Differential effects of gender on mismatch negativity to violations of simple and pattern acoustic regularities

Abstract Introduction The effects of gender on the mismatch negativity (MMN) potential have been studied using simple frequency deviants. However, the effects of gender on MMN to violations of abstract regularities have not yet been studied. Here, we addressed this issue and compared the effects of gender on simple and pattern frequency MMNs. Methods MMN response was recorded from 29 healthy young adults, 14 females (mean age = 26.20 ± 2.17) and 15 males (mean age = 27.57 ± 2.24), using 32 scalp electrodes during simple and pattern frequency oddball paradigms and the mean amplitude, peak latency, and scalp topography of MMN evoked by each paradigm were compared between the two genders. Results The peak latency of simple MMN was significantly longer in females (p < .05); however, its mean amplitude and topography were similar between the two genders (p > .05). There were no significant differences in peak latency, mean amplitude, and scalp topography of pattern MMN between the two genders (p > .05). Conclusions Based on the obtained results, gender differently affects simple and pattern MMN. These findings may provide preliminary evidence for distinct effects of gender on various types of MMN.


INTRODUCTION
Mismatch negativity (MMN) is a preattentive component of auditory event-related potentials (ERPs), which was first described by Näätänen et al. in 1978(Näätänen et al., 1978. This response is conventionally generated by presentation of an infrequent auditory stimulus (i.e., deviant) among a sequence of repetitive stimuli (i.e., standards) and its generation is mediated by two different neurophysiological mechanisms: (1) the formation of memory trace of standard stimuli and (2) change detection mechanisms that compare incoming sounds with the memory trace of previous stimuli (Bartha-Doering et al., This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Brain and Behavior published by Wiley Periodicals LLC 2015; Yu et al., 2015). Several studies, using more complex stimulus designs, have demonstrated that memory underlying MMN generation contains sophisticated sound encoding mechanisms that can extract any regular aspect of recent stimuli, even the abstract ones hidden in the ever-changing acoustic input (Herholz et al., 2009;Paavilainen, 2013). On the basis of these studies, change detection mechanisms detect any new event that violates the ongoing sound regularities (Garrido et al., 2009). Given the complex calculations needed for extracting abstract regularities, MMN is regarded as an index of primitive sensory intelligence in the auditory cortex (Näätänen et al., 2001(Näätänen et al., , 2010. MMN is among cognitive responses, which has found wide clinical utility (Duncan et al., 2009), providing a neurophysiological marker of auditory perceptual accuracy, general brain degeneration, and the gross functional state of the brain. Many studies have shown the applications of MMN in prediction of the functional state of patients with schizophrenia and coma outcome as well as diagnosis and treatment of impaired neurophysiological mechanisms in central auditory processing disorders, dyslexia, and Alzheimer's disease (Duncan et al., 2009;Gao et al., 2018;Kärgel et al., 2014 ;Näätänen, 2003;Näätänen & Escera, 2000;Näätänen et al., 2007Näätänen et al., , 2014Roberts et al., 2011).
A group of studies has demonstrated that elicitation of MMN with complex stimulus designs may improve its sensitivity and specificity and also can provide new insights into the pathophysiology and management of disorders (Näätänen et al., 2014;Paavilainen, 2013;Umbricht & Krljes, 2005). Pattern regularities, regular changes of acoustic cues between several stimuli, are among complex stimulus designs that can be easily incorporated for MMN elicitation in clinical settings (Paavilainen, 2013). Pattern MMN demonstrates the capacity of the brain for extracting abstract rules, which in turn underlies our adaptive behavior in demanding and complex environments (Bendixen & Schröger, 2008;Schröger et al., 2007). Furthermore, given the importance of pitch pattern processing in the perception of speech prosody and development of word segmentation, phonological skills, and literacy, several studies have provided evidence of decreased MMN amplitude to violations of pitch patterns in patients with dyslexia and schizophrenia (Gjini et al., 2010).  (Hall, 2007). However, to the best of our knowledge, no study has yet investigated the effects of gender on MMN to violations of complex regularities such as pitch patterns.
Processing of pitch pattern regularities entails the sound order encoding (Alain et al., 1998), which has been reported to be stronger in males than females in some behavioral studies (Fink et al., 2005;Szymaszek et al., 2006) and similar between the two genders in others (Shrivastav et al., 2008;van Kesteren & Wiersinga-Post, 2007). Therefore, it is important to explore whether these gender differences in auditory temporal ordering would also affect the preattentive processing of sound patterns reflected by MMN underlying mechanisms. In the present study, we investigated the effects of gender on pattern MMN evoked by the violation of a fixed pitch relation between three tones. Furthermore, given the inconsistency of scarce studies, which have addressed the gender effects on conventional MMN (Aaltonen et al., 1994;Kasai et al., 2002;Nagy et al., 2003), we also re-examined the effects of gender on the MMN evoked by simple frequency deviants.
Recording these two types of MMN in the same sample of females and males provided a unique opportunity for exploring any possible gender effect differences between various types of MMN.

Participants
The participants consisted of 14 females and 15 males ranging from 23 to 30 years old with the mean ages of 26.20 ± 2.17 and 27.57 ± 2.24 years, respectively. All the subjects were monolingual and right-handed, as measured by the Edinburg Handedness Questionnaire (EHQ), with hearing thresholds better than 20 dB HL within the frequency range of 250-8000 Hz in both ears, and with no history of otological, neurological, or psychiatric disorders and drug or alcohol abuse. Written informed consent was obtained from all participants.

Stimuli and procedure
Participants were seated comfortably in an electrically and acoustically shielded room. A silent movie with subtitles was played on a front monitor and the subjects were instructed to watch the movie and to ignore the auditory stimuli during the experiment.

