Gender differences in pre-attentive change detection for visual …or.nsfc.gov.cn/bitstream/00001903-5/235345/1/... · 2016-11-18 · larger aMMN in response to changes in intensity

Clinical Neurophysiology xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Clinical Neurophysiology

journal homepage: www.elsevier .com/locate /c l inph

Gender differences in pre-attentive change detection for visual but notauditory stimuli

http://dx.doi.org/10.1016/j.clinph.2015.05.0131388-2457/� 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

⇑ Corresponding authors at: No. 157, Baojian Road, Nangang District, Harbin150081, China. Tel.: +86 451 87502859 (Y. Yang); Bld 40-1-605, Yuzhong Xili,Xicheng District, Beijing 100029, China. Tel: +86 10 62383378 (L. Zhao).

E-mail addresses: [email protected] (Y. Yang), [email protected](L. Zhao).

1 These authors equally contributed to this work.

Xiuxian Yang a,1, Yunmiao Yu a,1, Lu Chen b, Hailian Sun c, Zhengxue Qiao a, Xiaohui Qiu a, Congpei Zhang d,Lin Wang a, Xiongzhao Zhu e, Jincai He f, Lun Zhao g,⇑, Yanjie Yang a,⇑a Department of Medical Psychology, Public Health Institute of Harbin Medical University, Harbin, Chinab Peking Union Medical College Hospital, Bingjing, Chinac The First Affiliated Hospital of Harbin Medical University, Heilongjiang, Chinad The First Special Hospital of Harbin, Harbin, Chinae The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, Chinaf Medical Psychological Institute, Second Xiangya Hospital, Central South University, Changsha, Hunan, Chinag Center for Visual Art & Brain Cognition, Beijing Shengkun YanLun Technology Co. Ltd., Beijing, China

a r t i c l e i n f o h i g h l i g h t s

Article history:Accepted 12 May 2015Available online xxxx

Keywords:GenderPre-attentive processingAuditory MMNVisual MMN

� We compared pre-attentive processing in males and females across visual and auditory modalities.� No significant gender differences were observed for auditory MMN (aMMN) amplitude, but visual

MMN (vMMN) amplitude was higher in males than females.� Increment duration vMMN could be a good indicator for pre-attentive processing.

a b s t r a c t

Objective: Despite ongoing debate about gender differences in pre-attention processes, little is knownabout gender effects on change detection for auditory and visual stimuli. We explored gender differencesin change detection while processing duration information in auditory and visual modalities.Method: We investigated pre-attentive processing of duration information using a deviant-standardreverse oddball paradigm (50 ms/150 ms) for auditory and visual mismatch negativity (aMMN andvMMN) in males and females (n = 21/group).Result: In the auditory modality, decrement and increment aMMN were observed at 150–250 ms afterthe stimulus onset, and there was no significant gender effect on MMN amplitudes in temporal orfronto-central areas. In contrast, in the visual modality, only increment vMMN was observed at180–260 ms after the onset of stimulus, and it was higher in males than in females.Conclusion: No gender effect was found in change detection for auditory stimuli, but change detectionwas facilitated for visual stimuli in males.Significance: Gender effects should be considered in clinical studies of pre-attention for visual stimuli.� 2015 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights

reserved.

1. Introduction

Pre-attentive detection of change in the environment is funda-mental for adapting to a rapidly changing environment and

ensuring survival (Nagy et al., 2003). Because mismatch negativity(MMN) can be generated under non-attentional conditions, it hasbeen considered to be an index of pre-attentive sensory memory.Relative to the memory template formed in response to standardauditory stimuli, MMN generated by deviant stimuli is a negativeevent-related potential (ERP) component that peaks at 100–250 ms and may have altered duration, frequency, and/or morecomplex properties (e.g., emotional words, syntax) (Naatanen,2000; Naatanen et al., 2007). Auditory MMN (aMMN) has beenwidely studied to elucidate the mechanisms and cortical networks

2 X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx

underlying pre-attentive processing. Indeed, aMMNs include twosub-components: (1) the supra-temporal subcomponent, gener-ated from the bilateral supra-temporal area, which is related todetection of pre-perceptual change, and (2) the frontal subcompo-nent, produced predominantly from the right frontal area, which isassociated with involuntary attention switch caused by inputinformation changes (Alho, 1995; Escera et al., 1998; Naatanenet al., 2007; Rinne et al., 2000).

In addition to the auditory modality, MMN has been observed inthe visual modality. Recent studies have demonstrated visualMMN (vMMN) for color (Czigler et al., 2004), motion direction(Pazo-Alvarez et al., 2004), orientation (Astikainen et al., 2008),spatial frequency (Kenemans et al., 2003), shape (Tales et al.,1999), duration (Chen et al., 2010; Qiu et al., 2011), and facialexpressions (Chang et al., 2011; Zhao and Li, 2006). vMMN is anegative ERP component measured at the temporo-occipitalelectrodes with variable latency between 150 and 350 ms afterstimulus onset. vMMN represents memory-based detection ofdeviant visual stimuli. Importantly, a recent study by Berti et al.further supported that vMMN reflected a pre-attentive process ofvisual deviance detection (Berti, 2011). Moreover, two recentreviews provided indirect evidence for the pre-attentive natureof vMMN (Kimura, 2012; Winkler and Czigler, 2012).

