R

Nh

CFa

b

c

d

e

a

ARR1AA

KLAH

1

etGrle

Emisn

SG

(h

0h

Brain Research Bulletin 91 (2013) 21– 30

Contents lists available at SciVerse ScienceDirect

Brain Research Bulletin

jo u r n al hom epage : www.elsev ier .com/ locate /bra inresbul l

esearch report

eural correlates of listening effort related factors: Influence of age andearing impairment

orinna Bernardinga,b, Daniel J. Straussa,b,c,∗, Ronny Hannemannd, Harald Seidlere,arah I. Corona-Straussa,b

Systems Neuroscience and Neurotechnology Unit, Saarland University, Faculty of Medicine, Building 90.5, 66421 Homburg/Saar, GermanyKey Numerics GbR, Saarbrücken, GermanyLeibniz-Institut for New Materials gGmbH, Campus D2 2, 66123 Saarbrücken, GermanySiemens Audiologische Technik GmbH, Gebbertstrasse 125, 91058 Erlangen, GermanyMediClin Bosenberg Kliniken, Am Bosenberg, 66606 St. Wendel, Germany

r t i c l e i n f o

rticle history:eceived 22 April 2012eceived in revised form8 November 2012ccepted 19 November 2012vailable online 29 November 2012

eywords:istening effort

a b s t r a c t

In the last years, there has been a rising interest to find an objective method to estimate listening effort.Previously, we have shown that the wavelet phase synchronization stability (WPSS), gained from theinstantaneous phase of auditory late responses (ALRs), could serve as a feasible measure for listeningeffort related factors.

In the current study, we examined if the WPSS reflects the listening effort in young as well as middle-aged subjects and in persons with a different degree of hearing loss. To evoke ALR sequences, we generatedsyllabic paradigms with a different level of difficulty to evoke ALR sequences. We expected, due to thevarying task demand, that the subjects require a measurable difference in the amount of effort to solve

uditory late responsesearing impairment

the paradigms. Additionally, a possible age and/or hearing loss related effect on the neural correlates oflistening effort was investigated.

The results indicate, that WPSS reflects the listening effort related factors needed to solve an auditorytask. A further finding was that the reaction time data and the N1 wave amplitude information hardlyyield any correlate of the invested listening effort. In addition, we noticed an age as well as hearing

n the

sensitivity related effect o

. Introduction

In recent years, the interest in finding an appropriate way tovaluate listening effort was constantly rising. So far, only subjec-ive procedures, like questionnaires (Gatehouse and Noble, 2004;atehouse and Akeroyd, 2006), rating scales (Humes, 1999) or self-

eports were applied in hearing aid fitting procedures to estimateistening effort. But an objective method to quantify this cognitiveffort in clinical settings is still not available.

Abbreviations: ALR, auditory late responses; DSP, difficult syllabic paradigm;RP, event related potentials; ESP, easy syllabic paradigm; manh, middle-aged nor-al hearing subjects; mild, mild hearing impaired subjects; mod, moderate hearing

mpaired subjects; HL, hearing level; LE, listening effort; RT, reaction time; SPL,ound pressure level; WPSS, wavelet phase synchronization stability; ynh, youngormal hearing subjects.∗ Corresponding author at: Systems Neuroscience and Neurotechnology Unit,

aarland University, Faculty of Medicine, Building 90.5, 66421 Homburg/Saar,ermany. Tel.: +49 6841 162 4091; fax: +49 6841 162 4092.

E-mail addresses: bernarding@snn-unit.de (C. Bernarding), strauss@snn-unit.deD.J. Strauss), ronny.hannemann@siemens.com (R. Hannemann),arald.seidler@mediclin.de (H. Seidler), corona@snn-unit.de (F.I. Corona-Strauss).

361-9230/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.brainresbull.2012.11.005

listening effort.© 2012 Elsevier Inc. All rights reserved.

There have been different approaches to estimate listen-ing effort objectively. One of the methodologies are dual-taskparadigms (e.g. Sarampalis et al., 2009; Downs, 1982; Rakerd et al.,1996; Gosselin and Gagné, 2010 for a review), which are based ona limited capacity theory of cognitive resources (Kahneman, 1973).The subjects have to perform two competing tasks: a primary lis-tening task and a secondary mostly visual or memory related task.The assumption is, that there is a competition for single limitedresources, so that the performance of the secondary task decreaseswith increasing difficulty level of the primary task. This reductionin secondary task efficiency serves as a measure of listening effort.However, this complex method is influenced by many factors andrequires a large amount of subjects cooperation. As further indicesof listening effort, the pupil response (Goldwater, 1972; Krameret al., 1997; Zekveld et al., 2010) and the galvanic skin response(Mackersie and Cones, 2011) were also taken into account. Krameret al. (1997) and Zekveld et al. (2010) showed in two consecu-tive studies, that the pupil dilates with increasing cognitive load.

In these studies sentences embedded in background noise at sev-eral signal-to-noise ratios had to be repeated by the subjects. Theincreasing pupil diameter is interpreted as an increase of listeningeffort.

2 Resear

fcsrirte

tfifldsa2tnbt(tcshaTMoe

eta(ieostcpn1aetcar2ltmstRtimasrA

2 C. Bernarding et al. / Brain

More recently, Mackersie and Cones (2011) considered dif-erent psychophysiological measures, such as heart rate, skinonductance, skin temperature, electromyographic activity as pos-ible indices for listening effort. The psychophysiological data wasecorded during a speech recognition task. As the task demandncreased, the electromyographic activity as well as the skinesponse were enhanced. They concluded that “the skin conduc-ance appears to be the most promising measure” for listeningffort (Mackersie and Cones, 2011).

Lately, Pichora-Fuller and Singh (2006) emphasized the impor-ance of knowledge about cognitive processes for the hearing aidtting procedure and for the field of rehabilitative audiology. One

ocus of their research was on the effects of age and hearing loss onistening effort. They mentioned that listening effort, which can beefined as the exertion listeners experience by processing naturalignals (e.g. speech) in demanding environments, requires alwaysttentional and cognitive resources (Pichora-Fuller and Singh,006; Gosselin and Gagné, 2010). Several authors assume, thathe cognitive effort can be influenced by two factors: First, by sig-al degradation, which can either result from environmental (e.g.ackground noise) or from biologic variances (e.g. elevated hearinghresholds) and secondly by age-related variations in cognitionPichora-Fuller and Singh, 2006; Akeroyd, 2008). In these cases,he listeners use hierarchically higher cognitive or top-down pro-esses (endogenous modulation), to perceive the attended (speech)ignal correctly. A number of authors investigated the effects ofearing loss and aging on speech/stimulus processing by measuringuditory evoked potentials (Goodin et al., 1978; Oates et al., 2002;remblay et al., 2003; Korczak et al., 2005; Bertoli et al., 2005;artin et al., 2008; Friedmann, 2010 for a review). To date, the main

bjectives of these studies were not linked to listening or cognitiveffort.

