Voice Guard and Soft Speech perception

CI Clinical and Scientific Research Group
Neuro One with
Coordinated Adaptive Processing
Evidence Pack
Version 1 - May 2017
Coordinated Adaptive Processing
Hearing Aid
Technologies
Cochlear Implant
Technologies
Inium Chip
Environment Detection
Group Evaluations - 1
Multicentric Evaluation, Italy
Group audiological evaluation -1- Italy
Study details
Patients:
10 patients (6f; 4m) aged 54,3 yrs on average.
Postlingual deaf adults, deprivation durations from 1 to 8 yrs.
Very homogeneous group.
Multicentric data collection at 3 months post-CI.
Centers: Piacenza 5, Lecco 3, Monza 1, Parma 1.
Mapping parameters:
Electrodes:
Strategy:
Rate:
N-M:
9: 20 electrodes / 1: 19 electrodes.
Crystalis xDP (preferred during habituation period)
346 to 520 Hz
6 to 8 peaks
VoiceTrack:
Soft
Group audiological evaluation -1- Italy
Study details
Test setup:
Tests ran according to the Italian “Common protocol for the evaluation of outcomes in
rehabilitative audiology, 1995”
Words: 10 open-set lists of 10 dissyllabic common words as “Pane”
bread, “Fuoco” fire. - 1 list leads score /10 - to percent. 1 correct item – 1 point.
Sentences: 10 closed-set lists of 10 * 6-7 words sentences as “Il saggio dice solo cose vere»
the wiseman only tells the truth.
Noise: Cocktail-party noise from back 55 dB SPL - target speech from front 65 dB SPL.
SNR = +10dB.
Group audiological evaluation -1- Italy
3 Months post-CI : Averaged Group Results
Average scores:
Words
Quiet:
Noise :
88.0% +/- 13.4
55.5% +/- 13.8
Sentences
Quiet:
Noise :
91.3% +/- 9.8
63.5% +/- 14.2
Group audiological evaluation -1- Italy
3 Months post-CI : Detailed Individual Results – Words Comprehension
Very good outcomes and fast benefits with a group of ‘ideal’ patients
(young – postlingual adults – short deprivation period)
Group audiological evaluation -1- Italy
Comparative Data: 1
Test setup:
Two groups in the study:
Elderly: 16 patients, mean age 66.8 yrs, mean
deafness duration of 21 months.
Young Controls: 14 patients, mean age 51.2 yrs, mean
deafness duration of 23 months.
Cochlear N22 or N24 devices not last generation
Measures at 70 dB SPL in sound field, 12 months after cochlear
implant surgery.
In comparison to Neuro patients: not bad match on certain
dimensions (type of measures, age… ), deprivation period a
little shorter, CI experience a little longer for Pasanisi patients or
device generation, not the last (mostly Cochlear Speak patients).
No “significant” difference just showing that we are in the upper
part of the expected performance range.
Group audiological evaluation -1- Italy
Comparative Data
(Pasanisi et al., 2003)
Group audiological evaluation -1- Italy
Comparative Data: 2
Test setup:
Patients: 26 Cochlear Nucleus Freedom users (CI24M or CI24R)
tested with Ace/Speak strategy of MP3000 stratgey.
CI experience at least 6 months.
Büchner et al., CII, 2011
No details for the Italian subgroup, but over the whole population:
Median age was 55 years; the range was 12–85 years.
Nine subjects were younger than 18 years.
Median severe-to-profound hearing loss (SPHL) duration
was 10 years. SPHL duration in 12% of the 209 subjects was
larger than 30 years. Twenty-six subjects had pre-lingual SPHL.
In comparison to Neuro patients: not bad match on certain
dimensions (type of measures, age… ), deprivation period
potentially longer in Cochlear group, but CI experience also a little
longer.
No “significant” difference just showing that we are in the upper
part of the expected performance range.
Group audiological evaluation -1- Italy
Comparative Data
(Büchner et al., 2011)
Neuro One
Büchner et al., 2011
Neuro One
Büchner et al., 2011
Group Evaluation - 2
Monocentric Evaluation, Germany
Hannover Medical School Hospital
Group evaluation – 1 – MHH, Germany
Study details
Neuro Patients:
10 patients aged 64.6 yrs on average.
Long deprivation duration: 18.2 yrs.
Postlingual deaf adults, variable etiologies.
Monocentric data collection at 3 to 12 months post Neuro-Zti surgery.
Matched Control Group:
174 patients from the MHH CI patient’s database matched for:
Speech perception in quiet - words scores (i.e., Spahr et al., EarHear, 2007)
Age: 61.7 yrs on average.
Deprivation duration: 17.4yrs.
Monocentric retrospective analysis, patients tested at 3 months post CI.
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Study details
MHH Unmatched group:
480 patients from the MHH CI patient’s database
Average deafness duration: 5.4 yrs.
Only newest implant generation (2014 analysis).
Monocentric retrospective analysis, patients tested at 3 months post CI.