Some evidence suggests that gender differences in incidentallearning and visual recognition memory could be due to greaterunconscious processing of environmental stimuli in females(McGivern et al., 1998), implying that pre-attentive processingmight explain gender differences in cognition. To test this hypoth-esis, several studies investigated gender effects on aMMN withinconsistent results (Aaltonen et al., 1994; Barrett and Fulfs,1998; Hansenne et al., 2003; Ikezawa et al., 2008; Kasai et al.,2002; Matsubayashi et al., 2008; Nagy et al., 2003; Schirmeret al., 2005). For example, compared to males, females exhibitedlarger aMMN in response to changes in intensity of pure tonalstimuli and vocal emotional expressions (Barrett and Fulfs, 1998;Ikezawa et al., 2008; Schirmer et al., 2005), and this effect wasstronger in the right hemisphere (Ikezawa et al., 2008). In contrast,one magneto-encephalography (MEG) study showed that maleshad stronger aMMN in the left hemisphere than females(Matsubayashi et al., 2008). Moreover, some studies reported nogender effects on aMMN amplitude in response to tonal, phonetic,or frequency changes (Aaltonen et al., 1994; Hansenne et al., 2003;Kasai et al., 2002). A possible explanation for these diverse resultscould be methodological variance among experiments. For exam-ple, dichotic listening tasks do not allow control of inattention pur-ity, which is critical to measure MMN. Furthermore, using thetraditional oddball paradigm to obtain MMN can create confounds.In this paradigm, aMMNs are derived by subtracting the ERP wave-forms elicited by standard stimuli from those of deviant stimuli.Because standard stimuli are presented more frequently, their neu-ronal processing likely has greater refractory effect than deviantstimuli. Thus, MMN could confound the detection of differencesin low-level processing between deviant and standard stimuli. Inaddition, when deviant and standard stimuli in the same oddballblock are compared directly, different physical stimuli can elicitoffset responses. This may also lead to errors in estimating thepre-attentive memory comparison-based MMN (Jacobsen andSchroger, 2003). Consequently, our study controlled inattentionpurity and refractory effects to measure MMN.

To our knowledge, only one study using the traditional oddballparadigm has reported no impact of gender on vMMN (Langrovaet al., 2012). However, the human perceptual system relies moreon vision than audition to characterize physical objects. The scalartiming model proposed by Gibbon and attention shift hypothesisproposed by Penney suggest that auditory stimuli are often expe-rienced as lasting longer than visual stimuli of equivalent duration,

and discrimination is more accurate for auditory than for visualstimuli, implying that auditory stimuli are easier to detect(Gibbon, 1991; Penney et al., 2000; Penney, 2003). Interestingly,Jausovec and Jausovec found that responses under attentional con-trol (e.g., early evoked gamma response, as well as P1 and P3 com-ponents), which were larger in females, showed more noticeablegender differences in the visual modality than in the auditorymodality (Jausovec and Jausovec, 2009a,b). We hypothesized thatgender differences in change detection would be evident in thevisual modality, but not in the auditory modality because auditorystimuli are more easily differentiated, thus creating a ceiling effect.

A critical factor in measuring MMN is the difference in refrac-tory response of the neural populations that respond to frequentand infrequent stimuli (Peter et al., 2010). Schroger and Wolffdeveloped an equal-probability sequence protocol that equalizesthe state of refractoriness for control and deviant stimuli(Schroger and Wolff, 1996). Specifically, they reported that theequal-probability sequence control condition produced the sameresult as the deviant-standard reverse oddball modality.Moreover, Evstigneeva and Aleksandrov explained these findings,reporting that deviant stimuli could elicit genuine MMN usingthe deviant-standard-reverse method with a ‘‘safe’’ presentationprobability of 15%, and a presentation probability of 20% can con-trol the refractoriness more effectively than 15% (Evstigneeva andAleksandrov, 2009). It is worthwhile to note that thedeviant-standard-reverse paradigm cannot completely excluderefractoriness due to the different percentages of standard anddeviant stimuli, but it can effectively control the refractoriness.Jacobsen and Schorger used this paradigm only for aMMN. Webelieve that the paradigm also applies to vMMN because therefractory effect likely arises due to different frequencies of stimu-lation, which should occur in all types of neurons. Because thedeviant-standard-reverse paradigm is easier to construct and caneffectively control refractoriness, the present study used this para-digm with a greater deviant stimulus presentation probability of20% to obtain genuine duration aMMN/vMMN, reflectingmemory-based mismatch detection.

2. Method

2.1. Participants

Participants included 22 female (25–45 y, mean 32.7 y) and 22male (24–46 y, mean 31.3 y) Chinese adults. One female and onemale were excluded from the analyses because of excessive arti-facts in the EEG recording, leaving 21 females and 21 males inthe statistical analyses. All participants were right-handed,reported normal auditory and normal or corrected-to-normalvision, had no history of current or past neurological or psychiatricillness, and used no medications known to affect the central ner-vous system. Informed consent was obtained from each subject,and experimental procedures were approved by the EthicsCommittee of the Department of Psychology, Harbin MedicalUniversity.

2.2. Stimuli and procedure

All participants were seated in a comfortable chair in front ofthe center of screen in a darkened, sound-attenuated, andelectrically-shielded room. The experiment consisted of two sepa-rate tasks with the order counterbalanced across participants.

Task1: Duration vMMNSubjects were instructed to focus their attention on a black

cross in the center of the screen, which was displayed throughoutthe stimulus blocks. Two solid black squares (1 cm � 1 cm) were

Fig. 1. Visual stimuli consisted of two solid black squares (1 cm � 1 cm) presented simultaneously for 50 ms or 150 ms in the periphery. The left panel illustrates deviantstimuli at 50 ms and standard stimuli at 150 ms, and right panel illustrates deviant stimuli at 150 ms and standard stimuli at 50 ms.

