OXTR Gene Polymorphisms and Event-Related Potentials in Humans: A Systematic Review

Abstract

Oxytocin receptor (OXTR) gene polymorphisms have been consistently associated with humans’ differences in sensitivity to social cues, social cognition, stress response, and brain activity. However, how social and affective neural processing differs across carriers of distinct OXTR gene polymorphisms remains unclear. This systematic PRISMA review is the first to examine the experimental literature on the relationship between OXTR polymorphisms and ERP components. Eight studies published between 2014 and 2019 were included. The rs53576 was the only OXTR gene polymorphism analyzed in all studies. The OXTR genetic variation explained significant changes in N1, P2, N2, P3, and late positive potential (LPP) components during social perception and empathy for pain tasks. OXTR genotypes were not related to P1, N170, N3, or any neural activity after 600 ms. The discussion is focused on the influence of OXTR genetics on neural processing, the development of brain neural networks implicated in social and emotional skills, cultural neuroscience of the oxytocinergic system, and methodological issues of this field. In conclusion, the evidence supports the hypothesis that genetic variations of the OXTR significantly influence neural activity related to emotional and social processing, except for the early phases of face recognition.

1. Introduction

Oxytocin is a neuropeptide that is implicated in relational phenomena such as maternal behaviors, social bonding, emotional communication, caring for others, and social cognition [1, 2]. Further, thanks to evidence from research involving nonhuman animal models [3], intranasal administration [3, 4], developmental psychopathology [5], genetic variations [6], and translational approaches [7], we now know that oxytocin supports the development of socioemotional skills such as perception of social cues, biobehavioral synchrony, regulation of emotions, prosocial concern, perspective taking, and empathy.

In the last decades, researchers have grown aware of the importance of oxytocin receptor (OXTR) gene polymorphisms in explaining individual and demographic variations in the development of social behaviors [6]. All OXTR polymorphisms identified thus far are single nucleotide polymorphisms (SNP), with the most studied SNP being a guanine (G) to adenine (A) substitution in the locus rs53576. These SNPs have been repeatedly associated with differences in sensitivity to social cues, regulation of stress response, social cognition, and brain activity during social tasks [8, 9]. However, some recent research has failed to find differences in emotional traits between carriers of the alleles for this polymorphism. However, some recent research has not found differences in emotional traits between carriers of the alleles of this polymorphism [10], which raises the question of whether genetic variations in the OXTR gene lead to variations in the brain’s processing of socioemotional signals.

Event-related potentials (ERPs) may lend a hand to this endeavor. ERPs are used as a reliable and safe method to examine the human neurophysiological activity related with psychological processing [11]. ERPs are changes in the electroencephalogram of a person, which are time-locked to cognitive, motor, affective, and social events presented during rigorously controlled tasks. ERPs require meticulous experimental control, making it easier to establish a fine-grained analysis of the time dynamics of the brain during the processing of specific aspects of stimuli, which can limit the establishment of broader functional analyses. However, it is justified to inquire about the ERP as it is one of the methodological approaches with the most outstanding reputation within cognitive neurosciences since it allows consolidating in greater detail how our brain constructs psychological phenomena [12].

ERPs are analyzed in terms of components that can be studied as a window into the neurophysiological mechanisms associated with ongoing mental activity. These components are built by using the polarity, latency, and scalp distribution of ERP waveforms and by considering the specific conditions of each experimental task [13, 14]. For example, in socio-affective tasks, frequently analyzed components include N170, early posterior negativity (EPN), and late positive potential (LPP). The first is an inferior negative wave, appearing approximately 170 ms after a face is shown; the second is occipital-parietal activity occurring around 200 ms after the display of images with affective content; and the last is anterior brain activity appearing shortly after 500 ms following the perception of arousing stimuli.

Accordingly, it is expected that carriers of distinct OXTR polymorphisms exhibit differences in neural processing, as measured by ERPs, during diverse psychological tasks. However, a systematic review of the studies that have examined the relationship between OXTR polymorphisms and ERP components has not been conducted so far. Such a review is valuable because it will allow for a better understanding of how genetic variations in the oxytocinergic system can lead to neurophysiological and behavioral differences. Additionally, it will be useful in guiding future investigations by identifying points of interest and questions to be answered. Therefore, our main aim is to review systematically the literature that has studied the relationship between OXTR polymorphisms and ERP components elicited during psychological tasks.

