Sound localization with bilateral cochlear implants investigated with a new
localization method 

B. Seeber(1), U. Baumann(2), H. Fastl(1)
(1) AG Technische Akustik, MMK, Munich Univ. of Technology, Arcisstr. 21, 
    Munich, D-80290 Germany
(2) Klinikum Großhadern/ENT/Audiology, Univ. of Munich, Marchioninistr. 15, 
    Munich, BY 81377 Germany,
 

To assess the localization ability of hearing impaired patients currently used
audiological tests apply the method of identification. For that purpose in a set
of loudspeakers the subject has to identify the perceived sound source direction
and name or show the speaker from which the sound was presented. Although this
method may be useful in clinical examinations due to its simplicity, the
applicability of source identification studies in audiological and
psychoacoustic research is limited, as identified sound directions are quantized
to the position of the speakers in the setup. To overcome this limitation a new
localization method was developed which allows an un-quantized, continuous
scaling of perceived sound direction by a movable light spot. By turning the
ball on a computer trackball the subject moves the spot of a laser pointer on a
horizontal track to the perceived auditory direction. The light spot is
projected onto a curtain which covers the loudspeakers. By using the trackball
the new method can be handled fast and intuitively and is therefore usable also
for patients of nearly all age groups.  

Using the new localization method experiments with 4 bilateral cochlear implant
patients were carried out to assess the ability and accuracy of sound
localization. All subjects showed localization ability with varying accuracy.
One bilateral supplied subject showed a localization accuracy of 5 degrees. For
this subject the absolute localization error is about 6 degrees and the
correlation between presented and perceived directions reaches 0.996. Using a
single CI this subject was also able to differentiate the side of sound origin
even with roving stimulus level.