Table 1The number of deviants and standards in per condition in two groups.

Type Female Male

Standards Deviants Standards Deviants

Decrement aMMN 562 ± 36 127 ± 13 571 ± 35 123 ± 15Increment aMMN 570 ± 29 136 ± 18 569 ± 31 133 ± 19Decrement vMMN 559 ± 23 125 ± 15 554 ± 29 128 ± 14Increment vMMN 564 ± 30 131 ± 17 560 ± 27 130 ± 13

Note: There are no significant group differences of the number of deviants andstandards in each condition.

X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx 3

simultaneously presented for 50 ms or 150 ms in the peripherywith a visual angle of 3.8� � 4.0� from a distance of 1 m. The stim-ulus onset asynchrony (SOA) was fixed at 600 ms. The size of thecross was altered randomly, and subjects were required to ignorethe peripheral stimuli and press either the left or the right buttonto indicate ‘‘big’’ or ‘‘small’’ as quickly and accurately as possiblewhen the size of the cross changed. Response hands were counter-balanced across subjects. At the end of the experiment, partici-pants were asked whether they noticed changes in the peripheralsquares. No participant reported awareness of these changes.

As shown in Fig. 1, there were two blocks of conditions in thistask: (a) decrement deviant (50 ms duration with a 150 ms stan-dard); (b) increment deviant (150 ms duration with a 50 ms stan-dard). The exposure probabilities of deviant and standard stimuluswere 20% and 80% respectively and both the stimulus and blockorder were counterbalanced in the two blocked conditions. Therewere six experimental sequences of 150 trials in each block inwhich the deviant occurred among standards in a pseudorandomfashion, with the constraint that each deviant stimulus was pre-ceded by at least two standards. In the present experiment the first15 stimuli of each sequence were all standards.

Task 2: Duration aMMNSubjects were instructed to watch a self-selected, sub-titled,

silent film while auditory stimuli were presented binaurallythrough headphones. To prevent subjects from attending to theauditory stimuli, they were informed to narrate the film after thetask, including characters, scene, incident, time and content of inci-dent. Participants were observed to ensure that they focused onthe film. The frequency of tonal stimuli was 800 Hz, and the SOAwas 600 ms. The task consists of two blocks of conditions, asdescribed above, a decrement deviant (50 ms) with a 150 ms stan-dard and an increment deviant (150 ms) with a 50 ms standard.Probabilities of the deviant and standard stimuli were 0.2 and

0.8, respectively. As in task 1, both the stimulus and block orderwere counterbalanced in the two blocked conditions, and stimuliwere presented in pseudorandom order.

2.3. EEG recording

EEG was recorded continuously with a Neuroscan 40-electrodecap (NuAmps amplifier) using channels based on the International10–20 system. Two electrodes were placed on the outer canthi ofboth eyes to record horizontal EOG, and another two electrodeswere placed above and below the right eye to record verticalEOG. The tip of nose was used as the reference electrode. The sam-pling rate was 500 Hz/channel. Impedances of all electrodes weremaintained below 5 kX throughout the experiment.

After EOG artifact correction with the Gratton method (Gratton,1998), the EEG was segmented in 500 ms epochs, time-locked tostimulus onset, and set to include a 100 ms pre-stimulus baseline.Trials contaminated with artifacts greater than ±100 lV wererejected before averaging. The EEG segments were averaged sepa-rately for standard and deviant stimuli in different conditions, andthe averaged ERP data were digitally filtered with a band-pass filterat 1–30 Hz, 24 dB/octave. The numbers of trials for each conditionin the two groups are shown in Table 1. There were no significantgroup differences in decrement or increment conditions.

2.4. Data analysis

In line with the MMN study by Jacobsen et al., MMNs reflectingautomatic sensory memory processes were obtained by subtract-ing the ERPs to a standard in one block from a deviant in thedeviant-standard-reversed block (Jacobsen and Schroger, 2003).Mean amplitudes of aMMN and vMMN components were analyzedfrom the grand average ERP waveforms (see Figs. 2 and 3).

Mean amplitudes of MMN were subjected to repeated measuresANOVAs. For vMMN, a 4-way ANOVA was conducted with gender(male and female) as a between-subject factor and stimulus type(50 ms duration MMN and 150 ms duration MMN), hemisphere(left hemisphere, midline, and right hemisphere), and site (P3/O1,Pz/Oz, P4/O2) as within-subject factors. For aMMN with negativedeflection distributed in fronto-central areas, a 4-way ANOVAwas conducted with gender as a between-subject factor and stim-ulus type, hemisphere, and site (F3/FC3/C3, Fz/FCz/Cz, F4/FC4/C4)as within-subject factors. For temporal aMMN or vMMN withpositive deflection, a 3-way ANOVA was conducted with genderas a between-subject factor and stimulus type and site (M1, M2)

4 X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx

as within-subject factors. In addition, topographic analyses wereperformed using repeated measures ANOVA as described above.Degrees of freedom for ANOVAs were corrected according to theGreenhouse–Geisser method.

3. Results

3.1. Behavioral performance

For a correct response, subjects had to press a button within150–800 ms after onset of the stimulus, which the cross target atthe center of the screen. Both hits and false alarms were measured.Males and females showed similar hit rates (95.2% and 94.1%,respectively; paired t-test: p > 0.1) and reaction times (456 msand 448 ms, respectively; paired t-test: p > 0.1), as well as falsealarms (2.79 and 3.15, respectively).

