Development of a test to measure listening effort with spoken digits in noise.

Modern hearing aids have several optional signal-processing features that are
meant to optimize the output signal. Although those features do not always
improve speech intelligibility users may prefer the activation of these
features over their inactivation. For example, many hearing aid users like
single-microphone noise reduction, while it does not improve speech
intelligibility scores. This preference might be caused by a reduction in the
effort required to obtain the same intelligibility scores. Although this
listening effort can be a prime factor determining the success of
rehabilitation with hearing aids, currently there is no method to clinically
measure listening effort. Recent results from our laboratory suggest that the
response time to spoken digits in noise can be used to measure listening effort
(Houben et al. 2012, in revision).

In the current study, we will further evaluate the use of response time to
digits for objective measurement of listening effort. The purpose is to
determine if there is an effect of single-microphone noise reduction on the
response time to spoken digits in noise, despite the lack of intelligibility
improvement.

This study is an experimental lab study, using the method as described by
Houben et al. (2012, in revision) where response times are measured to spoken
digit triplets in noise. The response time will be measured for two tasks
(crossover design). In the *identification* task, the participants have to
quickly identify the final digit of a triplet and in the *arithmetic* task they
have to quickly add the initial and the final digit of a triplet. The speech
signals are presented at four SNRs (-5, 0, 5dB and in quiet), to which three
types of processing are applied; *unprocessed* (no noise reduction), and two
types of noise reduction: minimum mean square error (*MMSE*, Wang 2005)
estimation and ideal binary masking (*IBM*, Ephraim and Malah 1984). We will
assess the effects of the two noise-reduction algorithms on the response times
of both normal hearing (NH) and hearing impaired (HI) listeners. Pure tone
audiometry will be performed to determine whether the subjects meet the
inclusion criteria. Speech intelligibility will be tested to ensure that there
is no effect of noise reduction on intelligibility, and listening effort rating
will be performed as a subjective measure of listening effort.

Since this study is observational, the burden for the subjects is minimal.
Several tests related or equal to those done in standard clinical practice will
be done during the two visits. This study forms an important step towards the
goal of being able to measure listening effort in a clinical setting and to use
this outcome for evaluation of hearing aid functions.