[PDF] - Evaluating New What are the parameters that impact the doing? PDF Document

SLIDE 1

11/1/2013 1

Evaluating New Technologies

RUTH BENTLER UNIVERSITY OF IOWA

1

Levels of Evidence*

APFs (Catherine Palmer, 2009)

What does the algorithm do?
What are the parameters that impact the doing?

Efficacy of the design

In a well-controlled (contrived?) environment, do we get an

effect?

Or, what is the effect of the feature in the lab?

Effectiveness of the design

In the real-world use of this design, do we get an effect?
Or, what is the effect of the feature in the real world?

*Ala Bentler http://www.uiowa.edu/~neuroerg/siren.html

SLIDE 2

11/1/2013 2

Levels of Evidence*

APFs (Catherine Palmer, 2009)

What does the algorithm do?
What are the parameters that impact the doing?

Efficacy of the design

In a well-controlled (contrived?) environment, do we get an

effect?

Or, what is the effect of the feature in the lab?

Effectiveness of the design

In the real-world use of this design, do we get an effect?

Or, what is the effect of the feature in the real world? Efficiency (not studied in my lab)

*Ala Bentler

Directional Microphones

8

APFs

THE FIRST STEP IS TO UNDERSTANDING THE BLACK BOX….

9 10 11 12

SLIDE 3

11/1/2013 3

13 14 15 16 17 18

SLIDE 4

11/1/2013 4

19 20 21 22 23 24

SLIDE 5

11/1/2013 5

25 26 27 28 29 30

SLIDE 6

11/1/2013 6

31 32 33 34 35 36

SLIDE 7

11/1/2013 7

37 38 39 40 41 42

SLIDE 8

11/1/2013 8

43 44 45

4.8 4.0 2.3 6.0 5.1 3.0 5.7 5.0 3.3

Theoretical FF (BTE) KEMAR (BTE)

Cardioid Hypercardioid Supercardioid 4.8 4.1 2.7 6.0 5.6 3.3 5.7 5.4 3.5

Theoretical FF (ITE) KEMAR (ITE)

Cardioid Hypercardioid Supercardioid

15
10
5

5 10 15 20 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 500 1000 2000 4000

48

SLIDE 9

11/1/2013 9

And so…

We are able to measure the acoustic and physical

facts (APFs) for all possible scenarios of test;

Such APF testing is necessary to develop our

hypotheses;

Newer technique for quantifying polar response

patterns and directivity indices (DI) helps us understand static function in a dynamic world of noise Wu & Bentler, 2009, 2010, 2012)!

Polargram

50

Data?

Plenty of efficacy data for all designs depending upon

Baseline used
Speaker arrangement
Noise type
Etc

Effectiveness data a bit harder to come by…

20 40 60 80 100 60 / 0 75 / +2 Speech Understanding in Noise Percent Correct OMNI DIR CST Test Condition Test Booth Field Ratings 10 Very Good 0 Very Poor p < .0001 p < .0001 8 6 4 2

Walden, Surr, & Cord, 2003

20 40 60 80 100 60 / 0 75 / +2 Speech Understanding in Noise Percent Correct OMNI DIR CST Test Condition Test Booth Field Ratings 10 Very Good 0 Very Poor p < .0001 p < .0001 8 6 4 2

SLIDE 10

11/1/2013 10 Research Question of Study #1

How do visual cues affect DIR benefit?

Laboratory Real world

Speech recognition test Speech Recognition Performance

SNR (dB)

10 6 2

2
6
10

Speech Recognition (%)

20 40 60 80 100

OMNI-AO OMNI-AV DIR-AO DIR-AV

Auditory-Only Wu & Bentler, 2010, Ear Hear

OMNI DIR

SLIDE 11

11/1/2013 11 Speech Recognition Performance

SNR (dB)

10 6 2

2
6
10

Speech Recognition (%)

20 40 60 80 100

OMNI-AO OMNI-AV DIR-AO DIR-AV

Audiovisual Auditory-Only Wu & Bentler, 2010, Ear Hear

Summary of Study 1

The advantage (benefit) of visual cues can overshadow

the measured benefit of directional mic schemes in real world environments.

