Neural activation to infant cry among Latina and non-Latina White mothers

Abstract

Cultural neuroscience is an emerging framework positing that culture (for example, values, beliefs, practices, and modes of emotional expression) critically informs socialization goals and desired behaviors, which are perhaps accompanied by differential patterns of brain activation. Using fMRI, the current study examines brain activation to infant cry stimuli and matched white noise among 50 first-time biological mothers identifying as Latina or White in the United States. Results showed that brain activation to infant cries in the right posterior insula, left cerebellum, and left auditory were higher for White mothers compared to Latina mothers, p’s < .05. White mothers showed greater activation to cry sounds compared to white noise in the right dorsolateral prefrontal cortex, left somatosensory, right and left premotor cortices, p’s < .05, whereas Latina mothers did not. These brain regions are involved in motor planning, movement, sensory processing, and social information processing. It is important to note that mothers in the two groups did not show differences in stress and behavioral parenting measures. Therefore, Latina and White mothers differentially recruiting brain regions related to infant parenting behaviors indicates the potential role of cultural context in shaping patterns of neural activation. Our exploratory analysis suggests that this difference might be due to greater pre-parenting exposure among Latina mothers to children compared to White mothers. Taken together, although our data did not completely explain the differences in brain activation between groups, findings suggest potential culture-related influences in brain activation occurring in the postpartum period.

Culture is often constructed as shared meanings and shared activities or values, beliefs, and practices that are transmitted through social contexts and behaviors. Cultural neuroscience is a discipline that acknowledges the rich ways in which culture is embedded across social contexts and thus shapes the brain (for a theoretical overview, see Kitayama & Uskul, 2011). Emerging evidence indicates that cultural context might shape neurobiological processes underlying cognitive and emotional responding as well as behaviors in social relationships (de Greck et al., 2012; Kim & Sasaki, 2014). As the human brain is built to adapt to contextual and social goals, attending to cultural ideals that individuals are socialized toward is important. For example, cultures emphasizing social hierarchy tend to have higher empathy toward other people. One neuroimaging study on individuals from Korea showed greater neural activity in an area involved in theory of mind and empathy, the temporoparietal junction, when viewing people in painful scenarios compared to individuals from the United States (Cheon et al., 2013).

Culture also organizes parenting beliefs and behaviors (Bornstein, 2012). When parents interact with children and direct them towards desired behaviors, they do so to socialize children toward an ideal; these ideals for child development vary across cultures (Keller & Kärtner, 2013). Different socialized cultural values may include individualism or collectivism, hierarchy or egalitarianism, orientation toward short- or long-term goals, and masculinity or femininity (Hofstede, 2001). For example, one study examining parental behavioral responses to infant non-distress vocalizations and nonverbal signals across six sociocultural contexts found that contingent responses to different infant senses (e.g., sight, touch, hearing) varied across cultural groups (Kärtner et al., 2008). Although hearing senses of infants were stimulated the most by parents in all cultures, parents living in an interdependent context (e.g., Nso in Cameroon) stimulated hearing and touch similarly compared to parents from more independent contexts (e.g., Berlin) who put more emphasis on face-to-face stimulation. Additionally, endorsing relational versus autonomous socialization goals explained a significant portion of the variance in contingent responses, suggesting that patterns of parental responses to infant cues are modulated by socialization goals. Given the importance of cultural context on parental behavior, it is important to begin examining how culture may shape the neurobiology of new parents to understand links between neurobiology and behavior. We utilized an existing neuroimaging study on maternal activation to infant cries as an initial proof-of-concept of how different cultural contexts may potentially shape differences in brain responses to infant stimuli; however, as the initial study was not designed to examine culture, we used a racial/ethnic grouping variable (along with acculturation among Latina mothers) as a helpful but imperfect proxy for culture.

1.1. Brain adaptations in the postpartum period

The postpartum period brings adaptations in the parental brain that are related to both environmental and biological forces (see Kim, 2021 for a detailed review). The postpartum period is associated with enhanced connectivity between motor, sensory and auditory brain networks (Rutherford et al., 2020) as well as subcortical structures such as amygdala and anterior insula (Paul et al., 2019). For example, both biological and adoptive parents show similar functional connectivity in empathy networks including the anterior insula when they watch themselves playing with their own infants (Abraham et al., 2018). Parenthood is also thought to increase the connectivity between subcortical and cortical structures that is related to increases in parental motivation to approach infants, understand their signals, and plans to respond to them (Feldman et al, 2019). Similarly, infant cues are associated with increased activation in the medial and lateral prefrontal cortex and anterior cingulate cortex, regions that are related to emotion regulation and cognitive control (Barrett & Fleming, 2011; Rutherford et al., 2015).

