Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
Application of electrophysiological Application of electrophysiological measurements in processor fitting;measurements in processor fitting;
parametric fittingparametric fitting
•• Guido F. SmoorenburgGuido F. Smoorenburg
•• Christina Christina WilleboerWilleboer
•• Johannes E. van Johannes E. van DijkDijk
QuestionQuestion•• Starting rehabilitation of a cochlear implantStarting rehabilitation of a cochlear implant
patient by assessing the Tpatient by assessing the T-- and Cand C--levels for 22levels for 22electrodes is quite cumbersome, particularlyelectrodes is quite cumbersome, particularlyin children or those individuals who have littlein children or those individuals who have littlehearing experience at all.hearing experience at all.
•• Can we simplify the procedure,Can we simplify the procedure,make it more patientmake it more patient--friendly,friendly,e.g.e.g. by using electrophysiological data,by using electrophysiological data,and / or by using speech stimuli ?and / or by using speech stimuli ?
ApproachApproach•• The literature (The literature (AbbasAbbas, Brown) shows that , Brown) shows that
the correlation between Tthe correlation between T--levels and NRTlevels and NRT--thresholds is weak.thresholds is weak.
•• However, the profile of the NRTHowever, the profile of the NRT--thresholds thresholds across the electrode array may be indicative across the electrode array may be indicative of the profile of the Tof the profile of the T-- and Cand C--levels.levels.
•• Take the NRT profile and adjust the overall Take the NRT profile and adjust the overall level of the profile to threshold and comfort level of the profile to threshold and comfort level using life speech.level using life speech.
MethodsMethods•• Adult subjectsAdult subjects•• SPEAK strategySPEAK strategy•• PostPost--operative NRT measurementsoperative NRT measurements•• NRT measurements at electrodes 3NRT measurements at electrodes 3--2222•• OverOver--all adjustment of NRT profile to Tall adjustment of NRT profile to T--
and Cand C--levels using life voice > NRTlevels using life voice > NRT--MAPMAP•• Speech reception tested with CVC wordsSpeech reception tested with CVC words•• Speech test immediately after assessing Speech test immediately after assessing
NRTNRT--MAP and after 2 weeks of optional MAP and after 2 weeks of optional use of the NRTuse of the NRT--MAP MAP
Results 1Results 1•• Postoperatively, we found NRT thresholds at Postoperatively, we found NRT thresholds at
all 20 electrodes in 13 < 27 subjects (48%). all 20 electrodes in 13 < 27 subjects (48%). •• The result of the Cochlear field study is The result of the Cochlear field study is
similar: 65 < 147 (44%).similar: 65 < 147 (44%).•• TT--NRT, the TNRT, the T--levels assessed by adjusting the levels assessed by adjusting the
overall level of the NRToverall level of the NRT--profile using life profile using life voice, appeared to be much lower than the voice, appeared to be much lower than the conventionally determined Tconventionally determined T--levels.levels.
Comparison of NRTComparison of NRT--thresholds thresholds to Tto T-- and Cand C-- levelslevels
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
NRT profile touching NRTNRT profile touching NRT--levelslevels
profiles
100
110
120
130
140
150
160
170
180
190
200
210
220
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRT
Comparison of NRTComparison of NRT--thresholds to Tthresholds to T--and Cand C-- levelslevels
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
Results 2Results 2•• The TThe T--NRT levels are much lower than the conventional NRT levels are much lower than the conventional
TT--levels.levels.•• The CThe C--NRT levels are somewhat higher than the NRT levels are somewhat higher than the
conventional Cconventional C--levels.levels.•• Thus, the dynamic range of the NRT MAP is about twice Thus, the dynamic range of the NRT MAP is about twice
the range of the conventional MAP.the range of the conventional MAP.•• For 1/3 the lower TFor 1/3 the lower T--NRT levels may be explained by the NRT levels may be explained by the
assumption of response of the 1assumption of response of the 1stst electrode.electrode.•• For 2/3 the lower TFor 2/3 the lower T--NRT levels may be explained by NRT levels may be explained by
spatial/frequencyspatial/frequency--band integration of impulse rate.band integration of impulse rate.
CVC scores with CVC scores with behaviouralbehavioural and and NRTNRT--based MAPS, 1based MAPS, 1stst sessionsession
0
10
20
30
40
50
60
70
80
90
100
PG JH MH NH PAK PK MO AR NS MS FT HT JW mean
phon
eme
scor
e (%
)
behavioural
ECAP-based
Hours of use of NRTHours of use of NRT--based MAPbased MAP
0
20
40
60
80
100
120
140
160
180
200
PG JH MH NH PAK PK MO AR NS MS FT HT JW mean
hours
CVC scores with CVC scores with behaviouralbehavioural and and ECAP based MAPS, 2ECAP based MAPS, 2ndnd sessionsession
0
10
20
30
40
50
60
70
80
90
100
PG JH MH NH PAK PK MO AR NS MS FT HT JW mean
phon
eme
scor
e (%
)
behavioural
ECAP-based
*
* *
* ** * *
*
*
*
* few hours of use * worse speech
Results 3Results 3•• On average, phoneme perception is little (7%) On average, phoneme perception is little (7%)
affected by a large decrease in overall Taffected by a large decrease in overall T--levels levels and some upward shift of the overall Cand some upward shift of the overall C--levels. levels.
