Abstract

We investigated attentional effects on human auditory signal-in-noise processing in a simultaneous masking paradigm using magnetoencephalography. Test signal was a monaural 1000-Hz tone; maskers were binaural band-eliminated noises (BENs) containing stopbands of different widths centered on 1000 Hz. Participants directed attention either to the left or the right ear. In an “irrelevant visual attention” condition subjects focused attention on a screen. Irrespective of attention focus location, the signal appeared randomly either in the left or right ear. During auditory focused attention (left- or right-ear attention), the signal thus randomly appeared either in the attended (“relevant auditory attention” condition) or the nonattended ear (“irrelevant auditory attention” condition). Results showed that N1m source strength was overall increased in the left relative to the right hemisphere, and for right-ear versus left-ear stimulation. Moreover, when attention was focused on the signal ear (relevant auditory attention condition) and the BEN stopbands were narrow, the right-hemispheric N1m source strength was increased, relative to irrelevant auditory attention. Such increments were neither observed in the left hemisphere nor for wide BENs. These novel results indicate 1) left-hemispheric dominance and robustness during auditory signal-in-noise processing, and 2) right-hemispheric assistance during attentive and demanding auditory signal-in-noise processing.

Introduction

Auditory signal-in-noise processing in humans has been investigated with “simultaneous masking” paradigms, by overlaying signals (e.g., tone, speech sound) with maskers (e.g., white noise), at behavioral (Zwislocki et al. 1968; Zwicker and Fastl 2007), and electrophysiological levels (Hari and Mäkelä 1988; Brancucci et al. 2004; Morita et al. 2006; Okamoto, Stracke, Ross, et al. 2007; Okamoto, Stracke, Wolters, et al. 2007). Using magnetoencephalography (MEG), Sams and Salmelin (1994) showed that the amplitude of the auditory N1m response (generated in nonprimary auditory cortex; Pantev et al. 1995; Eggermont and Ponton 2002) evoked by a tone was a function of the width of the frequency band that had been eliminated from simultaneously presented noise maskers overlapping the tone spectrum. Thus, auditory processing is systematically influenced by the signal-to-noise ratio of bottom-up neural inputs. However, top-down processes also modulate auditory signal-in-noise processing. We have shown recently (Okamoto, Stracke, Wolters, et al. 2007) that auditory focused attention not only amplified the amplitude of the auditory N1m response evoked by a tone during simultaneous masking, but also sharpened the frequency tuning in the human auditory system, indicating that bottom-up neural inputs and top-down processes interact.

Recent neuroscience research has revealed functional asymmetries of left and right auditory cortices during sound processing. On the one hand, studies indicated relatively increased activation in the left hemisphere during the processing of speech (Eulitz et al. 1995; Alho et al. 1998; Belin et al. 2000; Szymanski et al. 2001) as well as temporal acoustic features (Zatorre and Belin 2001; Jamison et al. 2006). Noteworthy, the speech processing-related leftward lateralization of brain activation is in accordance with works demonstrating a behavioral right-ear advantage for speech stimuli (e.g., Hugdahl and Andersson 1984; Bryden 1988). On the other hand, relatively increased activation in the right hemisphere during the processing of music (cf. Zatorre et al. 2002) as well as spectral acoustic features (Zatorre and Belin 2001; Jamison et al. 2006) has been observed.

However, functional asymmetries may not only depend on the cognitive quality (speech vs. music) or the composition of the sound signal on the acoustic feature level (temporal vs. spectral), but also on the presence or absence of “noise” as well as the attentional state of the listener. Regarding the role of noise, Shtyrov et al. (1998) observed significant increments (relative to a silent control condition) of both mismatch negativity and P2m amplitude (Shtyrov et al. 1999) in the right hemisphere during processing of consonant-vowel syllables masked by white noise. Results were interpreted as reflecting an increased right-hemisphere role in speech sound processing under noisy conditions, involving recruitment of additional right auditory cortex resources. Other studies observed relatively larger N1m amplitudes in the left auditory cortex compared with the right during processing of tonal signals masked by noises, during nonattentive listening as well as auditory focused attention (Okamoto, Stracke, Ross, et al. 2007; Okamoto, Stracke, Wolters, et al. 2007). Results indicated left-hemispheric dominance for auditory processing in noisy environments. Regarding the role of attention, it has for instance been demonstrated that directing attention to the right ear during the dichotic presentation of speech sounds increased the right-ear advantage, whereas directing attention to the left ear reduced the right-ear advantage (Hugdahl and Anderson 1986; Asbjørnsen and Hugdahl 1995). Moreover, utilizing functional magnetic resonance imaging it was found that focusing of attention on vowel sounds and spoken words (as compared with passive listening and attended pseudowords) increased temporal lobe activation with a leftward asymmetry (Hugdahl et al. 2003). Going one step further, Tallus et al. (2007) demonstrated that top-down attention as well as bottom-up sound intensity interactively influenced response laterality for speech sounds on the behavioral level.

However, at this point it appears unsettled whether, and if so how, hemispheric asymmetries during auditory nonspeech signal processing could be interactively influenced by both the signal-to-noise ratio and the attentional state of the listener. Here, investigating this issue by means of MEG, a tonal signal was delivered monaurally and simultaneously with binaural band-eliminated noises (BENs) containing either “wide” or “narrow” stopbands, thereby varying signal-to-noise ratio (or task difficulty, respectively). Furthermore, listeners were supposed to direct their attention either to the signal or away from the signal, thereby altering the attentional state (or the importance of correct signal processing, respectively). Based on aforementioned findings, results were expected to bear evidence for the 1) global left hemisphere dominance in noisy environments, and potentially 2) increased right auditory cortex activity under demanding and relevant conditions.

Materials and Methods

Subjects

Twenty-2 right-handed (assessed via the “Edinburgh Handedness Inventory”; Oldfield 1971) subjects (12 males; mean age 25.48 years [SD 1.73]) participated in the study. All participants had normal hearing in the frequency range from 250 to 8000 Hz, as tested by clinical audiometry. After having been informed about the nature of the study, subjects gave written consent. The study protocol has been approved by the Ethics Committee of the University Hospital Muenster, and the study was performed in accordance with the Declaration of Helsinki.