Research Question of Study #2

How does age impact DIR benefit?
Laboratory
Real world

Age

30 40 50 60 70 80 90

Laboratory DIR Benefit (%)

5

5 10 15 20 25 F(1, 21) = 1.21 p = 0.29 Wu, 2010, JAAA

Age

30 40 50 60 70 80 90

Real World DIR Preference (%)

20 40 60 80 100 F(1, 21) = 11.78 p = 0.003

Wu, 2010, JAAA

Summary of Study #2

Listeners of different ages obtain comparable benefits

from DIR in the laboratory.

Older users tend to perceive less DIR benefit than do

younger users in the real world.

Due to lifestyle differences, primarily
The focus of future efforts in the lab

SLIDE 12

11/1/2013 12

Example of unexpected function…

Front Back

Backward DIR Backward DIR Forward DIR Forward DIR

Condition

Conversation Listening

Directional Benefit (dB)

1 2 3 4 5 6 7

p = 0.17 p < 0.05

Wu, Stangl & Bentler, 2013

Our Data Manufacturer’s Data

http://www.despicableme.com/ Front Back

Big dogs can be dangerous. Backward DIR Backward DIR

Front Back

Big dogs can be dangerous. Backward DIR Backward DIR Forward DIR Forward DIR

Front Back

The boy fell from the window. Forward DIR Forward DIR

SLIDE 13

11/1/2013 13

Condition Conversation Listening Directional Benefit (dB)

1 2 3 4 5 6 7

p = 0.17 p < 0.05

Wu, Stangl & Bentler, 2013

Our Data Manufacturer’s Data

Briefly, for DIR

APFs are clear as to expected impact
Efficacy has been demonstrated

repeatedly; newer algorithms take special consideration

Effectiveness depends on many

factors

Environment, age, etc
…crud

Digital Noise Reduction

75

Analog NR (1980-90s)

Early spectral approaches

Switch
ASP (means low frequency compression)
Adaptive filtering
Frequency dependant input compression
Adaptive compressionTM
Zeta Noise BlockerTM

Today’s versions

Most are modulation-based with some algorithm for

where and how much gain reduction should occur;

At least one other (Oticon) first introduced a strategy

called “synchronous morphology” treating harminic inputs like speech;

Many are now implementing Wiener filters as well;
Many are now implementing impulse noise

reduction;

Many also use some mic noise reduction, expansion,

wind noise reduction, and even directional mics as part of the strategy they promote.

APFs

THE FIRST STEP IS TO UNDERSTANDING THE BLACK BOX….

78

SLIDE 14

11/1/2013 14

Siemens (TRIANO 3)

Frequency (Hz)

250 500 1000 2000 4000 8000

Difference (dB, 1/3 Octave)

12
10
8
6
4
2

2 SIREN TRAFFIC DINING

Frequency (Hz)

125 250 500 1000 2000 4000 8000

Difference (dB,1/3octave)

25
20
15
10
5

5

GN ReSound (CANTA 770-D)

a

Frequency (Hz)

125 250 500 1000 2000 4000 8000

Difference (dB,1/3octave)

25
20
15
10
5

5 ICRA Speech Random Noise Babble

Starkey (AXENT II AV MM)

b

Frequency (Hz)

125 250 500 1000 2000 4000 8000

ON versus OFF (output change)

16
14
12
10
8
6
4
2

2 SNR00 SNR05 SNR10 SNR15

70dB

Frequency (Hz)

125 250 500 1000 2000 4000 8000

ON versus OFF (output change)

16
14
12
10
8
6
4
2

2 SNR00 SNR05 SNR10 SNR15

85dB Starkey J13 Axent AV 75 dB

-SPEECH,RANDOM, MUSIC--

Frequency(Hz)

125 250 500 1000 2000 4000 8000

DIFFERENCE (dB,1/3octave)

25
20
15
10
5

5 Guitar Piano Saxophone with background music Random Noise Plain Speech

SLIDE 15

11/1/2013 15

What happens in the time domain?

