Upload
august-sentell
View
215
Download
0
Tags:
Embed Size (px)
Citation preview
Mild Hearing Loss is Serious Business
Harvey DillonSharon Cameron, Teresa Ching, Helen Glyde, Gitte Keidser,
David Hartley, Jorge Mejia
NAL, The Hearing CRC
IHCON, 2010
What is mild hearing loss?
• Four-frequency average (500, 1000, 2000, 4000 Hz) hearing loss in better ear between 20 and 40 dB HL
• Self-reported disability or handicap within a certain range
• SRT in noise loss of between 3 and 6 dB
Frequency Scatterp lot (Spreadsheet in 30000 audiograms 2010.stw 37v*30132c)
L4FA = 9.5175+ 0.9136*x
< = 0 (0 ,20] (20,40] (40,60] (60,80] (80,100] (100,120] (120,140] (140,160] (160,180] (180,200] (200,220] > 220 0 20 40 60 80 100 120
L3FA
0
20
40
60
80
100
120
L4F
A
4FAHL or 3FAHL?
How prevalent is mild hearing loss?
How prevalent is mild hearing loss?
• Davis (1995): 16% of adult population• Wilson (1990): 18% of adult population• Hartley et al (in press): 34% of people aged > 50 years
Blue Mountains 4FAHL better earH istogram (BM H S_audiogram_edited.sta 41v*2956c)
-10 0 10 20 30 40 50 60 70 80 90 100 110 120
4FA_Better
0
100
200
300
400
500
No of ob
s
Hartley et al (in press)
Aging population - Australia
0%
10%
20%
30%
40%
50%
2010 2020 2030 2040 2050
Year
Po
pu
lati
on
old
er t
han
50
Population aging, worldwide
0 5 10 15 20 25
Japan Italy
Greece Sweden
France United Kingdom
Canada Australia
New Zealand United States of America
Hong Kong (SAR of China) Korea, Republic of
Singapore China (exl. SARs and
World Viet Nam
Indonesia India
Malaysia South Africa
Philippines Papua New Guinea
Percentage > 65 years
C luster analysis of mild losses
C luster 1 C luster 2 C luster 3 C luster 4 C luster 5 C luster 6 C luster 7 C luster 8 C luster 9 C luster 10250 500 1000 2000 4000 6000 8000
Frequency (H z)
0
10
20
30
40
50
60
70
80
90
HL in ear w
ith better 4F
AH
L (dB H
L)
Mild losses for study
0
10
20
30
40
50
60
70
80
90
100
250 500 1000 2000 4000 8000
Frequency (Hz)
Th
resh
old
(d
B h
L)
Mild loss (4FA=29 dB)
Mild-mod loss (4FA=39 dB)
~ Minimum loss eligible for
government benefits
~ Median loss newly fitted in Australia
Mild losses for study
0
10
20
30
40
50
60
70
80
90
100
250 500 1000 2000 4000 8000
Frequency (Hz)
Th
resh
old
(d
B h
L)
Mild 9 dB
Mild-mod 39 dB
Moderate 49 dB
Mod-Sev 59 dB
Severe 69 dB
4FA HL
How common is hearing aid use amongst those with mild hearing loss?
Penetration by hearing loss
• Penetration as a function of hearing loss– Davis (1995)– Wilson et al (1998) – Hartley et al (in Press)– Kochkin
losshearingwithNumber
aidshearingowningNumbernPenetratio
consistent
Aid ownership by hearing lossBlue Mountains Population >55 years
0
20
40
60
80
100
5 15 25 35 45 55 65 75 85 95 105 115
4FA Hearing threshold (dB HL)
Pe
rce
nta
ge
of
pe
op
le
0
20
40
60
80
100
5 15 25 35 45 55 65 75 85 95 105 115
4FA Hearing threshold (dB HL)
Per
cen
tag
e o
f p
eop
lePopulation
Ownership
Ownership
Use
Hartley et al (in Press)
What are the characteristics of mild hearing loss?