Matched Control Group:
174 patients from the MHH CI patient’s database matched for:
Speech perception in quiet - words scores (i.e., Spahr et al., EarHear, 2007)
Age: 61.7 yrs on average.
Deprivation duration: 17.4yrs.
Monocentric retrospective analysis, patients tested at 3 months post CI.
patients.
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Study details
Test setup:
Monosyllables:
Freiburger Monosyllabic Test, 20 monosyllabic words at 65 dB HL.
1 correct word = 1 point – transformed to percent.
Sentences:
German HSM sentence test. 30 lists of 20 everyday
sentences, presented at 65 dB HL.
Noise:
Stationary speech-shaped noise was applied at +10 dB SNR.
Some of the patients presenting some residual hearing in the contralateral ear, the
tests were administered via cable connection, testing only the CI ear.
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Results
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Results
Monosyllables
Unmatched Controls: 60.0%
Matched Controls:
41.0%
Neuro Patients:
41.0%
HSM Sentences Quiet
Unmatched Controls: 80.0%
Matched Controls:
62.0%
Neuro Patients:
63.0%
HSM Sentences Noise
Unmatched Controls: 30.0%
Matched Controls:
12.0%
Neuro Patients:
24.4%
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Results
Monosyllables
Unmatched Controls: 60.0%
Neuro Patients:
41.0%
HSM Sentences Quiet
Unmatched Controls: 80.0%
Neuro Patients:
63.0%
HSM Sentences Noise
Unmatched Controls: 30.0%
Neuro Patients:
24.4%
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group evaluation – 1 – MHH, Germany
Discussion
• Interesting observation showing a potential advantage of Neuro system in noise
condition.
• Potential listening effort reduction with Coordinated Adaptive Processing, needs to
be further explored.
• Would first have to be reproduced on a larger group, with younger, shorter
deprivation period patients: Multicentric study in Germany started.
Group evaluation – 1 – MHH, Germany
Results
Modified from A. Büchner et al., 2017
Oticon Medical Scientific Meeting, KBN, Den.
Group Evaluations – 3
Multicentric Evaluation, France
Group audiological evaluation - 3 - France
Study details
Patients:
12 patients (6f; 6m) aged 20 to 75 yrs.
9 postlingual adults, deprivation durations up to 30 yrs.
3 prelingual adults, long deprivation durations.
Variable etiologies. Very heterogeneous group.
Multicentric data collection between 3 and 12 months post-CI.
Centers: Lille 7, Bordeaux 4, Brest 1
Patient withdrawn from analysis:
- 1 patient (Brest) was withdrawn from analysis because she was a very
long deprived, aged (75-85 y) prelingual patient, very low speech
comprehension in quiet and not testable in noise.
Group audiological evaluation - 3 - France
Study details
Test setup:
Words: Fournier’s dissyllabic words lists: 30 closed-set lists of 10 dissyllabic
common words as “Poulet” chicken, “Savon” soap. - 2 lists leads score /20 - to
percent. 1 correct word - 1 point. The lists are also balanced in phonemes, possible
to score in nb of phonemes.
Noise: Vocal audiometry: stationary speech shaped noise coming from the side
or front 55 dB SPL - target speech from front 65 dB SPL. SNR = +10dB.
Group audiological evaluation - 3 - France
Averaged Group Results
Detailed scores:
Words
Quiet:
Noise:
65.9% +/- 35.6
53.6% +/- 31.2
Note the small difference between quiet and noise (-12%),
Borger et al., 2015 showed -26% difference with Digisonic
after 1 year.
Group audiological evaluation - 3 - France
Individual Results
Group audiological evaluation - 3 - France
Individual Results
Brain Plasticity Matters
Two experienced CI users (re-implantations)
Contralateral Digisonic SP
Tested with their Neuro CI only !
Group audiological evaluation - 3 - France
Individual Results
Brain Plasticity Matters
Three patients with very long deprivation periods
(>30 years) and late CI surgery
Coordinated Adaptive Processing
Feature Evaluation
Multiple Case Studies
Lille & Bordeaux University Hospitals, France
Coordinated Adaptive Processing
Free Focus
Rationale
A typical user spends 70% of
time in Surround mode.
Focus on improving the
situations that matter the most
Optimised Omni vs. Speech Omni
Study details
Speech Omni is a light directional mode
that has a high frequency cut-off between
directional and omnidirectional at 1880 Hz.
This keeps the lower frequencies omnidirectional,
allowing environmental input from all directions,
and high frequencies in a more directional mode
to enhance focus on the source of interest when
positioned in front of the listener.
Optimised Omni vs. Speech Omni
Study details
Patients:
6 Neuro One users.
Words:
2 Fournier’s Lists: common disyllabic words.
Procedure:
Speech in Silence and in Spatialized noise.
Stationary Speech shaped noise at +10 dB SNR.
With Optimised Omni mode or Speech Omni mode.
Optimised Omni vs. Speech Omni
Results in Silence (N=6)
Speech Omni increased speech perception
scores in Quiet by +11%.
Speech Omni therefore provides a benefit in the
situation that users are most likely to spent time in.
