perception of auditory movement in children with poor ...€¦ · kelly j. shea-miller‡* anne m....

312

J Am Acad Audiol 16:312–326 (2005)

*New Jersey Neuroscience Institute, JFK Medical Center; †Seton Hall University, School of Graduate Medical Education;‡Montclair State University; §JFK Johnson Rehabilitation Institute

Ilse J.A. Wambacq, 65 James Street, Edison, NJ 08818; Phone: 732-321-7950; Fax: 732-632-1584; E-mail: [email protected]

Perception of Auditory Movement inChildren with Poor Listening Skills: AnERP Study

Ilse J.A. Wambacq*†Kelly J. Shea-Miller‡*Anne M. Eckert§Virginia Toth§

Abstract

Long-latency ERP components were examined for scalp activation differencesin children with poor and good listening skills in response to auditory movementcreated by IIDs. Eighteen children were grouped based on a parent questionnaire(CHAPS; Smoski et al, 1998) and clinical evaluation by a licensed audiologist.Obligatory cortical responses were recorded to an auditory movement and anauditory control task. Results showed greatest activation at fronto-centralelectrode sites. P1, N1, and P2 showed no significant effects. Significantdifferences in N2 amplitude and latency were present between groups at thelateral electrode sites (FC3, FC4) in the auditory movement but not in the auditorycontrol task. More specifically, good listeners exhibited predominance ofactivation over the right hemisphere for left-moving sounds, whereas the poorlisteners exhibited symmetric activation. These results suggest that abnormalhemispheric activation may be one of the reasons behind poor listening skillsobserved in some school-aged children.

Key Words: Auditory perceptual disorders, child, evoked potentials, soundlocalization

Abbreviations: CHAPS = Children’s Auditory Performance Scale; EEG =electroencephalographic; ERP = event-related potentials; IID = interauralintensity differences; ITD = interaural timing differences; N1 = negative peakbetween 80–120 milliseconds; N2 = negative peak between 180–290milliseconds; P1 = positive peak between 50–90 milliseconds; P2 = positivepeak between 140–190 milliseconds; P3 = positive peak at 300 millisecondsor later; VEOG = vertical electro-oculographic

Sumario

Se examinaron los componentes de los ERP de latencia prolongada, buscandodiferencias de activación en el cráneo, en niños con habilidades auditivas buenasy malas, en respuesta a movimientos auditivos creados por diferenciasinteraurales de intensidad (IID). Se agruparon dieciocho niños de acuerdo aun cuestionario llenado por los padres (CHAPS; Smoski y col., 1998) y a unaevaluación clínica realizada por un audiólogo certificado. Se registraronrespuestas corticales obligatorias ante un movimiento auditivo y una tarea decontrol auditivo. Los resultados mostraron una activación mayor en los sitiosde colocación de electrodos fronto-centrales. P1, N1 y P2 no mostraronefectos significativos. Se presentaron diferencias significativas entre gruposen la amplitud y la latencia del N2, con el sitio de electrodo lateral (FC3 y FC4),ante el movimiento auditivo pero no ante la tarea de control auditivo. Másespecíficamente, los sujetos con buena audición exhibieron predominancia enla activación en el hemisferio derecho, para sonidos moviéndose a la izquierda,mientras que aquellos con pobre audición exhibieron una activación simétrica.

PPeerrcceeppttiioonn ooff AAuuddiittoorryy MMoovveemmeenntt/Wambacq et al

313

According to the American Speech-Language Hearing Association (1996),

auditory processing involvesmechanisms responsible for sound localizationand lateralization, auditory discrimination,auditory pattern recognition, temporalaspects of audition, auditory performancewith competing acoustic signals, and auditoryperformance with degraded signals. Thisdefinition has given focus to research andclinical diagnosis of auditory processingdisorders (APD; Jerger and Musiek, 2000;Bellis, 2003) and has resulted in a betterunderstanding of the behavioral andunderlying neurological differences inchildren with APD (Estes et al, 2002; Purdyet al, 2002; Warrier et al, 2004). Therecontinues to be a lack of research examiningspatial localization and perception of soundmovement in children with problemsprocessing auditory information. Bellis (2003)states that there is a need for more effectivetests of localization and lateralization in thispopulation. The goal of this study was toobjectively investigate the neural processingof sound movement in children with poorlistening skills. For the purpose of this study,poor listeners are defined as children with (1)poor auditory performance in daily life asreported by parents on the Children’sAuditory Performance Scale (CHAPS; Smoskiet al, 1998) and (2) a clinical diagnosis ofAPD obtained by a certified audiologist basedon an auditory processing test battery.

Behavioral studies indicate that sound

localization is based on the processing ofbinaural cues from interaural time and leveldifferences and of monaural spectral cues fromthe filtering effects of the pinna, head, and body(Palmer, 1995). The ability to understandspeech in noise and to localize sound sources(i.e., complex auditory processing) depends inlarge part on the ability to use the binauralinformation in signals with spatial information(e.g., Durlach and Colburn, 1978). Althoughthere is a solid research tradition in thebehavioral measurement of basic and complexbinaural tasks (e.g., Durlach and Colburn,1978; Hausler et al, 1983; Koehnke et al, 1995;Koehnke and Besing, 2001), the application ofresearch on binaural processing to auditoryevent-related potentials (ERPs) has been lessextensive.

The initial coding of the binaural cues ofinteraural phase and intensity differences,which are used in sound localization, occursat the level of the superior olivary nuclei(Casseday and Neff, 1975; Polyakov andPratt, 1996; Pratt and Polyakov, 1999).However, the perception of the spatial locationof sound occurs at a higher level within thecentral auditory system (Itoh et al, 2000).Higher level processing of binauralinformation is illustrated by sensitivity ofthe primary and secondary auditory cortex tointeraural phase or intensity differences insingle cell recordings in mammals (Pickles,1988; Mendelson and Grasse, 2000; Xu et al,2000) and the measurement of ERP responsesto binaural cues in humans (Ungan et al,

Estos resultados sugieren que la activación hemisférica anormal puede seruna de las razones responsables de la habilidad pobre para escuchar,observadas en algunos niños de edad escolar.

