Some Factors affecting the prescription and adjustment of hearing aids

1
(Some) Factors affecting the prescription and
adjustment of hearing aids

• Harvey Dillon
• NAL, CRC Hear
• 1st International Congress on Gerontology
• Stockholm
• June 2004

2
To download this talk

• www.NAL.gov.au
• After June 16

3
Who really did the work ..

• Teresa Ching
• Frances Lockhart
• Emma Van Wanrooy
• Chris Brew
• Gitte Keidser
• Lydia Lai

• Wouter Dreschler
• Richard Katsch
• Karolina Smeds
• Liz Convery
• Justin Zakis

Dillon, NAL
4
Overview

• Real life versus prescriptions
• The intelligibility of speech
• Desired loudness
• Acclimatisation effects
• Binaural loudness summation
• NAL-NL2
• Client fine-tuning
• The Trainable Aid the pressures off
  prescription

Dillon, NAL
5
Fundamental differences

• Real life
• Supra-threshold levels
• Broad-band signals
• Hearing aid prescription
• Threshold-based
• Narrow-band signals (pure tones)

Dillon, NAL
6
Supra-threshold measures
Hearing threshold measurements
Optimal amplification characteristics
Estimated optimal amplification characteristics
Supra-threshold measurements
Supra-threshold characteristics
Dillon, NAL
7
Speech intelligibility

• Teresa Ching, Frances Lockhart, Emma van Wanrooy,
  Lydia Lai, Chris Brew

Dillon, NAL
8
Prescription and intelligibility

• NAL-NL response that maximises speech
  intelligibility
• Need to predict intelligibility for any
  gain-frequency response
• To prescribe accurately, we need to know how to
  predict speech intelligibility
• Prescription affected by other things
  (localization, tonal quality, detection, comfort,
  naturalness.

Dillon, NAL
9
30
1/3 octave SPL
Noise
Threshold
Freq
Dillon, NAL
10
The transfer function
Dillon, NAL
11
Speech intelligibility also depends on Level
distortion

• Normal hearing people perform poorer at high
  speech levels
• Recent evidence suggests that degradation is
  greater at high than at low frequencies (Molis
  Summers, 2003)

Dillon, NAL
12
Observed and Predicted performance
Dillon, NAL
Ching, Dillon Byrne, 1998
13
Modified speech intelligibility index
Hearing Loss Desensitization
Dillon, NAL
Ching, Dillon, Katsch Byrne (2001)
14
Frequency resolution
Dillon, NAL
15
Frequency resolution

• Severe and profound
• Outer hair cells mostly destroyed
• Frequency resolution much worse than normal
• Degradation is greater at high than at low
  frequencies

Dillon, NAL
16
Temporal resolution
Dillon, NAL
17
Temporal resolution

• Severe and profound
• Temporal resolution much worse than normal
• For the same degree of hearing loss, resolution
  degrades more at the high than at the low
  frequencies

Dillon, NAL
18
Summary
Speech intelligibility
Frequency resolution
Temporal resolution
Hearing thresholds
Dillon, NAL
19
What affects hearing loss desensitisation ?
.. Once we allow for the effects of hearing
threshold at the same frequency
TRIadj FRIadj HTL.35k- HTL1k- HTL2k-
HTL4k- Age CUI HTL HTL
HTL HTL
Supra-threshold testing not worthwhile !
Dillon, NAL
20
Correlation of frequency and temporal resolution
with speech intelligibility?

• Progressive loss of frequency and temporal
  resolution as hearing loss progresses - high
  correlation with speech intelligibility
• With hearing threshold allowed for (partialled
  out), little further correlation between
  resolution and intelligibility, so NAL-NL1 based
  on thresholds.

Then two new findings came along .
Dillon, NAL
21

• 1. Dead regions

RIP
NAL-NL1 only allows for hearing loss
desensitization on average
Dillon, NAL
22
Off-frequency listening TEN test
Basilar membrane vibration
Frequency
Dillon, NAL
Based on Moore (2004)
23
Off-frequency listening PTC
Basilar membrane vibration
Frequency
Dillon, NAL
24
Off-frequency listening
Basilar membrane vibration
Frequency
Dillon, NAL
25
HLD and Cochlear dead regions

• SII over-estimated performance of people with
  dead regions (Baer et al, JASA, 2002)

No dead regions
Measured
Predicted
Dillon, NAL
26
Audiograms with and without dead
regions(Vickers et al., 2001 Baer et al., 2002)
Dillon, NAL
27