Electroencephalography (EEG) recording
The EEG was recorded using a 32-electrode cap (Wavegaurd, ANT, according to the international 10-20 system. An electrode on the tip of the nose was used as the reference for all of the electrodes. The eye movements were monitored using two bipolar electrodes above and below the right eye and two electrodes at the outer canthi of each eye (vertical and horizontal electrooculograms). The impedance for all of the electrodes was maintained below 5 kΩ. The EEG signals were amplified using an ANT amplifier (Advanced Neuro Technology, Enschede, The Netherlands). The amplified signals were online bandpass filtered from 0.05 to 500 Hz and digitized with a sampling rate of 2048 Hz using ASA 4.7.1 software (ANT, Enschede, The Netherlands).

EEG analysis
EEG signals were offline filtered by a band-pass filter from 1 to 30 Hz.
All electrodes were re-referenced to the linked mastoids. Eye blinks, movement, and myogenic artifacts were removed by applying independent component analysis (Delorme & Makeig, 2004;Delorme et al., 2007). EEG data were segmented into epochs of 500 ms, including 100 ms pre-stimulus baseline time, and were averaged separately for deviant and standard stimuli. Trials with amplitude variations exceeding ±70 µv were automatically rejected. The frequency-change MMN response was obtained by subtracting the response of standard stimuli from that of deviant stimuli in frequency oddball condition. The pattern MMN was calculated by subtracting the response of the first tone in the standard pattern from that of the first tone in the deviant pattern (Sussman et al., 1999). MMN peak latencies were measured individually from the most negative peak at Fz at 100-250 ms poststimulus for each condition.
Individual MMN mean amplitudes were calculated from a 50-ms time window around the individual peak latencies obtained at Fz electrode for each condition.

Statistical analysis
For each oddball condition, a separate repeated-measure analysis of variance (ANOVA) including the factors stimulus type (deviant, standard) and gender (female, male) was calculated on the mean amplitudes in the MMN time range. To compare the effects of gender on simple and pattern MMNs, two separate repeated-measure ANOVAs were performed on MMN mean amplitudes and peak latencies using the within subject factor condition (frequency oddball, pattern oddball) and the between subject factor gender (female, male). Post hoc comparisons were made using paired and independent sample t-tests.
Differences were considered significant when p < .05. Statistical package for the social sciences version 16 (SPSS Inc., IL) was used for data analysis.
Global dissimilarity index (DISS) was computed in order to compare MMN scalp topographies between the two genders without considering signal strength differences (Murray et al., 2008

RESULTS
The grand-average responses of the standard and deviant stimuli and a clear difference waveform MMN were obtained in both frequency and pattern oddball conditions (Figure 1). The mean amplitudes and peak latencies for simple frequency-change and pattern MMN responses are given in Table 1. In both oddball conditions, repeated-measure ANOVA revealed a main effect of stimulus type (deviant, standard) on mean amplitudes in the time window of MMN (frequency oddball condition: F 1,27 = 98.31, p < .001, 2 P = 0.78; pattern oddball condition: F 1,27 = 51.43, p < .001, 2 P = 0.65) in terms of larger amplitudes elicited by deviant compared to standard stimuli, but no stimulus type X gender interaction was found for the mean amplitudes in any conditions

DISCUSSION
The present study examined the effects of gender on two types of MMN evoked by simple frequency changes and pitch pattern violations. In the frequency oddball condition, MMN peak latency was significantly increased in females compared to males, whereas its mean amplitude and topographic distribution was similar between the two genders. In the pattern oddball condition, no effect of gender on the mean amplitude, peak latency, and topographic map of MMN was found.  (Fink et al., 2005;Szymaszek et al., 2006). Given that the neural representation of pitch patterns is achieved by encoding the temporal order of pattern-composing stimuli (Alain et al., 1998) In the present study, the mean amplitude of MMN to frequency changes was similar between the two genders; however, its peak latency was longer in females. Nagy et al. (2003) and Tsolaki et al. (2015) also investigated the effects of gender on MMN evoked by frequency changes and found no differences in MMN mean amplitudes or peak latencies between females and males. Different effect of gender on MMN peak latency observed in the present study compared to the above-mentioned studies might origin from the small magnitude of change (two semitones) used in the current study. Previous studies have shown that neural generators of MMN evoked by small and large frequency changes are different and there is more in common between neural generators of N1 and MMN for large changes compared to small ones (Nagy et al., 2003;Opitz et al., 2002). So, opposed to Nagy et al. (2003) and Tsolaki et al. (2015) (Barrett & Fulfs, 1998;Kasai et al., 2002;Matsubayashi et al., 2008 ). For example, Aaltonen et al. (1994) found longer latencies for phonetic MMN in females than in males but Kasai et al. (2002) showed no difference in MMN amplitude and peak latency between the two genders using duration deviants. On the other hand, Schirmer et al. (2005) and Barrett and Fulfs (1998) reported larger MMN amplitudes in females compared to males with no latency differences between two sexes using emotional voice and intensity deviants, respectively.
These inconsistent findings might be explained by the fact that various types of deviants activate separate MMN neural networks (Escera et al., 2002;Molholm et al., 2005;Toufan et al., 2016), which can be affected differently by gender.
Altogether, our results showed differential effects of gender on MMN evoked by the violation of simple versus pattern acoustic regularities. The implications of these findings could be helpful for the interpretation of MMN results in healthy and/or clinical populations.
Furthermore, our findings provide preliminary evidence for the claim that gender might distinctly affect various types of MMN. More research is needed to further explore MMN gender differences by various stimulus designs.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.