3.2. ERP analysis

As shown in Figs. 2 and 3, grand-averaged ERPs were elicited bystandard and deviant stimuli in both males and females. In bothgenders, N1, P1, and P2 components were elicited by auditorystimuli at tempo-frontal areas, and N1, P1, and P2 componentswere elicited by visual stimuli at the posterior scalp, regardlessof whether stimuli were deviant or standard. Figs. 4 and 5 showedthe difference waveforms between ERPs in response to deviant andstandard (physically-identical) stimuli. Clearly, MMNs were

Fig. 2. Grand average ERPs in response to 50 ms and 150 ms de

evident from 150 to 300 ms after stimulus onset for aMMN andfrom 150 to 350 ms after stimulus onset for vMMN.

As shown in Fig. 4, grand-averaged waveforms of decrementaMMN and increment aMMN exhibited similar magnitudes in thefrontal and temporal regions for male and female participants. Asshown in Fig. 5, increment MMN was decreased in female subjectscompared with male subjects, while decrement MMN was similarin males and females in the parietal-occipital area. In the temporalarea, both increment and decrement MMN were similar acrossgenders. T-tests were conducted to identify intervals that weresignificantly below zero in all conditions (see Table 2).

3.2.1. Auditory MMNRepeated measures ANOVA showed no significant effect of

gender on aMMN amplitudes in fronto-central areas during the150–250 ms time window (F(1,42) = 0.152, p > 0.05). The effect ofsite was significant (F(2,84) = 3.619, p < 0.05, F(2,84) = 16.789,p < 0.001, respectively), showing a frontal distribution with a max-imum amplitude of �2.956 lV at FZ. The effect of hemisphere wassignificant (F(2,84) = 5.489, p < 0.05), indicating a predominantlyright-hemispheric frontal MMN (Left: �0.803 lV, Midline:�1.076 lV, Right: �1.645 lV). Furthermore, there were no signifi-cant interactions for type � region or gender � type � hem inter-action (Fs < 1, ps > 0.05), indicating that the laterality of aMMNwas similar between types for males and females (see Fig. 6). Fortemporal aMMN, there was also no effect of gender and no interac-tion effect with gender (Fs < 1, ps > 0.05). These data show nosignificant effect of gender on decrement or increment aMMN

viant and standard auditory stimuli in males and females.

X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx 5

amplitude. Mean amplitudes of aMMN in males and females infrontal and temporal areas are shown in Table 3.

3.2.2. Visual MMNBecause decrement vMMN was not observed in males or

females in this study, only increment vMMN was assessed. Therewere no significant effects of gender, interaction with gender, orsite on increment vMMN amplitude in the temporal area at anytime (Fs < 1, ps > 0.1). As shown in Fig. 5, vMMN amplitudes infemales were higher than in males during the time window of180–260 ms. As shown in Table 4, there was a significant gendereffect on vMMN in parietal-occipital areas during the 180–260 ms time window: females had significantly lower vMMNamplitudes than males. Furthermore, the effect of site was signifi-cant, showing an occipital distribution with a maximum amplitudeof �2.52 lV at O2. The effect of hemisphere was significant at 180–260 ms (F(2,84) = 6.930, p = 0.008), indicating a predominantlyright-hemispheric occipital vMMN (Left: �0.857 lV, Midline:�1.069 lV, Right: 2.015 lV; see Fig. 7).

Interestingly, the gender � hemisphere interaction was signifi-cant at 180–260 ms (F(1,42) = 9.637, p = 0.005). Further analysisrevealed that mean amplitudes of vMMN in the midline and righthemisphere were significantly smaller in females than males(F(1,42) = 4.239, p = 0.039; F(1,42) = 6.872, p = 0.015). Females showedsymmetrical distribution of vMMN amplitudes, which were similaracross the left hemisphere, midline, and right hemisphere.However, males showed asymmetrical distribution with greatervMMN amplitudes in the midline and right hemisphere relative

Fig. 3. Grand average ERPs in response to 50 ms and 150 ms d

to the left hemisphere (F(1,42) = 8.156, p = 0.002). In addition, thehem � site interaction was also significant at 180–260 ms(F(2,84) = 4.289, p = 0.031).

In addition, the vMMN peak latency was longer in females(226 ms) than in males (209 ms; F(1,42) = 4.866, p = 0.038). No othereffects reached significance (p > 0.1).

4. Discussion

The present study investigated gender effects on pre-attentiveprocessing by recording aMMNs and vMMNs for stimulus duration.In the auditory modality, there were no effects of gender on decre-ment or increment aMMN, which occurred at 150–200 ms afterstimulus onset. In contrast, in the visual modality, although decre-ment vMMN was not observed, the effect of gender was observedon increment vMMN at 180–260 ms after stimulus onset.Specifically, increment vMMN was enhanced in males relative tofemales. Moreover, increment vMMN was more prominent in theright hemisphere in males and was symmetrical in females.

These results support previous studies that used thedeviant-standard-reverse paradigm to elicit duration MMN tovisual or auditory stimuli (Jacobsen and Schroger, 2003; Qiuet al., 2011). As expected, we did not find significant gender effectsin auditory change detection. The aMMN generator in auditory cor-tex is known to be associated with detection of change in auditorystimuli. We hypothesized that males and females would recruit thesame processing resources to detect changes of auditory duration

eviant and standard visual stimuli in males and females.