Palabras Clave: Trastornos perceptuales auditivos, niño, potenciales evocados,localización del sonido

Abreviaturas: CHAPS = Escala de Desempeño Auditivo del Niño; EEG =electroencefalográfico; ERP = potenciales relacionados con el evento; IID =diferencias interaurales de intensidad; ITD = diferencias temporales interaurales;N1 = pico negativo entre 80 – 120 mseg.; N2 = pico negativo entre 180 – 290mseg.; P1 = pico positivo entre 50 – 90 mseg.; P2 = pico positivo entre 140 –190 mseg.; P3 = pico positivo a 300 mseg. o más; VEOG = electro-oculográficovertical

JJoouurrnnaall ooff tthhee AAmmeerriiccaann AAccaaddeemmyy ooff AAuuddiioollooggyy/Volume 16, Number 5, 2005

314

2001). For example, the N1 waveform inadults has been used to study the neuralprocessing of interaural time (ITD) andinteraural intensity (IID) differences in atrain of click stimuli (Ungan et al, 2001).Dipole analysis of the N1 obtained in theUngan et al’s study indicated contralateralprimary cortex dominance in response toboth binaural cues, specifically in thesupratemporal plane. Several studies haveused dipole source localization andneuroimaging techniques to determine howbinaural features are represented in thehuman brain (Griffiths et al, 1994; Mäkeläand McEvoy, 1996; Palomäki et al, 2000).These studies have confirmed two responsepatterns to sound movement: (1) greateractivation in the hemisphere contralateral tothe direction of the sound movement/localization (e.g., greater activation of theright hemisphere for left moving sounds), alsoreferred to as contralateral activation (Mäkeläand McEvoy, 1996; Pavani et al, 2002), and (2)overall more pronounced activation in theright hemisphere when compared to nonspatialstimuli, regardless of the direction of soundmovement (Griffiths et al, 1994; Anourova etal, 2001; Xiang et al, 2002).

Many auditory perceptual skills haveextended maturation time courses, such asauditory gap detection (Phillips, 1999),temporal ordering (Irwin et al, 1985), dichoticlistening (Jerger et al, 2002), and soundlocalization (Besing and Koehnke, 1995;Litovsky, 1997). In general, neuromaturationof sound processing can be seen in the age-related changes in the obligatory componentsof auditory ERPs (Ponton et al, 1996, 2000;Sharma et al, 1997; Ceponiene et al, 2002).The term “obligatory components” refers tocomponents in the ERP waveform that areelicited and vary in response to stimuluscharacteristics.

Although binaural cues in adults havebeen shown to affect the obligatory N1component (Ungan et al, 2001), in youngchildren P1 and N2 (sometimes referred to asN1b; e.g., Sharma et al, 1997) dominate thelong-latency obligatory auditory ERPs.According to Ponton et al (1996, 2000), N2experiences a more complex maturationpattern than the earlier responses. N2 latencyis significantly shorter in younger subjectsuntil age 14 (electrodes C3, C4, and Cz); theseshorter latencies continue at electrode C4until adulthood (age 18–20). These latency

changes are not evident at the Fz electrode(Ponton et al, 2000). In addition, N2demonstrates an interhemispheric latencydifference that increases with age; longerlatency over the hemisphere contralateral tothe stimulated ear (only the left ear wasstimulated in their study) is most notable inadulthood. N2 amplitude increases until ageten and then decreases in amplitude untiladulthood. Maturational changes were alsodiscussed in Ceponiene et al (2002), with theirobligatory responses showing the greatestactivation at the fronto-central scalp sites.

Several researchers have suggested thatthe generator sources for N1 and N2 are notlocated in the same area of the cortex andrepresent different aspects of auditoryprocessing (Ceponiene et al, 2002; Takeshitaet al, 2002). In children, N2 appears to betightly coupled to stimulus complexity level(Ceponiene et al, 2001). Thus, children withproblems processing auditory information maymanifest with differences in ERP responses(N1 vs. N2) that are dependent on the stimuliused and the cause of the processing problem(language impairment, auditory processingdisorders, and/or dyslexia).

In a series of articles by Kraus andcolleagues, late evoked potentials in responseto speech stimuli have demonstratedprocessing differences in children withlearning problems when compared to childrenwithout learning problems (Cunningham etal, 2000; King et al, 2002; Warrier et al, 2004;Wible et al, 2002). Cunningham et al (2000)found that, in addition to maturationalchanges occurring in P1/N1/N2, behavioralmeasures of auditory processing have apredictive relationship with the latency ofthe N2 component. While Kraus et al’sfindings have demonstrated no differences inthe overall development of P1/N1/N2 tospeech sounds presented in quiet betweenchildren with and without learning problems,differences have been obtained for speechsounds in noise (King et al, 2002). When theP1/N1/P2/N2 components generated in quietand in background noise are correlated,children with learning problems have areduction in the mean correlation (King et al,2002; Wible et al, 2002). The technique ofcross-correlation within individual subjectsdemonstrates a highly degraded response inthe presence of noise, particularly in the N2region, in listeners with learning problems(Warrier et al, 2004). This poor correlation


315

appears to be a result of a reduced N2 latencyin the learning-impaired children in thepresence of noise. The authors suggest thatthe N2 latency differences in these childrenwith learning problems may be due toauditory processing deficits (Warrier et al,2004). Current research from this group isfocusing on the effect of auditory-basedtraining on the response correlation in the N2region in children with learning problems(Warrier et al, 2004). The success of thistraining suggests that these children have anunderlying neurophysiologic problem withinthe auditory system.

Purdy et al (2002) have also demonstrateddifferences in the N2 component in childrenwith learning disabilities and a “possibleAPD” (Purdy et al, 2002, p. 370). They lookedat the obligatory components in response tosimple pure-tone stimuli presented in anoddball paradigm to children with possibleAPD as diagnosed by a psychologist, and tochildren with no learning or hearing problemsas reported by their parents. Children wereidentified as possibly having APD based ontheir performance on measures of auditoryimmediate memory/attention, auditory andvisual immediate memory/attention, auditorylong-term memory/attention, and auditorydiscrimination of speech phonemes. Inaddition to larger peak-to-peak amplitude ofP2-N2, shorter latencies and smalleramplitudes were observed for the earlierpeaks, P1 and N1, in the children withprocessing problems. The authors concludedthat the differences in ERP results betweenthe two groups of children might account forthe difficulties experienced by the childrenwith possible APD.