Siemens (Triano) Starkey (Axent)

APFs…10 years later

88

Any reason to expect SNR-50 would change?

Output SNR (re: Linear)

Miller et al. 2012

SLIDE 16

11/1/2013 16 Data?

Still, plenty of efficacy and effectiveness data for all

designs if you are asking the right question:

Walden et al (2000)
Boymans and Dreschler (2000)
Alcantara et al (2003)
Ricketts & Hornsby (2005)
Marcoux et al (2006)
Mueller et al (2008)
Bentler et al (2009)
Sarampalis et al (2009)
Bentler et al (2010)
Stelmachowicz et al (2010)
Pittman et al (2011):
And those are good outcomes

Briefly, for DNR

APFs are clear as to expected impact
Efficacy and Effectiveness have been

demonstrated…if you are asking the right question

Frequency Lowering

93

Not really a new concept

Four (sort of) choices on the market:

Frequency compression
Frequency transposition
Frequency “cueing”
Combination of above

Concept makes sense

Providing the widest input bandwidth possible
Data suggest this may be most important for children re:

speech and language development

APFs

THE FIRST STEP IS TO UNDERSTANDING THE BLACK BOX….

95

What is happening here?

Frequency compression hearing aid Default settings Steeply sloping loss Freq compression: OFF Assessed on 11/23/09 SN:0906H109W Input: 1s pure tones 100 Hz spaced with 500ms intervals (~75dB SPL) Upper graph: output of Hearing aid

96

SLIDE 17

11/1/2013 17

1st peak: 3468 Hz, 2nd peak: 4091 Hz, 3rd peak: 4700 Hz Input: 4091 Hz

97

1st peak: 3661 Hz, 2nd peak: 4306 Hz, 3rd peak: 4927 Hz Input: 4306 Hz

98

1st peak: 3765 Hz, 2nd peak: 4392 Hz Input 4392 Hz

99

1st peak: 4070 Hz, 2nd peak: 4694 Hz, 3rd peak: 5336 Hz Input 4694 Hz

100

1st peak: 5490 Hz Input 5490 Hz

101

1st peak: 5598 Hz Input 5598 Hz

102

SLIDE 18

11/1/2013 18

1st peak: 5457 Hz, 2nd peak: 6093 Hz Input: 6093 Hz

103

1st peak: 5553 Hz, 2nd peak: 6201 Hz Input: 6201 Hz

104

1st peak: 5665 Hz, 2nd peak: 5603 Hz Input 6309 Hz

105

1st peak: 1937 Hz, 2nd peak: 2562 Hz Input 6395 Hz

106

What is happening here?

Frequency compression hearing aid Default settings Steeply sloping loss Freq compression: ON Assessed on 11/23/09 SN:0906H109W Input: 1s pure tones 100 Hz spaced with 500ms intervals (~75dB SPL) Upper graph: output of Hearing aid

107

1st peak: 1071 Hz, 2nd peak: 1701 Hz, 3rd peak: 2346 Hz Input: 4091 Hz

108

SLIDE 19

11/1/2013 19

1st peak: 1894 Hz, 2nd peak: 1538 Hz Input: 4306 Hz

109

1st peak: 1207 Hz, 2nd peak: 1359 Hz, 3rd peak: 2851 Hz, 4th peak: 2001 Hz, 5th peak: 2482 Hz – Input 4392 Hz

110

1st peak: 1343 Hz, 2nd peak: 1656 Hz, 3rd peak: 1981 Hz, 4th peak: 2626 Hz – Input 4694 Hz