Fig. 4. ERPs to a standard in one block were subtracted from a deviant in a deviant-standard-reversed block to derive decrement and increment aMMN in males and females.Decrement aMMN was obtained by subtracting 50 ms stimuli in standard block from 50 ms stimuli in a deviant block. Similarly, increment aMMN was obtained bysubtracting 150 ms stimuli in standard block from 150 ms stimuli in a deviant block.

6 X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx

under non-attention conditions, in accordance with previousstudies (Kasai et al., 2002; Nagy et al., 2003). On the contrary,two previous studies have reported females to have a largerMMN in response to intensity changes of pure-tone stimuli andvocal emotional expressions (Barrett and Fulfs, 1998; Schirmeret al., 2005). The discrepancy between these two studies relative

to studies by our group and others may be partially due to the dif-ferent physical features used to elicit MMN (intensity versus dura-tion). Importantly, Schirmer chose emotional stimuli, which havegreater social relevance. Females have greater emotion recognition(Schirmer and Kotz, 2003), and hence, it is not surprising thatfemales had a larger MMN in response to vocal emotional

Fig. 5. ERPs to a standard in one block were subtracted from a deviant in a deviant-standard-reversed block to derive increment vMMN in males and females. DecrementvMMN was obtained by subtracting 50 ms stimuli in standard block from 50 ms stimuli in a deviant block. Similarly, increment vMMN was obtained by subtracting 150 msstimuli in standard block from 150 ms stimuli in a deviant block.

X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx 7

expression. Moreover, the present study showed a symmetricalhemispheric distribution of aMMN across genders. In contrast,Matsubayashi et al. observed that males had stronger durationMMNm over the left hemisphere than women (Matsubayashiet al., 2008). It should be noted that the stimuli used to elicitMMN by Matsubayashi differed in physical features (i.e., 1000 Hzvs 700 Hz), whereas the current experiment presented both

duration stimuli as standards and deviants to eliminate effects ofphysical differences by subtraction. In addition, the current studyused EEG, which detects MMN generated in both the frontal lobeand the superior temporal plane, while Matsubayashi used MEG,which predominantly detects signals in the superior temporalplane but not in the frontal lobe (Naatanen and Alho, 1995).Finally, the study by Matsubayashi had a ratio of males to females

Table 2Summary of the mean amplitudes of aMMN (female and male)in 150–250 ms andvMMN in 180–260 ms time window, with t-statistics for significant differences fromzero (N = 21).

Type/site Female t p Male t p

Auditory MMNIncrementFZ �2.209 5.971 0.0001 �2.175 5.858 0.0001FCZ �2.181 5.861 0.0001 �2.168 5.713 0.0001CZ �1.963 5.517 0.0001 �1.914 5.479 0.0001M1 0.956 2.813 0.0129 0.949 2.812 0.0129M2 0.839 2.717 0.0141 0.846 2.718 0.0151DecrementFZ �2.182 5.876 0.0001 �2.193 5.901 0.0001FCZ �2.073 5.624 0.0001 �2.142 5.869 0.0001CZ �1.959 5.579 0.0001 �2.011 5.587 0.0001M1 0.967 2.821 0.0128 0.896 2.764 0.0138M2 1.126 2.875 0.0098 1.108 2.869 0.0099Visual MMNIncrementO1 �0.943 2.917 0.0073 �1.269 3.015 0.0076OZ �0.996 3.061 0.0067 �1.304 3.276 0.0015O2 �1.018 3.512 0.0031 �1.397 3.849 0.0011DecrementO1 �0.462 1.201 0.311 �0.498 1.211 0.3081OZ �0.469 1.203 0.311 �0.503 1.216 0.307O2 �0.517 1.256 0.289 �0.483 1.207 0.3101

8 X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx

that was nearly 2:1, which is confound for statistical analysis ofgender effects. Hence, using the deviant-reverse paradigm to con-trol refractory effects, we concluded that there was no gender dif-ference in aMMN.

Previous studies have demonstrated vMMN, reflecting memorycomparison-based change detection (Kimura et al., 2009; Maekawaet al., 2009). Decrement vMMN was not elicited in males orfemales in the current study. According to the scalar timing model,duration memory traces form more readily in the auditory modal-ity than in the visual modality (Gibbon, 1991), which may explainwhy decrement aMMN was elicited while decrement vMMN was

Fig. 6. 2-D voltage maps of aMMN in response to pure tone duration sti

not. It is worth noting that increment duration vMMN could be agood indicator for investigating pre-attentive processing.Supporting this idea, our previous study found that only incrementvMMN was decreased in patients with major depressive disorder(Qiu et al., 2011).

To interpret gender differences in vMMN, it is important to con-sider the neural generator of vMMN: the right occipital visualextrastriate area (Kimura, 2012). A quantitative analysis of thecytoarchitecture of the human primary visual cortex showed thatvisual areas are sexually dimorphic, and the volumetric ratio ofthe motion-sensitive region (hOc5) to primary visual cortex(V1/Brodmann area 17) is significantly larger in males thanfemales (Amunts et al., 2007), indicating that males have more tis-sue to process visual information. The enhanced vMMN observedin males in the present study may provide indirect evidence of agender difference in occipital visual extrastriate areas. However,because ERPs do not directly reflect brain structure, the gender dif-ferences in occipital visual extrastriate require further investiga-tion by fMRI. Interestingly, data are mixed regarding whethermale or females show greater responses to visual stimuli, as somestudies have observed larger P300 and/or N400 components infemales (Steffensen et al., 2008), while others have observed largerN400 in males (Jausovec and Jausovec, 2009), and some haveshown no gender difference in P300 (Sangal and Sangal, 1996).Both P300 and N400 reflect the late stage of information process-ing, and the present study provides the first evidence that menhave an advantage in processing temporal information at thepre-attentive stage compared to women. However, Langrovaet al. reported no impact of gender on vMMN (Langrova et al.,2012). There are two possible explanations for these differentresults. First, the two studies used different paradigms: a tradi-tional oddball paradigm (Langrova et al., 2012) versus adeviant-standard reverse oddball paradigm (current study).Second, the physical attributes of the stimulus were different:direction (Langrova et al., 2012) versus duration (current study).Within the frame of a scalar timing model, Penney et al. proposed

muli from 160 to 220 ms after stimulus onset in males and females.