Auditory and Visual Stimuli

Five different auditory stimuli were used in order to evoke auditory magnetic fields. Four of these stimuli were spectrally complex overlays of BEN and tonal test stimulus (TS); the fifth stimulus was TS in isolation (No-BEN). BENs served as simultaneous maskers, TS was the test signal. TS was presented randomly monaurally to either the left or the right ear only. In contrast, BENs were presented strictly binaurally. During data analyses, No-BEN was used for source localization estimation purposes only, and thus not included in statistical analyses, because its frequency spectrum differs qualitatively from BEN-TS overlays. BENs were prepared by digitally eliminating (Gaussian filter) frequency bands of different widths (either 330 Hz [BEN330], 160 Hz [BEN160], 80 Hz [BEN80], or 40 Hz [BEN40], respectively) from 8000 Hz low-pass filtered white noise (sampling rate 48000 Hz). Resulting notches were spectrally located symmetrically around 1000 Hz (cf. Fig. 1). BEN330 and BEN160 can be considered as rather “wide” (i.e., outreaching the critical bandwidth of the auditory filter centered on 1000 Hz), whereas BEN80 and BEN40 are rather “narrow” (i.e., within the critical bandwidth of the auditory filter). TS was a 40-Hz amplitude-modulated pure tone (modulation depth 100%), with a carrier frequency of 1000 Hz. BENs had a duration of 3.0 s, TS had a duration of 700 ms. The sound onset asynchrony between subsequent TSs was fixed to 3.0 s; TS onset was 2.0 s delayed with respect to BEN onset (cf. Fig. 2).

Figure 1.

Amplitude spectra of the different BENs in the frequency range from 500 to 2000 Hz. BENs contain spectral notches of either 40 Hz (BEN40), 80 Hz (BEN80), 160 Hz (BEN160), or 330 Hz (BEN 330) width. Notches are spectrally located symmetrically around 1000 Hz.

Figure 1.

Amplitude spectra of the different BENs in the frequency range from 500 to 2000 Hz. BENs contain spectral notches of either 40 Hz (BEN40), 80 Hz (BEN80), 160 Hz (BEN160), or 330 Hz (BEN 330) width. Notches are spectrally located symmetrically around 1000 Hz.

Figure 2.

Concept and timing of the auditory stimulation. BENs are presented binaurally and in a random sequence. The TS is presented monaurally randomly either to the left or the right ear. In 90% of trials, TS is continuous (standard TS). In 10% of trials, TS contains a temporal gap (target TS) at variable positions (see text for details). BENs have duration of 3.0 s, TS has duration of 0.7 s. TS onset is 2.0 s delayed compared with BEN onset. The sound onset asynchrony between 2 subsequent TSs is fixed to 3.0 s.

Figure 2.

Concept and timing of the auditory stimulation. BENs are presented binaurally and in a random sequence. The TS is presented monaurally randomly either to the left or the right ear. In 90% of trials, TS is continuous (standard TS). In 10% of trials, TS contains a temporal gap (target TS) at variable positions (see text for details). BENs have duration of 3.0 s, TS has duration of 0.7 s. TS onset is 2.0 s delayed compared with BEN onset. The sound onset asynchrony between 2 subsequent TSs is fixed to 3.0 s.

In addition to the auditory standard TS introduced above, so-called “auditory target TSs,” which had to be detected by the subjects, were used in order to effectively implement selective attention to the left or the right ear, respectively. Auditory target TSs contained one of 6 different temporal gaps of 50-ms duration, which were variable in their positions (beginning at either 100, 200, 300, 350, 450, or 550 ms after TS onset, respectively) (cf. Fig. 2).

Visual stimuli were random configurations of one up to 9 “X”s, which could appear simultaneously in 9 predefined locations on the screen. One specific “X” served as fixation cross and was permanently visible in the center of the screen. The visual stimuli were solely used to distract attention from the auditory modality in a certain experimental condition (cf. Fig. 3). Visually evoked responses were not of interest and thus not analyzed.

Figure 3.

Concept of the visual stimulation. Stimuli are random configurations of one to 9 “X”s appearing simultaneously in predefined locations on the screen. Visual target configurations (e.g., G, H, and I) contain exactly one small square constituted of 4 “X”s appearing in one of the 4 corners of the screen (as indicated by the white dotted lines). Nontarget configurations (e.g., A to F) did either not contain a small square (e.g. D, E, and F) or more than one small square (e.g. A, B, and C).

Figure 3.

Concept of the visual stimulation. Stimuli are random configurations of one to 9 “X”s appearing simultaneously in predefined locations on the screen. Visual target configurations (e.g., G, H, and I) contain exactly one small square constituted of 4 “X”s appearing in one of the 4 corners of the screen (as indicated by the white dotted lines). Nontarget configurations (e.g., A to F) did either not contain a small square (e.g. D, E, and F) or more than one small square (e.g. A, B, and C).

In order to effectively implement selective attention to the screen, so-called ‘visual target configurations, which had to be detected by the subjects, were used in addition to the standard configurations. Visual target configurations contained exactly one small square constituted of 4 “X”s appearing in one of the 4 corners of the screen (cf. Fig. 3).

Manipulation of Attention

Before each experimental block, subjects were instructed to focus attention either on their left ear, right ear, or the screen only, and to ignore the other channels. Throughout the whole experiment, irrespective of the location of the attention focus, auditory and visual stimuli were presented simultaneously, but were uncorrelated.

The following 3 attention conditions were of interest. 1) Relevant auditory attention (REL_AUD): the auditory TS appeared in the attended ear. 2) Irrelevant auditory attention (IRR_AUD): the auditory TS appeared in the nonattended ear. 3) Irrelevant visual attention (i.e., distraction of attention from the auditory modality) (IRR_VIS): the auditory TS appeared in the left or the right ear (Table 1). The IRR_VIS condition served as baseline condition. Even though the state of attention differed between REL_AUD, IRR_AUD, and IRR_VIS, the auditory and visual bottom-up inputs were identical between these conditions.