Differences in the amplitudes of theobligatory components between childrensuspected of an auditory processing disorderand children not suspected of an auditoryprocessing disorder were also found by Liasiset al (2003). Children were suspected of APDbased on self-report and parental report,normal audiometric results, abnormal scoreson a screening test for APD, no CNSpathology, and normal scores on measures ofspeech and language or cognitive abilities.Liasis et al recorded auditory event-relatedpotentials to phonemes presented in anoddball paradigm and reported larger peak-to-peak P1-N1 and P2-N2 amplitudes andsmaller absolute amplitudes for N1 and P2in children suspected of auditory processing

disorders as compared to controls. Theapparent discrepancies between these twostudies (Purdy et al, 2002; Liasis et al, 2003)may be due to stimulus differences,differences in the definition of the subjectpopulation or in ERP measures utilized.

The only known study to date thatexamines moving sound stimuli in childrenwith auditory processing problems is fromJerger and colleagues (Estes et al, 2002).Jerger and Estes (2002) introduced a movingsound stimulus paradigm and measured theelectrophysiological response in children(9–12), younger adults (18–34), and olderadults (65–80). They found that the P3component (between 400 and 800 msec) tosound movement evoked a robust response innormal children with a greater activationover the right hemisphere than over the left.Children showed a substantially largerasymmetry than either the seniors or theyoung adults. When this same stimulusparadigm was applied to children who arepoor listeners (Estes et al, 2002), ERPresponses in the midline region were similarbetween good and poor listeners; however,there were group differences for the lateralelectrodes. Normal listeners had a largeractivation to the right while poor listeners hada smaller asymmetric activation to the left;differences in hemispheric activation patternswere only analyzed for the P3 component.This suggests that the poor listeners arelacking the right hemisphere specialization forsound movement found in the adult literature(Griffiths et al, 1994; Anourova et al, 2001;Xiang et al, 2002) and in normal children(Estes et al, 2002; Jerger and Estes, 2002).

In contrast to Estes et al, who evaluatedgroup effects of sound movement in the P3component, the current investigationdetermines whether the obligatory ERPcomponents in response to sound movementdistinguish between children who are poorlisteners and children who are good listeners.Based on previous research findings,perception of sound movement should resultin (1) greater amplitudes in the hemispherecontralateral to the direction of the soundmovement and (2) overall greater amplitudesin the right hemisphere for the good listeners,while the poor listeners would exhibit lesshemispheric specialization in the form ofsymmetrical bilateral activation patterns.This study focuses on the neural processingof IIDs, creating an illusion of sound


316

movement. Since auditory gap detection doesnot rely on the processing of binaural intensitycues, a gap detection task was included in thecurrent study as a control condition.

MMEETTHHOODDSS

SSuubbjjeeccttss

Eighteen children (ten females, eightmales; mean age 9.61, SD 1.65; age range7–13 years) participated in the study.Participants were recruited from theDepartment of Audiology at JFK JohnsonRehabilitation Institute or by advertisement.Participants were only included if the scoresobtained from the Children’s AuditoryPerformance Scale (CHAPS; Smoski et al,1998), answered by the parents, corroboratedwith results on an auditory processing testbattery administered by a licensed audiologist(see Table 1). The CHAPS is a behavioralquestionnaire that is effective in definingand evaluating listening behaviors that maygenerate a referral for Central AuditoryProcessing testing (Bellis, 2003). The CHAPSincludes questions specifically addressinglistening abilities involved in the perceptionof auditory movement, such as the detectionand processing of multiple auditory inputs.When the questionnaire was field tested inchildren with and without auditory processingdeficits, clear differences were obtainedbetween the groups with no overlap in therange of scores for each group (Smoski et al,1998).

Nine children (five males, four females;mean age 9.89, SD 1.69) were identified aspoor listeners in this study based on an at-risk score on the CHAPS (total conditionscore between -12 and -130) and abnormalperformance in the APD test battery on atleast two tests measuring the same auditoryprocess (Table 1). As abnormal scores across

all three auditory processes would pointtowards a global disorder (e.g., attentiondeficit, mental disability), we ascertainedthat none of the poor listeners had abnormalperformance on multiple tests in all of theprocesses tested. Nine children (three males,six females; mean age 9.33, SD 1.66) obtaineda passing score on the CHAPS (total conditionscore between +36 and -11) and had normalperformance in all or all but one task perauditory process on the APD test battery;they will be referred to as good listeners inthis study. All participants were right-handedbased on a handedness questionnaire (Annett,1970) and had normal hearing as measuredby an audiometric screening at octaveintervals between 500 and 4000 Hz (threshold≤25 dB HL).

SSttiimmuullii

Two auditory tasks were employed inthe current study, auditory movementdetection and auditory gap detection. Theauditory gap detection task was included asa control task because it does not requireintegration of auditory cues from both ears.In both tasks, the standard stimulus was a500 msec burst of broadband noise presentedbinaurally at 60 dB SPL. The tasks differedin the target stimuli that were to be identified.In the auditory movement task, participantswere to detect noise bursts that appeared tobe moving from the median plane to the leftor to the right. The illusionary soundmovement was solely created by IIDs (Figure1, adapted from Estes et al, 2002). In theauditory gap task, participants were toidentify silent gaps in the center of the noiseburst. The duration of the gap was either 30or 100 msec, and the overall length of thestimulus remained constant at 500 msecacross conditions (Figure 1). All stimuli werepresented with an interstimulus interval of2 seconds. The level of the IIDs and the gap

Table 1. Behavioral APD Test Battery Organized by Auditory Process

AAuuddiittoorryy PPrroocceessss CCoorrrreessppoonnddiinngg BBeehhaavviioorraall TTeesstt

Binaural Integration Dichotic DigitsSSW Staggered Spondaic Word Test

Temporal Ordering Duration Patterns (Verbal, Humming)Pitch Pattern Sequence Test (Verbal, Humming)Random Gap Detection

Monaural Closure NU-6 Low Pass Filtered SpeechNU-6 Cafeteria Noise


317

length in this study were selected to ensurehigh accuracy scores in both poor and goodlisteners. This allowed us to evaluate groupdifferences in neurophysiological processingof stimuli when the behavioral performanceis not compromised. A detailed description ofthe auditory stimuli and paradigm can befound in Estes et al (2002).