2. Method

This review was performed in accordance with the 2020 PRISMA guidelines [15]. Figure 1 shows a summary of all the procedures that led to the final selection of reviewed reports. Initially, 473 documents were found. Of them, eight studies fulfilled all selection criteria and were reviewed. Described below is each stage of the review process.

Figure 1.

This PRISMA flow chart depicts the stages of the review process.

2.1 Eligibility criteria

Eligible studies had to be scientific reports that inquired into the relationships between polymorphisms of the OXTR gene and ERPs components in human participants. All reports are needed to describe genotyping procedures, EEG acquisition procedures, ERP analyses, and statistical procedures. Considered studies had to be published or be in press anytime until April of 2023 in the English or the Spanish languages. No setting or time frame restrictions were imposed. Case series, case reports, conceptual literature, and literature not accomplishing the above criteria were excluded.

2.2 Search strategy

Identification, screening, and inclusion of studies were performed following the 2020 PRISMA guidelines (Figure 1). We used the Web of Science search engine to search the Web of Science Core Collection, Medline, the BIOSIS Citation Index, the Korean Journal, the Russian Science Citation Index, and the SciELO Citation Index databases. The search keywords were obtained from previous reports in the OXTR and ERP fields and from the MeSH thesaurus. These keywords were grouped into two lists. The OXTR list was composed of the terms [OXTR], [oxytocin receptor], [oxytocin polymorphism], [oxytocin single-nucleotide polymorphism], [oxytocin single nucleotide polymorphism], and [oxytocin SNP]. The ERP list was composed of the terms [event-related potential], [event related potential], [ERP], [electroencephalography], [EEG], [evoked potential].

Database searches were performed using the following query strategy: ([“OXTR list term 1”] OR … [“OXTR list term n”]] AND [[“ERP list term 1”] OR … [“OXTR list term n”]). A less restrictive query string (without enclosing the terms between quotation marks) was also used to confirm no relevant document was left out: ([OXTR list term 1] OR … [OXTR list term n]] AND [[ERP list term 1] OR … [OXTR list term n]). The terms had to be mentioned in the titles, abstracts, or keywords. Finally, we set the parameters of the searches to exclude books, conference reports, case reports, and editorials.

2.3 Studies selection and data collection

We read the titles and abstracts of all identified articles for evidence that they could meet the eligibility criteria. The reports that seemed to meet such criteria or could not be regarded as not meeting such criteria were further screened in full to verify inclusion criteria. Once fully scanned, the reports that indeed met the eligibility criteria were included in the review. Next, each article was read thoroughly, and relevant data was copied into a worksheet to be grouped with similar data from other studies. Variables for which data were extracted sought can be organized into five groups: report characteristics, sample characteristics, OXTR genotyping, ERPs and paradigms, and main findings (Table 1).

Paper	Country	Sample size	Participants	OXTR; SNPs	Genotypic groups	EEG sensors	Components	ERP paradigm	Main result
Sjaarda et al. [16]	Canada	167	ASD children and their parents	rs2254298, rs53576, rs7632287, rs1042778	GG, GA, AA	128	P1, N170, P300, N400	Face processing	Nonsignificant results.
Slane et al. [17]	USA	48	School children	rs53576, rs237897, rs1042778, rs2254298	GG, GA, AA	32	N170	Face processing	Nonsignificant results.
Choi et al. [18]	Japan	88	Healthy adult students	rs53576	GG, GA, AA	64	N1, N2, LPP	Affective image processing	Higher N1 and N2 amplitudes to human affective images in GG carriers than AA carriers.
Munk et al. [19]	Germany	150	Young adults	rs53576	GG vs. A+	32	N170	Face processing	Shorter N170 latency to upright angry faces in the right hemisphere of A carriers, but not in GG carriers.
Luo et al. [20]	China	48	Healthy adult students	rs53576	GG vs. AA	64	N1, P2, N2, P3	Empathy for pain	Higher P3 amplitude to sadistic painful faces in GG carriers than AA carriers.
Luo et al. [21]	China	50	Healthy adult students	rs53576	GG vs. AA	62	N1, P2, N2, P3	Empathy for pain	Higher P2 amplitude to suffering ingroup faces in, but not in AA carriers.
Peltola et al. [22]	Finnish	94	Healthy mothers and nonmother adult students	rs53576	GG vs. A+	21	N1, N170, EPN, LPP	Face processing	Shorter N1 latency to strong-intensity infant faces in GG carriers, but not in A carriers.
Fowler et al. [23]	USA	37	Young adults	rs53576	GG vs. A+	40	LPP	Affective image processing	Higher LPP amplitude to aversive images in GG carriers than A carriers.