In preliminary studies (Strauss et al., 2009, 2010; Bernardingt al., 2010a), we applied a new approach to the problem of lis-ening effort. As listening is an “active and dynamic process ofttending [...] to the information offered by other human beings”Purdy, 1991) it can be defined as “the process of hearing withntention and attention for purposeful activities demanding thexpenditure of mental effort” (Kiessling et al., 2003). Analogue,ur method is based on early stages of selective attention. Thesetages are endogenously modulated, i.e. they depend on atten-ional resources and require the aforementioned (higher-order)ognitive effort (Kahneman, 1973). Regarding auditory evokedotentials, selective attention effects were observed in an endoge-ous modulation of the N1 component (Hillyard et al., 1973,998; Hink et al., 1977; Kauramäki et al., 2007; Rao et al., 2010nd references therein). In these studies, the N1 amplitude wasnhanced due to an increased attentional effort. Besides the tradi-ional averaging procedures to analyze the amplitude and latencyharacteristics of evoked potentials in the time domain, recentpproaches investigate the phase characteristics of single-trialesponses (Ponjavic-Conte et al., 2012; Brockhaus-Dumke et al.,008; Bruns, 2004 and references therein). A measure for the phase

ocking or synchronization of single trials can be calculated inhe time–frequency space. In literature, different terms for this

easure have been proposed: phase locking (value/ index), phaseynchronization, phase consistency (see Bruns, 2004 and referencesherein) or intertrial phase locking (Ponjavic-Conte et al., 2012).ecently, Brockhaus-Dumke et al. (2008) showed in a study relatedo auditory gating in schizophrenia, that the phase of single-trialsn the lower EEG frequency range (alpha and theta band) provides

ore information on auditory information processing as the aver-

ged evoked potentials. Furthermore, they showed that the phaseynchronization in the latency range of the N1 component waseduced in schizophrenia patients compared to the control group.lso Ponjavic-Conte et al. (2012) observed attention effects on the

ch Bulletin 91 (2013) 21– 30

instantaneous phase in the time range of the N1 wave. As soon as aconcurrent auditory stimulus was introduced, the intertrial phaselocking in the theta band was reduced. Regarding the limited capac-ity model of (auditory) attention (Kahneman, 1973), the distractorconsumed attentional resources previously allocated to the targetstream. Kerlin et al. (2010) noticed that auditory selective atten-tion enhances the representation of the auditory speech signal inthe auditory cortex within the theta band (4–8 Hz). They assumethat selective attention acts like a gain mechanisms on the lowerfrequencies resulting in an increased auditory cortical activity. AlsoKlimesch et al. (2007) discussed in a review that the amplitudesof evoked responses reflect frequency characteristics of alpha aswell as theta activity. The relation between the electroencephalo-graphic phase reset due to auditory attention was investigatedin Low and Strauss (2009). It is hypothesized that modifying thephase at the alpha–theta border could modulate the attentionaleffects reflected in the N1 amplitude. In Strauss et al. (2010), weshowed that stabilizing artificially the phase of auditory evokedpotentials resulted in an increase of the N1 negativity. Thus, weassume that analyzing directly the phase of the EEG signals by cal-culating the phase stability reveals a more sensitive correlate oflistening effort as the traditional N1 amplitude measures. See alsoFig. 3 in which the decoupling of the WPSS and averaged amplitudeis noticeable. Furthermore, we were able to show that the instan-taneous phase of the N1 component could serve as an index of theamount of listening effort needed to detect an auditory event, e.g.a target syllable/toneburst. The instantaneous phase was extractedby applying the wavelet transform to the auditory late responses(ALR). We calculated the (wavelet) phase synchronization stabil-ity (WPSS) in the frequency range of the theta band. We assumethat a higher synchronization of the phase reflects a higher cogni-tive effort to solve the auditory task. We used as auditory stimulitonebursts (Strauss et al., 2009, 2010) or noise embedded syllables(Strauss et al., 2010; Bernarding et al., 2010a). These paradigms hada different degree of difficulty constructed by either the frequencyspacing of the tonebursts or by the combination of the syllables.In both studies the subjects had to pay attention to the auditoryparadigms in order to detect a previously defined target stimulus.Then, the WPSS of ALR sequences was calculated for sweeps evokedonly by the target stimuli. This was done in order to minimize theexogenous effects as these evoked potentials are also modulatedby physical attributes of the stimuli (e.g. intensity and frequency)(Hall, 1992). The prior achieved results indicate that the WPSS inthe time interval of the N1 wave of ALRs could serve as a feasiblemeasure of listening effort. For a graphical illustration and a moredetailed description of the previously gained results see AppendixA.

The current study is based on our preliminary results and takesthe following aspects into account: First, we wanted to examine, ifthe WPSS reflects the listening effort in young as well as middle-aged subjects and in persons with a different degree of hearingloss. For this reason, two syllabic paradigms with a different levelof difficulty were generated. The difficulty level was adjusted bythe combination of the syllables. Both paradigms contained thesame target syllable which had to be detected by the subjects. Weexpected that the subjects require a measurable different amountof effort to solve the paradigms due to the different task demand.Furthermore, behavioral data, namely reaction time (RT) and per-formance accuracy (d′prime), were taken into account in order tocomplement the electrophysiological data. The reaction time canbe interpreted as a possible index for the processing of the incom-ing stimuli (Gatehouse and Gordon, 1990). Whereas the d′prime

reflects the individual discrimination ability. Opposed to percentcorrect scores, also the false alarm rate of the subjects responsewas included (Yanz, 1984). The second purpose was to investigate,if there is a possible age and/or hearing loss related effect on the

Research Bulletin 91 (2013) 21– 30 23

lm

2

2

(

(

Siwm

dpccic

�

wcAi

wdt

g

(

(

E1iiihFd

2

/psbrebwpiIg

frequency [Hz]

60

50

40

30

20

10

0

ynh

manh

mild

mod

100 500 1k 2k 4k 6k

hearinglevel[dBHL]

standard

deviation

0

10

20

Fig. 1. Mean pure tone audiograms (top) and corresponding standard deviations(bottom) for each subject group (ynh, young normal hearing subjects (meanPTA4f: 10.25 ± 4.01 dB HL); manh, middle-aged normal hearing subjects; mild,

C. Bernarding et al. / Brain

istening effort and its neural correlates, respectively. Therefore, theain results of each subject group were compared to each other.

. Methods

.1. Participants and inclusion criteria

A total of 94 subjects participated in this study:

(a) 24 young subjects (13 male/11 female) with normal hearing levels (ynh; aged21–35 years, mean age: 25.25 ± 4.01 years),

b) 21 middle-aged subjects (9 male/13 female) with normal hearing levels (manh;aged 40–60 years, mean age: 51.15 ± 5.64 years),

(c) 25 middle-aged subjects (15 male/10 female) with mild hearing loss (mild; aged45–63 years, mean age: 51.87 ± 5.82 years) and

d) 24 middle-aged subjects (9 male/15 female) with moderate hearing loss (mod;aged 43–57 years, mean age: 51.12 ± 5.53 years).

The young subjects were student volunteers from the Saarland University andaarland University of Applied Sciences, the middle-aged subjects with normal hear-ng sensitivity were relatives or acquaintances, while the hearing impaired subjects

ere volunteers from the rehabilitation center MediClin Bosenberg. All measure-ents were conducted in accordance with the Declaration of Helsinki.

For an objective waveform analysis, a correlation waveform index (CWI) wasefined in a previous study (Bernarding et al., 2010b). The CWI was used to sup-ort objectively the visual detection of a clearly identifiable waveform of the N1–P2omplex. The ALR single sweep sequences were sorted into a matrix. The index wasalculated for each subject by a separation of the respective single sweep matricesnto odd x and even submatrices y. Then, the correlation coefficient �∈ [− 1, 1] wasalculated (Duda et al., 2001)

:=∑N

i=1(xi − x)(yi − y)√∑N

i=1(xi − x)2

√∑N

i=1(yi − y)2

, (1)

here x := N−1∑N

i=1xi and y := N−1

∑N

i=1yi . The CWI was achieved in our studies by

omparing different ALR waveforms with the corresponding correlation coefficients. value of 0.4 (in the range of the N1 wave (50–200 ms)) corresponds to a clearly

dentifiable ALR waveform visually analyzed by an expert.The subject inclusion criteria for the study were: (i) they had an identifiable

aveform of the N1–P2 complex in the ALRs and fulfilled the CWI and (ii) theyetected correctly at least 80% of the target syllables, which serves as a control ofhe cooperation of the subject.