Task-based fMRI techniques can indicate neural activation patterns that are relevant to human caregiving behavior by presenting caregiving-relevant stimuli. One paradigm that has advanced our understanding of parental brain function is presenting parents with infant cry sounds during magnetic resonance imaging, as cries are a very relevant cue signaling infant needs. Infant cry has been robustly associated with higher activation in cortical (e.g., premotor, auditory, medial orbitofrontal cortex, and dorsomedial and lateral prefrontal cortices) and subcortical (e.g., anterior insula, anterior thalamus, and posterior cingulate gyrus) regions among parents compared to non-parents (Rilling, 2013; Witteman et al., 2019). Collectively, these activation differences manifest as an integration in the sensorimotor network (i.e., auditory cortices, posterior and dorsal anterior insula, the right putamen, and pre- and postcentral gyrus). First-time mothers in predominantly White samples have been shown to exhibit activations in midbrain, basal ganglia, cingulate, amygdala, and insula in response to infant cries versus white noise (Lorberbaum et al., 2002). Compared to non-parents, parents in these samples show higher activation in the right amygdala, the right middle cingulate cortex, and bilateral insula in response to infant cry versus white noise (Seifritz et al., 2003).

Growing evidence indicates that brain activity is shaped by experiences that are related to adapting parenting roles and one potential factor contributing to these differences may be culturally embedded practices. Although regions associated with speech and auditory stimulation (e.g., supplementary motor area, inferior frontal, and superior temporal regions) were found to be activated among first-time mothers across cultures (Bornstein et al., 2017), potential cultural differences in brain activation patterns in response to infant stimuli are less clear. Examining how cultural contexts modulate differences in brain responses to infant cry sounds offers an opportunity to explore cultural influences on neurobiology. Addressing this gap can further our understanding of neurobiological systems of parenting and how they are shaped by interactions with the environment. Studies conducted in Western countries illustrate the need for including racially and ethnically diverse populations to balance the disproportionate number of studies conducted with predominantly non-Latinx White samples. Addressing this gap can further our understanding of neurobiological systems of parenting and how they are shaped by interactions with the environment. There is evidence showing that ethnic identity can potentially account for other contextual differences seen in cross-nation comparisons of maternal brain responses to infant stimuli (Bornstein et al., 2017). Hence, we focus on how brain activation might differ between Latinx and non-Latinx White first-time mothers living in the United States (US).

1.2. Considering differences in parenting shaped by Latinx and non-Latinx White culture

Latinx¹ culture has a higher collectivist orientation compared to White² culture that highlights individualistic features (Hofstede, 1991). Regarding parenting experiences, Latinx culture exhibits a more interdependent structure compared to European culture, which emphasizes independence in caregiving practices (Leyendecker et al., 1995). Demonstrated behaviors may shift depending on cultural values. For example, living in a multigenerational environment with non-biological children and division of labor in child rearing (Aubel, 2012; Guendelman et al., 1994) result in increases in alloparenting, or the care of offspring by individuals other than their biological parents. Indeed, Latinx youth have more experiences providing caregiving relative to White Americans (Siskowski, 2006; East & Weisner, 2009). Taken together, cultural differences in perceptions of caregiving and behavioral practices may reasonably be reflected and supported by differences in neural activation to infant signals.

1.3. Other factors that may systematically vary between groups

1.3.1. Prior caregiving experiences.

Prior experience with infants might reduce brain activation during the transition to parenting due to reduced preoccupations with caregiving, which tends to increase in the first months following birth for new mothers (Kim et al., 2013). Parental thoughts and actions may drive activation differences in sensory motor and social cue processing regions, such as the left superior temporal gyrus in response to infant cry sounds (Kim et al., 2015). Indeed, parental brain activation in response to infant cues has been found to be lower among multiparous mothers compared to primiparous counterparts (Bunderson et al., 2020).