EExxppeerriimmeennttaall PPrroocceedduurree

Participants were seated in a comfortablechair and stimuli were delivered toparticipants through earphones (left, right).A response panel was placed in front of theparticipant. The participant’s task was topush the button upon each targetpresentation. A standard oddball paradigmwas employed, with standard stimulioccurring in 50% of trials (100 trials) andthe two types of target stimuli in 25% oftrials (50 trials each).

EEEEGG RReeccoorrddiinnggss

Brain electrical activity was recordedusing the Neuroscan system. An elastic cap

(Neurosoft), equipped with 32 silver-silverchloride electrodes according to theInternational 10–20 system was affixed to theparticipant’s scalp. Vertical and lateralelectro-oculographic activity (VEOG) wasmonitored by placing electrodes above and atthe outer canthus of the left eye. Duringrecording, linked mastoid electrodes servedas reference electrodes, and APZ served as theground electrode. Impedances for eachelectrode did not exceed 5 kohm as measuredbefore the test session.

Electroencephalographic activity wasamplified and bandpass, analog filtered from.15 to 50 Hz (-12 dB per octave). Digitizationwas then carried out through NeuroscanAcquisition. Digitized individual sweeps of theelectroencephalographic (EEG) activity wererecorded and stored for off-line analysis,averaging, and digital filtering.

Auditory evoked potentials were recordedat a sampling rate of 1000 Hz. Sweep durationextended from -200 to +1400 msec relative tostimulus onset. A prestimulus baselineaverage was calculated based on the first200 data points (first 200 msec). Off-linesignal averaging was carried out after off-line

FFiigguurree 11.. Waveforms of auditory movement and auditory gap stimuli. Signal is broadband noise with dura-tion of 500 msec. In the auditory movement task, the amplitude levels in both channels are equal at onset butdiffer by 18 dB at offset (adapted with permission from Estes et al, 2002).


318

FFiigguurree 22.. Spatial movement detection. Grand averaged waveforms across subject groups and movement con-ditions for poor and good listeners.

FFiigguurree 33.. Auditory gap detection. Grand averaged waveforms across subject groups and gap conditions for poorand good listeners.


319

artifact rejection and baseline correctionprocedures. Individual sweeps were rejectedwhen the electro-oculographic activity(VEOG) exceeded ±50 microvolts (µV). Themean number of accepted epochs was 30 foreach auditory movement condition, and 33 foreach auditory gap condition. Upon completionof the previous steps, the averaged waveformswere linearly detrended, low-pass digitallyfiltered at 20 Hz with a filter slope of -48 dBper octave and re-referenced to a commonaveraged reference.

SSttaattiissttiiccaall AAnnaallyyssiiss

Latency and amplitude measures wereobtained with a software package (NeuroscanEdit 4.3) that identified both measures in theP1, N1, P2, and N2 peaks of the individualaveraged waveforms in response to targetstimuli. In agreement with previous findings(Ceponiene et al, 2002), the children’s obligatorycomponents were chiefly located at the fronto-central electrode sites. Therefore, the latencywindow for peak identification was based onvisual inspection of the grand averagedwaveforms across conditions and subject groupsat FCZ (Figures 2 and 3). The latency windowswere 50–90 msec for P1, 80–120 msec for N1,140–190 msec for P2, and 180–290 msec for theN2 component. Peak identification was basedon the point of largest amplitude within thepredetermined latency window.

Separate repeated measures ANOVAswere conducted for the auditory movementand auditory gap tasks. Within-subjectindependent variables included condition(auditory movement: movement direction leftor right; auditory gap: gap duration 30 or 100msec) and hemisphere (left, FC3; right, FC4).

Subject group was included as a between-subject independent variable. We wereultimately interested in the differences inbrain activation patterns between good andpoor listeners in response to sound movement.Therefore, we will report the results of theauditory gap detection task only in referenceto the results of the auditory movement task.

RREESSUULLTTSS

BBeehhaavviioorraall RReessuullttss

Based on the accuracy scores, all subjectsperformed well above chance-levels (>65%).Separate repeated measures ANOVAs ofpercent correct scores for the auditorymovement and auditory gap tasks withstimulus type (standard, target) as the within-subject variable and subject group as thebetween-subject variable revealed nosignificant effects of group (F [1,16] = 3.250;p = .090), stimulus type (F [1,16] = .855; p =.369), or stimulus type by group (F [1,16] =1.858; p = .226).

GGrraanndd AAvveerraaggeedd WWaavveeffoorrmmss

For auditory movement as well asauditory gap detection, the earliest peaksobserved reliably in the grand averagedwaveform across subject groups andconditions were the obligatory P1, N1, P2, andN2 components (from 50 to 290 msec),followed by a significantly later endogenousP3 component between 350 and 800 msec(Figures 2 and 3). It is important to notethat the obligatory components did notoverlap in time with the endogenous P3component. The goal of the current paperwas to evaluate group differences for onlythe obligatory components.

EERRPP AAnnaallyysseess

Analyses of the amplitude and latency ofthe P1, N1, and P2 peaks did not showsignificant findings for auditory movement inthe lateral fronto-central electrode sites (FC3,FC4). Analyses of the N2 component arereported below.

FFiigguurree 44.. Interaction effect of sound movement byhemisphere showing that sounds moving to the lefthemifield evoked greater responses over FC4 thanFC3. No such differences were observed for soundsmoving to the right hemifield. *Significant at p = .006.