Due to these inclusion criteria, some subjects had to be excluded. Finally, eachroup consisted of the following number of tested subjects:

(a) 20 included subjects: four subjects were excluded due to criterion (i),b) 20 included subjects: one subject was excluded due to criterion (i),

(c) 20 included subjects: five subjects were excluded due to criterion (i),d) 19 included subjects: three subjects were excluded due to criterion (i) and two

subjects were excluded due to criterion (ii).

The degree of hearing impairment was defined according to the definition of theuropean Commission as the pure tone average (PTA4f) of the frequencies at 0.5 kHz,

kHz, 2 kHz, 4 kHz (Shield, 2009). According to this definition, a PTA4f of 20 dB hear-ng level (HL) or less is classified as normal hearing sensitivity, a mild hearing losss defined with thresholds in the range of 21–39 dB HL and a moderate hearing losss defined as a PTA4f from 40 to 69 dB HL (Shield, 2009). Furthermore, most of theearing impaired subjects had a high-frequency or sloping hearing loss, respectively.ig. 1 depicts the mean pure tone audiograms (top) and the corresponding standardeviations (bottom) of the included subjects.

.2. Stimulus materials

Auditory stimuli were naturally produced consonant–vowel syllables (/ba/, /da/,pa/, /de/, /bi/). The consonants /b/ and /d/ were selected because of their differentlace of articulation as they are difficult to discriminate for hearing impaired per-ons (Oates et al., 2002). Also the bilabials /p/ and /b/ are difficult to distinguish,ased on their different voice onset times (Tremblay et al., 2003). Each syllable wasecorded by a female speaker (sampling frequency: 16 kHz). The stimuli were short-ned to a total duration of 200 ms and their amplitudes were normalized, followedy the application of a window. The window was built by a rise and fall time, which

ere the first and second halves of a Gaussian window (duration: 50 ms), and alateau part (duration: 150 ms) with a flat amplitude of 1. The syllables were cal-

brated according to the norms (European Committee for Standardization, 2007;nternational Organization for Standardization, 2007) as the stimuli are in the cate-ory of test signals of short duration in audiometric equipment.

middle-aged subjects with mild hearing impairment (mean PTA4f: 28.12 ± 5.95 dBHL); mod, middle-aged subjects with moderate hearing impairment (mean PTA4f:45.26 ± 5.06 dB HL)).

2.3. Experimental design

Two paradigms with a different degree of difficulty were constructed. The diffi-culty level was achieved by the combination of the syllables.

Difficult syllabic paradigm (DSP) consisted of the syllables /pa/, /da/ and /ba/. Thisparadigm was expected to be more difficult to solve because the syllables had thesame vowel /a/ and the consonants are difficult to distinguish (c.f. Section 2.2).

Easy syllabic paradigm (ESP) was built by the syllables /pa/, /de/ and /bi/. Thisparadigm should be easier to solve, because different vowels were presented.

Both paradigms had the same target syllable /pa/, which had to be detected bythe subjects and evoked the analyzed ALR sequences. Thus, we assume to mini-mize the exogenous effects on the ALR components due to the different physicalcharacteristics of the stimuli. In the paradigms, each syllable was presented 100times.

In both paradigms, the syllables had a randomized order and a randomizedinterstimulus interval ranging from 1 to 2 s. The randomized order was used tomaximize the entropy of the experiment that the solution requires an effortful task.The subjects were instructed to pay attention to the stimulus and to press a but-ton responding to the target syllable. Furthermore, the subjects were asked to keeptheir eyes closed and to minimize movements to avoid muscular artifacts. The audi-tory stimuli were monaurally presented at a fixed intensity of 65 dB sound pressurelevel (SPL) to the subjects with normal hearing sensitivity and mild hearing loss,respectively.

A loudness level adjustment was performed for the subjects with moderatehearing loss in order to ensure an equal audibility of the syllables. This loudnessadjustment was performed in two steps: First, a recorded sentence (In German: Bittegleichen Sie die Lautstärke von [...] an die des Satzes an. Translation: Please adjust theloudness of [...] to the loudness of the sentence.) was presented and adjusted to a com-fortable loudness level by the subject. In the second step, the gap [...] was filledwith each syllable (/ba/, /da/, /pa/, /de/, /bi/). Then, the loudness of each syllable wasgradually adjusted to the loudness of the sentence. After this, a short sequence ofeach paradigm was generated and presented to every subject in order to guarantee,that each syllable could be perceived and distinguished. Note that the target syllablehad always the same intensity in both testing conditions. The selected intensitiesof the loudness adjustment for each syllable are shown in Table 1. The experimentlasted for 30 min and a short break was made between each paradigm to avoidfatigue. After a detailed explanation of the procedure, all subjects signed a consentform.

2.4. Data acquisition

Syllable-evoked ALRs were recorded with a commercially available amplifier(g.tec USBamp, Guger Technologies Austria), using a sampling frequency of 512 Hz(bandpassfilter: 1–30 Hz). Ag/AgCl-electrodes were attached as follows: ipsilateralto the stimulus at the right or left mastoid, common reference at the vertex andground at the upper forehead. Electrodes impedances were kept below 10 k� in all

24 C. Bernarding et al. / Brain Research Bulletin 91 (2013) 21– 30

Table 1Mean and standard deviation values of the individually adjusted intensities for each syllable for the 19 included subjects with moderate hearing impairment.

mst

2

2

abiaetsnsNlcTbpe(

2

ctpmfei�iiczsc(

tsse

TR

yp

Syllable /ba/ /da/

Intensity [dB SPL] 74.4 ± 4.5 74 ± 3.4

easurements. Artifacts were rejected by an amplitude threshold of 50 �V. The ALRingle sweeps were gained from the continuous electroencephalographic data in aime window of 900 ms starting 100 ms prestimulus.

.5. Data analysis

.5.1. Behavioral measuresFor both experimental paradigms, the median reaction time (RT) (Gatehouse

nd Gordon, 1990) and the performance accuracy d′prime, which was proposedy the theory of signal detection (Yanz, 1984), were determined for each subject

n order to complement our electrophysiological data. These two measures werelso applied in other event-related potential studies (e.g. Oates et al., 2002; Korczakt al., 2005; Martin et al., 1999) to support the gained results. The median reac-ion time can be interpreted as a possible index for the processing of the incomingtimuli (Gatehouse and Gordon, 1990). Whereas the d′prime reflects the discrimi-ation ability and the perception of information. Here, opposed to percent correctcores, also the false alarm rate of the subjects response is included (Yanz, 1984).ote that two subjects could have had equal performance but a different amount of

istening effort to understand speech (Edwards, 2007). The performance accuracyan be calculated as d′ = z(false alarm rate) − z(hit rate) (Yanz, 1984; Stanislaw andodorov, 1999), where z is the standard score. The hit rate is defined as the proba-ility to detect correctly a target syllable, and the false alarm rate corresponds to therobability to respond to a distractor. The subjects were asked which paradigm theyxperienced more effortful. Almost all of them experienced the DSP more effortfulexcept for the young normal hearing subjects).

.5.2. Wavelet phase synchronization stability and listening effort levelsFor the analysis of the single sweep sequences, we used the wavelet phase syn-

hronization stability (WPSS) that was introduced in Strauss et al. (2005, 2008)o assess attention correlates in ALR sequences. The stability of the instantaneoushase, which solely regards the phasing of the sweeps (Bruns, 2004), can be used toonitor auditory attention (Low et al., 2007). For this purpose, the wavelet trans-

orm is applied to each data set in order to extract the instantaneous phase ofvery sweep. The instantaneous phase of a signal x ∈ L2(R) can be achieved by tak-ng the complex argument from the complex wavelet transform with the signal:a,b = arg((W x)(a, b)). The circular mean of the instantaneous phase �a,b , depend-

ng on a particular time and scale (“frequency”) parameter, of these sweep sequencess called phase synchronization stability and is a value between 0 and 1. A valuelose to one means that we have perfectly coherent phases, whereas a value close toero represents a more non-phase locked activity. In literature the (wavelet) phaseynchronization stability is also known as phase locking (value/index), as phaseonsistency (see Bruns, 2004 and references therein) or intertrial phase lockingPonjavic-Conte et al., 2012).