Table 1.

Summary of relevant findings.

In the report characteristics, data was sought for publication year, country in which the study was carried out, study design, and general topic of research. In the sample characteristics, data was sought for sample size, age, sex, country, and clinical diagnoses. In relation to OXTR genotyping, data was sought for OXTR polymorphisms, genotyping method, tissue for DNA extraction, other genes under study, genotypic groups, allelic frequencies, genotypic frequencies, and Hardy–Weinberg equilibrium (HWE). Regarding ERPs and paradigms, data was sought for preprocessing procedures, characteristics of ERP components (number, family, latency range, number of electrodes, and region of interest), procedures used for ERP calculation, and experimental paradigms. Finally, the main findings were extracted and summarized. We focused solely on the results addressing ERP component comparisons between OXTR genotypic groups, so other results were not analyzed.

3. Results

3.1 Report characteristics

Eight studies were included in the final review; a summary of these papers is shown in Table 1. They were published between 2014 and 2019 in the English language. Two studies were conducted in the United States, two in China, one in Canada, one in Japan, one in Germany, and one in Finland. There was no study from South America, Central America, Africa, Eastern Europe, Asia (other than East Asia), or Oceania.

In general, researchers examined whether OXTR polymorphisms were related to neural differences during socioemotional processing (Table 1). Four studies examined the differences between OXTR genotypic groups in ERPs during face processing tasks; two studies looked for associations between ERPs elicited by perception of others’ pain tasks and OXTR genotypes; and the two remaining studies were concerned with the differences in neural activity between genotypic groups during emotional cue processing.

3.2 Sample characteristics

Sample size was 85 people on average, with only two investigations engaging over a hundred participants. Only one study mentioned its sample source, which was individuals with autism and their relatives [16]. Two studies recruited participants from both sexes [16, 17], two recruited only males [18, 19], three selected only females [20, 21, 22], and one did not state the sexes of the participants [23]. Most of the participants were young adults [75%], while two studies included school-aged children (7–12 years) [16, 17]. No sample included infants, toddlers, preschoolers, adolescents, or elderly (Table 1).

3.3 OXTR genotyping

Besides the OXTR gene, one study examined genes linked with monoaminergic neurotransmission, such as DR1, COMT, and SLC6A4 [16]; while another explored the CD38 [19], a gene implicated in oxytocin release. For the OXTR gene, polymorphisms rs53576, rs237897, rs1042778, and rs2254298 were studied, being rs53576 analyzed in all studies. Only one study explored the interactions between different OXTR SNPs and although they did not find any significant effect on ERP components, they did on social cognition measures [17].

The genotypic frequencies for the rs53576 variants across the reviewed studies are shown in Figure 2. Some studies did not report the genotypic or allelic frequencies [19, 20, 23], so we calculated these values from their HWE report. The three studies conducted with Asian participants revealed lower GG genotype frequencies, while two of the five studies with Caucasian participants found a greater ratio of GG homozygotes. The three remaining studies did not satisfy the HWE probe, pointing to a large proportion of heterozygotes (Figure 2).

Figure 2.

Report of genotypic frequencies for rs53576 polymorphism in each study. GG = carriers of the GG genotype; GA = carriers of the GG genotype; AA = carriers of the AA genotype. *Studies that did not satisfy the HWE.

3.4 ERPs and paradigms

We detected large discrepancies between the ERP procedures used across the studies. EEG acquisition and ERP preprocessing protocols varied widely in terms of sensor number (Table 1), referencing, filtering, artifact detection, artifact correction, and epochs selection. For example, the four studies that ran algorithms to correct artifacts did not justify their decision. EPR construction was also heterogeneous across the reports. Indeed, the reports generally provided little justification about the procedures that led to electrode number reduction, selection of latency ranges, and potential measurements. Most authors remarked that they selected latency ranges by means of visual inspection of the grand average waveform, and even one study specified a range for each participant [16]. Therefore, latencies for each family of components varied greatly (see component lines in Figure 3). Regarding the ERP measures, four studies calculated the latency peak and the amplitude peak for each latency period [16, 17, 19, 22], while the other four studies used the amplitude average for each latency window [18, 20, 21, 23]. All paradigms were time-locked with the stimuli, meaning no paradigm was time-locked with the responses. All studies employed visual affective or visual social stimuli; four used faces, two used affective images, and two used pictures of people suffering (Table 1).