Table 1

Overview regarding experimental conditions

Condition Attention focus/auditory signal BEN 
Relevant auditory attention Left ear/left ear and right ear/right ear Both ears 
Irrelevant auditory attention Left ear/right ear and right ear/left ear Both ears 
Irrelevant visual attention Screen/left ear and screen/right ear Both ears 
Condition Attention focus/auditory signal BEN 
Relevant auditory attention Left ear/left ear and right ear/right ear Both ears 
Irrelevant auditory attention Left ear/right ear and right ear/left ear Both ears 
Irrelevant visual attention Screen/left ear and screen/right ear Both ears 

Note: In the relevant auditory attention condition, the spatial locations of attention focus and signal are identical (either left or right ear). In the irrelevant auditory attention condition, attention is focused on either the left or the right ear, and the signal appears in the nonattended ear. In the irrelevant visual attention condition, attention is focused on the screen; the signal appears either in the left or the right ear. BENs are presented strictly binaurally.

In addition to the mere instruction to focus attention on a certain channel and ignore other channels, auditory and visual target detection tasks were applied in order to effectively implement selective attention. Subjects were instructed to press a response button as quickly as possible with their right index finger when detecting targets (10% probability) while ignoring standard stimuli (90% probability). In case of auditory focused attention (i.e., left-ear or right-ear attention), “auditory target TSs” had to be detected in the attended ear; in case of attention focused on the screen, specific “visual target configurations” had to be detected.

Design

As dependent variable, N1m source strength was measured, whereby BEN type (Wide [BEN330, BEN160], Narrow [BEN80, BEN40]), attention (REL_AUD, IRR_AUD, IRR_VIS), and hemisphere (Left, Right) served as factors. For each subject, 2 MEG recording sessions on different days were performed. In each session, 1200 auditory TSs were presented in total. BEN type was delivered randomly within subject and session. Attention was manipulated in blocks within subject and session (9 blocks per session; 3 blocks per attention condition per session). The block order was pseudorandomized and counterbalanced between subjects. One block lasted approximately 6 min, and the total duration of one session was approximately 1 hour.

Procedure, Data Acquisition, and Data Analysis

In the beginning of each MEG recording session, the hearing threshold for TS was determined for each subject and each ear individually. TS had loudness of 35 dB above individual threshold, whereas the power of BENs was 15 dB larger than TS power. BENs were always presented binaurally, TS was presented monaurally, randomly either to the left or the right ear (Table 1). Auditory evoked fields were recorded with a helmet-shaped 275-channel whole head neuro-gradiometer (OMEGA, CTF Systems, Inc. Port Coquitlam, Canada) in a silent magnetically shielded room. Participants were comfortably seated upright. Head position was fixed with cotton pads, and subjects were instructed not to move. Head position and compliance were monitored continuously by video camera during the MEG recordings. Auditory evoked fields were digitally sampled at a 600-Hz rate and 150-Hz low-pass filtered during acquisition. Artifact epochs containing field changes larger than 3 picotesla were rejected. Data epochs elicited by standard TS, including a 100-ms pre-TS-onset baseline interval and a 500-ms post-TS-onset interval, were averaged selectively for each session, BEN type, and attentional condition. Source locations and orientations were determined in a head-based Cartesian coordinate system, with the origin at the midpoint of the medio-lateral axis (y-axis) joining the center points of the entrances to the ear canals (positive toward the left ear). The posterior–anterior axis (x-axis) ran between nasion and origin, the inferior–superior axis (z-axis) ran through the origin perpendicularly to the xy-plane.

The auditory evoked N1m response is generated in a relatively restricted cortical area (posterior temporal plane and lateral aspects of Heschl's gyrus; Pantev et al. 1995; Eggermont and Ponton 2002). Therefore, N1m source locations and orientations were estimated for each subject and session individually, by means of single equivalent current dipoles (one per hemisphere) based on the grand-averaged No-BEN condition, using a spherical head model. A previous MEG study (Sams and Salmelin 1994) had shown that estimated location and orientation of the N1m component elicited by a tonal stimulus are unaffected by simultaneously presented BENs. The averaged evoked magnetic fields were 30-Hz low-pass filtered, followed by a baseline correction relative to the 100-ms prestimulus interval. The time point of maximal global field power, measured as root-mean square across all sensors around 100 ms after stimulus onset, was identified. Afterward, the 10 ms time window prior to the peak was used for fixed source estimation. The goodness-of-fit of the underlying dipolar source model was above 90% for all included subjects and sessions (mean 96.33% [SD 1.74]), legitimating the utilization of the single dipole source model. The estimated source for each hemisphere of each subject and each session was fixed in location and orientation, and source strength was calculated for all time points for each BEN type (BEN330, BEN160, BEN80, and BEN40) and for each attentional state (REL_AUD, IRR_AUD, and IRR_VIS), respectively.

Results

Clearly identifiable auditory evoked fields were observed from 17 out of the 22 subjects measured. For 5 subjects, it was difficult to estimate reliable dipolar sources (goodness-of-fit ≥ 90%). Therefore, data from these subjects were not included into further analyses. As in our previous study (Okamoto, Stracke, Wolters, et al. 2007), the subjects in the present study could not identify the different BENs.

Contour maps and estimated source locations and orientations of the N1m overlaid onto the MRI of one representative subject for the grand-averaged No-BEN-condition are displayed in Figure 4. Clear dipolar patterns over both hemispheres were observed. The grand-averaged N1m cortical source waveforms across 17 subjects (time range −100 to +300 ms) are displayed in Figure 5, demonstrating the clear N1m response peaking at around 100 ms after TS onset for the BEN330 condition. N1m responses for narrower BEN-TS combinations are to a certain degree delayed and smaller in amplitude.

Figure 4.

Contour maps (A, B) and dipole source locations and orientations (C, D) for the grand-averaged No-BEN condition of one session of one representative subject based on the boundary element head model created on basis of the structural individual magnetic resonance image. Black lines indicate inward flow of magnetic fields; gray lines denote outward flow (A, B). The black dipoles represent left- (C) respectively right-hemispheric (D) sources.

Figure 4.

Contour maps (A, B) and dipole source locations and orientations (C, D) for the grand-averaged No-BEN condition of one session of one representative subject based on the boundary element head model created on basis of the structural individual magnetic resonance image. Black lines indicate inward flow of magnetic fields; gray lines denote outward flow (A, B). The black dipoles represent left- (C) respectively right-hemispheric (D) sources.