In Strauss et al. (2010) we have shown that the WPSS in the time interval ofhe N1 wave, which is modulated by auditory attention (Hillyard et al., 1973), canerve as a possible objective measure for listening effort. We assume that a higherynchronization stability, i.e. a value close to 1 represents an augmented cognitiveffort, resulting from an increased attention to solve an auditory task. In order to

able 2esults of the (one-way) ANOVA (F- and p-values).

Effect: paradigm DSP × ESP DSP × ESP

Subject group ynh manh

N1 amplitude 0.49 (p = 0.48) 0.8 (p = 0.37N1 latency 0.07 (p = 0.7961) 0.23 (p = 0.6WPSS 0.03 (p = 0.8728) 5.93 (p = 0.0d′prime 0.02 (p = 0.9023) 11.08 (p = 0.0RT 0.66 (p = 0.42) 0.31 (p = 0.5

Effect: subject group ynh × manh ynh × manh ynh × manhParadigm DSP ESP DSP & ESP

N1 amplitude 2.05 (p = 0.1601) 2.43 (p = 0.1271) 4.48 (p = 0.0N1 latency 1.99 (p = 0.1699) 1.4 (p = 0.2438) 3.41 (p = 0.0WPSS 4.76 (p = 0.0354)* 0.24 (p = 0.6246) 4.15 (0.048d′prime 0.77 (p = 0.3852) 2.09 (p = 0.1562) 0.13 (p = 0.7RT 0.25 (p = 0.6207) 0.61 (p = 0.4401) 0.82 (p = 0.3

nh, young normal hearing subjects; manh, middle-aged normal hearing subjects; mild,

aradigm; RT, reaction time.* Significantly different, p < 0.05.a For the comparison of the WPSS between the subject groups, the difference was mad

/pa/ /de/ /bi/73.8 ± 4.1 74.7 ± 5.1 75.6 ± 4.5

compare different listening conditions, we defined in Bernarding et al. (2010a) socalled listening effort (LE)-levels, representing the mean of the WPSS in the rangeof the N1 interval (calculated for a specific scale and wavelet type). The limits ofthe N1 interval were taken from the minimum and maximum latency values ofall measured subjects. So that the interval boundaries were set to 70 and 160 ms,respectively. For a more detailed explanation of the extraction of the WPSS and itsrelation to listening effort we refer to Appendix B.

Furthermore, the difference between the LE-level of both experimentalparadigms (DSP and ESP) was calculated to show possible intergroup changes(Bernarding et al., 2011). Here, the WPSS (or LE-level, respectively) of the ESP, servesas a “baseline” condition. Thus, we assume to minimize the exogenous effects result-ing from elevated hearing thresholds. In this way, we are able to focus the analysison endogenous components which are rather linked to listening effort.

3. Results

Table 2 shows an overview of the results of the (one-way)ANOVA conducted on behavioral and electrophysiological data.

3.1. Behavioral analysis: reaction time and d′prime

Fig. 2 depicts the means and standard deviations of median reac-tion time (left y-axis, black) and d′prime values (right y-axis, gray)for the different subject groups and conditions (DSP, difficult syl-labic paradigm; ESP, easy syllabic paradigm). For all middle-agedsubjects, the averaged d′prime value is smaller for the DSP com-pared to the ESP. This difference is not noticeable for the youngersubjects. Here the d′prime showed almost the same values for bothparadigms. Furthermore, an increase of the reaction time for bothhearing impaired subject groups can be observed. There is also asmall decrease of the reaction time for the ESP compared to theircorresponding RT of the DSP in all subject groups except for themild hearing impaired subjects. The ESP shows also a large standarddeviation.

3.2. Electrophysiological analysis

The mean and standard deviation values of the N1 amplitudes

and latencies are shown in Table 3. In general, there is a tendencyfor elevated N1 amplitudes and for a decrease of N1 latenciesfor the DSP compared to the ESP. However these observationswere not significantly different. Furthermore, only for mild hearing

DSP × ESP DSP × ESPmild mod

74) 0.16 (p = 0.6925) 0.72 (p = 0.4014)377) 0.57 (p = 0.4558) 0.06 (p = 0.8012)197)* 5.71 (p = 0.0219)* 4.5 (p = 0.0408)*

019)* 4.59 (p = 0.0387)* 9.81 (p = 0.0034)*

789) 0.01 (p = 0.912) 2.57 (p = 0.1178)

manh × mild manh × mild manh × mildDSP ESP DSP & ESP

374)* 0.5 (p = 0.4821) 1.05 (p = 0.311) 1.55 (p = 0.2168)685) 9.59 (p = 0.0037)* 2.9 (p = 0.0968) 11.48 (p = 0.0011)*

6)*,a 4.59 (p = 0.0386)* 0.4 (p = 0.5288) 1.07 (p = 0.3077)a

147) 0.65 (p = 0.4266) 1.19 (p = 0.2818) 1.4 (p = 0.2401)669) 1.51 (p = 0.2269) 1.39 (p = 0.2464) 2.9 (p = 0.0923)

mild hearing impaired subjects; DSP, difficult syllabic paradigm; ESP, easy syllabic

e between the WPSS of the two syllabic paradigms.

C. Bernarding et al. / Brain Research Bulletin 91 (2013) 21– 30 25

DSPynh

ESPynh

DSPmanh

ESPmanh

DSPmild

ESPmild

DSPmod

ESPmod

350

400

450

500

550

600

650

subject group

2

2.5

3

3.5

4

4.5

5

5.5

6

reactiontime[ms]

d’prim

e

*

* **

*

*

Fig. 2. Mean and standard deviation values of the reaction time and the d′prime for each paradigm (DSP and ESP) and subject group (ynh, young normal hearing subjects;manh, middle-aged normal hearing subjects; mild, middle-aged subjects with mild hearing impairment; mod, middle-aged subjects with moderate hearing impairment).Note that the intensity of the syllables was individually adjusted to each moderate hearing impaired subject. The asterisk (*) denotes significantly different d′prime values(effect: DSP × ESP, significance level p = 0.05).

Table 3Mean and standard deviation values of the N1 amplitudes and latencies for each subject group and condition. Note, that the moderate hearing impaired subjects hadindividually adjusted stimulation intensities (c.f. Section 2.3).

Subject group ynh manh mild mod

N1 amplitude [�V]DSP −7.27 ± 3.05 −6.14 ± 1.78 −6.61 ± 2.33 −6.35 ± 1.47ESP −6.67 ± 2.27 −5.58 ± 2.14 −6.31 ± 2.33 −5.90 ± 1.78

± 13 ± 11

iD

scBt2qt

Flm

N1 latency [ms]DSP 109.76 ± 13.97 103.71ESP 111.05 ± 17.31 105.56

mpaired subjects the latencies are larger for the ESP than for theSP (p < 0.05).

The WPSS was calculated for each subject and condition for thecale a = 40 using M = 70 ALR sweeps. These sweeps were evoked byorrectly detected target syllables and free of artifacts (c.f. Appendix). The scale a = 40 was selected due to a good temporal resolu-

ion of the WPSS in the time interval of the N1 wave (Low et al.,007; Strauss et al., 2010). Note that this scale equals to the fre-uency domain of a “pseudo” frequency of 6.4 Hz corresponding tohe upper theta band).

ig. 3. Individual result for a representative subject of the manh, middle-aged normal heaine)). Right: corresponding WPSS for both paradigms (DSP (black line) and ESP (gray lin

uch clearer than the averaged amplitude.