1.3.3. Stress exposure.

Emerging neuroimaging literature suggests that stressors can affect adaptive neural changes seen in the transition to parenting (Kim, 2021). Relevant stress exposures can occur in the past, such as parental experience of adverse childhood events (e.g., divorce, incarcerated parent, unstable caregivers), and in the present. Current increased exposure to socioeconomic, environmental, and psychosocial stressors can decrease brain activation seen in circuits relevant to emotional and social information processing (Kim et al., 2020). Relevant to the current study, epidemiological studies indicate that Latinx Americans, on average, experience higher levels of stressors (Myers, 2008) due to historical marginalization and systemic racism (Bailey et al., 2017). Both electrical brain activity and blood-oxygen-level-dependent (BOLD) signal studies show that mothers exhibit greater response to infant cues with higher anxiety symptom levels (Finnegan et al., 2021; Malak et al., 2015). Mothers with elevated perceived stress also showed higher activation in the right dorsolateral prefrontal cortex in response to aversive images and this activation was associated with parenting stress. However, it is unclear how the presence of stressors intersects with cultural factors to shape neural response patterns to infant stimuli among Latinx mothers: they may show lower neural activation in parenting-relevant circuits to the degree that stressors are high, or perhaps greater neural activation in reward processing regions if other-focused actions are more rewarding compared to self-focused actions (e.g., Telzer et al., 2010). In the current study, we focus on recent stressors, such as negative life events and income-to-needs ratio as a proxy for financial strain.

1.3.3. Parental mental health.

We elected to focus on maternal depressive symptoms as an important indicator of mental health, with prevalence rates of 17% worldwide and 13% in the US following birth (Shorey et al., 2018; United Health Foundation, 2021; Woody et al., 2017). The literature presents inconsistent findings on the brain activation patterns to infant signals among mothers who report elevated depressive symptoms. While there is evidence showing that mothers with greater depression symptoms show lower or no BOLD activation in response to infant cries in reward processing regions related to parental motivation (e.g., nucleus accumbens, caudate, and thalamus; Laurent & Ablow, 2012) or infant faces in regions related to motivation and regulation (e.g., dorsal anterior cingulate cortex. orbitofrontal cortex, insula, left prefrontal and insula/striatal; Laurent & Ablow, 2013), other studies documented positive relations between perinatal depression symptoms and functional activity in response to infant cues (Noll et al., 2012; Rutherford et al., 2016). Anhedonia may account for the positive association between reduced quality in parenting behaviors and increased activation to infant cries (Rodrigo et al., 2011). Given that postpartum depression rates are higher among Latinx populations compared to non-Latinx Whites (Ceballos et al., 2017), we explore how maternal mental health is potentially related to brain activation differences among Latina and White groups of mothers.

1.4. Study rationale

To our knowledge, only one study to date has examined brain activation among parents to infant stimuli across cultures, as indicated by nationality (Bornstein et al., 2017). The current study aims to extend this literature by investigating neural differences in maternal responses to infant cries between Latina³ and White groups within a single nation (US) while attending to relevant factors that may meaningfully differ between these groups and that have conceptual relationships with neural activation during the postpartum period.

We examined differences in neural activation to parenting-relevant stimuli between two racial/ethnic groups as a proxy for culture. We additionally examined how endorsement of cultural orientation (i.e., Anglo, Latinx) within Latina mothers related to brain activation patterns. Consistent with prior studies on the postpartum period, we hypothesized that prefrontal regions, sensory processing circuits, and emotion processing areas would emerge as significant regions of interest. The degree and pattern of brain activation within these regions of interest may vary as a function of cultural values endorsed due to different socialization goals and division of caregiving responsibilities across social networks. There is evidence showing both higher and lower brain activation in response to infant cries with increasing caregiving experience. While Maupin and colleagues (2019) found lower activation to infant cues with more children, Parsons et al. (2017) documented higher activation with increasing age of the child (and thus greater experience with that child, although this is also conflated with child developmental stage). Moreover, high levels of stress might be related to decreased activation (Kim et al., 2020), whereas anxious thoughts about child rearing can predict increases in activation (Finnegan et al., 2021). Therefore, given the dearth of parental neuroimaging literature examining cultural differences, we anticipate differences in brain activation between Latinx and White mothers; however, we do not have an expected direction of effects. Further, we explored whether broad patterns of activation differences were related to prior caregiving experience (as represented by non-biological children in the household), stressful life events, mental health, and cultural orientation.

Method

2.1. Participants

Eligibility criteria among mothers for this study were: speaking English, not having a pregnancy-related illness or an infant with a medical condition, no current or historical psychiatric/neurological illness other than depression or anxiety diagnoses, no psychoactive drug use, and no magnetic metal in the body.