320

AAuuddiittoorryy MMoovveemmeenntt DDeetteeccttiioonn

NN22 AAmmpplliittuuddee aatt FFCC33 aanndd FFCC44

Significant interaction effects ofmovement by hemisphere (F [1,16] = 5.138;p = .038) and of group by movement (F [1,16]= 7.580; p = .014) were found for the auditorymovement task. Post hoc paired samples t-tests for the movement by hemisphereinteraction resulted in a significant effect ofhemisphere for sound movement to the left(t = 3.109, df = 17, p = .006) but not to theright; that is, when groups were combined,greater activation was observed for leftmovement over FC4 as compared to FC3; nohemispheric differences were obtained for

right movement (Figure 4). Post hoc pairedsamples t-tests for the group by movementinteraction revealed that N2 amplitude inthe good listeners was greater for leftmovement than for right movement (t = -2.772,df = 8, p = .024); no significant differencebetween left and right movement was foundin the poor listeners. Furthermore,independent samples t-tests showed that N2amplitude in response to left movement wassignificantly greater in the good listeners ascompared to the poor listeners (t = -2.384, df = 16, p = .030; Figure 5). No significantfindings were obtained for N2 amplitude inresponse to right movement. As the observedeffects were mainly caused by differences inresponse to left-moving sounds, repeatedmeasures ANOVAs were conducted includingonly the left movement, with hemisphere asthe within-subject variable and subject groupas the between-subject variable. Main effectsof hemisphere (F [1,16] = 9.475; p = .007)and of group (F [1,16] = 5.684; p = .030) wereobtained. Subsequent paired samples t-testsfor each subject group showed that in goodlisteners the N2 amplitude was greater atFC4 than at FC3 (t = 2.654, df = 8, p = .029;Figure 6) while no significant differencebetween hemispheres was observed in thepoor listeners in response to left movingsounds. Finally, separate independentsamples t-tests for each hemisphere showeda significant difference between groups atFC4 (t = -2.492, df = 16, p = .024; Figure 6).

NN22 LLaatteennccyy aatt FFCC33 aanndd FFCC44

Repeated measures ANOVA with N2latency as the dependent variable revealed asignificant interaction effect of hemisphere bygroup (F [1,16] = 4.712, p = .045). Post hocpaired samples t-tests for each subject groupshowed good listeners to have shorter N2latencies over FC3 than FC4 (t = 3.313, df =8, p = .011). N2 latencies over FC3 and FC4did not differ significantly in poor listeners.As sound movement did not contribute tothe observed interaction effect, the grandaveraged waveforms in Figure 7 representresponses averaged across left and rightmoving sounds. As stated earlier, N2 peakedearlier over FC3 than FC4. Interestingly, atthe FC3 peak latency, the N2 amplitude wasobserved to be equal over FC3 and FC4.However, an additional negative shift wasobserved at the FC4 scalp site with a peak

FFiigguurree 55.. Interaction effect of group by movementshowing that in good listeners sounds moving to theleft hemifield evoked greater responses than soundsmoving to the right hemifield. Furthermore, forsounds moving to the left hemifield, good listeners hadgreater N2 amplitude than poor listeners. +Signifi-cant at p = .030. *Significant at p = .024.

FFiigguurree 66.. Repeated measures ANOVA for left move-ment only. In good listeners, the N2 amplitude wasgreater over FC4 than FC3. N2 amplitude at FC4 wasgreater for good listeners than poor listeners. +Significant at p = .029. *Significant at p = .024.


321

latency occurring around 280 msec poststimulus onset. We will refer to this additionalnegative shift as the N2′ for the remainderof the paper. The N2′ was not present in thegrand averaged waveform of the poorlisteners.

NN22 AAmmpplliittuuddee aanndd LLaatteennccyy aatt FFCCZZ

To ascertain that the observed effectstruly represented differences in scalpactivation patterns for auditory movementand were not a mere reflection of overallERP differences between subject groups, weconducted separate repeated measuresANOVAs for N2 amplitude and latency at

midline electrode FCZ. No significantdifferences between groups were found.

AAuuddiittoorryy GGaapp DDeetteeccttiioonn

In contrast to the movement task, nosignificant group differences were obtainedin the statistical analyses of N2 amplitudeand latency at FC3, FC4, and FCZ for theauditory control task.

SSUUMMMMAARRYY

Left-moving sounds evoked a clearhemispheric asymmetry in good listeners,

by way of greater N2 amplitude at FC4

FFiigguurree 77.. (a) Grand averaged waveforms across movement conditions for good listeners at FC3 and FC4. Notethe presence of the N2′ at FC4. A significant difference was obtained between FC3 amplitude and FC4 ampli-tude each measured at the peak latency observed at FC4 (t = -2.483, df = 8, p = .038); (b) Grand averaged wave-forms across movement conditions for poor listeners at FC3 and FC4. Note that the N2′ is absent and that nosignificant amplitude difference was obtained between FC3 amplitude and FC4 amplitude each measured atthe peak latency observed at FC4 (t = -1.347, df = 8, p = .215).


322

compared to FC3, while no hemisphericdifferences were found in poor listeners.Furthermore, N2 amplitude at FC4 for left-moving sounds was greater in good listenersthan in poor listeners. The hemisphericasymmetry in good listeners wasaccompanied by longer latencies at FC4 thanat FC3 regardless of the direction of thesound movement; this may be due to anadditional negative shift at FC4, which wehave labeled “N2′.” N2′ was absent in thepoor listeners. Finally, the correspondingmain effects and interaction effects in theauditory control task were not significant.

DDIISSCCUUSSSSIIOONN

The current study adds to the literature ofauditory processing disorders in that few

studies have investigated obligatory ERPcomponents for sound movement in childrenwith poor listening skills. In addition to thedetection of sound movement, children wereasked to perform an auditory gap detectiontask (control task), which involved stationarysound images. ERP measures revealed groupdifferences in the auditory movement taskthat were not noted in the auditory gapdetection task, nor in the behavioral measuresof auditory movement detection. The goal ofthis study was to evaluate group differencesrelated to the processing of sound movementin the horizontal plane. Therefore, we willfocus our discussion on the pattern of resultsunique to sound movement distinguishingpoor listeners from good listeners.

The overall impression of the presentfindings is that scalp activation patterns forsound movement differed statisticallybetween good and poor listeners in thatasymmetry in scalp activation was present ingood listeners but absent in poor listeners.Poor listeners did not have greater difficultythan the good listeners with the tasks athand; statistical analysis of the behavioralaccuracy scores showed no group differences.This indicates that the ERPs obtained in thisstudy reflect underlying neurophysiologicaldifferences when behavioral performance isnot compromised. In addition, no groupdifferences were obtained in the analysis ofthe N2 amplitude and latency measures atmidline electrode FCZ, suggesting that ourresults truly represent different scalpactivation patterns for poor listeners andgood listeners.

