OzarksTechnical Community College
 Threshold tests
 Supra-threshold tests
 Speech-in-noise tests
 Speech reception threshold (SRT)
 This may also be called speech recognition
threshold
 Word recognition score (WRS)
 This may also be called speech discrimination
score
 If the patient is not a native English speaker, you
may need to skip speech audiometry and note
why you didn’t perform speech testing on the
audiogram.
 If you have hearing loss (as the tester) you may
have difficulty hearing what the patient said.
 Don’t be afraid to ask the patient for clarification on a
word or to use the word in a sentence if you aren’t
sure what they said.
 Have the patient face you so you can read their lips
(just make sure they can’t see your lips if you aren’t
using a recorded list)
 Speech reception threshold (SRT)
 The softest level (dB HL) at which a patient can
accurately repeat spondees (two-syllable words; i.e.
baseball, hotdog, birthday) 50% of the time
 SRT is primarily used as a reliability check
▪ In comparing the SRT to the PTA, they should be within 10 dB
of each other
▪ The patient should be able to repeat words about as softly as they
can hear tones at .5, 1, and 2 kHz
▪ If you have poor SRT-PTA agreement, then the reliability of
your results should be considered to be questionable
 Same procedure as determining thresholds for tones
 Tell the patient: “You will hear some soft two-syllable words. Please
repeat the words as best as you can. If you’re not sure of a word, please
try to take a guess.”
 Start at the patient’s most comfortable level
 Use the same bracketing technique as puretone testing
 If they repeat word correctly, decrease intensity by 10 dB
 If they miss the word, increase intensity by 5 dB
 SRT is the softest level at which the patient accurately repeats
spondees 50% of the time with at least 3 correct responses at
that intensity
 Just like masking for puretones
 Determine the need for masking and use plateau method to determine
effectively masked threshold for speech
 Consider the interaural attenuation (IA):
 40 dB for headphones
 70 dB for inserts
 0 dB for bone (you probably won’t do a bone-conducted speech test, though it
is possible and very useful in children)
 The same rules apply for speech
 Take your SRT in theTE minus the best bone conduction threshold in the NTE.
If that value is greater than the IA, you need to mask.
▪ Example, if your SRT using traditional headphones is 50 dB in theTE and the best BC
threshold in the NTE is 5 dB (50dB-5dB=45 dB), you need to mask since 45 dB is greater
than 40 dB.
▪ ***Note: usually you have not yet performed BC when you are performing SRT, so
assume air=bone. Alternatively, you can subtract the SRT of the NTE from the SRT of
the TE to determine the need for masking. Just keep in mind that the SRT is an air-
conduction test and you mask based on the BC scores of the non-test ear.
 How much noise?
 If you’ve determined the need for masking, add 10 dB
to the SRT of the non-test ear and present a spondee
at threshold
 Just like puretone masking utilizing the plateau
method, if they get the word correct, then increase
masking by 5 dB and present another spondee and
repeat until your 15 dB masking plateau is achieved.
Or, if they cannot repeat the spondee, increase the
presentation level in 5dB steps until threshold is
reached and plateau your masking level accordingly.
 Do NOT waste too much time on SRT
 As far as clinical utility is concerned, SRT is
low on the priority list
 Puretone thresholds andWRS are much more
important
 Get it and go 
 If a patient is unable to repeat spondees for any
reason (i.e. poor speech discrimination, non-
verbal, non-English speaking), a speech
awareness threshold (SAT) or speech detection
threshold (SDT) can be obtained.
 This is the softest level that the patient can hear or
detect speech
 Task=push the button when you are aware that
speech is present
 Usually, the SAT is equal to the patient’s best
puretone threshold on the audiogram in theTE
 Word recognition score (WRS)
 The percentage of phonetically-balanced, monosyllabic
words that a patient can accurately repeat
 Presented at either MCL (most comfortable level) or MIL
(most intelligible level)
▪ This is generally thought to be approximately 30-40 dB SL re: SRT
 A pre-recorded list of 25 to 50 words should be presented
to each ear
▪ Most common word lists: CIDW-22 and NU-6
 Each word should be preceded by a carrier phrase:
▪ “Say the word pick”
▪ “Say the word room”
 When scoring a patient’s performance onWR
testing, you must remember that wrong is
wrong!