Figure 3.

Differences in ERP components across carriers of distinct OXTR rs53576 polymorphisms. At the top, the heads display the brain regions where components were recorded. Their colors represent a family component: blue = early, red = middle, and orange = late. At the bottom, the leftmost column contains the studied components. The middle column shows latencies in milliseconds, where lines represent the components examined in their respective paper and their lengths indicate the latency range used to calculate the component. Black dotted lines represent nonsignificant differences between genotypes, whereas colored solid lines point out significant differences. Doble lines indicate a higher component amplitude in GG carriers, while single lines denote a higher component amplitude in A-allele carriers. The rightmost column shows the source study.

3.5 Main findings

Figure 3 summarizes the ERP component amplitude differences between OXTR rs53576 genotypic groups, as reported by the reviewed studies. On the one hand, OXTR genotypes were not significantly associated with the P1 or the N170 components during the socio-affective tasks. Similarly, neuronal responses occurring after 600 ms were not explained by any genetic variation of the OXTR. On the other hand, the studies together found the N1, P2, N2, P3 and LPP potentials to have a significantly higher amplitude in GG carriers during the experimental tasks. Such genotype-dependent components can be grouped into three general time ranges of the neural dynamics (colored heads in Figure 3). The earliest differences in neural activity were found between 50 and 200 ms in centro-parietal and frontal zones (respectively N1 and P2; blue head). The OXTR also affected middle neural processing in the latency range between 200 to 350 ms in posterior regions, specifically, the EPN component (red head). The last differences in neural activity were observed in the central regions of both frontal and parietal lobes in latencies between 300 and 500 ms (P3 and LPP; orange head).

Face processing was generally not different across OXTR genotypes. Specifically, amplitudes of neural responses to upright or inverted faces did not vary significantly in healthy adults [19], mothers [22], children [17], or autistic children [16]. However, a significant interaction was observed in the N1 latency. Peltola’s team found that, unlike A carriers, GG carriers had a shorter N1 latency to strong-intensity infant faces than they had to mild-intensity ones [22]. Likewise, Munk et al. [19] found a significant interaction in the N170 latency, where A-allele carriers had shorter latencies in the right hemisphere in reaction to upright angry faces. Nevertheless, such a difference was not observed in a replication sample [19].

The processing of affective images varied in function of the rs53576 genotypes. GG homozygotes exhibited a more negative posterior N1 (50–200 ms) during the perception of affective images of humans and objects and a more negative posterior N2 (200–320 ms) when they perceived affective images of humans only [18]. In the same study, no effects of OXTR genotypes on LPP (600–1000 ms) were found. However, Fowler and colleagues found independently that GG participants responded with a higher parietal LPP amplitude (at 300–500 ms) during the perception of aversive pictures. That said, it should be noted that the LPP latencies and regions of interest were considerably different between the two studies, making it difficult to compare the results (Figure 3).

Neural responses to the suffering of others were significantly mediated by OXTR genotypes. Specifically, GG carriers exhibited a heightened frontal P2 amplitude (at 136–176 ms) when they were shown painful expressions adopted by people that the participants perceived as themselves [21]. Additionally, in a previous study [20], frontal P3 was found to be more positive for sadistic painful stimuli in the GG group, which may indicate a bigger excitatory activity in the amygdala and insula.

4. Discussion

4.1 OXTR polymorphisms and neural processing

Our results provide evidence that variations in the OXTR gene are associated with differences in brain activity during socioemotional tasks (Table 1). These differences were distributed along several time intervals and brain regions (Figure 3). In particular, the main differences were found in rs53576 GG homozygotes, who produce wider neural potentials in the face of salient social stimuli such as people expressing pain.