•• The MAP may change considerably without The MAP may change considerably without affecting speech reception.affecting speech reception.
•• The results show marked The results show marked intersubjectintersubject variability; variability; some subjects perform just as well with the NRTsome subjects perform just as well with the NRT--based map, some perform up to 25% worse.based map, some perform up to 25% worse.
•• WHY ?WHY ?
Examples of individual profiles Examples of individual profiles displayed in terms of the two factorsdisplayed in terms of the two factors
-2
-1
0
1
2
-2 -1 0 1 2
factor 1
fact
or 2
100
120
140
160
180
200
220
240
20 18 16 14 12 10 8 6 4 2
# electrode
curre
nt u
nits
Profile factor 1: overall levelProfile factor 1: overall level
-4
-3
-2
-1
0
1
2
100 120 140 160 180 200 220
avg. level
fact
or 1
T-level
C-level
NRT-level
T-NRT
T-NRT-T
C-NRT
Profile factor 2: slopeProfile factor 2: slope
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5
slope in current units / electrode
fact
or 2 T-level
C-level
NRT-level
Correlation between overCorrelation between over--all Tall T--, C, C-- and NRTand NRT--levelslevels
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3C-level
NR
T-le
vel
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3T-level
NR
T-le
vel
scale unit equals ~ 20 current units
Corr. coeff = 0.64 Corr. coeff = 0.39
Speech perception Speech perception reredifference in overall leveldifference in overall level
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
-4
-3
-2
-1
0
-3 -2 -1 0 1T-level
T-NR
T le
vel
worse speech similar speech-2
-1
0
1
2
-2 -1 0 1 2C- level
C-N
RT
leve
l
worse speech similar speech
scale unit equals ~ 20 current units
Correlation between tilts Correlation between tilts in Tin T--, C, C-- and NRTand NRT--levelslevels
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3C-level
NR
T-le
vel
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3T-level
NR
T-le
vel
Corr. coeff = 0.36
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
Corr. coeff = 0.80
scale unit equals ~ 2/3 cu/electrode
profiles
100
110
120
130
140
150
160
170
180
190
200
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
# electrode
cure
nt u
nits
T-levelC-levelNRT-levelT-NRTC-NRTT-NRT-T
Speech perception Speech perception reredifference in tiltdifference in tilt
-2
-1
0
1
2
-2 -1 0 1 2 3T-level
T-N
RT
leve
l
worse speech similar speech-2
-1
0
1
2
-2 -1 0 1 2 3
C-level
C-N
RT
leve
l
worse speech similar speech
scale unit equals ~ 2/3 cu/electrode
Individual example of poor Individual example of poor speech perceptionspeech perception
80
120
160
200
240
22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3
# electrode
curr
ent u
nits
T-level
C-level
NRT-level
T-NRT
C-NRT
Results 4Results 4•• The TThe T--, C, C-- and NRTand NRT--profiles can de described by profiles can de described by
only two factors: only two factors: overall leveloverall level and and tilttilt..•• For For overall leveloverall level the correlation between Tthe correlation between T--level level
and NRT threshold is weak (0.64), and NRT threshold is weak (0.64), for Cfor C--level and NRT threshold it is poor (0.39).level and NRT threshold it is poor (0.39).
•• For For tilttilt the correlation between Tthe correlation between T--level and NRT level and NRT threshold is satisfactory (0.80), threshold is satisfactory (0.80), for Cfor C--level and NRT threshold it is poor (0.36).level and NRT threshold it is poor (0.36).
Results 5Results 5•• NRT predicts the TNRT predicts the T--profile accurately but this profile accurately but this
may be unimportant to speech perception.may be unimportant to speech perception.•• NRT does not accurately predict the CNRT does not accurately predict the C--profile profile
whereas the Cwhereas the C--profile may be important to profile may be important to speech perception. Phoneme perception speech perception. Phoneme perception decreases with a marked increase in the tilt of decreases with a marked increase in the tilt of the Cthe C--level profiles.level profiles.
•• However, these results may have been affected However, these results may have been affected considerably by habituation to the initial MAP.considerably by habituation to the initial MAP.
•• Acceptance of stronger highAcceptance of stronger high--frequency frequency stimulation might stimulation might favourfavour speech perception.speech perception.
Parametric fittingParametric fitting
120
140
160
180
200
22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3
# electrode
curr
ent l
evel
-2
-1
0
1
2
-2 -1 0 1 2
overall level
tilt
An implementation of An implementation of parametric fittingparametric fitting