Figure 5.

Source waveforms grand-averaged across all included subjects. The graphs show clear N1m responses for all conditions. The left panel (A, C) displays left-hemispheric responses; the right panel (B, D) displays right-hemispheric responses. The top row (A, B) shows the relevant auditory attention condition, and the bottom row (C, D) shows the irrelevant auditory attention condition.

Figure 5.

Source waveforms grand-averaged across all included subjects. The graphs show clear N1m responses for all conditions. The left panel (A, C) displays left-hemispheric responses; the right panel (B, D) displays right-hemispheric responses. The top row (A, B) shows the relevant auditory attention condition, and the bottom row (C, D) shows the irrelevant auditory attention condition.

Paired t-tests applied to the dipole source locations of the N1m response revealed significant hemispheric differences in posterior–anterior (x-axis: t33 = −3.588, P = 0.001) and inferior–superior (z-axis: t33 = 4.174, P < 0.0001) dimensions. Hence, estimated locations of the measured neural activities slightly differed between hemispheres. These differences most likely reflect anatomical differences between hemispheres (Morosan et al. 2001; Rademacher et al. 2001).

Planned contrasts were calculated. Notably, the contrasts [(REL_AUDLeft, Wide − IRR_AUDLeft, Wide) − (REL_AUDLeft, Narrow − IRR_AUDLeft, Narrow)] ((7) in Table 2) and [(REL_AUDRight, Wide − IRR_AUDRight, Wide) − (REL_AUDRight, Narrow − IRR_AUDRight, Narrow)] ((13) in Table 2) were calculated to test the interactions between attention (REL_AUD vs. IRR_AUD) and BEN type (Wide vs. Narrow) for the left and right hemispheres separately. All performed contrasts including corresponding F and P values are shown in Table 2.

Table 2

Overview regarding calculated planned contrasts on N1m source strength and N1m latency

Contrast Source Strength
 
Latency
 
 F P F P 
(1) [LeftRight25.3 0.001* 7.9 0.013* 
(2) [REL_AUDLeft – IRR_AUDLeft0.5 0.48 3.1 0.095 
(3) [REL_AUDLeft – IRR_VISLeft5.9 0.027* 23.3 0.001* 
(4) [IRR_AUDLeft – IRR_VISLeft4.6 0.048* 4.6 0.048* 
(5) [(REL_AUDLeft – IRR_AUDLeft) – IRR_VISLeft] 5.6 0.031* 20.1 0.001* 
(6) [WIDELeft – NARROWLeft] 40.3 0.001* 123.9 0.001* 
(7) [(REL_AUDLeft, Wide – IRR_AUDLeft, Wide) – (REL_AUDLeft, Narrow – IRR_AUDLeft, Narrow)0.8 0.39 0.15 0.703 
(8) [REL_AUDRight – IRR_AUDRight9.2 0.008* 1.4 0.261 
(9) [REL_AUDRight – IRR_VISRight12.3 0.003* 22.5 0.001* 
(10) [IRR_AUDRight – IRR_VISRight2.5 0.137 12.9 0.002* 
(11) [(REL_AUDRight – IRR_AUDRight) – IRR_VISRight] 6.9 0.018* 18.5 0.001* 
(12) [WIDERight – NARROWRight] 24.7 0.001* 61.8 0.001* 
(13) [(REL_AUDRight, Wide – IRR_AUDRight, Wide) – (REL_AUDRight, Narrow – IRR_AUDRight, Narrow)6.8 0.019* 0.992 
Contrast Source Strength
 
Latency
 
 F P F P 
(1) [LeftRight25.3 0.001* 7.9 0.013* 
(2) [REL_AUDLeft – IRR_AUDLeft0.5 0.48 3.1 0.095 
(3) [REL_AUDLeft – IRR_VISLeft5.9 0.027* 23.3 0.001* 
(4) [IRR_AUDLeft – IRR_VISLeft4.6 0.048* 4.6 0.048* 
(5) [(REL_AUDLeft – IRR_AUDLeft) – IRR_VISLeft] 5.6 0.031* 20.1 0.001* 
(6) [WIDELeft – NARROWLeft] 40.3 0.001* 123.9 0.001* 
(7) [(REL_AUDLeft, Wide – IRR_AUDLeft, Wide) – (REL_AUDLeft, Narrow – IRR_AUDLeft, Narrow)0.8 0.39 0.15 0.703 
(8) [REL_AUDRight – IRR_AUDRight9.2 0.008* 1.4 0.261 
(9) [REL_AUDRight – IRR_VISRight12.3 0.003* 22.5 0.001* 
(10) [IRR_AUDRight – IRR_VISRight2.5 0.137 12.9 0.002* 
(11) [(REL_AUDRight – IRR_AUDRight) – IRR_VISRight] 6.9 0.018* 18.5 0.001* 
(12) [WIDERight – NARROWRight] 24.7 0.001* 61.8 0.001* 
(13) [(REL_AUDRight, Wide – IRR_AUDRight, Wide) – (REL_AUDRight, Narrow – IRR_AUDRight, Narrow)6.8 0.019* 0.992 

Note: F and P values (*P < 0.05) are provided. Left/Right = left/right hemispheres, REL_AUD/IRR_AUD/IRR_VIS = relevant auditory attention/irrelevant auditory attention/irrelevant visual attention conditions, WIDE, Wide/NARROW, Narrow = wide/narrow BENs.

Overall, N1m source strength was significantly larger in the left compared with the right auditory cortex (Table 2, contrast (1)). In both hemispheres, N1m source strength was significantly larger in case of auditory focused (REL_AUD, IRR_AUD) compared with visually focused attention (IRR_VIS) (Table 2, contrasts (5), (11)). Moreover, N1m source strength was significantly larger for wide (160 Hz, 330 Hz) than for narrow (40, 80 Hz) BENs in both hemispheres (Table 2, contrasts (6), (12)). In the left hemisphere, there was neither a significant N1m source strength difference between REL_AUD and IRR_AUD (Table 2, contrast (2)), nor a significant interaction between attention (REL_AUD vs. IRR_AUD) and BEN type (Wide vs. Narrow) (Table 2, contrast (7)). In contrast, the main effect (Table 2, contrast (8)) and the interaction were both present for the right hemisphere, where N1m source strength was significantly larger during REL_AUD than IRR_AUD, but only for narrow BENs (Table 2, contrast (13)) (cf. Fig. 6).