Characteristics of mild loss
• Threshold elevation √• Loss of OHC/OAE• Loss of frequency resolution – TEN, PTC, FRI• Loss of temporal (envelope) resolution• Loss of fine temporal information• Loss of spatial processing• Loss of SRT in noise• Increased disability and handicap
Loss of active process in cochlear:OHC and OAE
-20 0 20 40 60 80 100 120
H L at 2 kH z (dB)
-35
-30
-25
-20
-15
-10
-5
0
5
10
OA
E at 2 kH
z (dB S
PL
)
Loss of frequency resolution
f
A
f
AFRI
HTL (dB HL)
FR
I (d
B)
-30
-20
-10
0
10
20
30
40
50
60
0 20 40 60 80 100 120
4 kHz
Ching & Dillon (unpublished data)
HTL (dB HL)
AF
RI
(dB
)
-20
0
20
40
60
0 20 40 60 80 100 120
350 Hz
HTL (dB HL)
AF
RI (
dB
)
-20
0
20
40
60
0 20 40 60 80 100 120
1 kHz
HTL (dB HL)
AF
RI
(dB
)
-20
0
20
40
60
0 20 40 60 80 100 120
2 kHz
HTL (dB HL)
AF
RI
(dB
)
-20
0
20
40
60
0 20 40 60 80 100 120
4 kHz
Frequency resolution
Loss of temporal resolution
A
t t
HTL (dB HL)
TR
I (d
B)
-20
0
20
40
60
0 20 40 60 80 100 120
4 kHz
TRI
Temporal resolution
HTL (dB HL)
AT
RI (d
B)
-20
0
20
40
60
0 20 40 60 80 100 120
350 Hz
HTL (dB HL)
AT
RI (d
B)
-20
0
20
40
60
0 20 40 60 80 100 120
1 kHz
HTL (dB HL)
AT
RI (d
B)
-20
0
20
40
60
0 20 40 60 80 100 120
2 kHz
HTL (dB HL)
AT
RI (d
B)
-20
0
20
40
60
0 20 40 60 80 100 120
4 kHz
Ching & Dillon (unpublished data)
Loss of fine temporal information
• 20 synapses per IHC • Synapse loss or IHC loss reduced
averaging temporal jitter
Inspiration:• Bodian, Lieberman, Moore,
Pichora-Fuller, Spoendlin,
A digression into “normal” hearing
Spatial Processing Disorder
26
Noise
Noise
Noise
Noise
Speech
Sharon Cameron
Listening in Spatialised Noise - Sentences(LiSN-S) Conditions
Total Advantage
Talker Advantage
Sp
atial A
dva
ntag
e
High Cue
Low Cue
Same voices Different voices
Same direction
Differentdirections
Cameron & Dillon (2009)
LiSN-S Diagnostic Screen
6 7 8 9
10
11
12
-13
14
-15
16
-17
18
-19
20
-24
25
-29
30
-39
40
-49
50
-60A ge Group
8
9
1 0
1 1
1 2
1 3
1 4
1 5
1 6
1 7
1 8
Spatia
l Advanta
ge (d
B)
Spatial Advantage (≡ Spatial Release from Masking)
Australia
Nth America
Spatial AdvantageB
ette
r
Results profile: spatial processing disorder
Spatial processing remediation Pre vs. Post (n=9)
LC SRT - p = 0.158
Talker Advantage - p = 0.981
HC SRT - p = 0.0002
Spatial Advantage - p = 0.0002
Total Advantage - p = 0.001
L C S RT HC S RT T a l ke r S p a ti a l T o ta l
L iS N-S Co n d i ti o n
-4
-3
-2
-1
0
1
2
LiS
N-S
Sco
re(P
op
ula
tion
Sta
nd
ard
De
via
tion
Un
its)
P re -T ra i n i n g P o st-T ra i n i n g 3 M P o st-T ra i n i n g
Application to people with hearing loss
Spatial hearing loss in hearing-impaired people:LiSN-S Prescribed Gain Amplifier
Deficit in SRTn with hearing loss
Cameron, Glyde & Dillon, unpublished data)
Low C ue defic it (dB) = 0.0147-0.0631*xH igh cue defic it (dB) = 4.0175-0.3029*x
20 25 30 35 40 45 50 55 60
4FA H L w orse ear (dB H L)
-16
-14
-12
-10
-8
-6
-4
-2
0
De
ficit re n
orm
al (d
B)
Low C ue defic it (dB ) H igh cue defic it (dB )