111

1st peak: 1351 Hz, 2nd peak: 1672 Hz, 3rd peak: 1981 Hz, 4th peak: 2626 Hz – Input 5490 Hz

112

1st peak: 1287 Hz, 2nd peak: 1916 Hz, 3rd peak: 2410 Hz, 4th peak: 2562 – Input 5598 Hz

113

1st peak: 1624 Hz, 2nd peak: 2260 Hz, 3rd peak: 2907 Hz Input: 6093 Hz

114

SLIDE 20

11/1/2013 20

1st peak: 1720 Hz, 2nd peak: 2368 Hz Input: 6201 Hz

115

1st peak: 1830 Hz, 2nd peak: 2466 Hz Input 6309 Hz

116

1st peak: 1937 Hz, 2nd peak: 2562 Hz Input 6395 Hz

117

500 1500 2500 3500 4500 5500 6500 7500 500 1500 2500 3500 4500 5500 6500 7500 Input (frequency in Hz) Output (frequency in Hz)

CF2.6; CR1.7 CF1.7; CR1.6 Uncompressed CF6.0; CR1.5 CF5.9; CR2.1 CF4.7; CR2.0 CF3.8; CR1.9 CF3.2; CR1.8 CF2.2; CR1.7 CF1.9; CR1.5 CF1.5; CR1.5 CF1.5; CR2.0 CF1.5; CR2.5 CF1.5; CR3.2 CF1.5; CR4.1 Graph from A Perreau dissertation, 2011 118 990.53 1528.86 1507.32 1356.59 1550.39 1550.39 1744.19 1765.72 1787.26 500 750 1000 1250 1500 1750 2000 2250 2500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Input (frequency in Hz) Output (frequency in Hz)

*N=7/11: Lowered Output <1500Hz *

Graph from A Perreau dissertation, 2011 119

Microphone Low-pass filter High-pass filter FFT Bin 1 Bin 2 Bin 3 . . . Bin 24 Σ Oscillators Bin 1 Bin 2 Bin 3 . . . Bin 24 Processing of data blocks delay = 9 ms

Σ

Receiver Cutoff Frequency Microphone

Graph from A Perreau dissertation, 2011 120

SLIDE 21

11/1/2013 21

Sound quality

Guitar Original Guitar Max Comp Singer Original Singer Compressed Piano Original Piano Compressed

Evidence (efficacy here):

Better speech-sound perception
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech

perception; 1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more

(group) benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

SLIDE 22

11/1/2013 22 Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

Evidence:

Better speech perception/satisfaction
Simpson et al (2005) 8/17 improvement phoneme recognition
Simpson et al (2006) 1/7 (words) 1/5 (sentences) improved speech perception;

1/6 better APHAB

Kuk et al (2007; 2009) improved consonant recognition (group)
Gifford et al (2007): 2/6 improved sentence recognition in Q and N; more (group)

benefit on EC, BN and RV subscales of APHAB

Robinson et al (2007) 4/7 improved affricates; 5/7 improved /s and /z/
Nyffeler (2008) improved (group) satisfaction (11 adults)
Robinson et al (2009) 1/5 improved affricates; 1/5 improved /s and /z/
Glista et al (2009) 5/11 children, 5/13 adults benefit for /s/ and /z/ detection
O’Brien et al (2010) initial improvement in speech perception (23 adults)
Wolfe at al (2010) group improvement for tokens /asa/ and /ada/ in quiet (15

children)

Wolfe et al (2011) group improvement for tokens /asa/, /ata/ and /ada/ in quiet;

after 6 mo. of use improvement on nonsense syllable SRT in quiet, 13/15 improved on plural test

And now for the conflicting evidence:

Worse performance or no change
Simpson et al (2005) 8/17 no improvement phoneme recognition; 1/17 poorer
Simpson et al (2006) 4/7 (words) 4/5 (sentences) no improvement speech perception; 2/7 (words)

poorer; 4/6 APHAB preference for conventional amplification, 1/6 APHAB no preference