• 2.
• Desensitization in noise

Dillon, NAL
28
Different HLDs for quiet vs noise?
Yes!
Source Hogan Turner, 1998 Turner Henry, 2002
Quiet
Noise
Dillon, NAL
29
Hearing loss desensitization and frequency
resolution
Ability to extract Speech information From an
audible signal
Dillon, NAL
30
Subjects

• 20 adults with normal hearing
• 59 adults with sensorineural hearing loss
• mild to profound
• Experienced hearing aid users

Dillon, NAL
31
Method

• Hearing threshold levels
• Outer hair cell function
• click-evoked otoacoustic emissions
• Frequency resolution
• psychophysical tuning curves
• cochlear dead regions TEN test
• Speech perception in quiet and noise
• consonants
• sentences

Dillon, NAL
32
Speech perception

• Stimuli Filtered speech
• CUNY sentences
• VCV syllables
• Shaping
• POGO prescription
• Conditions
• Quiet at high and low sensation levels
• Babble Noise
• Headphones Sennheiser HD25

Dillon, NAL
33
Audibility and Speech intelligibility
Preliminary analysis
Preliminary analysis
Dillon, NAL
34
Dillon, NAL
35
Proficiency SA / S
100
80
Proficiency 0.3/0.6 0.5
60
Percent Correct
40
?
20
0
0
0.2
0.4
0.6
0.8
1
Speech Intelligibility Index (SII)
Dillon, NAL
36
Psychoacoustical tuning curve Q10
Fc
Q Fc / BW Q10 1000/340 3.4
Dillon, NAL
37
Healthy PTC no dead region
Dillon, NAL
38
Poor PTC Dead region at 4 kHz
Dillon, NAL
39
Psychophysical tuning curve and cochlear dead
region 2 kHz
Dillon, NAL
Sharply tuned
40
Psychophysical tuning curve and cochlear dead
region 4 kHz
Dillon, NAL
41
Does frequency selectivity affect speech
proficiency?
Dillon, NAL
42
Speech proficiency vs tuning curve sharpness low
pass
Dillon, NAL
Poor tuning
Sharp tuning
43
Speech proficiency vs tuning curve sharpness low
pass
Poor tuning
Sharp tuning
Dillon, NAL
44
Speech proficiency vs tuning curve sharpness
high pass
Dillon, NAL
45
Yes speech proficiency decreases as frequency
selectivity gets broader
There is at least as much need to allow for
hearing loss desensitization in noise as in quiet
Dillon, NAL
46
Does hearing threshold loss affect speech
proficiency?
Dillon, NAL
47
Speech proficiency versus hearing loss low pass
Dillon, NAL
48
Speech proficiency versus hearing loss low pass
Dillon, NAL
49
Speech proficiency versus hearing loss high pass
Dillon, NAL
50
Yes speech proficiency decreases as thresholds
increase
Dillon, NAL
51
Does age affect speech proficiency?
Dillon, NAL
52
Distribution of ages
Dillon, NAL
53
But .

• Age

Hearing threshold
Frequency selectivity
Dillon, NAL
54
Effect of age

• Multi-linear step-wise regression
• Dependent variable
• Proficiency factor
• Independent variables
• Hearing thresholds
• Age category
• Frequency selectivity (Q10 from tuning curve)
• Otoacoustic emission strength
• Elevation in TEN test

Dillon, NAL
55
Multi-linear regression results
Dillon, NAL
56
Effect of age HP 700 Hz
Dillon, NAL
57
Effect of age HPF 2800 Hz
Dillon, NAL
58
Effect of age LPF 700 Hz
Dillon, NAL
59
Multi-linear regression results
Dillon, NAL
60
In most cases age did not affect speech
intelligibility, once hearing loss had been
allowed for
. But these are research subjects!
Dillon, NAL
61
Results summary

• Ability to use speech information decreases with
  increased hearing loss
• Degradation greater at high than at low
  frequencies
• Degradation in noise at least as great as in
  quiet
• Higher proficiency associated with better
  frequency resolution (Q10)
• Partial relation between Q10 and cochlear dead
  region
• Only a minor effect of age demonstrated

Dillon, NAL
/ desired loudness
62

• Desired Loudness
• Karolina Smeds, Justin Zakis, Gitte Keidser, Liz
  Convery, Frances Grant