 The patient should only be given one attempt at
each word (oftentimes if they miss a word the first
time, they will ask you to repeat it…don’t)
 Close isn’t correct
▪ i.e. If they say the word “eats” for “eat”
 Patient’s with normal hearing or conductive
hearing loss will perform normally onWR
testing
 Patient’s with only high-frequency or a mild,
flat SNHL will also generally perform near-
normal
 The more severe the SNHL, the poorer a
patient will perform onWR testing
 Neural losses result in poor performance
 AsymmetricalWRS, especially in the
presence of puretone asymmetries, should be
considered a “red flag” for an acoustic
neuroma
 In patient’s with a retrocochlear lesion, roll-
over may also be present onWR testing. This
means that with increased presentation level,
the patient’s performance will actually
decrease (i.e. 56% at 85 dB decreasing to 12%
at 95 dB).
 Ideally, recorded word lists should be used
when performing speech audiometry
 Limits variability from test-to-test
 If the clinician says the words while
monitoring his/her voice to make sure they
are peaking their voice at 0 on theVU meter,
this is called a monitored-live voice (MLV)
presentation method.
 You will be much more likely to mask for WRS than SRT since WRS
is a supra-threshold test
 Consider the interaural attenuation (IA):
 40 dB for headphones
 70 dB for inserts
 0 dB for bone (you probably won’t do a bone-conducted speech test,
though it is possible and very useful in children)
 The same rules apply for speech
 Take your presentation level in theTE minus the best bone conduction
threshold in the NTE. If that value is greater that the IA, you need to
mask.
▪ Example, you are performingWR testing using traditional headphones at 75 dB
in theTE and the best BC threshold in the NTE is 20 dB (75dB-20dB=55 dB). So,
you know you need to mask since 55 dB is greater than 40 dB.
▪ ***Note: usually you have not yet performed BC when you are performing
WRS, so assume air=bone. Alternatively, you can subtract the SRT of the NTE
from the presentation level to determine the need for masking, but, again, the
SRT is an air-conduction test.
 How much noise?
 If you’ve determined the need for masking, add 10
dB to the SRT of the non-test ear + an extra 15 dB
for your plateau. So…
 Noise in NTE = SRT of NTE + 25 dB
▪ Ex:The SRT of the NTE is 30 dB, so you add 25 dB to
arrive at your masking level of 55 dB in the NTE.
 Both tests measure a patient’s ability to
understand speech in the presence of
background noise (#1 complaint of hearing-
impaired)
 Useful in making treatment
recommendations and in counseling re:
realistic expectations
 See supplemental materials re:QuickSIN and
ANL

Speech audiometry & masking

  • 1.
  • 2.
     Threshold tests Supra-threshold tests  Speech-in-noise tests
  • 3.
     Speech receptionthreshold (SRT)  This may also be called speech recognition threshold  Word recognition score (WRS)  This may also be called speech discrimination score
  • 4.
     If thepatient is not a native English speaker, you may need to skip speech audiometry and note why you didn’t perform speech testing on the audiogram.  If you have hearing loss (as the tester) you may have difficulty hearing what the patient said.  Don’t be afraid to ask the patient for clarification on a word or to use the word in a sentence if you aren’t sure what they said.  Have the patient face you so you can read their lips (just make sure they can’t see your lips if you aren’t using a recorded list)
  • 6.
     Speech receptionthreshold (SRT)  The softest level (dB HL) at which a patient can accurately repeat spondees (two-syllable words; i.e. baseball, hotdog, birthday) 50% of the time  SRT is primarily used as a reliability check ▪ In comparing the SRT to the PTA, they should be within 10 dB of each other ▪ The patient should be able to repeat words about as softly as they can hear tones at .5, 1, and 2 kHz ▪ If you have poor SRT-PTA agreement, then the reliability of your results should be considered to be questionable
  • 7.
     Same procedureas determining thresholds for tones  Tell the patient: “You will hear some soft two-syllable words. Please repeat the words as best as you can. If you’re not sure of a word, please try to take a guess.”  Start at the patient’s most comfortable level  Use the same bracketing technique as puretone testing  If they repeat word correctly, decrease intensity by 10 dB  If they miss the word, increase intensity by 5 dB  SRT is the softest level at which the patient accurately repeats spondees 50% of the time with at least 3 correct responses at that intensity
  • 8.
     Just likemasking for puretones  Determine the need for masking and use plateau method to determine effectively masked threshold for speech  Consider the interaural attenuation (IA):  40 dB for headphones  70 dB for inserts  0 dB for bone (you probably won’t do a bone-conducted speech test, though it is possible and very useful in children)  The same rules apply for speech  Take your SRT in theTE minus the best bone conduction threshold in the NTE. If that value is greater than the IA, you need to mask. ▪ Example, if your SRT using traditional headphones is 50 dB in theTE and the best BC threshold in the NTE is 5 dB (50dB-5dB=45 dB), you need to mask since 45 dB is greater than 40 dB. ▪ ***Note: usually you have not yet performed BC when you are performing SRT, so assume air=bone. Alternatively, you can subtract the SRT of the NTE from the SRT of the TE to determine the need for masking. Just keep in mind that the SRT is an air- conduction test and you mask based on the BC scores of the non-test ear.