In the good listeners, sounds movingfrom the median plane horizontally to the lefthemifield evoked greater N2 amplitude overthe right hemisphere (over FC4) as comparedto the left hemisphere (over FC3). Althoughwe did not conduct source analyses of theERP activation patterns in the current study,source localization studies and MEG studiesto auditory movement and localization haveestablished that the contralateral, and not theipsilateral, auditory pathway processes soundmovement (Palomäki et al, 2000; Ungan etal, 2001; Fujiki et al, 2002). According toUngan et al (2001), this was found to beparticularly true for sound stimuli in whichthe illusionary motion was created by IIDs,as was the case in our study. Mäkelä andMcEvoy (1996) also utilized IID stimuli tocreate a perception of sound motion. UtilizingMEG as a recording technique in youngadults, they registered greater N1mamplitudes in the hemisphere contralateralto the ear receiving the intensity increase.Similarly, a greater N1m response was foundcontralateral to the location of stationarysound images created with head-relatedtransfer functions (Palomäki et al, 2000).

In addition to contralateral activationpatterns for sound perception, predominanceof the right hemisphere for spatial hearinghas been documented in behavioral andneuroimaging studies. Greater localizationaccuracy has been reported for soundspresented in the left hemifield as comparedto the right hemifield (Burke et al, 1994).Over and above the greater activation in thehemisphere contralateral to the locus ofsounds, Palomäki et al (2000) reported overallgreater amplitude in the electromagneticfield responses (EMFs) of the righthemisphere than the left hemisphere andattributed this difference to the specializedfunction of the right hemisphere for spatialorganization of auditory events. Thus, incontrast to sounds moving to the righthemifield, sounds moving to the left hemifieldare more likely to evoke greater activationover the right hemisphere, not only becausethe right hemisphere is specialized for soundmovement but also because the hemispherecontralateral to the sound movement (i.e.,the right hemisphere for left-moving sounds)is specifically involved in processing of theauditory information. Similarly, for right-moving sounds, right hemispherespecialization for sound movement in


323

combination with contralateral activation(i.e., the left hemisphere for right-movingsounds) could result in bilateral activation.Indeed, in the current study, the asymmetryin N2 scalp activation observed in responseto left-moving sounds is lacking in responseto right-moving sounds. Instead, wedocumented a bilateral activation patternfor sounds moving to the right (Figure 4).

As high resolution EEG is needed foraccurate source localization, we cannot inferthat the neural sources evoking right scalpactivation patterns originated in the righthemisphere. However, there is evidence fromevoked magnetic field recordings (Xiang et al,2002; Mathiak et al, 2003) as well as evokedpotential recordings in adults (Ducommun etal, 2002) suggesting that responses evoked atthe same time interval as the N2 componentidentified in the current study are generatedin the right cortex during the perception ofsound motion and sound localization. Righthemisphere activation has been observed inthe right parietal lobe (Griffiths et al, 1998;Ducommun et al, 2002; Xiang et al, 2002;Zatorre et al, 2002) as well as the right insula(Griffiths et al, 1994). Baumgart andcolleagues (1999) identified activation of theright planum temporale in response toillusionary sound motion resulting from ITDmanipulation. Finally, acoustic cues toauditory distance perception such as soundintensity variations specifically activate theright supratemporal plane (Seifritz et al,2002; Mathiak et al, 2003).

Interestingly, N2 generators in childrenhave been identified in the supratemporalauditory cortex, frontal cortex, and parietalcortex (Ceponiene et al, 2002). Thus, the N2component seems to be particularly suited tomeasuring children’s brain responses tomoving sounds. Indeed, scalp asymmetry inthe N2 amplitude was only present inresponse to sound movement. Responses toauditory gaps evoked a bilateral scalpactivation pattern in both groups.

N2 amplitude in response to leftmovement was significantly greater in thegood listeners as compared to the poorlisteners, which could conceivably reflect (1)immature auditory processing in the poorlisteners, (2) better perception of left movingsounds by good listeners than by poorlisteners, and (3) better synchrony along theauditory pathway in the good listeners. Duringchildhood, N2 amplitude increases with age

and reaches its maximal value around 10–11years of age. Afterwards it decreases graduallyand reaches adultlike values consistently at17 years of age (Ponton et al, 2000).Furthermore, N2 amplitude increases as thestimulus becomes more readily discernable(Ungan and Özmen, 1996). It is important tonote, however, that further research involvingyounger children is needed to document thematurational time course of auditorymovement detection before one could assertthat the differences demonstrated here aredue to maturational changes.

Initially the effect of longer latenciesover the right hemisphere, where N2amplitude is largest, seems peculiar in thatlong latencies are typically associated withsmall amplitudes if both measures representthe same neurophysiological mechanism(Ungan and Özmen, 1996). One possibleexplanation for this difference is that twoneurophysiological mechanisms arerepresented. Based on the grand averages forthe observed latency interaction effect ofhemisphere by group (Figure 7), we noticedthat the longer latency at FC4 reflected anadditional negative shift, earlier referred toas N2′, which was only noted in the goodlisteners. Given that this effect was combinedacross directions of sound movement, we could hypothesize that N2 ′ is themanifestation of the right hemispherepredominance for sound localizationremarked upon in the literature (Baumgartet al, 1999; Ducommun et al, 2002; Zatorreet al, 2002; Mathiak et al, 2003). The absenceof N2′ in our poor listeners, combined withbilateral N2 voltage distribution, suggests alack of right hemisphere dominance for soundlocalization in this subject group. Bilateralrepresentation of function has been thoughtto reflect a lack of maturation (Bryden, 1982;Estes et al, 2002). This finding is inconcordance with previous accounts on brainactivation patterns for children with learningproblems (Cunningham et al, 2000; King etal, 2002; Wible et al, 2002). A difference in themyelination process and synaptogenesis hasbeen put forward as one reason why childrenwith poor listening skills might show differentpatterns in their evoked potentials (Liasis etal, 2003). Based on this premise, it is notsurprising that we registered differencesbetween subject groups in the N2 and N2′amplitude and latency as age-related changesin the N2 component appear to be caused by

“biological brain maturation in terms ofsynaptic count decrease, myelination,developing inhibitory control, rather than bysubstantial changes in the N2 function”(Ceponiene et al, 2002, p. 881).