Kuk et al (2007; 2009): no change in vowel recognition (group data; n=13, 8)
Gifford et al (2007): 4/6 no diff in sentence recognition in Q and N; more (group) aversiveness on

APHAB

Robinson et al (2007) 3/7 no effect affricates; 2/7 no improvement /s and /z/
Nyffeler (2008) no improvement (group) in sentence recognition in noise
Robinson et al (2009) 2/5 decreases performance affricates; 4/5 no improvement /s/ and /z/; 4/5

preferred control (no compression) condition 1/5 had no clear preference

Glista et al (2009) 5/11 children, 6/13 adults no benefit for /s/ and /z/ detection; 1/11, 1/13 showed

poorer performance

O’Brien et al (2010) initial improvement in speech perception (23 adults) disappeared after 8
weeks. No difference/improvement on SSQ.
Wolfe at al (2010) no improvement (group) for sentence recognition in noise or for tokens /afa/,

/aka/, /asha/ or /ata/ in quiet

Wolfe et al (2011) no improvement (group) for sentence recognition in noise or for tokens /afa/,

/aka/, or /asha/, no effect for 2/15 who performed at ceiling on plural test

More recent data (still efficacy)

Mussoi pre-dissertation project:
Less is more
Musical training makes the distortion more negative

NH-NT NH-T HL-NT HL-T Moderate compression NH-NT NH-T HL-NT HL-T

Max. compression

Slight preference Moderate preference Strong preference NH-NT NH-T HL-NT HL-T No compression % Preference 10 20 30 40 50 60 Group

More recent data (still efficacy)

Perreau dissertation
Adults tend to opt for conventional technology as the bimodal
ption to CI
No objective evidence of better localization

Perreau, Bentler & Tyler, 2013

More recent data (still efficacy)

Perreau dissertation
Adults tend to opt for conventional technology as the bimodal
ption to CI
No objective evidence of improved speech perception

Perreau, Bentler & Tyler, 2013

SLIDE 23

11/1/2013 23 Effectiveness data

Perreau dissertation
Adults tend to opt for conventional technology as the bimodal
ption to CI…..

Perreau, Bentler & Tyler, 2013

“OCHL” Study (real effectiveness data)

Outcomes of Children with Hearing Loss
Co-PIs Mary Pat Moeller, J Bruce Tomblin
Multi-site (UIowa, UNC, Boys Town)
Using accelerated longitudinal design
Recruited children 6 mos-7 years of age
Follow same children for 3+ years
Lengthy burden tables resulting in many data points!

NIH/NIDCD R01 DC009560

Recruitment

Sampling Regions
Iowa, Nebraska, Eastern

Kansas/Northern Missouri, Illinois, Southern Virginia, North Carolina, Minnesota

Sampling Method
Referral from Newborn

Hearing Screening

Children identified in EHDI via

follow up clinics

Children identified via

audiology or medical service providers

Children identified through

school screening

135

Sample

321 children with hearing loss
182 children with normal hearing
Ages 6 months to 7 years, 3 months
Speaks English in the home
No major secondary disabilities
Permanent Bilateral Mild to Severe Hearing

Loss –PTA of 25-75 dB HL (500, 1k, 2k, 4 kHz)

136

Domains of Study

137

Child and Family Outcomes

Background characteristics

f child/family

Hearing & Speech Perception Speech Production Language Skills Academic Abilities

Psychosoci al and Behavioral

Interventions (clinical, educational, audiological)

Opportunity to observe:

What hearing aids children wear;
How they are fit;
How long they wear them (i.e., use time);
What kind of audibility is provided;
If any of the above impact outcomes in speech and

language.