.16 118.16 ± 16.17 116.67 ± 18.39

.43 113.97 ± 18.88 118.10 ± 16.41

The wavelet used in this study was the 4th-derivative of thecomplex Gaussian function. This wavelet was chosen becauseof its symmetry, i.e. it acts as a linear phase filter and avoidsphase distortions (Bradley and Wilson, 2004). This waveletwas also applied in previous studies Strauss et al. (2008a,2010).

Fig. 3 shows the individual result for a representative subjectof the middle-aged normal hearing subject group. It is noticeablethat the WPSS illuminates the differences between DSP and ESPmuch clearer than the averaged amplitude. This figure shows that

ring subject group. Left: ALRs for the two conditions (DSP (black line) and ESP (graye). It is noticeable that the WPSS illuminates the differences between DSP and ESP

26 C. Bernarding et al. / Brain Research Bulletin 91 (2013) 21– 30

DSPynh

ESPynh

0 200 400 600 800

0 200 400 600 800

DSPmanh

ESPmanh

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

0 200 400 600 800

DSPmild

ESPmild

0.0

0.2

0.4

0.6

0.8

1.0

time [ms]

time [ms]

time [ms]

LE-level

DSP ESP DSP-ESP0.0

0.2

0.4

0.6

0.8

1.0

0 200 400 600 800

DSPmod

ESPmod

DSP ESP DSP-ESP

DSP ESP DSP-ESP

DSP ESP DSP-ESP

LE-level

LE-l evel

LE-level

time [ms]

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

averageoftheWPSS

0.0

0.2

0.4

0.6

0.8

1.0

a)

b)

c)

d)

averageoftheWPSS

averageoftheWPSS

averageoftheWPSS

Fig. 4. Grand average of the WPSS (left) and the corresponding LE-level for each paradigm (right, DSP (black) and ESP (gray), difference (DSP–ESP, light gray)) and subjectg ing sum inten

tp

si

roup (a) ynh, young normal hearing subjects, (b) manh, middle-aged normal heariddle-aged subjects with moderate hearing impairment and individually adjusted

here can be a decoupling between the averaged amplitude and the

roposed circular WPSS measure.

Fig. 4 illustrates the normalized grand averages (over all theubjects) of the WPSS (left panel) and the corresponding listen-ng effort (LE)-levels (right panel, mean of the WPSS in the time

bjects, (c) mild, middle-aged subjects with mild hearing impairment, and (d) mod,sities of the syllables).

interval of the N1 wave, c.f. Section 2.5.2) for each subject group,

from top to bottom: (a) young normal hearing subjects, (b) middle-aged normal hearing subjects, (c) middle-aged subjects with mildhearing impairment, and (d) middle-aged subjects with moderatehearing impairment. The result of the difficult paradigm is always

C. Bernarding et al. / Brain Research Bulletin 91 (2013) 21– 30 27

0.02

0.06

0.10

0.14

4-15 dB HL(11.06±3.04 dB HL,

n=27)

16-25 dB HL(

n=20)19.92±2.69 dB HL,

26-42 dB HL

n=13)(29.78±4.50 dB HL,

difference

oftheLE-level

0

10

20

30

40

50

meanage[yrs]

±0.0896

±0.1743

±0.1392

±12.8656

±49.65

±54.1538

FP

dTf(tt7rifjeis2n

eLt(jb1

L1miirho2yaT

0

5

10

15

20

21-35 yrs(25.70±4.25 yrs,

n=20)

41-50 yrs(47.00±2.68 yrs,

n=19)

51-63 yrs(56.61±2.99 yrs,

n=21)

meanPTA[dBHL]

difference

oftheLE-level

0.02

0.06

0.10

0.14

±0.1906

±0.1033

±0.0973

±3.8028

±5.7518

±7.4458

ig. 5. Top: grand average of the difference of the LE-level (DSP–ESP) over the threeTA7f intervals. Bottom: corresponding grand average of the age over the PTA7f .

isplayed in black, whereas the easy conditions are plotted in gray.he light gray bars in the graphics of the LE-levels represent the dif-erence of the LE-level between the difficult and the easy conditionDSP–ESP). It is noticeable, that the WPSS is larger for the DSP thanhe ESP for all middle-aged subjects (p < 0.05, (one-way) ANOVA) inhe time interval of interest corresponding to the N1 wave (approx.0–160 ms). In the case of the young normal hearing subjects, theesult of the WPSS revealed almost the same values. Furthermore,t can be seen that the difference of the LE-level is slightly enhancedor the mild hearing impaired subjects compared to the coeval sub-ect group with normal hearing sensitivity. Note that a listeningffort related comparison of both hearing impaired subject groupss not appropriate due to the individual loudness adjustment of theyllables for the moderate hearing impaired subjects (c.f. Section.3). Nevertheless, a significantly enhanced WPSS for the DSP isoticeable for the moderate hearing impaired subjects.

In order to make a statement about the effects of age andlevated hearing thresholds on the WPSS, the difference of theE-levels (c.f. Section 2.5.2) was calculated for each subject of thehree groups, which were tested using a fixed intensity of 65 dB SPLyoung normal hearing subjects, middle-aged normal hearing sub-ects, mild hearing impaired subjects). This difference was orderedy the pure tone average (PTA7f of the frequencies: 125, 250, 500,

k, 2 k, 4 k and 6 kHz) and the age of the subject, respectively.At the top of Fig. 5 the grand average of the difference of the

E-levels divided into three PTA7f ranges (4–15 dB HL: including7 ynh and 10 manh subjects, 16–25 dB HL: including 1 ynh, 12anh, and 7 mild subjects, 26–42 dB HL: including 13 mild subjects)

s illustrated. An increase of the difference of the LE-levels withncreasing PTA7f is visible. Shown in the graphic below is the cor-esponding grand average of age. A dependency between age andearing threshold can be noticed. In Fig. 6 (top) the grand averagef the difference of the LE-level over three age ranges (21–35 years:

0 ynh subjects, 41–50 years: 11 manh and 8 mild subjects, 51–63ears: including 9 manh and 12 mild subjects) is displayed. Heren increasing slope of the LE-difference (DPS–ESP) can be observed.he lower part of this figure shows the corresponding mean PTA7f

Fig. 6. Top: grand average of the difference of the LE-level (DSP–ESP) over the threeage intervals. Bottom: corresponding grand average of the PTA7f over the age.

of these age ranges. Here an increase of the hearing threshold withincreasing age is noticeable.

4. Discussion

This study was designed in order to extract neural correlatesof listening effort in young, middle-aged, and hearing impairedsubjects. Furthermore, behavioral data was taken into account tocomplement the electrophysiological results.

The “accuracy data” suggests, that for all middle-aged subjectsthe perception of information, namely the detection of the targetsyllable, was significantly more challenging in the case of the dif-ficult paradigm. A comparison of the d′prime values of the mildhearing impaired and of middle-aged subjects with normal hearingsensitivity revealed similar results. However, as already mentioned,the accuracy measure does not reflect the effort needed to solve atask. Edwards (2007) and Downs (1982) remarked that individualswith hearing impairment could have similar speech discriminationscores as normal hearing persons, but mostly they experience a dif-ferent amount of cognitive effort. The increased reaction times forthe hearing impaired persons could be interpreted in a way thatthese individuals needed a longer processing time as compared tothe normal hearing control groups – regardless of their age. AlsoOates et al. (2002) noticed this enhancement of the reaction time inan ERP-study with hearing impaired persons. They mentioned thatthis observation could be related to a possible slowing of the cogni-tive processing. Additionally, the reaction time is slightly enhancedfor the difficult condition for all subject groups, besides the mildhearing impaired subjects. Nevertheless, neither the inter- nor theintragroup comparison of the reaction time showed significant dif-ferences.