R = -0.87
P = 0.000006
Cause of deficit in SRTn T alker Adv defic it (dB) = 3.9432-0.0061*x
Spatia l Adv defic it (dB ) = -3.4051+ 0.2273*x
20 25 30 35 40 45 50 55 60
4FA H L w orse ear
0
2
4
6
8
10
12
14
De
ficit re n
orm
al (d
B)
T alker Adv defic it (dB ) Spatia l Adv defic it (dB)
R = -0.86
P = 0.00001
Cameron, Glyde & Dillon, unpublished data)
Talker advantage deficit versus age
Cameron, Glyde & Dillon, unpublished data)
T alker Adv defic it (dB) = 1.6806+ 0.0407*x
0 10 20 30 40 50 60 70 80 90
A ge
0
1
2
3
4
5
6
7
8
Ta
lker A
dv d
eficit (d
B)
Binaural processing
ILDITD
L R
Executive
control
xx
CAPD
~Sensorineural
hearing loss
ILDITD
CN CN
SO / IC / A1
Loss of SNR in understanding speech
“Basic” loss of 0.6 dB per 10 dB of loss
+
Loss of Spatial release from masking of 2.3 dB per 10 dB of loss
+
Loss of Talker cue release from masking of 0.5 dB per 10 years of age
-10 0 10 20 30 40 50 60 70 80
Test ear 4FA (500 to 4k Hz)
-24
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
2
4
6
SR
Tn
Version 1 Version 2
Loss of SRT in noiseCommonly 1.5 dB increase in SNR per 10 dB of hearing loss
1 dB / 10 dB
1.8 dB / 10 dB
SNR = -4 dB
Carter, Zhou & Dillon, unpublished data)
Should mild hearing loss interfere with speech perception?
Speech and noise levels
30
40
50
60
70
80
90
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
lev
el (
dB
A)
Source: Pearsons, Bennett and Fidell (1977)
SNR
Calculation of SII
20
40
60
80
100
100 1000 10000
Frequency (Hz)
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
Noise = 60 dBA Speech = 64 dBA SIInh=0.65 SIIhi=0.46
Effective audibility
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70
Sensation level (dB)
Eff
ecti
ve a
ud
ibil
ity 40
60
80
100
20
Transfer functionSII Percent correct
0102030405060708090
100
0 0.2 0.4 0.6 0.8 1
SII
Per
cen
t co
rrec
t
ConnectedSpeechTest
Mild losses for study
0
10
20
30
40
50
60
70
80
250 500 1000 2000 4000 8000
Frequency (Hz)
Thre
shol
d (d
B h
L)
Mild loss (4FA=29 dB)
Mild-mod loss (4FA=39 dB)
Predicted speech intelligibility
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Normal hear
Mild unaided
Mild-modunaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
in
tell
(C
ST
%)
Norm hear
Mild unaided
Mild-mod unaided
Greatest problems in noisy places !!
Predicted variation of SRTn with hearing loss
-5
-4
-3
-2
-1
0
0 20 40 60 80
4FA hearing threshold (dB HL)
SR
T (
dB
) 50607080
Modified SII model predicts only 0.4 dB loss per 10 dB of hearing loss
Speechlevel
Modification of SII
• Assume normal hearers get 6 dB advantage from spatial separation of speech and noise
• Assume hearing impaired listeners lose spatial advantage at a rate of 1.3 dB per 10 dB of loss (above SII predictions)
total loss of SNR is 1.7 dB per 10 dB of loss
Calculation of intelligibility
Speech spectrum
Noise spectrum
Threshold
Sensation Level
Effective audibility
Importance function
SII
(Information received)
Percent correctMax
Do hearing aids help people with mild hearing loss?