The sample consisted of 59 first-time biological mothers. Those who did not identify as either Latina or White were excluded from the analyses (n = 9). The final sample had 50 participants (52% Latina; M_{mother age} = 26.11 years ± 5.75; M_{infant age} = 4.48 months ± 2.06). Most Latina mothers (70%) reported having parents who immigrated (i.e., were first-generation Americans) and 86% were born in the US. All non-Latinx White participants were born in the US. See Table 1 for further demographic characteristics of the sample.

Table 1.

Sample demographics.

	M (SD) or percentage
	White (n = 24)	Latina (n = 26)	Group Differences
Maternal age (years)	27.92 (6.37)	24.47 (4.63)	t(48) = 2.20, p = .03
Child age (months)	4.14 (2.04)	4.78 (2.06)	t(48) = 1.10, p = .28
Child sex (boy)	33%	54%	χ2(1, 50) = 2.13, p = .14
Highest grade completed	14.88 (2.54)	13.38 (2.33)	t(48) = 2.16, p = .04
Relationship status
Single	4%	12%
Dating	8%	4%
Engaged	8%	16%	χ2(4, 49) = 4.76, p = .31
Married	58%	32%
Long-term relationship or Common Law Marriage	21%	36%
Income-to-needs ratio	3.57 (1.98)	2.76(1.69)	t(48) = 1.55, p = .13
Maternal sensitivity	5.29 (1.21)	5.15 (1.29)	t(48) = 0.41, p = .69
Number of non-biological children in household	1.17 (0.48)	1.73 (1.25)	t(48) = 2.07, p = .04
Negative life events (count)	4.14 (2.85)	4.00 (2.64)	t(39) = 0.17, p = .87
Negative life events(valence)	2.39 (0.50)	2.55 (0.65)	t(47) = −.99, p = .33
Negative life events (count & valence composite)	10.89 (8.54)	11.08 (7.77)	t(39) = −.08, p = .94.
Depression symptoms	9.0 (5.45)	6.27 (4.64)	t(48) = 1.91, p = .06
Acculturation
Mexican Orientation	-	3.15 (1.51)	-
Anglo Orientation	-	3.86 (0.56)	-
Country of origin
United States	100%	86%
Mexico	0%	10%	χ2(2,40) = 2.93, p = .23
Venezuela	0%	5%

2.2. Procedures

Recruitment occurred via fliers and brochures distributed in the Denver metro area through the Women, Infant, and Children (WIC) program and Colorado state Prenatal Plus programs. Trained research staff visited eligible mothers at their home and administered questionnaires and interviews following written consents. They also recorded infants’ voice when they cried during the visit to use in fMRI visits. Mothers then conducted an fMRI visit three weeks after the home visit, on average. Mothers were reimbursed for their participation and offered childcare and transportation assistance.

2.3. Measures

2.3.1. Demographics.

Mothers self-reported maternal age, child age, years of completed formal education, relationship status, daily responsibilities (employment), and household income (See Table 1). Income-to-needs ratio (INR) was calculated by dividing mother reported household income over the relevant federal poverty threshold specified by the U.S. Census Bureau (2020), where a ratio of 1 or below indicates living at or below the poverty line. The average INR in this sample was 3.15 (SD = 1.86).

2.3.2. Maternal sensitivity.

Mothers and infants engaged in a naturalistic interaction for 15 minutes. Mothers were instructed to interact with the infant in a typical manner without using toys. Maternal sensitivity during this naturalistic interaction task was coded using the Emotional Availability Scales, 4th edition (EAS; Biringen, 2008). The EAS includes Sensitivity, Structuring, Non-Intrusiveness, and Non-Hostility subscales. The EAS has indicated high internal consistency in another study’s sample of Mexican American toddlers, α = .95, with no statistically significant differences in EAS dimension scores across Latinx, White, and Black mother-toddler dyads (Derscheid et al., 2019). In this sample, internal consistency across subscales was .86. Among those four subscales, we focused on the sensitivity dimension due to conceptual and statistical relations to parent-child attachment (Biringen et al., 2005). The sensitivity subscale is a global measure of parental behaviors and reflects the extent to which parents are affectively sensitive, accepting, clear and prompt in their interactions with their infants. Maternal sensitivity scores can range between 1 and 7, with 7 indicating a highly sensitive, 1 indicating a highly insensitive parent, and mid-range scores representing ‘apparent’ sensitivity (i.e., incongruent affect and behavior).

2.3.3. Number of non-biological children in household.

Mothers reported on the number of total children living in the household. As all participants were first-time mothers, this represents the number of non-biological children in the household.

