Mental representations of and adaptation to speech transmitted via cochlear implants: How the impoverished signal finds its way to the mental lexicon

Cochlear implants (CIs) allow deaf people to re-gain their speech perception.
After implantation listeners need to adapt to a very different sensory
experience, facilitated through the electric stimulation of the auditory nerve.
While this adaptation is utmost important for a successful outcome of the CI,
there are substantial individual differences in the success of this adaptation
and the factors that cause such variation are not well understood. This project
aims at examining causes of such individual variation by relating speech
processing with CIs to models of speech perception and lexical access. These
models explain differences in processing efficiency with degrees of match
between the input signal and the mental representations on various levels. As
long as the representations are accurate there is a constant flow of
information between levels of processing, such as semantic, lexical and
sub-lexical levels. This interaction secures the efficiency of speech
processing, it minimizes the listening effort, and offers means to compensate
for mishearing and to build up predictions about upcoming events. However, such
a perfect match is rarely the case in CIs. The acoustic input does not match
the representations because only parts of the acoustic information from the
natural signal are transmitted in CIs, and, more importantly, listeners' mental
representations and their processing of speech may have been altered through
long-term exposure to degraded signals.
This project will examine how speech processing with CI fits within the
architecture of models of speech perception. The study aims to give a
comprehensive perspective on speech processing in CI and to track the way in
which the CI signal maps onto the lexicon. If long-term exposure to degraded
signals alters the cognitive-perceptual processing of speech then individuals
may continue to use such non-optimal type of processing even after their
sensory processing has been improved. The questions in this project are thus
relevant for our basic understanding of speech perception, and for development
of new methods in hearing rehabilitation, especially for CI end-users.

The project has three main scientific objectives:
(1) To investigate how speech processing with CIs affects listeners* mental
representations at lexical and sub-lexical levels, as well as the flow of
information between these levels;
(2) To study how mechanisms of speech perception relate to measures of cortical
processing, at the level of sensory processing and at the level of predictive
computing;
(3) To investigate the reverse plasticity of mental representations as a
response to long-term sensory deprivation in long-term users of CI, and the
recovery plasticity as a response to the processing of a different sensory
information in new users of CIs.

The project will consist of two parts. Part I will combine behavioural
experiments with event related potential (ERP) recordings and an eye-tracking
experiment. This part will compare the performance of long-term CI users with
normal hearing (NH) listeners, in normal and in CI-simulated situations. Part
II will be a longitudinal study of new users of CIs after the fitting with a
new CI. These participants will be tested on the set of experiments used and
developed in Part I, at three points in time: within 2 weeks, within 2 months
and 3-5 months after the fitting. Comparisons will be drawn across groups
(CI-users, NH, NH with CI simulation), and correlations across experiments will
be made within subjects. The longitudinal experiments will lead to
within-subject comparisons across the testing sessions.

There is no known risk, nor benefit associated with participation. All
experiments are non-invasive in nature. A test session will last for a maximum
of 4 hours per listener, including breaks. Session duration will be
accommodated on the request of the participant as needed. The sound level will
always be adjusted to a comfortable listening level for the participant.