current
^
Benefit of hearing aids
Predict increase in speech intelligibilitywith
the modified Speech Intelligibility Index
Calculation of aided benefit
Noise = 50 dBA
Speech = 58 dBA 0
10
20
30
100 1000 10000Inse
rtio
n g
ain
(d
B)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
Speech intelligibility (mild loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hearDirectOmniUnaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
inte
ll (
CS
T %
)
Norm hear
Direct
Omni
Unaided
• Hearing aid “helps” in quiet places
• Tiny additional benefit from directivity
Conditions:
DI = 3 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -3.7 dB
Speech intelligibility (mild-moderate loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hearDirectOmniUnaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
inte
ll (
CS
T %
)
Norm hear
Direct
Omni
Unaided
• Hearing aid “helps” in quiet places
• Tiny additional benefit from directivity
Conditions:
DI = 3 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -5.1 dB
Speech intelligibility (moderate loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hearDirectOmniUnaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
inte
ll (
CS
T %
)
Norm hear
Direct
Omni
Unaided
Conditions:
DI = 3 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -6.5 dB
Speech intelligibility (moderate-severe loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hearDirectOmniUnaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
inte
ll (
CS
T %
)
Norm hear
Direct
Omni
UnaidedConditions:
DI = 3 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -7.7 dB
Speech intelligibility (severe loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hearDirectOmniUnaided
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Sp
eech
inte
ll (
CS
T %
)
Norm hear
Direct
Omni
UnaidedConditions:
DI = 3 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR loss re n.h. = -9 dB
Summary of benefit versus HL
0
20
40
60
80
100
0 20 40 60 80
4FA Hearing Threshold (dB HL)
Pe
rce
nt
co
rre
ct
0
20
40
60
80
100
0 20 40 60 80
4FA Hearing Threshold (dB HL)
Pe
rce
nt
co
rre
ct
0
20
40
60
80
100
0 20 40 60 80
4FA Hearing Threshold (dB HL)
Pe
rce
nt
co
rre
ct
40 dB A
60 dB A
80 dB A
Background noise level
Why don’t directional microphones help more in noise?
(and adaptive noise reduction)
1. Impact of open fittings on directivity
-20
-10
0
10
20
30
125 250 500 1k 2k 4k 8kFrequency (Hz)
Inse
rtio
n ga
in (
dB) Combined path
Amplif iedpath
Ventpath
DI = 3 dB 1.5 dB
DI = 10 dB 5 dB
DirectionalOmni-directional
Frequency range of directional mic (mild loss)
0
10
20
100 1000 10000
Frequency (Hz)
Inse
rtio
n g
ain
(d
B) 50
5560657075808590
50
60
70
80
Effect of dynamic
noise reduction on
directional mic
Effect of aiding at 60 dBA (mild loss)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
Noise = 60 dBA
Speech = 64 dBA
20
40
60
80
100
100 1000 10000
Frequency (Hz)
0
10
20
30
100 1000 10000
Inse
rtio
n g
ain
(d
B)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
Noise = 80 dBA
Speech = 76 dBA
20
40
60
80
100
100 1000 10000
Frequency (Hz)
0
10
20
30
100 1000 10000
Inse
rtio
n g
ain
(d
B)
Effect of aiding at 80 dBA (mild loss)
2. Impact of reverberation on directivity (and vice versa)
50
60
70
80
90
0 1 2 3 4 5
Distance from source (m)
So
un
d l
evel
(d
B S
PL
)
DirectReverberant
Total
50
60
70
80
90
0 1 2 3 4 5
Distance from source (m)
So
un
d l
evel
(d
B S
PL
)
DirectReverberant
Total
60
057.0RT
VQQdistcritical ms
effective
2. Impact of reverberation on directivity (and vice versa)
Benefit of directional microphones
Benefit affected by:– Directivity pattern of microphone– Distance and direction of talker and noise sources– Acoustics of the room– Frequency range over which the hearing aid is
directional– Frequency range over which the wearable has usable
hearing
– Nothing else
(OK, Measurement error!!)
Conclusion of acoustic analysis
1. People with mild loss need help in noisy places
2. Hearing aids increase the speech information available mostly in quiet places!
3. Mics directional only where there is gain high frequencies and lower levels Where audibility is limited by threshold, not noise
So, objective benefit questionable, and increasing with degree of loss, but …….
What do hearing aid users say ?