Dillon, NAL
63
Desired loudness

• NAL-NL1 aims at normal overall loudness.
• What do people want compared to NAL-NL1?
• Can also compare to NAL-RP

Dillon, NAL
64
Preferred gain relative to prescription
Clinician survey 39 clinicians - estimate the
proportion of clients who change gain
Too soft 8 Just right 71 Too loud 21
Dillon, NAL
65
Desired loudness

• Direct experiments Smeds et al
• Adjustment of gain in lab, and in the field

Dillon, NAL
66
Rated loudness - lab
Dillon, NAL
67
Desired loudness re normal in lab
Dillon, NAL
68
Loudness at desired volume setting
Dillon, NAL
69
Desired loudness field
Hearing impaired subjects
Dillon, NAL
70
Desired loudness in field
Normal hearing subjects
Dillon, NAL
71
Desired loudness
Dillon, NAL
72
Gain deviation from NAL-NL1 Zakis
Dillon, NAL
73

• Acclimatization to gain
• Liz Convery

Dillon, NAL
74
Acclimatization to gain

• Literature search

Dillon, NAL
75

• Binaural effects

Dillon, NAL
76
Binaural interactions

• Binaural interference
• minority of elderly (perhaps 10) Jerger et al
  1993 Siegenthaler Craig, 1981
• may arise from cochlear distortion
• simulations Hood Prasher 1990
• diplacusis Markides 1977, 1982
• loss of efferent control of OHC LePage 1989
  Rajan 1995 Scharf et al 1994
• loss of transfer between hemispheres (in corpus
  callosum)
• left ear deficiency on diotic and dichotic word
  scores,
• symmetrical pure tone thresholds, ABR wave V
  latencies amplitudes
• Chmiel et al 1997

To NAL-NL2
/ fine tuning
Dillon, NAL
77

• Patient fine-tuning
• Wouter Dreschler, Gitte Keidser, Liz Convery

Dillon, NAL
78
Patient fine-tuning

• Two loudspeakers and a TV-screen
• Binaural presentation
• With or without hearing aids

Dillon, NAL
79
Controller types
Dillon, NAL
80
Dillon, NAL
81
Preferred gain relative to NAL-RP prescription
Reasonable similarities between the overall
results for different controllers
Consistent differences across listening
environments
Dillon, NAL
82
The starting baseline affects the apparent
preference!
Flatter starting response
Steeper starting response

• The effect occurs almost equally for all videos

Dillon, NAL
83
The starting baseline affects the apparent
preference!

• The effect occurs almost equally for all
  controllers

Dillon, NAL
84
But not all preferred results are equally good!
Dillon, NAL
/ trainable aid
85

• NAL-NL2
• Richard Katsch, Teresa Ching

Dillon, NAL
86
NAL-NL2

• Altered loudness targets less than normal
  loudness, especially for inputs gt 65 dB SPL
• Altered desensitisation amount unknown
• May add supplementary supra-threshold tests?
• Prescriptions at all levels, frequencies and
  hearing losses
• Higher gain for congenitally impaired children?
• Probably still no acclimatization steps

Dillon, NAL
To end
/ Trainable aid
87
Moving away from prescriptions

• The trainable hearing aid
• Justin Zakis, Gitte Keidser,
• Hugh Mcdermott, Liz Convery

Dillon, NAL
88
Aid user adjusts settings...
Trainable Aid
Dillon, NAL
89
Aid user adjusts settings...
Trainable Aid
Dillon, NAL
90
...then presses voting button
Trainable Aid
Dillon, NAL
91
Process repeats for other sounds
Trainable Aid
Dillon, NAL
92
After training, preferred settingsare
automatically applied...
Trainable Aid
Dillon, NAL
93
After training, preferred settingsare
automatically applied...
Trainable Aid
Dillon, NAL
94
After training, preferred settingsare
automatically applied...
Trainable Aid
Dillon, NAL
95
After training, preferred settingsare
automatically applied...
Trainable Aid
Dillon, NAL
96
First and Second Comparison Trial Results
p lt 0.05 p lt 0.01 Only S201 had a
statistically significant difference in scores
between comparison trials (p 0.0176)
Dillon, NAL
97
Trainable aid philosophy

• Prescribe what can be prescribed, automatically
  and in minimum clinical time
• Leave the rest to the client (and the
  intelligence of the hearing aid)

Dillon, NAL
98
thanks for listening
www.nal.gov.au