Figure 6.

Interaction plots illustrating N1m source strength grand-averaged across all included subjects with respect to relevant auditory attention and irrelevant auditory attention conditions. The left panel (A) displays left-hemispheric responses, the right panel (B) displays right-hemispheric responses. The circles (A, B) denote the relevant auditory attention condition (i.e., attention focused on the stimulated ear), and the triangles (A, B) denote the irrelevant auditory attention condition (i.e., attention focused on the not stimulated ear). Error bars indicate the 95% confidence limits of the mean.

Figure 6.

Interaction plots illustrating N1m source strength grand-averaged across all included subjects with respect to relevant auditory attention and irrelevant auditory attention conditions. The left panel (A) displays left-hemispheric responses, the right panel (B) displays right-hemispheric responses. The circles (A, B) denote the relevant auditory attention condition (i.e., attention focused on the stimulated ear), and the triangles (A, B) denote the irrelevant auditory attention condition (i.e., attention focused on the not stimulated ear). Error bars indicate the 95% confidence limits of the mean.

So far, both left- and right-ear stimulation and left- and right-ear attention had been pooled together in the analysis on attentional relevance effects, which were of particular interest in this study. However, recent studies (Bryden et al. 1983; Hugdahl et al. 2000; Saetrevik and Hugdahl 2007) indicated that stimulated ear and attended ear should be separated as factors influencing auditory processing. We thus evaluated whether N1m source strength varied systematically as function of stimulation side (Left, Right) or attention side (Left, Right). Therefore, source waveforms were grand-averaged across BEN types (excluding No-BEN), hemispheres, and the 17 included subjects. Planned contrasts showed that N1m source strength did not differ between left- or right-ear attention (F1, 33 = 0.302, P = 0.586). However, N1m source strength was significantly larger for right-ear than for left-ear stimulation (F1, 33 = 8.305, P = 0.007).

Motivated by the latter results, we performed an additional behavioral test on 17 age-matched subjects, to check for potential differences between stimulation sides. This test was conducted in the MEG room, and the auditory stimuli used were identical to those used during the MEG measurement. BEN type (BEN330, BEN160, BEN80, BEN40) and stimulation side (Left, Right) served as independent factors, whereas response accuracy (hit rate) and speed (reaction time) served as dependent variables. As in the MEG measurement, BENs were presented binaurally, whereas TS was presented randomly to the left or the right ear. As opposed to the MEG measurement, TS was target (or nontarget) in 50% of trials. Subjects were instructed to direct attention (block-wise) to their left or to their right ear only, and to press a response button as quickly as possible, with their right index finger, when detecting a target TS in the attended ear. The repeated-measures ANOVA calculated on hit rates revealed a significant main effect of BEN type (F3, 48 = 59.278, P < 0.001), indicating that response accuracy became worse with stopband becoming narrower, as had been shown earlier (Okamoto, Stracke, Wolters, et al. 2007). Crucially, there was no main effect of stimulation side (F1, 16 = 0.783, P = 0.389) and no interaction between BEN type and stimulation side (F3, 48 = 0.087, P = 0.967). A similar pattern was found for reaction times: although there was a significant main effect for BEN type (F3, 48 = 35.606, P < 0.001; reaction times became longer with stopband becoming narrower), there was no main effect for (F1, 16 = 1.725, P = 0.208) nor an interaction (F(3, 48) = 0.907, P = 0.444) with stimulation side.

Discussion

The present study yielded 3 major findings. 1) An overall hemispheric difference was observed, with N1m source strength significantly larger in the left than in the right hemisphere. 2) Overall, N1m responses were larger for right-ear than for left-ear stimulation, even though the TS was not a speech signal. 3) Crucially, interactive effects were found for hemisphere, attentional relevance, and BEN type. Whereas N1m source strength was similar for relevant auditory attention and irrelevant auditory attention conditions in the left hemisphere, an interaction between attentional relevance and BEN type was observed in the right hemisphere. Relative to the irrelevant auditory attention condition, N1m source strength was larger for narrow, but not for wide BENs in the relevant auditory attention condition.

The effects observed in the present study partly seem to result from the interactive interplay of bottom-up neural inputs and top-down processes. The characteristic bottom-up input in this experiment was the spectral overlap between masker and signal, which was operationalized as function of the eliminated bandwidth of the different BENs. In case of wide BENs, spectral overlap between BEN and TS was rather small, and therefore the signal (TS) was relatively easy to detect and to process by the auditory system. In contrast, in case of narrow BENs (stopband widths ½ or ¼ critical bandwidth, respectively), spectral overlap was quite large, and it appears very likely that part of the noise passed through the auditory filter centered on the signal (Patterson 1976), resulting in a reduced effective signal-to-noise ratio. Possibly, TS-related neural activity was attenuated by lateral inhibitory processes consequent on the noise (Okamoto, Stracke, Wolter, et al. 2007), and therefore TS was harder to process under these conditions.

The characteristic top-down manipulation concerned the locus of the attentional focus. In the irrelevant visual attention condition, attention resources were allocated to the visual modality. Hence, top-down processing regarding the auditory signal was weak. In the relevant auditory attention condition, attention resources were optimally allocated and focused with respect to modality (auditory) and location (stimulated ear). Finally, in case of the irrelevant auditory attention condition, attention resources were also allocated to the auditory modality, but the focus was not on the stimulated ear (Table 1).

In the present experiment, N1m source strength was overall larger in the left compared with the right auditory cortex (hemispheric asymmetry effect; Fig. 6A vs. B). Moreover, N1m source strength was larger during auditory focused attention than during visually focused attention in both hemispheres (intermodal attentional gain effect). Additionally, in both hemispheres, wide BENs favored larger N1m source strengths than narrow BENs (spectral overlap-dependent masking effect; Fig. 6A,B). With regard to overall hemispheric differences, the results indicate basal left auditory cortex dominance during auditory signal-in-noise processing.