Experimental evaluation of self-reported benefit
• 400 clients sampled from national database– 41,521 new clients fitted Feb to Sept, 2004
• Audiometric and other details obtained from clients’ files
• Questionnaire sent to clients 5 to 12 months after fitting
– International Outcome Inventory for Hearing Aids
– Plus 6 purpose-designed questions
• Non-responders followed up by phone or additional mail to get a high response rate (effectively 86%)
Hearing loss distribution
0
10
20
30
5 15 25 35 45 55 65 75 85 95 105 115
4FA Hearing threshold (dB HL)
Pe
rce
nta
ge
of
pe
op
le
0
10
20
30
40
5 15 25 35 45 55 65 75 85 95 105 115
4FA Hearing threshold (dB HL)
Per
cen
tag
e o
f p
eop
leSample distribution
Population distribution
Usage of hearing aids
21%
10%
35%
19%
13%
None< 1 hr/day
1-4 hrs /day4-8 hrs /day
> 8 hrs /day
Q 3: Daily us age
0
20
40
60
80
100
120
Num
ber o
f clie
nts
Factor analysis of questionnaireFactor Loadings
Factor 1 Factor 2 Factor 3
Q1: want aids 0.69 0.30 0.21
Q2: difficulty unaided 0.70 0.41 0.25
Q3: use 0.74 -0.18 0.08
Q4: benefit 0.82 -0.32 0.00
Q5: residual difficulty 0.03 -0.76 -0.33
Q6: Worth it 0.83 -0.33 -0.00
Q7: Residual handicap -0.29 -0.56 -0.06
Q8: Bother to others -0.18 -0.68 -0.26
Q9: Quality of life 0.82 -0.32 0.02
Q10: Replace them 0.34 -0.15 -0.23
Q11: Face vision -0.29 -0.42 0.74
Q12: paper vision -0.22 -0.47 0.70
Proportion of variance 0.32 20 0.12
Composite benefit Composite difficulty Vision
International Outcomes Inventory
for Hearing Aids
Effect of hearing loss on benefitCurrent effec t: F (3, 281)= 1.1951, p= .31194
V ert ic al bars denote 0.95 c onfidenc e intervals
10-19 20-29 30-39 40-49
4FA H L in better ear (dB H L)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Com
posite benefit
If hearing loss does not determine benefit, then what does?
Wishes And Needs Tool
1. How strongly did you want to get hearing aids? Wanted it very much Wanted it quite a lot Wanted it moderately Wanted it slightly Did not want it
2. Overall how much difficulty do you have hearing when you are not wearing your hearing aids?
Very much difficulty Quite a lot of difficulty Moderate difficulty Slight difficulty No difficulty
No
of o
bs
Q 2: No diffic ulty unaided
1 2 3 4 50
10
20
30
40
50
60
Q 2: S light diffic ulty unaided
1 2 3 4 5
Q 2: M oderate diffic ulty unaided
1 2 3 4 5
Q 2: Q uite a lot of diffic ulty unaided
1 2 3 4 50
10
20
30
40
50
60
Q 2: V ery m uc h dific ulty unaided
1 2 3 4 5
Difficulty hearing unaided and wish to get hearing aids
Unaided difficulty related to wish to get
hearing aids
Want hearing aidsVery muchNot at all
Need increases with hearing lossC urrent e ffect: F (3, 275)= 8.5017, p= .00002
Vertica l bars denote 0 .95 confidence in tervals
10-19 20-29 30-39 40-49
B etter ear 4FA range (dB H L)
2 .2
2 .4
2 .6
2 .8
3 .0
3 .2
3 .4
3 .6
3 .8
Ne
ed
sum
ma
ry
Cu rre n t e ffe ct: F(8 , 1 0 9 )=7 .9 1 0 9 , p = .0 0 0 0 0
1 1 .5 2 2 .5 3 3 .5 4 4 .5 5
S tre n g th o f n e e d
1 .0
1 .5
2 .0
2 .5
3 .0
3 .5
4 .0
4 .5
5 .0
Co
mp
osite
be
ne
fit: =("Q
3: u
se" +
"Q4
: be
ne
fit" + "Q
6: W
orth
it" +
"Q9
: QO
L")/4
Benefit versus need strength
Why don’t more people with mild hearing loss even try hearing aids?
Factors affecting benefit experienced (and hence the reports of others)
• Degree of pure-tone loss
• Self-reported disability and handicap
• Acceptable Noise Level
• Stigma / cosmetic concern
• Manipulation and management
• Age
• Tinnitus
• Personality ….
Personality
People more likely to acquire hearing aids are:– Open– Non-obsessive– Non-neurotic– Internal locus of control
People more likely to report benefit are:– Extroverted– Agreeable
Health Belief Model
People act rationally, in their best interests, based on their beliefs
Weighing up of beliefs for and against a health decision
What are the disadvantages of the solution?
Do I have a problem?
Is it serious enough for me to want to remove it?
Is there a solution
that works?AdvantagesDisadvantages
“My hearing loss is not bad enough to need them” (Kochkin, 1993)
Motivation comes from ….