Table 3Mean amplitude of auditory MMN (lV) in males and females in fronto-central and temporal areas.

Type 150–200 ms 200–250 ms 250–300 ms

Female Male Female Male Female Male

Fronto-central MMNDecrement-MMN �0.78 (0.56) �1.17 (0.69) �1.27 (0.89) �1.71 (1.10) 0.34 (0.31) �0.32 (0.36)Increment-MMN �0.80 (0.75) �0.64 (0.47) �1.36 (0.78) �1.12 (0.96) �0.09 (0.10) �0.16 (0.27)Temporal-MMNDecrement-MMN 0.18 (0.28) 0.40 (0.37) 0.62 (0.43) 0.68 (0.67) 0.06 (0.12) �0.19 (0.17)Increment-MMN 0.53 (0.65) 0.26 (0.27) 0.75 (0.57) 0.73 (0.71) �0.17 (0.15) 0.27 (0.30)

Data are given as mean (standard deviation).

Table 4Mean amplitude of visual MMN (lV) in males and females in parietal-occipital andtemporal areas during 180–260 ms.

Type Female Male F P

Parietal-occipital vMMNDecrement-vMMN 0.81 (0.65) 0.93 (0.86) 1.790 0.187Increment-vMMN �1.25 (0.68) �1.96 (0.93) 5.216 0.029Temporal vMMNDecrement-vMMN 0.76 (0.59) 0.85 (0.59) 1.245 0.216Increment-vMMN �0.69 (0.57) �0.75 (0.61) 1.148 0.243

Data are given as mean (standard deviation), ⁄P < 0.05.

X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx 9

an ‘‘attentional switch’’ theory, which suggests that the modeswitch would be different in the closed state for the differentmodalities. The fact that our results differ from Langrova’s mayextend the attentional switch theory, suggesting that stimuli withdifferent physical attributes may have different closed states.Future studies should investigate this theory further.Furthermore, by using low-resolution brain electromagnetictomography, Pazo-Alvarez reported that the neural generators ofdirection for visual MMN included the occipital visual extrastriateareas (Pazo-Alvarez et al., 2004). Using fMRI, Urakawa et al. foundthat color deviant stimuli elicited magnetic responses in the

Fig. 7. 2-D voltage maps of vMMN in response to

occipital visual extrastriate areas and ventrolateral prefrontal areas(Urakawa et al., 2010). In particular, Shi et al. found that the cur-rent sources of duration vMMN were located in the frontal lobeand parietal lobe (Shi et al., 2013). Thus, different attributes ofstimuli may elicit responses from different cortical areas, whichcould explain why our results differ from those of Langrova et al.

Visual perceptual functions are less dependent on the righthemisphere in women than in men (Halpern, 2000). The moststriking findings in the present study were gender differences inthe functional organization of the right hemisphere areas involvedin visual duration change detection. Previous studies have shownthat processing of emotional visual stimuli and facial expressionsis more strongly lateralized in men, showing right hemisphericdominance (Godard and Fiori, 2010; Proverbio et al., 2006; Wraseet al., 2003). Gender differences have also been found in the later-alization of visual-spatial processes, such as mental rotation tasks,in which males are typically right hemisphere dominant andfemales have bilateral distribution (Johnson et al., 2002).Furthermore, MRI analysis of 290 brain structures revealed thatmales are more asymmetric in the majority of brain regions(Kovalev et al., 2003). Electrophysiology studies also indicate moreasymmetric scalp distribution ERP components in men. Forinstance, the N1 component relevant to face processing peakedearlier in the right hemisphere in men, suggesting that gender

visual duration stimuli from 200 to 260 ms.

10 X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx

effects on hemispheric asymmetry appear very early (Proverbioet al., 2006). Roalf suggested that the gender difference originatesearlier in the information processing stream than the P300 (Roalfet al., 2006). Our findings that vMMN distribution showed righthemispheric dominance in men support the view that femalevisual structures are more bilaterally symmetric even at thepre-attentive stage of information processing.

Overall, the present study indicated that gender effects onpre-attentive processing were modulated by perceptual modality.Specifically, gender effects were observed for vMMN but notaMMN. Recently, Jausovec and Jausovec found that responsesunder attentional control (early evoked gamma response, as wellas P1 and P3 components), which were larger in females, alsoshowed more noticeable differences in the visual modality thanin the auditory modality (Jausovec and Jausovec, 2009). The atten-tional switch hypothesis proposes that auditory temporal informa-tion is processed relatively automatically, whereas visual temporalinformation requires controlled attention. The present studyextends this hypothesis by providing electrophysiological evidencethat auditory temporal information is processed automatically inboth males and females, while visual duration memory tracesmay be more easily formed in males. However, small amplitudedifferences in the already small visual increment MMN responses,together with absent decrement MMN response, do not necessarilysupport the idea that visual duration memory traces are facilitatedin males. Additional studies will be required to replicate thesefindings.