This is further substantiated by the observed right-ear advantage (reflected in the overall increased N1m source strength for right-ear as compared with left-ear stimulation), given that auditory projections are contralaterally dominant in the auditory cortex. Notably, this right-ear advantage could not be confirmed on the behavioral level. However, in the behavior test, stimulation side and attention side necessarily were perfectly confounded, and thus only responses to the attended ear could be analyzed. In the MEG measurement, in contrast, it was possible to decouple effects of attention side and stimulation side by means of selective signal averaging (irrelevant auditory attention condition). Thus, response levels are not fully comparable. Moreover, the mean overall hit rates (left ear: 86.7%; right ear: 85.3%) indicate that the stimuli might not have been hard enough to uncover differences at the behavioral level in the relevant auditory attention condition. However, stimuli were optimized for the MEG measurement (which was of primary interest for this study), and for the sake of comparability stimuli were identical between MEG and behavioral tests.

Crucially, an interactive hemispheric difference was revealed in addition to the overall hemispheric asymmetry. Although the left auditory cortex showed no N1m source strength differences between relevant auditory attention and irrelevant auditory attention conditions, N1m source strengths in the right hemisphere differed significantly as a function of attentional relevance, but only for narrow, not for wide BENs. Thus, for the left hemisphere it did not seem to make a difference whether attention was focused on the correct or a wrong location, as long as attentional resources were allocated to the auditory modality (Fig. 6A). This indicates the “robustness,” in addition to the dominance, of the left auditory cortex during attentive auditory signal-in-noise processing. For the right hemisphere, in contrast, it did not seem to make a difference whether attention was distracted to the visual modality, or wrongly focused within the auditory modality. If, however, precise auditory signal-in-noise processing was clearly required (relevant auditory attention condition: attention optimally allocated and focused) and circumstances were demanding (narrow BENs: large spectral overlap between BEN and TS), neuronal activity in the right auditory cortex was significantly larger, compared with situations where such precise auditory signal-in-noise processing was not required (irrelevant auditory attention and irrelevant visual attention conditions) (Fig. 6B).

The N1m source strength difference between the relevant auditory attention and irrelevant auditory attention conditions, observed in the right hemisphere for narrow BENs, could reflect either a decrement in the irrelevant auditory attention condition, or an increment in the relevant auditory attention condition. However, because the irrelevant visual attention and irrelevant auditory attention conditions in the right hemisphere did not differ, increased neuronal activity in the relevant auditory attention condition seems more likely. This interpretation is supported by close inspection of the graphs in Figure 6, showing a monotonic decrease of the cortical N1m source strength with narrowing BENs, in all conditions except the relevant auditory attention condition in the right hemisphere for BEN40 compared with BEN80.

N1m source strength as measured by MEG depends on the number of activated neurons and on the degree of synchronicity of activity of the involved neurons (Pantev et al. 1998). Therefore, the increased neuronal activity observed in the right hemisphere for narrow BENs during the relevant auditory attention condition might reflect recruitment of additional neurons, improved phase-locking of involved neurons, or most likely, a combination of both.

The novel result pattern observed here may reflect an “interhemispheric support mechanism.” Basically, the left auditory cortex seems to be both dominant and robust during auditory signal-in-noise processing (as indicated by the overall larger neuronal activity and the similarity of the relevant auditory attention and irrelevant auditory attention conditions). If, however, accurate performance under demanding circumstances is requested, the right auditory cortex offers assistance (as indicated by increased neuronal activity in exactly such instances). Previous experimental findings have been interpreted in a similar vein. Relative to a complete silence condition, Shtyrov et al. (1998, 1999) observed increments in both mismatch negativity and P2m amplitude in the right compared with the left hemisphere during the processing of syllables simultaneously masked by white noise. These results point to an increased role of the right hemisphere during speech sound processing in noisy conditions, taken to be a reflection of the consumption of “supplemental” right-hemispheric resources. Furthermore, Liikkanen et al. (2007) observed a right-hemispheric augmentation in N1m source strength during the processing of vowels, which were “degraded” by means of uniform scalar quantization. Results were interpreted in terms of right-hemispheric compensation for poor speech-signal quality. The signal applied in the present study was not a speech sound. Instead, a simple 1000-Hz tone had to be detected and processed during the presence of interfering noise. Nevertheless, the present major finding is completely in line with the results obtained in previous experiments focusing on speech processing (Shtyrov et al. 1998, 1999; Liikkanen et al. 2007). Hence, it appears highly probable that the right-hemispheric supplementation observed here is not unique for the processing of masked or degraded speech, but rather reflects a basic brain mechanism that enables reliable auditory signal-in-noise processing. Thus, the present data probably uncovered a proficiency of the brain that is highly relevant in many day-to-day situations.

The amplitude-modulated tone, which was used as signal in this study, is known to generate the so-called auditory steady-state response (Rees et al. 1986; Hari et al. 1989; Pantev et al. 1993). The auditory steady-state response, as opposed to the N1m, is known to be of primary auditory cortex origin (Pantev et al. 1996; Draganova et al. 2008). Therefore, the present study theoretically has the potential to assess whether the hypothesis of hemispheric specialization applies to primary as well as nonprimary auditory cortex areas. However, similar to the study of Okamoto, Stracke, Wolters, et al. (2007), it proved to be impossible to extract auditory steady-state response source waveforms for the BEN conditions, due to insufficient signal-to-noise ratio. Thus, the conclusions drawn here are confined to nonprimary auditory cortex areas.

Here, the signal (as opposed to the BENs) was presented monaurally, causing a larger response in the contralateral compared with the ipsilateral hemisphere. However, an equal number of trials with left-, respectively, right-ear stimulation contributed to each experimental condition (cf. Table 1), making sure that potential monaural stimulation-related hemispheric asymmetries were eliminated during selective signal averaging. Therefore, monaural signal presentation cannot account for the functional hemispheric asymmetries observed in this experiment. In addition, our interpretation of left-hemispheric dominance in noisy environments is supported by a previous study employing a very similar paradigm, but binaural signals (Okamoto, Stracke, Wolters, et al. 2007).