GG homozygotes exhibited more negative early activity (N1) in posterior cerebral regions when they perceived socially salient cues (Figure 3). Psychophysiological studies have previously linked enhancement of the posterior N1 with an attentional shift to prominent stimuli and implicit visual discrimination [24, 25]. Therefore, GG carriers may detect and recognize socio-affective signals more readily. This enhanced activity may originate in interneurons in the occipital, parietal, and temporal areas implicated in social perception. Additionally, it is possible that these areas are excited by afferences from projection neurons in the limbic area, which is rich in OXTRs.

GG homozygotes also had stronger early anterior positive potentials (P2) when they watched people suffering (Figure 3). This early frontal excitation may be associated with the involuntary allocation of attentional resources to the processing of the highly arousing images [26, 27], suggesting GG carriers may bear neural mechanisms that enable them to have a more sensitive perception of emotional signals from others.

Moreover, OXTR SNPs produced variations in the N2-EPN potentials in posterior brain regions, with GG homozygotes having wider negative potentials to affective human pictures and AG-AA carriers doing so to affective pictures without humans (Figure 3). The N2-EPN component is a typical index of attentional engagement to emotional pictures [28, 29]. It may be generated from interneuron activity in brain regions such as the amygdala, the hippocampus, and the superior and inferior parietal lobes [30, 31], which contain an abundance of OXTRs [32, 33]. Therefore, it seems that OXTR rs53766 polymorphisms influence the activity of posterior corticolimbic networks, which enhances GG carriers’ and AA-AG carriers’ discrimination and categorization of social cues and nonsocial arousing stimuli, respectively.

The late components modulated by the OXTR rs53576 polymorphisms were positive potentials at frontal (P3; Figure 3) and parietal sites (LPP; Figure 3). Both ERP components were higher in GG carriers when they perceived pain faces and aversive pictures. The P3 is usually interpreted as an indicator of effortful decision-making during demanding tasks [34, 35], as was the case in Luo’s experiment [20]. The LPP component is elicited by highly arousing stimuli and has been associated with enhanced memory encoding, appraisal, and suppression of response to affective pictures [36, 37]. These findings may indicate that GG homozygotes’ greater sensitivity to social images allows them to execute enhanced top-down processes to evaluate arousing social contexts and to control their responses in accordance with such contexts. However, these interpretations are necessarily preliminary, considering that six of the reviewed studies failed to find any significant differences in the LPP component between OXTR genotypes.

Overall, these findings are consistent with the social salience hypothesis, which contends that an increase in oxytocinergic neurotransmission influence neural activity in such a way that the processing of social information is enhanced [38, 39]. Animal research has proven that SNPs in the noncoding region of the OXTR gene have an impact on the number of OXTR receptors in brain regions implicated in social behaviors [40]. Likewise, neuroimage studies have confirmed that OXTR SNPs are associated with variations in brain activity in areas responsible for the processing of social information [8, 41]. In brief, GG homozygotes for the rs53576 SNP may have a higher density of receptors in specific critical areas, which may enhance oxytocinergic neurotransmission in neural networks implicated in the processing of social cues [40].

The null results on the face-specific N170 component are surprising because there exists a long-standing theoretical association between oxytocinergic neurotransmission and face processing [42]. Indeed, Skuse et al. found a link between the OXTR rs237887 alleles and face recognition in families with autistic children [43]. Moreover, fMRi studies have also found significant effects on face recognition involving OXTR SNPs. For example, the Westberg team reported that rs7632287 genotypes differ in recognition of faces and amygdala activity [44]. Similarly, O’Connell and coworkers linked the rs2268498 SNP with inferior occipital gyrus activity due to perception of fear expressions [45]. Nevertheless, these results should be analyzed cautiously given that the Skuke study used a very particular sample and the fMRi studies used tasks sensitive to other psychological functions such as emotional processing and mental inference. What is more, the lack of association between OXTR SNPs and the N170 component, is in line with the nonsignificant relationship between many OXTR SNPs and several face recognition tasks, which was found in an exhaustive study by [46]. Therefore, this evidence together seems to indicate that oxytocinergic neurotransmission is not essential to the early stages of discrimination of facial configurations.

