondees.

These words are phonetically balanced.

Word recognition is performed at a suprathreshold level.

10

Method for Obtaining SRS

Decide method of delivery (MLV, recorded).
Choose materials to be used (word lists etc).
Inform patient with regards to method of response.
Select intensity.
Decide if multiple levels should be tested.
Decide if test will be presented with noise in the background.

Word Recognition Testing

Open set : pt can respond with any word he/she can think of.
Closed set: response options are provided for the pt

(multiple choice test).

Free response-client is free to respond or not.
Forced Response-client must say something.
[Forced choice = closed set forced response.]

11

Word recognition scores

Quantifies pts ability to discriminate speech:
It determines the extent of speech-recognition difficulty.
It aids in diagnosis of the site of the disorder in the auditory

system.
It assists in the determination of the need for and proper selection

of amplification systems.
It helps the clinician to make a prognosis for the outcome of the

treatment efforts.
Hearing aid fitting.

Contribution of speech evaluation to differential

diagnosis
Discrepancy in performance between scores from a patients

two ears or between word and sentence recognition.

Rollover effect: Reduction in speech recognition (more than
20% from maximum performance) with increases in
intensity.
Occurs with retrocochlear pathological conditions

12

Performance-Intensity Functions
Curve reaches a peak (Pbmax), and then
Either remains high (normal), or
Drops at higher levels (Rollover)

Rollover Index = (PBmax Pbmin)/PBmax

Rollover Indices for the preceding examples

Normal: (100 - 100) / 100 = 0.0
Rollover: (44 - 20) / 44 = 0.54
Cochlear: (80 - 70)/80 = 0.125
Rollover Indices of 0.45 or greater indicate a neural (VIIIth

nerve) problem.

13

 The use of PI functions is a way of sensitizing speech by

challenging the auditory system at high-intensity levels.

Because of its ease of administration, many audiologists use it
routinely as a screening measure for retrocochlear disorders.
The most efficacious clinical strategy is to present a word list
at the highest intensity level (usually 80 dB HL). If the patient
scores 80%, then potential rollover of the function is
minimal, and testing can be terminated. If the patient scores
below 80%, then rollover could occur, and the function is
completed by testing at lower intensity levels.

14

 Test-specific:

each test much

have its values
to determine
max word
recognition

Predicting Speech Recognition

Word-recognition ability is predictable from the audiogram

in most patients.This has been known for many years.

Essentially, speech recognition can be predicted based on the
amount of speech signal that is audible to a patient.The
original calculations for making this prediction resulted in
what was referred to as an articulation index, a number
between 0 and 1.0 that described the proportion of the
average speech signal that would be audible to a patient based
on his or her audiogram (French & Steinberg, 1947).

15

Predicting WRS from the audiogram: The AI

The Articulation Index
Audibility Index
Count the dot audiogram
Each dot = 1%
# of dots above a pts threshold

is the audibility index revealing

the percentage of
conversational speech energy
audible to the pt at a distance of
3-6 feet.
If word recognition is poorer than

prediction: neural hearing loss or

central disorder.

Counting Dots: (Mueller & Killion

Killion))
Calculating AI from this

audiogram is simple.This Figure

shows a patients audiogram
superimposed on the count-thedots audiogram.
Those components of average
speech that are below (or at higher
intensity levels than) the
audiogram are audible to the
patient, and those that are above
are not.
To calculate the AI, simply count
the dots that are audible to the
patient. In this case, the AI is 60
(or 0.6).This essentially means
that 60% of average speech is
audible to the patient.

16

AI clinical applications
The AI has at least three useful clinical applications.
It can serve as an excellent counseling tool for explaining to a

patient the impact of hearing loss on the ability to understand

speech.
The AI has a known relationship to word-recognition ability.
Thus word-recognition scores can be predicted from the AI or,
if measured directly, can be compared to expected scores based
on the AI.
AI can be useful in hearing aid fitting in serving as a metric of
how much average speech is made audible by a given hearing
aid.

17