2.3.4. Negative life events.

Mothers completed the 63-item Crisis in Family Systems (CRISYS) measure, which asks about several life stressors that occurred in the past 6 months (Shalowitz et al., 1998). These stressors include 11 domains, including legal, relationships, medical, and housing issues. Respondents dichotomously rate whether each event occurred, rate distress using a 1–4 Likert scale, rate whether it was positive, negative or neutral, and ongoing. Test-retest reliability for this measure has been estimated at r = .93 (Shalowitz et al., 1998). A negative life events count score, and its valence were calculated for each participant to correspond to the total number of events that were rated as negative and whether the events were difficult to go through. We also created a composite variable by multiplying total number of negative life events with their valence.

2.3.5. Maternal depression.

Mothers completed the Beck Depression Inventory, 2^nd edition (BDI II; Beck et al., 1996). Psychometric properties of the BDI-II suggest measurement invariance across Latinx and White American populations, with consistent factor structure and high internal consistency, α’s > .82 (Wiebe & Penley, 2005). Internal consistency of the BDI-II was previously found to be high, α = .90 (Wang & Gorenstein, 2013). Within our sample, internal consistency was .80.

2.3.6. Acculturation.

Latina mothers completed 12 items from the Acculturation Rating Scale for Mexican Americans, 2^nd edition (ARSMA-II; Cuellar et al., 1995) that assessed preferred language (e.g., “I enjoy speaking Spanish”) and friendships (e.g., “My friends are of Anglo origin”). The ARSMA was initially developed to focus on Mexican Americans, however, the ARSMA-II was revised to allow assessment among other Latinx cultural groups (Cuéllar et al., 1995) and it has become a widely used scale with mixed Latinx samples (Gamst et al., 2002; Unger et al., 2007) as well as Latina parents (Chapman & Perez-Escamilla, 2013; Halgunseth et al., 2006). Response options used a Likert scale from Not at all (1) to Almost Always (5). Items were averaged to create two subscales: Mexican orientation and Anglo orientation. Although internal consistency was excellent for the Mexican orientation subscale in our sample (α = .97), it was low for the Anglo orientation subscale (α = .45).

2.4. fMRI paradigm

Mothers heard four infant cry stimuli: (1) their own infant cry, (2) other infant cry, (3) white noise matched to their own infant cry sample, and (4) white noise matched to the other infant cry sample. Own infant cry samples were recorded as they naturally occurred during the home visit and largely related to being hungry, seeking attention, or a normal range of discomfort. The control infant cry was recorded from a non-participating mother whose infant cried during a diaper change. The own infant and other infant cry sounds were matched for volume. White noises were matched to the gross temporal envelope of the own infant and control infant cry sounds by generating a spectral average of the cry.

The task consisted of two functional runs. Blocks of cry stimuli and control sounds lasted 20 seconds and each stimulus block was separated by an average 10-second rest period during which only background scanner noise could be heard. Each run contained blocks of the four infant cry stimuli. The order of the blocks was randomized, and each block was repeated five times, thus a total of 10 times per condition. Block design is the most commonly used design for infant cry paradigms (see Swain et al., 2014) and allows for presentation of an extended cry, which may allow more time for parents to mentalize on how to care for crying infants in comparison to the 2-second cry stimulus that is warranted by event-related potential designs. Consistent with studies utilizing the infant cry paradigm (Kim et al.2020; Lorberbaum et al., 2002), we elected to elicit a more naturalistic interaction with infant cries and to capture variance in the subjective experience of cries with post-scan cry ratings.

2.5. fMRI data acquisition and processing

Data were initially collected in a 3.0 T Siemens magnet scanner (n = 31) with a standard 32-channel head coil, which was upgraded to Prismamat scanner (n = 19) midway through the study. Functional data were acquired with parameters matched across scanners (540 T2*-weighted echo-planar-imaging (EPI) volumes; TR = 2300 ms; TE =27 ms; flip angle = 73; field of view = 192mm; matrix size, 64×64; 36 axial slices; voxels = 3 mm³). High-resolution anatomical T1-weighted images were gathered using the 3D magnetization-prepared rapid gradient-echo (MPRAGE) protocol. Analysis of Functional Neuroimages software (AFNI; Cox, 1996; Cox & Hyde, 1997) was used for preprocessing and statistical analysis. The first two images of each run were discarded to account for magnetic equilibrium. Slice time correction was applied and images within each run were realigned to the third image for motion correction. Images with motion greater than 0.5mm in any direction were censored. After motion correction, realigned functional images were co-registered to anatomical images and functional images were anatomically normalized to Talairach space. Images were spatially smoothed with a 6-mm root-mean-square deviation Gaussian blur. No participants had more than 20% of their TRs removed and were thus all included. We examined potential systematic scanner effect on collected brain imaging data by comparing the average temporal signal-to-noise ratio (TSNR) for the 3.0 T Siemens magnet scanner (M = 228.08, SD = 31.10) and the Prismamat scanner (M = 226.06, SD = 44.96), which indicated no statistically significant differences, t(48) = .57, p = .57.