The current study did not assess personality traits, which havebeen shown to be associated with MMN amplitude (Matsubayashiet al., 2008). We also did not assess the menstrual cycle in females,and a previous study found that menstrual cycle stage and gonadalsteroid hormone levels affect functional lateralization (Hausmannet al., 2002). Further studies will be necessary to determinewhether these factors affect pre-attentive processing.

In conclusion, the present study supported previous findings ofautomatic auditory and visual temporal information processing.Auditory temporal change detection was not modulated by gender,while visual temporal memory traces were facilitated in males. Thegender differences of cross-modality effects on pre-attentivechange detection emphasize the importance of considering genderas a factor in the study of cognitive function.

Acknowledgements

This research was supported by the National Natural ScienceFoundation of China (31271093) to Prof. Yanjie Yang and theNational Natural Science Foundation of China (81202213) toXiuxian Yang.

Conflict of interest: The authors report no biomedical financialinterests or potential conflicts of interest.

References

Aaltonen O, Eerola O, Lang AH, Uusipaikka E, Tuomainen J. Automaticdiscrimination of phonetically relevant and irrelevant vowel parameters asreflected by mismatch negativity. J Acoust Soc Am 1994;96:1489–93.

Alho K. Cerebral generators of mismatch negativity (MMN) and its magneticcounterpart (MMNm) elicited by sound changes. Ear Hear 1995;16:38–51.

Amunts K, Armstrong E, Malikovic A, Homke L, Mohlberg H, Schleicher A, Zilles K.Gender-specific left-right asymmetries in human visual cortex. J Neurosci2007;27:1356–64.

Astikainen P, Lillstrang E, Ruusuvirta T. Visual mismatch negativity for changes inorientation–a sensory memory-dependent response. Eur J Neurosci2008;28:2319–24.

Barrett KA, Fulfs JM. Effect of gender on the mismatch negativity auditory evokedpotential. J Am Acad Audiol 1998;9:444–51.

Berti S. The attentional blink demonstrates automatic deviance processing in vision.NeuroReport 2011;22:664–77.

Chang Y, Xu J, Shi N, Pang X, Zhang B, Cai Z. Dysfunction of preattentive visualinformation processing among patients with major depressive disorder. BiolPsychiatry 2011;69:742–7.

Chen Y, Huang X, Luo Y, Peng C, Liu C. Differences in the neural basis of automaticauditory and visual time perception: ERP evidence from an across-modaldelayed response oddball task. Brain Res 2010;1325:100–11.

Czigler I, Balazs L, Pato L. Visual change detection: event-related potentials aredependent on stimulus location in humans. Neurosci Lett 2004;364:149–53.

Escera C, Winkler I, Alho K, Naatanen R. Neural mechanisms of involuntaryattention to acoustic novelty and change. J Cogn Neurosci 1998;10:590–604.

Evstigneeva MD, Aleksandrov AA. Mismatch negativity: the contribution ofdifferences in the refractoriness of stimulus-specific neuron populations.Neurosci Behav Physiol 2009;39:833–40.

Gibbon J. Origins of scalar timing. Learn Motiv 1991;22:33–8.Godard O, Fiori N. Sex differences in face processing: are women less lateralized and

faster than men? Brain Cogn 2010;73:167–75.Gratton G. Dealing with artifacts: the EOG contamination of the event-related brain

potential. Behav Res Meth Instrum Comput 1998;30:44–53.Halpern D. Sex differences in cognitive abilities. Mahwah, NJ: Lawrence Erlbaum

Associates, Inc.; 2000.Hansenne M, Pinto E, Scantamburlo G, Renard B, Reggers J, Fuchs S, Pitchot W,

Ansseau M. Harm avoidance is related to mismatch negativity (MMN)amplitude in healthy subjects. Personality Individ Differ 2003;34:1039–48.

Hausmann M, Becker C, Gather U, Gunturkun O. Functional cerebral asymmetriesduring the menstrual cycle: a cross-sectional and longitudinal analysis.Neuropsychologia 2002;40:808–16.

Ikezawa S, Nakagome K, Mimura M, Shinoda J, Itoh K, Homma I, Kamijima K. Genderdifferences in lateralization of mismatch negativity in dichotic listening tasks.Int J Psychophysiol 2008;68:41–50.

Jacobsen T, Schroger E. Measuring duration mismatch negativity. Clin Neurophysiol2003;114:1133–43.

Jausovec N, Jausovec K. Gender related differences in visual and auditory processingof verbal and figural tasks. Brain Res 2009;1300:135–45.

Johnson BW, McKenzie KJ, Hamm JP. Cerebral asymmetry for mental rotation:effects of response hand, handedness and gender. NeuroReport2002;13:1929–32.

Kasai K, Nakagome K, Iwanami A, Fukuda M, Itoh K, Koshida I, Kato N. No effect ofgender on tonal and phonetic mismatch negativity in normal adults assessed bya high-resolution EEG recording. Cogn Brain Res 2002;13:305–12.

Kenemans JL, Grent T, Verbaten MN. Detection of visual change: mismatch orrareness? NeuroReport 2003;14:1239–42.

Kimura M. Visual mismatch negativity and unintentional temporal-context-basedprediction in vision. Int J Psychophysiol 2012;83:144–55.