As stated earlier, sound localization andmovement are based on binaural cues (i.e.,ITDs and IIDs) as well as monaural cues (e.g.,intensity changes). On the other hand, gapdetection does not rely on the processing ofbinaural or monaural intensity cues. Whileno group differences were found for the N2 inresponse to auditory gaps, the N2 in responseto sound movement differed between poor andgood listeners. Based on our findings, we cannotdiscern whether the ERP differentiations arein response to the monaural or binauralcharacteristics of the stimuli. Monauralintensity changes have been found to evokelarger responses in the right hemisphere, morespecifically in the right supratemporal plane,one of the N2 generators in children. Strictlyspeaking, “purely intensive mechanisms”(Ungan and Özmen, 1996, p. 486) could besolely responsible for larger activity in theright hemisphere (Belin et al, 1998; Mathiaket al, 2003). Nonetheless, brain activationpatterns still differ between good listeners andpoor listeners in response to auditorymovement. Whether differences in the N2component between poor and good listeners areprimarily evoked by the monaural or binauralintensity cues within the stimulus should beaddressed in subsequent studies specificallydesigned to evaluate this question.

CCOONNCCLLUUSSIIOONN

Scalp activation patterns in children withpoor listening skills show less right

hemisphere predominance for auditorymovement perception in the long-latencyobligatory N2 component when compared toage-matched children with good listeningskills. These disparities could be due tomaturational differences between poor andgood listeners, better synchrony along theauditory pathway, or increased salience formoving sounds in good listeners.

AAcckknnoowwlleeddggmmeennttss.. The authors would like to thankthe audiologists of JFK Johnson Rehabilitation foradministering the auditory test battery, the parentsand children for their participation in the study, JamesJerger and Rebecca Estes for sharing their stimuliand paradigm, Eric Sandler for assisting us with the

analysis of the data, and Brett Martin and AbuhuziefaAbubakr for reviewing a previous draft of the man-uscript.

RREEFFEERREENNCCEESS

American Speech-Language Hearing Association.(1996) Central auditory processing: current status ofresearch and implications for clinical practice. Am JAudiol 5:41–54.

Annett M. (1970) A classification of hand preferenceby association analysis. Br J Psychol 61:303–321.

Anourova I, Nikouline VV, Ilmoniemi RJ, Hotta J,Aronen HJ, Carlson S. (2001) Evidence for dissocia-tion of spatial and nonspatial auditory informationprocessing. Neuroimage 14:1268–1277.

Baumgart F, Gaschler-Markefski B, Woldorff MG,Heinze H-J, Scheich H. (1999) A movement sensitivearea in auditory cortex. Nature 400:724–725.

Belin P, McAdams S, Smith B, Savel S, Thivard L,Samson S, Samson Y. (1998) The functional anatomyof sound intensity discrimination. J Neurosci18:6388–6394.

Bellis TJ. (2003) Assessment and Management ofCentral Auditory Processing Disorders in theEducational Setting: From Science to Practice. 2nded. Clifton Park, NY: Delmar Learning.

Besing J, Koehnke J. (1995) A test of virtual auditorylocalization. Ear Hear 16:220–229.

Bryden MP. (1982) Laterality: Functional Asymmetryin the Intact Brain. Perspectives in Neurolinguistics,Neuropsychology, and Psycholinguistics. New York:Academic Press.

Burke KA, Letsos A, Butler RA. (1994) Asymmetricperformances in binaural localization of sound inspace. Neuropsychologia 32:1409–1417.

Casseday JH, Neff WD. (1975) Auditory localiza-tion: role of auditory pathways in brain stem of thecat. J Neurophysiol 38:842–858.

Ceponiene R, Rinne T, Näätänen R. (2002)Maturation of cortical sound processing as indexedby event-related potentials. Clin Neurophysiol113:870–882.

Ceponiene R, Shestakova A, Balan P, Alku P,Yiaguchi K, Näätänen R. (2001) Children’s auditoryevent-related potentials index sound complexityand speechness. Int J Neurosci 109:245–260.

Cunningham J, Nicol T, Zecker S, Kraus N. (2000)Speech-evoked neurophysiologic responses in chil-dren with learning problems: development andbehavioral correlates of perception. Ear Hear21:554–568.

Ducommun CY, Murray MM, Thut G, Bellmann A,Viaud-Delmon I, Clarke S, Michel CM. (2002)Segregated processing of auditory motion and audi-tory location: an ERP mapping study. Neuroimage16:76–88.

Durlach NI, Colburn HS. (1978) Binaural Phenomena.In: Carterette EC, Friedman MP, eds. Hearing. Vol.


324

4 of Handbook for Perception. New York: AcademicPress.

Estes RI, Jerger J, Jacobson G. (2002) Reversal ofhemispheric asymmetry on auditory tasks in chil-dren who are poor listeners. J Am Acad Audiol13:59–71.

Fujiki N, Riederer KA, Jousmäki V, Mäkelä JP, HariR. (2002) Human cortical representation of virtualauditory space: differences between sound azimuthand elevation. Eur J Neurosci 16:2207–2213.

Griffiths TD, Bench CJ, Frackowiak RSJ. (1994)Human cortical areas selectively activated by appar-ent sound movement. Curr Biol 4:892–895.

Griffiths TD, Rees G, Rees A, Green GGR, Witton C,Rowe D, Buchel C, Turner R, Frackowiak RSJ. (1998)Right parietal cortex is involved in the perception ofsound movement in humans. Nat Neurosci 1:74–79.

Hausler R, Colburn S, Marr E. (1983) Sound local-ization in subjects with impaired hearing:spatial-discrimination and interaural-discriminationtests. Acta Otolaryngol 400(Suppl.):1–62.

Irwin RJ, Ball AKR, Kay N, Stillman JA, Rosser J.(1985) The development of auditory temporal acuity.Child Devel 56:614–620.

Itoh K, Yumoto M, Uno A, Kurauchi T, Kimitaka K.(2000) Temporal stream of cortical representation forauditory spatial localization in human hemispheres.Neurosci Lett 292:215–219.

Jerger J, Estes R. (2002) Asymmetry in event-relatedpotentials to simulated auditory motion in children,young adults, and seniors. J Am Acad Audiol 13:1–13.