Furthermore the amplitudes and latencies of the N1 wavewere analyzed. The enhanced amplitudes and decreased laten-cies of the N1 component for the DSP suggest that there was an

increased attention in order to detect the target syllable (Pictonand Hillyard, 1974). However, this difference did not reach anystatistical significance. Also Rao et al. (2010) reported increasedamplitudes for a more difficult task.

2 Resear

ttifahwtTehafsTbtswWanipc(ttiml

Ladtn2yuglgsfgieHs(pMiafsatfb

mrb

8 C. Bernarding et al. / Brain

The main focus of this study was to examine, if the WPSS in theime interval of the N1 wave could serve as an indicator for lis-ening effort in different age groups and in persons with hearingmpairment. An objective discrimination between a more effort-ul, i.e. DSP and an easier listening situation, i.e. ESP could bechieved in all middle-aged subject groups, regardless of theirearing sensitivity. The WPSS in the expected N1 time intervalas significantly enhanced for the difficult task. We assume that

he subjects required more effort to solve the difficult paradigm.his is in line with the findings of our previous studies (Strausst al., 2009, 2010; Bernarding et al., 2010a). Only the young normalearing group constitutes an exception. Here, the WPSS showedlmost the same morphology for both paradigms. In one of ourormer studies (Strauss et al., 2010; Bernarding et al., 2010a), theame conditions were tested in young normal hearing persons.here the paradigms were embedded in a competing multitalkerabble noise. We assume that in this noisy surrounding, the detec-ion of the target syllable exerted more effort from the youngerubjects, especially for the DSP. Here, an increased listening effortas reported and could objectively determined with the help of thePSS. In this study the noise is removed, so that both paradigms

re easy to solve for the younger subjects. Also Murphy et al. (2000)oticed in a memory-related recall task that young persons tested

n noise have a similar performance as their older counterpartsroved in quiet environments. They argue that the additional noiseould be “associated with a reduction in processing resources”Murphy et al., 2000), so that attention was divided. If we comparehe results of the d′prime with the results of the WPSS, the expectedendency is visible in the middle-aged normal hearing as well asn the middle-aged hearing impaired subject groups: the DSP was

ore difficult to solve (reflected in the d′prime) and required moreistening effort from the subjects (represented in the WPSS).

In order to reveal an intergroup comparison, the difference of theE-level (DSP–ESP) was calculated. By applying this simple linearpproach, we assume to reduce the exogenous components (e.g.ifferent stimulation intensities resulting from elevated hearinghresholds). In this way, we are able to focus the analysis on endoge-ous parts which are rather linked to listening effort (c.f. Section.5.2). Regarding the small mean difference of the LE-level for theoung subjects, we can interpret analogue to the mean WPSS val-es, that both paradigms were equally easy to solve by this subjectroup. It can be noticed that the averaged difference of the LE-evel is slightly increased for the mild hearing impaired subjectroup compared to the young and the middle-aged normal hearingubjects. On the other hand, a small decrease of this averaged dif-erence is visible in the results of the moderate hearing impairedroup. This last finding is not unexpected as the WPSS is analyzedn the time interval of the N1 wave, i.e. the WPSS is not only influ-nced by endogenous factors (e.g. attention effects (Woldorff andillyard, 1991)). It is also affected by exogenous factors (e.g. inten-

ity effects (Adler and Adler, 1989), different interstimulus intervalsHall, 1992)). It is not transparent to which proportions the N1 com-onent is influenced by increased attention (Picton et al., 1974;uller-Gass and Schröger, 2007) or by the different stimulation

ntensities (Hall, 1992; Schadow et al., 2007) due to the loudnessdjustment of the syllables, i.e. the stimulation intensity was higheror the moderate hearing impaired subjects. Due to the higher pre-entation level it can be argued that the speech intelligibility wasdditionally improved resulting in a decreased task difficulty. Inhis case a less endogenous (effortful) modulation is present. There-ore, the results of the moderate hearing impaired subjects cannote compared with the other groups.

Moreover it can be argued that related to the loudness adjust-ent, the difficult paradigm was eased, so that listening effort was

educed. Nevertheless, the objective discrimination via the WPSSetween the two testing conditions could also be achieved in this

ch Bulletin 91 (2013) 21– 30

hearing impaired group. Due to this variation of the intensity levelof the paradigms, the moderate hearing impaired subject group wasnot considered for the further analysis.

To examine a possible age and/or hearing impairment relatedeffect on the listening effort, the LE-level difference was calcu-lated separately for each subject. First, the data was separatedaccording to hearing threshold (PTA7f of all 7 tested frequencies),aligned to three PTA7f groups, and finally averaged. A clear increaseof the LE-difference with increasing PTA7f is evident. It could behypothesized that with decreasing hearing sensitivity, the listeningeffort is enhanced. The same can be noticed, if the data is arrangedby age. As the age increases, the difference of the LE-levels showsthe same tendency. A further observation is that with increasingage also the hearing threshold raises, in accordance to the standards(International Organization for Standardization, 2000).

In other studies using dual-task paradigms to determine lis-tening effort, the authors noticed a similar age (Tun et al., 2009;Gosselin and Gagné, 2011) and hearing loss (Tun et al., 2009) relatedeffect on the listening effort. The listening effort was enhancedfor the hearing impaired and the elder subjects. The age relatedincrease of perceptual effort was interpreted on the basis of areduced cognitive performance (see Pichora-Fuller and Singh, 2006for an overview). However, in these studies the age differencebetween the subject groups was more pronounced. In our case,we can suggest that due to the small age difference between theyoung and the middle-aged subject groups, there is a tendency ofpredominantly signal degradation related effect on listening effort.

5. Conclusions

We presented a possible method to extract correlates of listen-ing effort related factors using electroencephalographic data. Thismethod was tested in young as well as middle-aged subjects and inpersons with a different degree of hearing loss. The results indicatethat the WPSS of ALR sequences reflects the listening effort neededto solve an auditory task. A further finding was that reaction timedata or N1 wave amplitude information hardly reflect the investedeffort. Moreover we noticed an age as well as a hearing sensitiv-ity related effect on the listening effort by intergroup comparisonof the WPSS difference. Regarding these findings, we can concludethat analyzing the instantaneous phase of EEG signals could be anappropriate step to develop a clinical tool for the objective estima-tion of listening effort related factors in future.

Acknowledgements

The authors wish to thank Dr. Ulrich Jobst for his help withthe coordination of the study and Markus Landwehr, Ahmed Bel-lagnech, and Birgit Seidler-Fallböhmer for their assistance withsubject recruitment.