In the present experiment, N1m latency was overall shorter in the right compared with the left hemisphere. Similar findings had already been observed in previous MEG studies investigating auditory responses evoked by tonal signals (e.g., Gabriel et al. 2004). Moreover, N1m latencies were shorter in case of auditory compared with visually focused attention; this has also been observed earlier (cf. Okamoto, Stracke, Wolters, et al. 2007). Finally, N1m latencies were shorter for wide compared with narrow BEN conditions (cf. Sams and Salmelin 1994).

Previous functional magnetic resonance imaging studies (e.g., Petkov et al. 2004) had also suggested correlates of attention in lateral areas of auditory cortex involved in N1m generation. Due to the poor temporal resolution capabilities of fMRI, fine-scale timing of neural responses cannot be obtained with this method. On the contrary, MEG principally offers the possibility to unravel these attentional effects with high temporal precision, and may therefore be utilized to uncover potential early attention effects (cf. Woldorff et al. 1993). However, the N1m response, contrariwise to other responses arising earlier (e.g., P1m) or later (e.g., P2m), is robust and stable. The present experiment is based on this N1m robustness and stability, and was explicitly designed for N1m evaluation. It was not intended to analyze other auditory evoked responses.

In conclusion, the present results provide evidence for increased neuronal activity in the right auditory cortex during auditory signal-in-noise processing, when signal-to-noise ratio is poor and adequate performance is explicitly relevant. The increased neuronal activity presumably reflects increased involvement of the right auditory cortex in the sense of additional consumption of right-hemispheric resources in particularly demanding conditions. These findings could be interpreted as an indication of “right-hemispheric support” for the (basically dominant and robust) left auditory cortex during difficult and relevant auditory signal-in-noise processing.

Funding

Deutsche Forschungsgemeinschaft (Pa 392/10-2).

We thank Maximilian Bruchmann, Pascal Belin, and Pienie Zwitserlood for helpful discussions on previous versions of this manuscript. We thank Andreas Wollbrink for technical assistance, and Karin Berning, Ute Trompeter, and Hildegard Deitermann for help during data acquisition. Conflict of interest: None declared.

References

Alho
K
Connolly
J
Cheour
M
Lehtokoski
A
Huotilainen
M
Virtanen
J
Aulanko
R
Ilmoniemi
R
Hemispheric lateralization in preattentive processing of speech sounds
Neurosci Lett.
 , 
1998
, vol. 
258
 (pg. 
9
-
12
)
Asbjørnsen
A
Hugdahl
K
Attentional effects in dichotic listening
Brain Lang.
 , 
1995
, vol. 
49
 (pg. 
189
-
201
)
Belin
P
Zatorre
R
Lafaille
P
Ahad
P
Pike
B
Voice-selective areas in human auditory cortex
Nature.
 , 
2000
, vol. 
403
 (pg. 
309
-
312
)
Brancucci
A
Babiloni
C
Babiloni
F
Galderisi
S
Mucci
A
Tecchio
F
Zappasodi
F
Pizzella
V
Romani
G
Rossini
P
Inhibition of auditory cortical responses to ipsilateral stimuli during dichotic listening: evidence from magnetoencephalography
Eur J Neurosci.
 , 
2004
, vol. 
19
 (pg. 
2329
-
2336
)
Bryden
M
Correlates of the dichotic right-ear effect
Cercor
 , 
1988
, vol. 
24
 (pg. 
313
-
319
)
Bryden
M
Munhall
K
Allard
F
Attentional biases and the right-ear effect in dichotic listening
Brain Lang.
 , 
1983
, vol. 
18
 (pg. 
236
-
248
)
Draganova
R
Ross
B
Wollbrink
A
Pantev
C
Cortical steady-state responses to central and peripheral auditory beats
Cereb Cortex.
 , 
2008
, vol. 
18
 (pg. 
1193
-
1200
)
Eggermont
J
Ponton
C
The neurophysiology of auditory perception: from single units to evoked potentials
Audiol Neurootol.
 , 
2002
, vol. 
7
 (pg. 
71
-
99
)
Eulitz
C
Diesch
E
Pantev
C
Hampson
S
Elbert
T
Magnetic and electric brain activity evoked by the processing of tone and vowel stimuli
J Neurosci.
 , 
1995
, vol. 
15
 (pg. 
2748
-
2755
)
Gabriel
D
Veuillet
E
Ragot
R
Schwartz
D
Ducorps
A
Norena
A
Durrant
J
Bonmartin
A
Cotton
F
Collet
L
Effect of stimulus frequency and stimulation site on the N1m response of the human auditory cortex
Hear Res.
 , 
2004
, vol. 
197
 (pg. 
55
-
64
)
Hari
R
Hämäläinen
M
Joutsiniemi
S
Neuromagnetic steady-state responses to auditory stimuli
J Acoust Soc Am.
 , 
1989
, vol. 
86
 (pg. 
1033
-
1039
)
Hari
R
Mäkelä
J
Modification of neuromagnetic responses of the human auditory cortex by masking sounds
Exp Brain Res.
 , 
1988
, vol. 
71
 (pg. 
87
-
92
)
Hugdahl
K
Andersson
L
A dichotic listening study of differences in cerebral organization in dextral and sinistral subjects
Cercor
 , 
1984
, vol. 
20
 (pg. 
135
-
141
)
Hugdahl
K
Law
I
Kyllingsbaek
S
Brønnick
K
Gade
A
Paulson
O
Effects of attention on dichotic listening: an 15O-PET study
Hum Brain Mapp.
 , 
2000
, vol. 
10
 (pg. 
87
-
97
)
Hugdahl
K
Thomsen
T
Ersland
L
Rimol
L
Niemi
J
The effects of attention on speech perception: an fMRI study
Brain Lang.
 , 
2003
, vol. 
85
 (pg. 
37
-
48
)
Jamison
H
Watkins
K
Bishop
D
Matthews
P
Hemispheric specialization for processing auditory nonspeech stimuli
Cereb Cortex.
 , 
2006
, vol. 
16
 (pg. 
1266
-
1275
)
Hugdahl
K
Anderson
L
The “forced-attention paradigm” in dichotic listening to CV-syllables: a comparison between adults and children
Cortex.
 , 
1986
, vol. 
22
 (pg. 
417
-
432
)
Liikkanen
L
Tiitinen
H
Alku
P
Leino
S
Yrttiaho
S
May
P
The right-hemispheric auditory cortex in humans is sensitive to degraded speech sounds
Neuroreport.
 , 
2007
, vol. 
18
 (pg. 
601
-
605
)
Morita
T
Fujiki
N
Nagamine
T
Hiraumi
H
Naito
Y
Shibasaki
H
Ito
J
Effects of continuous masking noise on tone-evoked magnetic fields in humans
Brain Res.
 , 
2006
, vol. 
1087
 (pg. 
151
-
158
)
Morosan
P
Rademacher
J
Schleicher
A
Amunts
K
Schormann
T
Zilles
K
Human primary auditory cortex: cytoarchitectonic subdivisions and mapping into a spatial reference system
Neuroimage.
 , 
2001
, vol. 
13
 (pg. 
684
-
701
)
Okamoto
H
Stracke
H
Ross
B
Kakigi
R
Pantev
C
Left hemispheric dominance during auditory processing in a noisy environment
BMC Biol.
 , 
2007
, vol. 
5
 pg. 
52
 