Optimised Omni vs. Speech Omni
Results in +10 dB SNR Noise (N=6)
Speech Omni increased speech perception
scores in Noise by +10%.
Speech Omni seems to improve speech reception in
CI users in quiet environments and in situations of
light noise level.
Optimised Omni vs. Speech Omni
Individual Results (N=6)
Silence
Noise
Speech Omni did not cause any significant
decrease in performance in any user.
In some patients, the improvement was large.
Optimised Omni vs. Speech Omni
Discussion - Conclusions
Speech Omni is a light directional mode that has a high frequency cut-off between
directional and omnidirectional at 1880 Hz.
Speech Omni improved speech perception scores in Quiet and in a light noise
environment compared to Optimised Omni in a group of 6 Neuro One users.
Patients seemed to react positively to Speech-Omni, and none showed a decrease in
performance in the two conditions tested.
This observation must be confirmed with a larger group, and in real-life situations,
however it seems that Speech-Omni can reduce listening effort and improve speech
perception in CI users.
Coordinated Adaptive Processing
Voice Guard and Soft Speech perception
Rationale
Voice Guard is Neuro One’s post-spectral
analysis adaptive compression system,
placed under the command of the Inium
environment detector.
Voice Guard will adapt its kneepoint in order to
always keep 95% of the speech dynamic onto the
less compressive part of the XDP function.
In particular, Voice Guard will allow Neuro One
patients to get access to soft speech, presented
at low intensity levels.
Voice Guard and Soft Speech perception
Study details
Patients:
6 Neuro One users.
Words:
2 Fournier’s Lists: common disyllabic words.
Procedure:
Speech in Silence at 40 dB A or 65 dB A
Voice Guard set to static mode (laboratory testing)
In Soft-setting or in Medium-setting.
Voice Guard and Soft Speech perception
Results (N=6)
For speech presented at a soft level (40 dB
A), adapting the kneepoint of the Voice
Guard output compression function from
the Medium setting, to the Soft setting
improved speech perception by 16% in a
group of Neuro One users.
Voice Guard is able to maintain speech
perception in difficult listening situations, here soft
speech and will automatically adapt based on the
environment analysis of the Inium Chip platform.
Voice Guard and Soft Speech perception
Individual Results
All six patients showed some benefit of adapting
the compression kneepoint.
None of the tested Neuro One users showed a
decrease in performance after adjusting the
compression kneepoint of the Voice Guard
Function.
Voice Guard and Soft Speech perception
Discussion - Conclusions
Voice Guard is Neuro One’s post-spectral analysis adaptive compression system,
placed under the command of the Inium environment detector.
Voice Guard improved speech perception scores for Soft speech perception (40 dB A presentation
level) in a group of 6 Neuro One users, when adapting the compression kneepoint from the Medium
setting to the Soft setting.
This observation must be confirmed with a larger group, however it seems that Voice Guard can
reduce listening effort and improve speech perception in CI users for otherwise difficult soft speech
perception.
Coordinated Adaptive Processing
Voice Track evaluation
Rationale
Voice Track is Neuro One’s noise reduction
system based on Wiener Filter technology.
Voice Track actively reduces noise in 64
frequency bands based on a dynamic analysis of
sound energy.
Voice Track is particularly efficient to suppress
static noise.
Voice Track evaluation
Study details
Patients:
9 Neuro One users.
Words:
2 Fournier’s Lists: common disyllabic words.
Procedure:
Speech in Noise at +10 dB SNR.
Stationary Speech-derived noise from front direction.
Voice Track turned Off or in Medium setting.
Voice Track evaluation
Group results (N=9)
Voice Track noise reduction system
improved speech in noise perception by
18% in a group of 9 Neuro One users.
Voice Track is able to maintain speech perception
in noisy listening environments.
Voice Track evaluation
Individual results (N=9)
Most of the patients showed a benefit of using
Voice Track for speech-in-noise perception,
benefit ranging from +5% to +50%, except for
one patient at ceiling (100% in noise without
Voice Track).
None of the tested Neuro One users showed
a decrease in performance after adjusting the
Voice Track noise-reduction feature to the
medium setting.
Voice Track evaluation
Discussion - Conclusions
Voice Track is Neuro One’s noise reduction system based on Wiener Filter technology,
operating in 64 frequency bands.
Voice Track improved speech-in-noise perception scores in a group of 9 Neuro One
users, when activating the feature in preset mode Medium.
This observation must be confirmed with a larger group, however it seems that Voice
Track can reduce listening effort and improve speech-in-noise perception in Neuro One
users.
Neuro One audiological outcomes
First audiometric observations with Neuro One
and Coordinated Adaptive Processing
• Results from first speech audiometry on single cases or small groups suggest
competitive outcomes for speech audiometry in Silence and Noise.
• Difficult cases seem to behave among the good range of expected outcomes for
their patient’s group.
• Coordinated Adaptive Processing features including Free Focus directionality,
Voice Guard output compression and Voice Track noise reduction system all
provide measurable benefits to Neuro One users.