Appendix A. Results of the pre-study for the objectiveestimation of listening effort

Fig. A1 shows the partial results of a preliminary study(Bernarding et al., 2010a) using syllabic stimulation paradigms,which were embedded in multitalker babble noise at a SNR of +5 dB(total of 20 young subjects; 12 male/9 female, aged 20–34 years,mean age 25 ± 3.52 years). On the left side: the grand normalizedaverage of the WPSS (over 18 included young subjects, scale a = 40)for the difficult (DSPnoise) and the easy paradigm (ESPnoise). In the

centre: the corresponding LE-levels (mean of the WPSS in the rangeof the N1 wave). On the right side: the result of the time-resolved(one-way) ANOVA. It is noticeable, that the WPSS and the LE-levelsare much larger for the paradigm with the higher task demand

C. Bernarding et al. / Brain Research Bulletin 91 (2013) 21– 30 29

averageoftheWPSS

1.0

0.8

0.6

0.4

0.2

0.00 200 400 600 800

0.1

0.01

time [ms]0 200 400 600 800

0.0

0.2

0.4

0.6

0.8

1.0DSPnoise

ESPnoise

time [ms]

p-value(log)

ListeningEffort

DSPnoise ESPnoise

significance level (p=0.05)

N1 wave

F r 18 su( lts fo

((t

Al

nt iwL(�b

�

s(

L

i

R

A

A

B

B

B

B

B

B

B

ig. A1. Results of the pre-study. Left: grand normalized average of the WPSS (oveSNR: +5 dB). Centre: LE-level (mean WPSS in the range of the N1 wave). Right: resu

DSPnoise) compared to the paradigm with the lower task demandESPnoise). Furthermore, a significant difference is evident betweenhe two paradigms in the time interval of the N1–P2-complex.

ppendix B. Wavelet phase synchronization stability andistening effort

As a general definition, for the determination of the WPSS weeed an adaptation of the derived phase locking measure betweenwo signals to our problem. Let a,b(·) = |a|−1/2 ((· − b)/a) where

∈ L2(R) is the wavelet with 0 <∫R

|�(ω)|2|(ω)|−1dω < ∞ (�(ω)s the Fourier transform of the wavelet and a, b ∈ R, a /= 0). The

avelet transform W : L2(R) −→ L2(R2, (dadb)/a2) of a signal x ∈2(R) with respect to the wavelet is given by the inner L2-productW x)(a, b) = 〈x, a,b〉L2 . We define the synchronization stability

a,b of a sequence X = {xm ∈ L2(R) : m = 1, . . . , M} of M ALR sweepsy

a,b(X) := 1M

∣∣∣∣∣M∑m=1

eıarg((W xm)(a,b))

∣∣∣∣∣ . (B.1)

Note that (B.1) yields a value in [0, 1]. For a fixed a and b anduitable experimental paradigm, we have suggested in Strauss et al.2008) that

istening effort ∝ �a,b(X)

n the time interval of the N1 wave.

eferences

dler, G., Adler, J., 1989. Influence of stimulus intensity on AEP components in the80-to-200-millisecond latency range. Audiology 28, 316–324.

keroyd, M.A., 2008. Are individual differences in speech reception related to indi-vidual differences in cognitive ability? A survey of twenty experimental studieswith normal and hearing-impaired adults. International Journal of Audiology 47(Suppl 2), 53–71.

ernarding, C., Corona-Strauss, F.I., Latzel, M., Strauss, D.J., 2010a. Auditory stream-ing and listening effort: an event-related potential study. In: Conf. Proc. IEEEEng. Med. Biol. Soc, pp. 6817–6820.

ernarding, C., Latzel, M., Strauss, D.J., Corona-Strauss, F.I., 2011. On the objectiveelectrophysiological estimation of listening effort in different age groups. In:2011 5th International IEEE/EMBS Conference on Neural Engineering NER 2011,pp. 671–674.

ernarding, C., Strauss, D.J., Latzel, M., Corona-Strauss, F.I., 2010b. Non-listeningeffort related parameters in auditory discrimination paradigms. In: Conf. Proc.IEEE Eng. Med. Biol. Soc, pp. 6682–6685.

ertoli, S., Smurzynski, J., Probst, R., 2005. Effects of age, age-related hearing loss, andcontralateral cafeteria noise on the discrimination of small frequency changes:psychoacoustic and electrophysiological measures. Journal of the Associationfor Research in Otolaryngology 6, 207–222.

radley, A.P., Wilson, W.J., 2004. On wavelet analysis of auditory evoked potentials.Clinical Neurophysiology 115, 1114–1128.

rockhaus-Dumke, A., Mueller, R., Faigle, U., Klosterkoetter, J., 2008. Sensory gatingrevisited: relation between brain oscillations and auditory evoked potentials inschizophrenia. Schizophrenia Research 99, 238–249.

runs, A., 2004. Fourier-, Hilbert-, and wavelet based signal analysis: are they reallydifferent approaches. Journal of Neuroscience Methods 137, 321–332.

bjects) for the two paradigms (DSPnoise and ESPnoise) embedded in babble noiser the time resolved ANOVA (DSPnoise vs. ESPnoise) in logarithmic scale.

Downs, D.W., 1982. Effects of hearing aid use on speech discrimination and listeningeffort. Journal of Speech and Hearing Disorders 47, 189–193.

Duda, R.O., Hart, P.E., Stork, D.G., 2001. Pattern Classification. Wiley-Interscience,New York.

Edwards, E., 2007. The future of hearing aid technology. Trends in Amplification 11,31–45.

European Committee for Standardization, 2007. Electroacustics – AudiometricEquipment. Part 3: test signals of short duration. Technical Report. The EuropeanStandard. EN 60645-3:2007.

Friedmann, D., 2010. The components of aging. In: Luck, Kappermann (Eds.), OxfordHandbook of Event-Related Potential Components. Oxford University Press,New York.

Gatehouse, S., Akeroyd, M., 2006. Two-eared listening in dynamic situations. Inter-national Journal of Audiology 45 (Suppl. 1), S120–S124.

Gatehouse, S., Gordon, J., 1990. Response times to speech stimuli as measures ofbenefit from amplification. British Journal of Audiology 24, 63–68.

Gatehouse, S., Noble, W., 2004. The speech, spatial and qualities of hearingscale(SSQ). International Journal of Audiology 43, 85–99.

Goldwater, B.C., 1972. Psychological significance of pupillary movements. Psycho-logical Bulletin 77, 340–355.

Goodin, D.S., Squires, K.C., Henderson, B.H., Starr, A., 1978. Age-related variationsin evoked potentials to auditory stimuli in normal human subjects. Electroen-cephalography and Clinical Neurophysiology 44, 447–458.

Gosselin, P.A., Gagné, J., 2011. Older adults expend more listening effort than youngadults recognizing speech in noise. Journal of Speech, Language, and HearingResearch 54, 944–958.

Gosselin, P.A., Gagné, J.P., 2010. Use of a dual-task paradigm to measure listeningeffort. Canadian Journal of Speech–Language Pathology and Audiology 34 (1),43–51.

Hall, J.W., 1992. Handbook of Auditory Evoked Responses. Allyn and Bacon, NeedhamHeights, MA.

Hillyard, S.A., Hink, R.F., Schwent, V.L., Picton, T.W., 1973. Electrical signs of selectiveattention in the human brain. Science 182, 177–180.

Hillyard, S.A., Vogel, E.K., Luck, S.J., 1998. Sensory gain control (amplification) as amechanism of selective attention: electrophysiological and neuroimaging evi-dence. Philosophical Transactions of the Royal Society of London, Series B:Biological Sciences 353, 1257–1270.

Hink, R.F., Van Voorhis, S.T., Hillyard, S.A., 1977. The division of attention and thehuman auditory evoked potential. Neuropsychologia 15, 597–605.

Humes, L.E., 1999. Dimensions of hearing aid outcome. Journal of the AmericanAcademy of Audiology 10, 26–39.

International Organization for Standardization, 2000. Acoustics – Statistical distri-bution of hearing thresholds as a function of age. Technical Report. InternationalOrganization for Standardization. ISO 7029:2000.

International Organization for Standardization, 2007. Acoustics – Reference zero forthe calibration of audiometric equipment. Part 6: reference threshold of hearingfor test signals of short duration. Technical Report. International Standards forBusiness. ISO 389-6:2007.

