linical & Experimental
Audiology
Speech-in-noise screening tests by internet; improving test sensitivity for noise-induced hearing loss
Monique Leensen
Wouter Dreschler
Academic Medical Centre, Amsterdam, The Netherlands
linical & Experimental
Audiology
2
27-2-2010
Screening
Early detection hearing impairment
Raising awareness about hearing loss
Self-assessment screeningtests:
• National Hearing Test by telephone (Smits et al. ’04)
• Online speech-in-noise tests
Easy to administer
Introduction
linical & Experimental
Audiology
3
27-2-2010
Online speech-in-noise tests
Online speech-in-noise tests
Adjustable presentation level
Spectrum of masking noise is equal to speech spectrum
• Independent of presentation level
Adaptive procedure
Fixed noise level
Speech Reception Treshold (SRT)
• 50% correct word recognition
Introduction
linical & Experimental
Audiology
4
27-2-2010
Oorcheck (LUMC ’04)
Target population:
youngsters (12-24 years)
Closed set of 9 CVC words
Introduction
linical & Experimental
Audiology
5
27-2-2010
Oorcheck and NIHL
Aims at youngsters
• Risk for hearing loss after exposure to loud music
• Raising awareness about hearing loss
However, NIHL does not greatly affect SRT in stationary
noise (Smoorenburg 1992)
• Closed set of speech stimuli
• Hearing threshold levels at low and middle frequencies are
(near) normal
Introduction
linical & Experimental
Audiology
6
27-2-2010
Objectives of this study
1. Investigate the potential of Oorcheck to discover NIHL
in an early stage
2. Examining ways to enhance the test sensitivity for mild
NIHL using different forms of noise modifications
Make existing speech-in-noise screening tests (like
Oorcheck) applicable for NIHL
Objectives
linical & Experimental
Audiology
7
27-2-2010
Objectives of this study
1. Investigate the potential of Oorcheck to discover NIHL
in an early stage
2. Examining ways to enhance the test sensitivity for mild
NIHL using different forms of noise modifications
Make existing speech-in-noise screening tests (like
Oorcheck) applicable for NIHL
Objectives - 1
linical & Experimental
Audiology
8
27-2-2010
Subjects
n=100 (2 excluded)
NH: 49 (age 27 yrs ± 8.5)
with HTLs ≤ 15 dB HL
HI: 49 (age 56.3 yrs ± 9.4)
- 25 with narrow dip (ND)
- 24 with broad dip (BD)
HTL2kHz affected
Methods - 1
linical & Experimental
Audiology
9
27-2-2010
Oorcheck
SRT (dB snr)
-16-15-14-13-12-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
HI : BD
HI : ND
NH
good poorin-sufficient
moderate
Test sensitivity for NIHL
Sensitivity 51%
49% HI is classified as NH
Specificity 90%
10% NH is classified as HI
Results - 1
linical & Experimental
Audiology
10
27-2-2010
Objectives of this study
1. Investigate the potential of Oorcheck to discover NIHL in an
early stage
2. Examining ways to enhance the test sensitivity for mild
NIHL using different forms of noise modifications
Make existing speech-in-noise screening tests (like
Oorcheck) applicable for NIHL
Objectives - 2
linical & Experimental
Audiology
11
27-2-2010
Noise modifications
Spectral filtering
• Low pass filtered noise
Less masking of high frequencies
Temporal modulation
• NH benefit more from interrupted noise than HI (e.g. De Laat & Plomp 1983, Festen & Plomp 1990, Versfeld & Dreschler ‘02)
Both types of modifications combined
Noise parameters were based on model predictions (ESII)
and pilot measurements
Methods - 2
linical & Experimental
Audiology
12
27-2-2010
Experimental set of noises
test noise filtering modulation noise floor
1 Oorcheck - - -
2 16 Hz - 16 Hz -15 dB
3 LP LP 1.4 kHz - -15 dB
4 LP & 16 Hz LP 1.4 kHz 16 Hz -15 dB
5 HP HP 1.4 kHz - -15 dB
6 HP & 16 Hz HP 1.4 kHz 16 Hz -15 dB
7 OC -15 - - -15 dB
LP HP
Methods - 2
linical & Experimental
Audiology
13
27-2-2010
Results – group differences
Results - 2
linical & Experimental
Audiology
14
27-2-2010
Results – group differences
Results - 2
linical & Experimental
Audiology
15
27-2-2010
Results – group differences
Results - 2
linical & Experimental
Audiology
16
27-2-2010
Results – group differences
Results - 2
linical & Experimental
Audiology
17
27-2-2010
Results – group differences
Results - 2
linical & Experimental
Audiology
18
27-2-2010
Results - frequency dependence
Correlations between individual SRTs and HTLS
Highest correlation for LP noise and 2-6 kHz
Results - 2
linical & Experimental
Audiology
19
27-2-2010
Validation of OC LP
Increased sensitivity &
specificity:
sensitivity 95%
specificity 98%
Oorcheck - Low Pass noise
SRT (dB snr)
-26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0
HI: BD
HI : ND
NH
good in-sufficient
poor
Results - 2
linical & Experimental
Audiology
20
27-2-2010
Discussion
To be investigated further:
• Extrapolation of lab results
ambient noise
system settings
transducers
• Level dependency
Discussion
linical & Experimental
Audiology
21
27-2-2010
Conclusions
1. The original Oorcheck showed only small differences
between NH and HI subjects with mild NIHL (narrow dips)
2. Hearing-impaired subjects deviated more clearly from
normal performance when using low-pass filtered
masking noise
This suggests that this noise condition is more sensitive
to discover NIHL in an early stage
Conclusions
linical & Experimental
Audiology
22
27-2-2010
Generalization
We expect that the approach used will be valid for other
types of speech-in-noise tests as well
If this can be proven, manipulation of the masking noise will
increase the sensitivity and specificity with respect to NIHL
for different types of speech-in-noise screening tests
Conclusions
linical & Experimental
Audiology
23
27-2-2010
23
Thank you!
Questions?
linical & Experimental
Audiology
24
27-2-2010
linical & Experimental
Audiology
25
27-2-2010