Acknowledge loss
Experience difficulty
Experience handicap
Self-image
Expect benefit
Expect to manage them
$ Cost OK
Try them
Health belief model
Hearingloss
Cost
Difficulties experienced:
frequency, severity
Self-image
Ability to manage
Hearing aideffectiveness
Externalimage
Incon-venience
Health belief model
Hearingloss
Cost
Difficulties experienced: frequency, severity
Self-image
Ability to manage
Hearing aideffectiveness
External image
Inconven-ience
Health belief model
Hearing
loss
Cost
Difficulties
experienced:
frequency, severity
Self-
image
Ability to
manageHearing aid
effectivenessExternal
image
Inconven
-ience
How can the balance of benefits to disadvantages be improved?
Improving advantages and removing disadvantages
Problem 1 Solution 1
Problem 2 Solution 2
Problem 3 Solution 3
• WDRC
• Slim-tube, miniaturization
• Feedback cancellation
• Low-level expansion
• Re-chargeable batteries
• Auto telecoil
• Frequency lowering
• Bandwidth extension
• Trainable responses
• Adaptive noise reduction
• Transient noise reduction
• Directional microphones
Cost
Working better in noise
Speech intelligibility in noise
Binaural-Processing Super-directional Microphone (Mejia et al., 2007)
W3 W4
∑
∑
Rout
K
Rear- directional array
Masking threshold
Cross-fading process
Outputs with spatial reconstruction
HRTFL HRTFR
Z-d
d = 3 ms delay
(Precedence effect)
DOA- reconstruction
Q1 Q2 Q3 Q4
∑
Binaural beamformer
Main directional signal
W1 W2
∑
∑
Lout
K
Subsidiary signal
Super-directional microphones
Speech reception threshold in noiseReverberant room: crit dist = 0.4m, radius = 1 m
-14.0
-12.0
-10.0
-8.0
-6.0
-4.0
-2.0
0.0
NH Dir BF NH Dir BF NH Dir BF
2-speaker 4-speaker 140-babble
Spee
ch r
ecep
tion
thr
esho
ld (d
B)
Mejia and Johnson, unpublished data
ListenerListener
Hearing Aid
Linked binaural hearing aid technology
Female talker
Male talker
Childrenplaying
Blind-source separation binaural noise reduction
Implication: People with mild or moderate hearing loss were not making much use of spatial cues.
n.h. benefit from re-insertion of spatial cues
h.i.gain most from
directivity
Effect of super-directivity (mild loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hear
Direct
Unaided
Super-directional
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Spe
ech
inte
ll (C
ST
%)
Norm hear
Direct
Unaided
Super-directional
• Super-directional mic not directional at all over a broader and broader range as noise levels rise
Conditions:
DI = 6 dB when REIG > 3 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -3.7 dB
So super-directivity alone is no use ……..
What if we could achieve directivity at low frequencies?
Effect of low-frequency directivity (mild loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hear
Direct
Unaided
Extended direct
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Spe
ech
inte
ll (C
ST
%)
Norm hear
Direct
Unaided
Extended direct
• Now directional over entire frequency range in noisy places
Conditions:
DI = 3 dB when REIG ≥ 0 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -3.7 dB
Occlusion with closed moldsVent or leak transmission in
Active Occlusion Reduction
HearAid C
Σ
BA
+
-
Active occlusion reduction
Mejia, Dillon, & Fisher (2008)
In combination?