Kahneman, D., 1973. Attention and Effort. Prentice Hall, Englewood Cliffs, NJ.Kauramäki, J., Jääskeläinen, L.P., Sams, M., 2007. Selective attention increases both

gain and feature selectivity of the human auditory cortex. PLoS ONE 2, 909.Kerlin, J.R., Shahin, A.J., Miller, L.M., 2010. Attentional gain control of ongoing cor-

tical speech representations in a “cocktail party”. Journal of Neuroscience 30,620–628.

Kiessling, J., Pichora-Fuller, M.K., Gatehouse, S., Stephens, D., Arlinger, S., Chisolm,T., Davis, A.C., Erber, N.P., Hickson, L., Holmes, A., Rosenhall, U., von Wedel, H.,2003. Candidature for and delivery of audiological services: special needs ofolder people. International Journal of Audiology 42 (Suppl. 2), 92–101.

Klimesch, W., Sauseng, P., Hanslmayr, S., Gruber, W., Freunberger, R., 2007.Event-related phase reorganization may explain evoked neural dynamics. Neu-roscience & Biobehavioral Reviews 31, 1003–1016.

Korczak, P.A., Kurtzberg, D., Stapells, D.R., 2005. Effects of sensorineural hearing lossand personal hearing aids on cortical event-related potential and behavioral

measures of speech–sound processing. Ear and Hearing 26, 165–185.

Kramer, S.E., Kapteyn, T.S., Festen, J.M., Kuik, D.J., 1997. Assessing aspects of auditoryhandicap by means of pupil dilatation. Audiology 36, 155–164.

Low, Y.F., Corona-Strauss, F.I., Adam, P., Strauss, D.J., 2007. Extraction of auditoryattention correlates in single sweeps of cortical potentials by maximum entropy

3 Resear

L

M

M

M

M

M

O

P

P

P

P

P

R

R

0 C. Bernarding et al. / Brain

paradigms and its application. In: Proceedings of the 3rd International IEEE EMBSConference on Neural Engineering, Kohala Coast, HI, USA, pp. 469–472.

ow, Y.F., Strauss, D.J., 2009. EEG phase reset due to auditory attention: an inversetime-scale approach. Physiological Measurement 30, 821–832.

ackersie, C.L., Cones, H., 2011. Subjective and psychophysiological indexes of lis-tening effort in a competing-talker task. Journal of the American Academy ofAudiology 22, 113–122.

artin, B., Kurtzberg, D., Stapells, D., 1999. The effects of decreased audibility pro-duced by high-pass noise masking on n1 and the mismatch negativity to speechsounds /ba/ and /da/. Journal of Speech, Language, and Hearing Research 42 (2),271–286.

artin, B.A., Tremblay, K.L., Korczak, P., 2008. Speech evoked potentials: from thelaboratory to the clinic. Ear and Hearing 29, 285–313.

uller-Gass, A., Schröger, E., 2007. Perceptual and cognitive task difficulty has dif-ferential effects on auditory distraction. Brain Research 1136, 169–177.

urphy, D.R., Craik, F.I.M., Li, K.Z.H., Schneider, B.A., 2000. Comparing the effects ofaging and background noise on short-term memory performance. Psychologyand Aging 15, 323–334.

ates, P.A., Kurtzberg, D., Stapells, D.R., 2002. Effects of sensorineural hearing losson cortical event-related potential and behavioural measures of speech–soundprocessing. Ear and Hearing 23, 399–415.

ichora-Fuller, M.K., Singh, G., 2006. Effects of age on auditory and cognitiveprocessing: implications for hearing aid fitting and audiologic rehabilitation.Trends in Amplification 10, 29–59.

icton, T., Hillyard, S., Krausz, H., Galambos, R., 1974. Human auditory evoked poten-tials. Audiology and Neurootology 36, 179–190.

icton, T.W., Hillyard, S.A., 1974. Human auditory evoked potentials. II.Effects of attention. Electroencephalography and Clinical Neurophysiology 36,191–199.

onjavic-Conte, K.D., Dowdall, J.R., Hambrook, D.A., Luczak, A., Tata, M.S., 2012. Neu-ral correlates of auditory distraction revealed in theta-band EEG. Neuroreport23, 240–245.

urdy, M., 1991. What is listening? In: Listening in Everyday Life: A Per-sonal and Professional Approach. Univ. Press of America, Lanham, Maryland,

pp. 1–20.

akerd, B., Seitz, P., Whearty, M., 1996. Assessing the cognitive demands of speechlistening for people with hearing losses. Ear and Hearing 17, 97–106.

ao, A., Zhang, Y., Miller, S., 2010. Selective listening of concurrent auditory stimuli:an event-related potential study. Hearing Research 268, 123–132.

ch Bulletin 91 (2013) 21– 30

Sarampalis, A., Kalluri, S., Edwards, B., Hafter, E., 2009. Objective measures of listen-ing effort: effects of background noise and noise reduction. Journal of Speech,Language, and Hearing Research 52, 1230–1240.

Schadow, J., Lenz, D., Thaerig, S., Busch, N.A., Fründ, I., Herrmann, C.S., 2007. Stim-ulus intensity affects early sensory processing: sound intensity modulatesauditory evoked gamma-band activity in human EEG. International Journal ofPsychophysiology 65, 152–161.

Shield, B., 2009. Evaluation of the Social and Economic Costs of Hearing Impairment– A Report for HEAR-IT. Technical Report. HEAR-IT.

Stanislaw, H., Todorov, N., 1999. Calculation of signal detection theory measures.Behavior Research Methods, Instruments, and Computers 31 (1), 137–149.

Strauss, D.J., Corona-Strauss, F.I., Bernarding, C., Latzel, M., Froehlich, M., 2009. On thecognitive neurodynamics of listening effort: a phase clustering analysis of large-scale neural correlates. In: Conf. Proc. IEEE Eng. Med. Biol. Soc, pp. 2078–2081.

Strauss, D.J., Corona-Strauss, F.I., Froehlich, M., 2008a. Objective estimation of the lis-tening effort: towards a neuropsychological and neurophysical model. In: Conf.Proc. IEEE Eng. Med. Biol. Soc, pp. 1777–1780.

Strauss, D.J., Corona-Strauss, F.I., Trenado, C., Bernarding, C., Reith, W., Latzel, M.,Froehlich, M., 2010. Electrophysiological correlates of listening effort: neurody-namical modeling and measurement. Cognitive Neurodynamics 4, 119–131.

Strauss, D.J., Delb, W., D’Amelio, R., Falkai, P., 2005. Neural synchronization stabilityin the tinnitus decompensation. In: Proceedings of the 2nd International IEEEEMBS Conference on Neural Engineering, Arlington, VA, USA, pp. 186–189.

Strauss, D.J., Delb, W., D’Amelio, R., Low, Y.F., Falkai, P., 2008. Objective quantificationof the tinnitus decompensation by measures of auditory evoked single sweeps.IEEE Transactions on Neural Systems and Rehabilitation Engineering, 74–81.

Tremblay, K.L., Piskoz, M., Souza, P., 2003. Effects of age and age-related hearingloss on the neural presentation of speech cues. Clinical Neurophysiology 114,1332–1343.

Tun, P.A., McCoy, S., Wingfield, A., 2009. Aging, hearing acuity, and the attentionalcosts of effortful listening. Psychology and Aging 24, 761–766.

Woldorff, M.G., Hillyard, S.A., 1991. Modulation of early auditory processing dur-ing selective listening to rapidly presented tones. Electroencephalography andClinical Neurophysiology 79, 170–191.

Yanz, J.L., 1984. The application of the theory of signal detection to the assessmentof speech perception. Ear and Hearing 5 (2), 64–71.

Zekveld, A.A., Kramer, S.E., Festen, J.M., 2010. Pupil response as an indication ofeffortful listening: the influence of sentence intelligibility. Ear and Hearing 31,480–490.