Kimura M, Katayama J, Ohira H, Schroger E. Visual mismatch negativity: newevidence from the equiprobable paradigm. Psychophysiology 2009;46:402–9.

Kovalev VA, Kruggel F, von Cramon DY. Gender and age effects in structural brainasymmetry as measured by MRI texture analysis. Neuroimage2003;19:895–905.

Langrova J, Kremlacek J, Kuba M, Kubova Z, Szanyi J. Gender impact onelectrophysiological activity of the brain. Physiol Res 2012;61(Suppl.2):119–27.

Maekawa T, Tobimatsu S, Ogata K, Onitsuka T, Kanba S. Preattentive visual changedetection as reflected by the mismatch negativity (MMN)–evidence for amemory-based process. Neurosci Res 2009;65:107–12.

Matsubayashi J, Kawakubo Y, Suga M, Takei Y, Kumano S, Fukuda M, Itoh K, YumotoM, Kasai K. The influence of gender and personality traits on individualdifference in auditory mismatch: a magnetoencephalographic (MMNm) study.Brain Res 2008;1236:159–65.

McGivern RF, Mutter KL, Anderson J, Wideman G, Bodnar M, Huston P. Genderdifferences in incidental learning and visual recognition memory: support for asex difference in unconscious environmental awareness. Personality IndividDiffer 1998;25:223–32.

Naatanen R. Mismatch negativity (MMN): perspectives for application. Int JPsychophysiol 2000;37:3–10.

Naatanen R, Alho K. Generators of electrical and magnetic mismatch responses inhumans. Brain Topogr 1995;7:315–20.

Naatanen R, Paavilainen P, Rinne T, Alho K. The mismatch negativity (MMN) in basicresearch of central auditory processing: a review. Clin Neurophysiol2007;118:2544–90.

Nagy E, Potts GF, Loveland KA. Sex-related ERP differences in deviance detection. IntJ Psychophysiol 2003;48:285–92.

Pazo-Alvarez P, Amenedo E, Cadaveira F. Automatic detection of motion directionchanges in the human brain. Eur J Neurosci 2004;19:1978–86.

Penney TB. Modality differences in interval timing: attention, clock speed, andmemory. In: Meck WH, editor. Functional and neural mechanisms of intervaltiming. Boca-Raton, FL: CRC Press; 2003. p. 209–33.

Penney TB, Gibbon J, Meck WH. Differential effects of auditory and visual signals onclock speed and temporal memory. J Exp Psychol Human Percept Perform2000;26:1770–87.

Peter V, McArthur G, Thompson WF. Effect of deviance direction and calculationmethod on duration and frequency mismatch negativity (MMN). Neurosci Lett2010;482:71–5.

Proverbio AM, Brignone V, Matarazzo S, Del Zotto M, Zani A. Gender differences inhemispheric asymmetry for face processing. BMC Neurosci 2006;7:44–50.

Qiu X, Yang X, Qiao Z, Wang L, Ning N, Shi J, Zhao L, Yang Y. Impairment inprocessing visual information at the pre-attentive stage in patients with a major

X. Yang et al. / Clinical Neurophysiology xxx (2015) xxx–xxx 11

depressive disorder: a visual mismatch negativity study. Neurosci Lett2011;491:53–7.

Rinne T, Alho K, Ilmoniemi RJ, Virtanen J, Naatanen R. Separate time behaviors ofthe temporal and frontal mismatch negativity sources. Neuroimage2000;12:14–9.

Roalf D, Lowery N, Turetsky BI. Behavioral and physiological findings of genderdifferences in global-local visual processing. Brain Cogn 2006;60:32–42.

Sangal B, Sangal JM. Topography of auditory and visual P300 in normal adults. Clin.Electroencephalogr 1996;27:145–50.

Schirmer A, Kotz SA. ERP evidence for a sex-specific Stroop effect in emotionalspeech. J Cogn Neurosci 2003;15:1135–48.

Schirmer A, Striano T, Friederici AD. Sex differences in the preattentive processing ofvocal emotional expressions. NeuroReport 2005;16:635–9.

Schroger E, Wolff C. Mismatch response of the human brain to changes in soundlocation. NeuroReport 1996;7:3005–8.

Shi L, Wu J, Gang S, Dang L, Zhao L. Visual mismatch negativity in the ‘‘optimal’’multi-feature paradigm. J Integr Neurosci 2013;12:1–12.

Steffensen SC, Ohran AJ, Shipp DN, Hales K, Stobbs SH, Fleming DE. Gender-selectiveeffects of the P300 and N400 components of the visual evoked potential. VisionRes 2008;48:917–25.

Tales A, Newton P, Troscianko T, Butler S. Mismatch negativity in the visualmodality. NeuroReport 1999;10:3363–7.

Urakawa T, Inui K, Yamashiro K, Kakigi R. Cortical dynamics of the visual changedetection process. Psychophysiology 2010;47:905–12.

Winkler I, Czigler I. Evidence from auditory and visual event-related potential (ERP)studies of deviance detection (MMN and vMMN) linking predictive codingtheories and perceptual object representations. Int J Psychophysiol2012;83:132–43.

Wrase J, Klein S, Gruesser SM, Hermann D, Flor H, Mann K, Braus DF, Heinz A.Gender differences in the processing of standardized emotional visual stimuli inhumans: a functional magnetic resonance imaging study. Neurosci Lett2003;348:41–5.

Zhao L, Li J. Visual mismatch negativity elicited by facial expressions under non-attentional condition. Neurosci Lett 2006;410:126–31.