SLIDE 24

11/1/2013 24 This Data Set

Three age levels (3-, 4- and 5-yr olds)
All children had 1+ yrs. experience with aids and ~equal

number in each group:

Nonlinear Frequency Compression (NLFC)
Conventional signal processing
Data from one site only since that site fit majority of

subjects using NLFC, using “best-practice” verification protocol.

Questions

Are children using nonlinear frequency compression

(NLFC) in their hearing aids getting better access to the speech signal than children using conventional processing schemes?

Questions

Are children using nonlinear frequency compression

(NLFC) in their hearing aids getting better access to the speech signal than children using conventional processing schemes?

We hypothesized that children whose hearing aids

provided wider input bandwidth would have more access to the speech signal, as measured by an adaptation of the Speech Intelligibility Index (SII, ANSI S3.5-1997, R2007)

Questions

Are speech and language skills different for children who

have been fit with the two different technologies; if so, in what areas?

Questions

Are speech and language skills different for children who

have been fit with the two different technologies; if so, in what areas?

We hypothesized that if the children were getting

increased access to the speech signal as a result of their NLFC hearing aids (Question 1), we would see improved performance in areas of speech production, morphosyntax, and speech perception compared to the group with conventional processing.

Demographics

No significant differences between groups (NLFC and

conventional processing) at any age (3, 4, 5):

Age loss confirmed
Age began intervention
Months using hearing aids
Reported daily use time
Datalogged use time
Mother’s education
Family income
All children wore current hearing aids > 1 year

SLIDE 25

11/1/2013 25 Outcome Measures, Age 3

Goldman-Fristoe Test of Articulation-2 (GFTA-2,

Goldman & Fristoe, 2000) is a standardized measure of speech production;

Vineland Adaptive Behavior Scales-II (VABS-II;

Sparrow, Cicchetti, & Balla, 2005), parent-report questionnaire of personal/social behavior;

Comprehensive Assessment of Spoken Language

(CASL 3-4; Carrow-Woolfolk, 1999), standardized measure of global language development.

Outcome Measures, Age 4

VABS-II also administered in the 4-year old protocol;
Test of Preschool Early Literacy (TOPEL; Lonigan et

al., 2007), standardized measure of early literacy, specifically phonological processing and print knowledge;

CASL 3-4 also administered in the 4-year old protocol;
Wechsler Preschool and Primary Scales of

Intelligence-III (WPPSI-III; Wechsler, 2002), standardized measure of verbal and nonverbal intelligence

Outcome Measures, Age 5

Goldman-Fristoe Test of Articulation-2 also administered in the 5-

year old protocol;

Peabody Picture Vocabulary Test-4 (PPVT-4; Dunn & Dunn, 2007),

standardized measure of receptive vocabulary;

TOPEL also administered in the 5-year old protocol;
CELF-4 Word Structure. Subtest of the Clinical Evaluation of

Language Fundamentals-4 (CELF-4; Semel, Wiig, & Secord, 2003), assesses morphological development using picture stimuli;

Comprehensive Test of Phonological Processing (CTOPP;

Wagner, Torgesen, & Rashotte, 1999), standardized measure of phonological processing;

Preschool Language Assessment Instrument (PLAI-2; Blank et al,

2003), standardized measure of expressive and receptive discourse;

PBKs for speech perception

Frequency (Hz)

250 500 1000 2000 4000 8000

Hearing Level (dB)

10

10 20 30 40 50 60 70 80 90 100 110 Non-Compressed Compressed

Three-year olds

3 year olds

NLFC Conventional P value

GFTA

88.9 99.6 .07

Vineland

94.8 97.1 .66

CASL

82.1 95.5 .02

Better ear PTA

56.5 50.8 .23

Better ear aided SII (50)

.52 .59 .37

Better ear aided SII (65)

.70 .78 .16

Better ear unaided SII

.21 .20 .44 Frequency (Hz)

250 500 1000 2000 4000 8000

Hearing Level (dB)