Okamoto
H
Stracke
H
Wolters
C
Schmael
F
Pantev
C
Attention improves population-level frequency tuning in human auditory cortex
J Neurosci.
 , 
2007
, vol. 
27
 (pg. 
10383
-
10390
)
Oldfield
R
The assessment and analysis of handedness: the Edinburgh inventory
Neuropsychologia.
 , 
1971
, vol. 
9
 (pg. 
97
-
113
)
Pantev
C
Bertrand
O
Eulitz
C
Verkindt
C
Hampson
S
Schuierer
G
Elbert
T
Specific tonotopic organizations of different areas of the human auditory cortex revealed by simultaneous magnetic and electric recordings
Electroencephalogr Clin Neurophysiol.
 , 
1995
, vol. 
94
 (pg. 
26
-
40
)
Pantev
C
Elbert
T
Makeig
S
Hampson
S
Eulitz
C
Hoke
M
Relationship of transient and steady-state auditory evoked fields
Electroencephalogr Clin Neurophysiol.
 , 
1993
, vol. 
88
 (pg. 
389
-
396
)
Pantev
C
Oostenveld
R
Engelien
A
Ross
B
Roberts
L
Hoke
M
Increased auditory cortical representation in musicians
Nature.
 , 
1998
, vol. 
392
 (pg. 
811
-
814
)
Pantev
C
Roberts
L
Elbert
T
Ross
B
Wienbruch
C
Tonotopic organization of the sources of human auditory steady-state responses
Hear Res.
 , 
1996
, vol. 
101
 (pg. 
62
-
74
)
Patterson
R
Auditory filter shapes derived with noise stimuli
J Acoust Soc Am.
 , 
1976
, vol. 
59
 (pg. 
640
-
654
)
Petkov
C
Kang
X
Alho
K
Bertrand
O
Yund
E
Woods
D
Attentional modulation of human auditory cortex
Nat Neurosci.
 , 
2004
, vol. 
7
 (pg. 
658
-
663
)
Rademacher
J
Morosan
P
Schormann
T
Schleicher
A
Werner
C
Freund
H
Zilles
K
Probabilistic mapping and volume measurement of human primary auditory cortex
Neuroimage.
 , 
2001
, vol. 
13
 (pg. 
669
-
683
)
Rees
A
Green
G
Kay
R
Steady-state evoked responses to sinusoidally amplitude-modulated sounds recorded in man
Hear Res.
 , 
1986
, vol. 
23
 (pg. 
123
-
133
)
Saetrevik
B
Hugdahl
K
Endogenous and exogenous control of attention in dichotic listening
Neuropsychology.
 , 
2007
, vol. 
21
 (pg. 
285
-
290
)
Sams
M
Salmelin
R
Evidence of sharp frequency tuning in the human auditory cortex
Hear Res.
 , 
1994
, vol. 
75
 (pg. 
67
-
74
)
Shtyrov
Y
Kujala
T
Ahveninen
J
Tervaniemi
M
Alku
P
Ilmoniemi
R
Näätänen
R
Background acoustic noise and the hemispheric lateralization of speech processing in the human brain: magnetic mismatch negativity study
Neurosci Lett.
 , 
1998
, vol. 
251
 (pg. 
141
-
144
)
Shtyrov
Y
Kujala
T
Ilmoniemi
R
Näätänen
R
Noise affects speech-signal processing differently in the cerebral hemispheres
Neuroreport.
 , 
1999
, vol. 
10
 (pg. 
2189
-
2192
)
Szymanski
M
Perry
D
Gage
N
Rowley
H
Walker
J
Berger
M
Roberts
T
Magnetic source imaging of late evoked field responses to vowels: toward an assessment of hemispheric dominance for language
J Neurosurg.
 , 
2001
, vol. 
94
 (pg. 
445
-
453
)
Tallus
J
Hugdahl
K
Alho
K
Medvedev
S
Hämäläinen
H
Interaural intensity difference and ear advantage in listening to dichotic consonant-vowel syllable pairs
Brain Res.
 , 
2007
, vol. 
1185
 (pg. 
195
-
200
)
Woldorff
M
Gallen
C
Hampson
S
Hillyard
S
Pantev
C
Sobel
D
Bloom
F
Modulation of early sensory processing in human auditory cortex during auditory selective attention
Proc Natl Acad Sci USA.
 , 
1993
, vol. 
90
 (pg. 
8722
-
8726
)
Zatorre
R
Belin
P
Spectral and temporal processing in human auditory cortex
Cereb Cortex.
 , 
2001
, vol. 
11
 (pg. 
946
-
953
)
Zatorre
R
Belin
P
Penhune
V
Structure and function of auditory cortex: music and speech
Trends Cogn Sci.
 , 
2002
, vol. 
6
 (pg. 
37
-
46
)
Zwicker
E
Fastl
H
Psychoacoustics. Facts and models
 , 
2007
Berlin Heidelberg
Springer
Zwislocki
J
Buining
E
Glantz
J
Frequency distribution of central masking
J Acoust Soc Am.
 , 
1968
, vol. 
43
 (pg. 
1267
-
1271
)

Author notes

1
These authors contributed equally.