Furthermore, these results could help to understand the conflicting findings on the relationship between genetic variations in the oxytocinergic system and human psychological phenotypes [10]. Variations between genotypes in neural processing during exposure to social cues with high emotional content observed during intermediate and late latencies (100–600 ms) indicate that the oxytocinergic system is central in facilitating the implicit processing of emotional signs during social interactions. Still, this system would have a less relevant and direct role in the awareness of one’s affective states; therefore, a minimal effect could be expected when self-report questionnaires or verbal responses are used. In this sense, future studies would benefit from using experimental tasks instead of questionnaires and psychological tests to assess social and emotional phenotypes.

4.2 OXTR polymorphism, neural functioning, and human development

OXTR polymorphisms influence brain development by modulating the density of OXTRs, by modifying the sensibility toward the social environment, and by orchestrating fine-tuned social transactions during critical periods of brain maturation. Consequently, OXTR rs53576 GG individuals tend to be more sensitive to their social environment, developing larger phenotypic variability [47]. For instance, in G allele carriers, protective and synchronic caring favors the development of better emphatic, prosocial, and emotional skills, while childhood adversity leads to more avoidant behaviors and poor social skills. In contrast, A carriers, who are less sensitive, show fewer developmental variations [48, 49]. Moreover, effects of OXTR SNPs on the modulation of developmental plasticity have also been found in neuroimage studies, where limbic and frontal networks have been shown to be more plastic [8, 50, 51]. Plasticity in fronto-limbic networks is consistent with the differences in the ERP components observed in this review.

The developmental plasticity associated with some OXTR genotypes may be due to epigenetic mechanisms. Indeed, numerous studies have reported an association between OXTR DNA methylation and differences in social cognition, emotional behaviors, and neuroendocrine functioning in several moments of human and animal development [52, 53, 54, 55]. A recent systematic review found that an increase in OXTR gene methylation was linked with a reduction in receptor expression, social sensitivity, and developmental plasticity, leading to poor social skills and affective dysregulation in healthy and psychopathologic samples [56]. A leading hypothesized mechanism is that people with more G alleles have more CpG islands, facilitating epigenetic modulation and major developmental plasticity, with early adverse experiences as the main predictor of OXTR hypermethylation.

4.3 OXTR polymorphism, neural functioning, and cultural differences

Figure 2 shows a large variation in the allelic frequencies of the OXTR rs53576 SNP between Caucasian and Asian samples, which is in line with previous findings that Asian populations have higher A-allele frequencies [57]. This geographic distribution of OXTR genotypes is associated with cultural patterns, including collectivistic values, control of emotional expressions, emotional support seeking, social interdependence, empathy for pain, altruism motivation, prevalence of depression, and brain functioning [58, 59, 60]. These cultural and genetic differences may have been shaped in human societies throughout history. Selection for A alleles in collectivistic nations could be the result of son favoritism, prolonged infanticide, and marriage patterns, whereas G allele accumulation may have been facilitated by mothers’ investment in childcare and male cooperation in everyday life activities.

These cultural and demographic factors should be considered to better interpret neural functioning differences. In the first place, genotypic frequencies may limit statistical analysis. For example, since Caucasian samples have fewer AA carriers, such studies require larger samples to find significant results. Further, as cultural values and relational tendencies have important effects on various genotypes, it is indispensable to include measures for these variables. Finally, samples that do not satisfy the HWE must be carefully analyzed because the sampling could be biased, or evolutionary pressures could be affecting the genetics of these populations [61].

4.4 Limitations and future directions

4.4.1 Samples

Three important sample issues can be identified. First, no studies included participants from South America, Central America, Africa, Eastern Europe, the Middle East, or Oceania. There exist important variations in the allelic frequencies between these regions, which may be related to differences in behavior and brain function. Therefore, generalization of these results to diverse geographic and cultural regions is limited, such as in Conner et al. [10] report, who did not find an association between OXTR rs53576 SNP and emotional trait, but they only included a sample from New Zealand. Second, studies have focused on young adults, and there are no inquiries on infants, adolescents, or elderly. Since adolescence is considered a sensitive period for developing social skills [62], the lack of studies at this age range restricts our understanding of how OXTR polymorphisms influence the development of brain activity and social behavior from childhood to adulthood. There is evidence that during adolescence OXTR polymorphisms produce high social sensitivity, opening a critical period to rewiring brain networks and reorganizing behavior, as the oxytocin system interacts with pubertal hormones to create age- and sex-dependent developmental trajectories [63]. And third, the small sample sizes used in the studies limit the possibility of finding significant results. Guidelines to ERPs recommend including over 40 participants per condition [64], meaning researchers may need to incorporate samples of more than 120 participants to analyze the effect of variables such as sex, age, and other polymorphisms. Finally, larger samples are necessary to get lower p-values and more marked effect sizes, which are ideal conditions to reproduce results and consolidate this scientific field, although if large samples of different populations and cultural contexts are included, so much genetic and phenotypic variation can be added that it can be challenging to get satisfactory results.