10

10 20 30 40 50 60 70 80 90 100 110 Non-Compressed Compressed

Four-year olds

SLIDE 26

11/1/2013 26

NLFC Conventional

P value

Vineland

90.29 95.39 .31

CASL

99.77 102.16 .74

TOPEL Phono

85.55 91.67 .40

WPPSI Block

10.67 9.81 .55

WPPSI Reasoning

11.89 10.53 .40

WPPSI Vocab

7.44 8.13 .60

Better Ear PTA

53.0 47.9 .29

4-year olds

Frequency (Hz)

250 500 1000 2000 4000 8000

Hearing Level (dB)

10

10 20 30 40 50 60 70 80 90 100 110 Non-Compressed Compressed

Five-year olds

5-year olds

NLFC Conventional P value

GFTA

93.7 95.0 .84

PPVT

100.8 100.3 .94

TOPEL

104.2 105.7 .74

CELF

9.1 8.2 .57

PLAI

110.9 106.4 .54

PBK

79.0 78.6 .93

Better Ear PTA

52.4 51.6 .85

Limitations

Not a true comparison of impact of NLFC on bandwidth

(i.e., audibility) in that this was a between-groups analysis;

Reflects best-case fitting methods, which may not be

representative of other clinics;

The audiometric data of the subjects did not support

assumption that NLFC would be more readily fit to children with more sloping configuration of loss.

Summary of OCHL findings

In this study, audiograms and unaided audibility (ala SII)

same for both groups at each age;

Aided audibility was not different for the two groups

(NLFC and Conventional) for soft or average inputs;

As an expected consequence, speech and language
utcomes were not different for the two groups.
Emerging data suggest that detection may be enhanced

for some children, but there is still little evidence of broader advantage for children of this audiometric profile.

More longitudinal data of this sort necessary.

OCHL Team Members

156

University of Iowa

J. Bruce Tomblin, Ph.D. (Co-PI)

Marlea O’Brien, Program Coordinator Rick Arenas (IT) Ruth Bentler, Ph.D. Lenore Holte, Ph.D. Elizabeth Walker, Ph.D., CCC-A/SLP Connie Ferguson, M.S., CCC-SLP Marcia St. Clair, SLP Examiner Wendy Fick Jacob Oleson, Ph.D. (biostatistics) BTNRH Mary Pat Moeller, Ph.D. (Co-PI) Patricia Stelmachowicz, Ph.D. Meredith Spratford, Au.D. Lauren Berry, M.S., CCC-SLP Emilie Sweet, M.S., CCC-SLP Sophie Ambrose, Ph.D. (LENA) University of North Carolina-Chapel Hill Melody Harrison, Ph.D. Patricia A. Roush, Au.D. Shana Jacobs, Au.D.

M. Thomas Page, M.S., CCC-SLP

SLIDE 27

11/1/2013 27

Briefly, for frequency lowering

APFs are manageable, but different for

different algorithms;

Efficacy has been demonstrated repeatedly

in terms of sibilant detection & discrimination for adults and children;

Little effectiveness data not very

encouraging.

What can we do?

i.e., we as in clinicians, not me as in researcher

What can we do?

Know the black box (APFs)

DIR/DNR: test it!
Frequency Lowering: Verify it!

Look at efficacy measures:

Have high ecological validity
Represent individual’s listening environments
Include a variety of test situations

Look at effectiveness

COSI, e.g.
Self-report measures

..and the “evidence” will have the strength (both in level and grade) to impact decision-making in the clinics.

Buzz words…

Evidence-based design Evidence-based practice Evidence Evidence Evidence

So, how does this all go?

Three prongs

Empirical evidence
Clinician experience/evidence
Patient needs and characteristics

Acknowledgements

National Institute on Disability and Rehabilitation Research (NIDRR) National Institute on Health (NIH/NIDCD) ASHFoundation AAA Foundation Starkey laboratories, Inc. Siemens Hearing Instruments, Inc. Research participants