  • 9.
     How muchnoise?  If you’ve determined the need for masking, add 10 dB to the SRT of the non-test ear and present a spondee at threshold  Just like puretone masking utilizing the plateau method, if they get the word correct, then increase masking by 5 dB and present another spondee and repeat until your 15 dB masking plateau is achieved. Or, if they cannot repeat the spondee, increase the presentation level in 5dB steps until threshold is reached and plateau your masking level accordingly.
  • 10.
     Do NOTwaste too much time on SRT  As far as clinical utility is concerned, SRT is low on the priority list  Puretone thresholds andWRS are much more important  Get it and go 
  • 11.
     If apatient is unable to repeat spondees for any reason (i.e. poor speech discrimination, non- verbal, non-English speaking), a speech awareness threshold (SAT) or speech detection threshold (SDT) can be obtained.  This is the softest level that the patient can hear or detect speech  Task=push the button when you are aware that speech is present  Usually, the SAT is equal to the patient’s best puretone threshold on the audiogram in theTE
  • 13.
     Word recognitionscore (WRS)  The percentage of phonetically-balanced, monosyllabic words that a patient can accurately repeat  Presented at either MCL (most comfortable level) or MIL (most intelligible level) ▪ This is generally thought to be approximately 30-40 dB SL re: SRT  A pre-recorded list of 25 to 50 words should be presented to each ear ▪ Most common word lists: CIDW-22 and NU-6  Each word should be preceded by a carrier phrase: ▪ “Say the word pick” ▪ “Say the word room”
  • 14.
     When scoringa patient’s performance onWR testing, you must remember that wrong is wrong!  The patient should only be given one attempt at each word (oftentimes if they miss a word the first time, they will ask you to repeat it…don’t)  Close isn’t correct ▪ i.e. If they say the word “eats” for “eat”
  • 15.
     Patient’s withnormal hearing or conductive hearing loss will perform normally onWR testing  Patient’s with only high-frequency or a mild, flat SNHL will also generally perform near- normal  The more severe the SNHL, the poorer a patient will perform onWR testing  Neural losses result in poor performance
  • 16.
     AsymmetricalWRS, especiallyin the presence of puretone asymmetries, should be considered a “red flag” for an acoustic neuroma  In patient’s with a retrocochlear lesion, roll- over may also be present onWR testing. This means that with increased presentation level, the patient’s performance will actually decrease (i.e. 56% at 85 dB decreasing to 12% at 95 dB).
  • 17.
     Ideally, recordedword lists should be used when performing speech audiometry  Limits variability from test-to-test  If the clinician says the words while monitoring his/her voice to make sure they are peaking their voice at 0 on theVU meter, this is called a monitored-live voice (MLV) presentation method.
  • 18.
     You willbe much more likely to mask for WRS than SRT since WRS is a supra-threshold test  Consider the interaural attenuation (IA):  40 dB for headphones  70 dB for inserts  0 dB for bone (you probably won’t do a bone-conducted speech test, though it is possible and very useful in children)  The same rules apply for speech  Take your presentation level in theTE minus the best bone conduction threshold in the NTE. If that value is greater that the IA, you need to mask. ▪ Example, you are performingWR testing using traditional headphones at 75 dB in theTE and the best BC threshold in the NTE is 20 dB (75dB-20dB=55 dB). So, you know you need to mask since 55 dB is greater than 40 dB. ▪ ***Note: usually you have not yet performed BC when you are performing WRS, so assume air=bone. Alternatively, you can subtract the SRT of the NTE from the presentation level to determine the need for masking, but, again, the SRT is an air-conduction test.
  • 19.
     How muchnoise?  If you’ve determined the need for masking, add 10 dB to the SRT of the non-test ear + an extra 15 dB for your plateau. So…  Noise in NTE = SRT of NTE + 25 dB ▪ Ex:The SRT of the NTE is 30 dB, so you add 25 dB to arrive at your masking level of 55 dB in the NTE.
  • 21.
     Both testsmeasure a patient’s ability to understand speech in the presence of background noise (#1 complaint of hearing- impaired)  Useful in making treatment recommendations and in counseling re: realistic expectations  See supplemental materials re:QuickSIN and ANL