4.4.2 Sex differences

None of the studies showed sex-dependent effects of OXTR polymorphisms on the ERP components, which was unexpected considering previous findings involving OXTR polymorphism–sex interaction effects on behavior and brain function [65]. For the OXTR rs53576 SNP, there is evidence of sex-dependent effects on the volume of limbic structures, such as the hypothalamus and the amygdala [41], and on functional connectivity of the prefrontal cortex [66, 67]. Furthermore, it has been asserted that sex-dependent differences in the development of brain function and social behavior across different OXTR genotypes are likely linked to the role of sexual hormones regulating OXTR expression [68, 69]. It is likely that, as sample sizes increase, sex-dependent effects will be detected.

4.4.3 ERP procedures

As all studies were interested in behavioral tasks tapping social or emotional processing (Table 1), there were no inquiries about ERP components for other psychological functions such as object perception, attention, memory, language, executive functioning, and motor control. This is a major gap to fill, as behavioral and fMRI studies have shown that OXTR polymorphisms may partly explain the differences in the development of these cognitive processes [6, 70].

The experimental paradigms used in the studies hinder comparison between ERPs components such as P3, Nc, LPP, slow wave potentials, and error-related negativity among OXTR genotypes. In relation to EEG acquisition and preprocessing, we detected wide variation in the number of electrodes, referencing, filtering, sampling rate, artifact detection, elimination, and correction, which makes it hard to compare ERPs results [71]. In general, the reports lacked substantial explanations about the selection of ERP measurements, latency ranges, electrode reduction methods, and statistical procedures. Most researchers used visual inspection of the grand average to select latency ranges, employed amplitude, and latency peaks and chose electrodes autonomously to calculate each component, being all these procedures discouraged by ERP guidelines [27, 72].

4.4.4 Publication bias

Our objective was to carry out a systematic review but not a meta-analysis because of the heterogeneity in ERP procedures, psychological paradigms, statistics for hypothesis testing, and number of subgroups compared. All these make very difficult to run quantitative analyses for publication bias, sensitivity, and subgroups. However, in the exploratory analysis we did not find publication bias solid evidence; three of the eight studies report negative results with nonsignificant differences between OXTR genotypes. Moreover, of the 23 components analyzed (Figure 3), seven [30%] showed significant differences between genotypes, and in the 70% of comparison, the differences were nonsignificant. However, small sample sizes (Table 1), small effect sizes, and high p-value (<0.05) in two studies with positive results could mean a possible bias in these publications [22, 23]. In the future, editors could demand higher effect sizes and narrower confidence intervals in the reports of quantitative results, as suggested in the APA guidelines, which may facilitate the comparison of studies.

In short, future studies will benefit from the use of larger samples, more heterogenic aged populations in the samples, and the recruitment of participants from different geographical places around the world. Additionally, it will be helpful for studies to include sex or gender as a control variable, especially when the samples include adolescents. Also, it may be advantageous for the field to demand that studies use reliable, conventional, and standardized protocols to acquire and process EEG signals. Accumulating evidence from future studies will help to establish, refute, and clarify how EPR components function as a window into the oxytocinergic processing of human psychological functions.

5. Conclusion

In this systematic review, we have found suggestive, preliminary evidence supporting the hypothesis that the rs53576 polymorphism of the OXTR gene significantly influences ERP components elicited during socioemotional tasks. Prominently, larger N1, P2, EPN, P3, and LPP components in GG homozygotes seem to be associated with an increase in the sensitivity to salient social cues in social and emotional situations. Moreover, genetic variations in the OXTR do not affect the neural activity during the earlier moments of perceiving faces.

Acknowledgments

We thank the Universidad de los Andes and the Universidad Externado de Colombia for their support in the publication of this work.

Conflict of interest

The authors declare no conflict of interest.