Jerger J, Musiek F. (2000) Report of the consensusconference on the diagnosis of auditory processingdisorders in school-aged children. J Am Acad Audiol11:467–74.

Jerger J, Thibodeau L, Martin J, Mehta J, TillmanG, Greenwald R, Britt L, Scott J, Overson G. (2002)Behavioral and electrophysiological evidence of audi-tory processing disorder: twin study. J Am Acad Audiol13:438–460

King C, Warrier CM, Hayes E, Kraus N. (2002)Deficits in auditory brainstem pathway encoding ofspeech sounds in children with learning problems.Neurosci Lett 319:111–115.

Koehnke J, Besing J. (2001) The effects of aging onbinaural and spatial hearing. Semin Hear 22:241–253.

Koehnke J, Culotta CP, Hawley ML, Colburn HS.(1995) Effects of reference interaural time and inten-sity differences on binaural performance in listenerswith normal and impaired hearing. Ear Hear16:331–353.

Liasis A, Bamiou D-E, Campbell P, Sirimanna T, BoydS, Towell A. (2003) Auditory event-related potentialsin the assessment of auditory processing disorders:a pilot study. Neuropediatrics 34:23–29.

Litovsky R. (1997) Developmental changes in theprecedence effect: estimates of minimum audibleangle. J Acoust Soc Am 102:1739–1745.

Mäkelä JP, McEvoy L. (1996) Auditory evoked fieldsto illusory sound source movements. Exp Brain Res110:446–454.

Mathiak K, Hertrich I, Kincses WE, Riecker A,Lutzenberger W, Ackermann H. (2003) The rightsupratemporal plane hears the distance of objects:neuromagnetic correlates of virtual reality.Neuroreport 14:307–311.

Mendelson JR, Grasse KL. (2000) Auditory corticalresponses to the interactive effects of interaural inten-sity disparities and frequency. Cereb Cortex 10:32–39.

Palmer AR. (1995) Neural signal processing. In: MooreB, ed. The Handbook of Perception and Cognition.2nd ed. San Diego: Academic Press, 75–121.

Palomäki K, Alku P, Mäkinen V, May P, Tiitinen H.(2000) Sound localization in the human brain: neu-romagnetic observations. Neuroreport 11:1535–1538.

Pavani F, Macaluso E, Warren JD, Driver J, GriffithsTD. (2002) A common cortical substrate activated byhorizontal and vertical sound movement in the humanbrain. Curr Biol 12:1584–1590.

Phillips DP. (1999) Auditory gap detection, percep-tual channels, and temporal resolution in speechperception. J Am Acad Audiol 10:343–354.

Pickles JO. (1988) An Introduction to the Physiologyof Hearing. New York: Academic Press.

Polyakov A, Pratt H. (1996) Evidence for spatio-topicorganization of binaural processing in the humanbrainstem. Hear Res 94:107–115.

Ponton CW, Don M, Eggermont JJ, Waring MD,Masuda A. (1996) Maturation of human cortical audi-tory function: differences between normal-hearingchildren and children with cochlear implants. EarHear 17: 430–437.

Ponton CW, Eggermont JJ, Kwong B, Don M. (2000)Maturation of human central auditory system activ-ity: evidence from multi-channel evoked potentials.Clin Neurophysiol 111:220–236.

Pratt H, Polyakov A. (1999) Evoked potentials insound localization: timing of activity along the audi-tory pathway. In: Comi G, Lucking CH, Kimura J,Rossini PM, eds. Clinical Neurophysiology: FromReceptors to Perception. Electroencephalogr (Suppl.50):235–242.

Purdy SC, Kelly AS, Davies MG. (2002) Auditorybrainstem response, middle latency response, andlate cortical evoked potentials in children with learn-ing disabilities. J Am Acad Audiol 13:367–382.

Seifritz E, Neuhoff JG, Bilecen D, Scheffler K,Mustovic H, Schächinger H, Elefante R, Di Salle F.(2002) Neural processing of auditory looming in thehuman brain. Curr Biol 12:2147–2151.

Sharma A, Kraus N, McGee TJ, Nicol TG. (1997)Developmental changes in P1 and N1 central audi-tory responses elicited by consonant-vowel syllables.Electroencephalogr Clin Neurophysiol 104:540–545.


325

Smoski W, Brunt M, Tannahill C. (1998) C.H.A.P.S.:Children’s Auditory Performance Scale InstructionManual. Tampa, FL: Educational Audiology Society.

Takeshita K, Nagamine T, Thuy DHD, Satow T,Matsuhashi M, Yamamoto J, Takayama M, FujiwaraN, Shibasaki H. (2002) Maturational change of par-allel auditory processing in school-aged childrenrevealed by simultaneous recordings of magnetic andelectric cortical responses. Clin Neurophysiol113:1470–1484.

Ungan P, Özmen B. (1996) Human long-latencyresponses to brief interaural disparities of intensity.Electroencephalogr Clin Neurophysiol 99:479–490.

Ungan P, Yagcioglu S, Goksoy C. (2001) Differencesbetween the N1 waves of the responses to interauraltime and intensity disparities: scalp topography anddipole resources. Clin Neurophysiol 112:485–498.

Warrier CM, Johnson KL, Hayes EA, Nicol T, KrausN. (2004) Learning impaired children exhibit timingdeficits and training-related improvements in audi-tory cortical responses to speech in noise. Exp BrainRes 157(4):431–441.

Wible B, Nicol T, Kraus N. (2002) Abnormal neuralencoding of repeated speech stimuli in noise in chil-dren with learning problems. Clin Neurophysiol113:485–494.

Xiang J, Chuang S, Wilson D, Otsubo H, Pang E,Holowka S, Sharma R, Ochi A, Chitoku S. (2002)Sound motion evoked magnetic fields. ClinNeurophysiol 113:1–9.

Xu L, Furukawa S, Middlebrooks JC. (2000) Corticalmechanisms for auditory spatial illusions. ActaOtolaryngol 120:263–266.

Zatorre RJ, Bouffard M, Ahad P, Belin P. (2002) Whereis ‘where’ in the human auditory cortex? Nat Neurosci5:905–909.


326

perception of auditory movement in children with poor ...€¦ · kelly j. shea-miller‡* anne m....

Documents