97 / 162
95
5
M-levels. Pfingst and Xu [2005] reported a strong correlation between DRs within subjects with bipolar or
monopolar stimulation and showed a mean DR of 8.1 dB for those using monopolar stimulation. Although
some level data concerning MED-EL recipients were published [Sainz et al., 2003; Boyd, 2010], no T-level
distribution or DR could be derived from these studies. Bonnet et al. [2012] reported a T-/M-level ratio
between 14 and 21% (8.5–6.8 dB DR) for CII/HiRes 90K implants, depending on the rate used.
During the initial period of cochlear implant use, regular fittings are common, but thereafter, fitting can also
be necessary – and, indeed, changes over time in levels have been described. Smoorenburg [2007] reported
increases in Tand M-levels for both adults and children. In contrast to Smoorenburg [2007], Walravens et al.
[2006], Hughes et al. [2001] and Wesarg et al. [2010] showed stable T-levels in adults after initial fitting. In
children, Hughes et al. [2001] showed that Tand M-levels continued to increase months after implantation
and first fitting. The eCAP thresholds also increased over time, suggesting changes beyond simple learning
effects [Hughes et al., 2001]. Zwolan et al. [2008] and Henkin et al. [2006] showed increases in M-levels
for children, with the largest increase in the first months. No comparable quantification of change in levels
over time can be distilled from the studies mentioned above.
All effort put into fitting is meant to maximize speech understanding for the individual recipient. However,
Shannon [2002] showed that speech understanding is relatively unaffected by amplitude changes such
as peak or center clipping or amplitude compression. Modest effects of average fitting levels on speech
perception have been found [Pfingst et al., 2004; Pfingst and Xu, 2005], with significant predictive values
of M-levels and DR. This confirmed other work, showing that a larger DR correlated with better speech
perception [Blamey et al., 1992].
Although it was predicted by Pfingst et al. [2004] that M-levels would correlate better with speech
understanding than wouldT-levels, across-site variance ofT-levels correlatedmore with speech understanding
than did across-site variance of M-levels [Pfingst and Xu, 2005]. The fact that across-site variation in T-levels
is correlated with speech understanding gives strength to the hypothesis that excluding electrodes with
aberrant patterns of neural stimulation could improve speech recognition. With tripolar pulses, Bierer et al.
[2010] demonstrated considerable across-site variation and showed that electrodes with higher thresholds
have broader tuning curves and smaller DRs. With monopolar stimulation, this across-site variation was
much smaller. However, even with monopolar stimulation, numerous researchers report higher thresholds
at the basal end of the electrode array [Thai-Van et al., 2001; Smoorenburg et al., 2002; Sainz et al., 2003;
Miller et al., 2008; Lai et al., 2009; Botros and Psarros, 2010; Wesarg et al., 2010]. Nowadays, fitting at the
basal part of the cochlea receives extra attention as electric acoustic stimulation is emerging [Adunka et al.,
2010] and shorter electrodes are being used [Gantz et al., 2009; Lenarz et al., 2009].
The aim of this study is to predict fitting parameters on the basis of behavioral levels for a group of 151
cochlear implant recipients, all implanted with a Clarion HiFocus 1/1 J electrode. The Tand M-levels
obtained during regular follow-up 1 year after implantation provided the basis for this prediction. To allow