Active occlusion reduction
(closed mold)
Super-directivity = ?+
Low-frequency super-directivity (mild loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hear
Direct
Unaided
Extend/super
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Spe
ech
inte
ll (C
ST
%)
Norm hear
Direct
Unaided
Extend/super
• Super-directivity over entire frequency range
super-normal hearing
Conditions:
DI = 6 dB when REIG ≥ 0 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -3.7 dB
0
20
40
60
80
100
120
100 1000 10000
Frequency
Th
resh
old
(d
B H
L)
Low-frequency super-directivity (mild-moderate loss)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hear
Direct
Unaided
Extend/super
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Spe
ech
inte
ll (C
ST
%)
Norm hear
Direct
Unaided
Extend/super
• Super-normal hearing for the median hearing aid wearer
Conditions:
DI = 6 dB when REIG ≥ 0 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -5.1 dB
0
20
40
60
80
100
120
100 1000 10000
Frequency
Th
resh
old
(d
B H
L)
Low-frequency super-directivity (moderate loss, 4FA = 49.8 dB HL)
0
0.2
0.4
0.6
0.8
1
30 40 50 60 70 80 90
Background noise level (dBA)S
II
Norm hear
Direct
Unaided
Extend/super
0
20
40
60
80
100
30 40 50 60 70 80 90
Background noise level (dBA)
Spe
ech
inte
ll (C
ST
%)
Norm hear
Direct
Unaided
Extend/super
• Super-normal hearing in very noisy places for even a moderate-severe loss
0
20
40
60
80
100
120
100 1000 10000
Frequency
Th
resh
old
(d
B H
L)
Conditions:
DI = 6 dB when REIG ≥ 0 dB
n.h. spatial adv = 6 dB
∆SNR re n.h. = -6.5 dB
Adaptive noise reduction
• Gain dependent on SNR correct in principle, but room for improvement:– Gain reduction can reduce directional mic
effectiveness– No point in reducing noise below threshold– Gain increase where SNR is best is just as
important
Cost
Jump to summary
Self-fitting hearing aid
Audiometer
NAL-NL2 Prescription
Formula
Real-ear to coupler difference
Adjust Hearing Aid
Automatic
Developing countries: 1 audiologist per 500,000 people, to 1 per 6,000,000
Australia, USA: 1 audiologist per 10,000 people
Plus tra
inability
-10 0 10 20 30 40 50 60 70
M anual thresho lds (dB H L)
-10
0
10
20
30
40
50
60
70
Au
tom
atic th
resh
old
s (dB H
L)
Automatic versus manual audiometry
1 kHz
Test-retest standard deviations
0
1
2
3
4
5
250 1000 4000
Frequency (Hz)
Tes
t-re
test
sta
nd
ard
d
evia
tio
n (
dB
)
Manual audiometry (5 dB Hughson-Westlake)
Automatic audiometry (2 dB final step size)
NAL-NL2
RECD
Adjust
Auto Aud
Real-Ear to Dial Difference:Inter-subject standard deviations
0123456789
10
100 1000 10000
Frequency (Hz)
Sta
nd
ard
dev
iati
on
(d
B) Insert
ClosedOpen
Real-Ear to Dial Difference:Inter-subject standard deviations
0123456789
10
100 1000 10000
Frequency (Hz)
Sta
nd
ard
dev
iati
on
(d
B)
Insert
Closed
Open
Saunders & Morgan,InsertValente et al, ER3A
Valente et al, TDH39
Hawkins etal, TDH39
Hawkins et al, ER3A
Insert
Saunders & Morgan
Valente et al
Hawkins et al
Supra-aural
Valente et al
Hawkins etal
Trainable Hearing Aids
0
5
10
15
20
25
30
30 40 50 60 70 80 90
Input level (dB SPL)
Gai
n (
dB
) CRGain
CT
In summary…
In summaryvery
not very
many, including spatial hearing loss
yes, in noise
only in quiet places
….. expected benefit too small re need
closed-ear, binaural processing
How prevalent is mild hearing loss?
How common is hearing aid use amongst those with mild hearing loss?
What are the characteristics of mild hearing loss?
Is mild hearing loss a problem to people?
Do hearing aids help people with mild hearing loss?
Why don’t more people with mild hearing loss even try hearing aids?
How can hearing aids provide greater benefit where it is most needed?
Messages for ….
Public health authorities:
• Increase hearing awareness (prevention, rehabilitation)
• Increase hearing screening opportunities
Messages for ….
Clinicians:
• Discern primary reasons why unmotivated clients are unmotivated
• Provide information to change unrealistic beliefs
• Understand and diagnose the fundamental problem that clients are presenting with SRT loss
Messages for ….
Researchers:
• Better understanding of the components and causes of SNR loss
• Prescription procedures for adaptive noise suppression
• Time constants• Relationship with thresholds• Relationship with noise spectrum and level
Messages for ….
Manufacturers:• Achieve better performance in noise
Binaural processing Closed fittings Wireless Smarter adaptive noise suppression
Hearing aids of the future
Convergence:
• hearing aid/enhancer,
• phone interface,
• hearing protector,
• computer interface (in and out),
• music player,
• GPS interface
?
Thanks for listening
For the slides from this talk ..…
www.nal.gov.au
Amplification and directivity
20
40
60
80
100
100 1000 10000
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
1/3
oc
t le
ve
l a
t T
M (
dB
SP
L)
20
40
60
80
100
100 1000 10000
Frequency (Hz)
20
40
60
80
100
100 1000 10000
50 dBA 80 dBA
Unaided
Aided