General linear models estimated the shape of the hemodynamic response to each stimulus (own infant cry, control infant cry, own infant cry matched white noise, and control infant cry matched white noise) for each individual participant. Regressors (four conditions and six motion parameters) were convolved with a gamma-variate hemodynamic response function. The beta images, representing each participant’s and each condition’s estimated activation, were then used in group-level analyses.

2.6. Data analytic plan

In order to estimate the sample size for group comparisons, we performed a power analysis with G*Power, 3.1 (Faul et al., 2007) to test the interaction between ethnicity, sound, and identity conditions. The test revealed a sample size of 36 for all two-way interactions (ethnicity vs. sound, ethnicity vs. identity, sound vs. identity) in the study to detect an a medium to high effect size (f = 0.25) with 90% power at 0.05 alpha level. Thus, we decided that the current sample size (n = 50) had sufficient power. AFNI’s 3dLME was used to employ a whole-brain linear mixed effects model testing ethnicity (Latina vs non-Latina White) as between subjects and sound (cry vs. white noise) and identity (own vs. other) as within-subjects factors. Child age represents postpartum months and was included as a covariate in the whole-brain model due to potential influences on maternal brain responses to infant cry (e.g., Kim et al., 2011). Scanner type was entered into the whole-brain model to account for any potential systemic differences in imaging data by scanner. Latina and non-Latina White groups were compared in terms of demographic variables: maternal age, income-to-needs ratio, maternal education, employment, relationship status. Maternal age and maternal education emerged as demographic variables that differed between groups and were subsequently entered into the whole-brain model.

Brain regions that were related to ethnicity status in response to cry and white noise sounds were examined by setting the cluster size threshold to k ≥ 36.5 with a per-voxel threshold of p < .001, equivalent to a whole brain corrected false positive probability of p < .05, as calculated by 3dClustSim. Post-hoc analyses to decompose significant interactions were then tested in SPSS software, version 25.0 (IBM Corp., 2017) using values extracted from clusters identified in the whole-brain analyses and applying Bonferroni corrections to adjust for multiple comparisons. To probe the relationship of variables that may account for differences in brain activation patterns to infant cries versus white noise between Latina and non-Latina White mothers, we examined several variables: maternal sensitivity, number of non-biological children in the household, negative life events, maternal depression, and acculturation. We first examined if these variables significantly differed between Latina and non-Latina White mothers. Variables that demonstrated statistically significant group differences were then examined using correlations with clusters identified in the whole-brain analyses.

Results

3.1. Sample characteristics

The overall sample was mixed low- and middle-income, with 32% of participants endorsing an income-to-needs ratio of one or below. Demographics were compared between Latina and White groups (see Table 1 for descriptives and statistical results). Groups were similar in income-to-needs ratio, child age, relationship status, and daily responsibilities. On average, Latina mothers were younger and had fewer years of completed formal education compared to White mothers. See Table 2 for the correlation between study variables and Supplementary Appendix for further analysis of age in the whole-brain analysis.

Table 2.

Correlations among study variables.

	1	2	3	4	5	6	7	8	9
1. Child age
2. Maternal age	.07
3. Maternal education	.05	.74***
4. Maternal Sensitivity	−.12	.26+	.35*
5. Depressive symptoms	.08	−.004	.07	.13
6. Negative life events	−.28+	−.18	−.29+	−.18	.28+
7. Number of non-biological children	−.004	−.34*	−.43**	−.12	−.12	.22
8. Income-to-needs-ratio	−.01	.33*	.37**	.15	−.02	−.21	−.23
9. Mexican Orientation	.07	−.04	−.15	.30	−.05	.07	.23	−.12
10. Anglo Orientation	.01	.38	.38+	.13	−.20	−.14	−.31	.22	−.40*

Note:

^***p < .001

^**p < .01

^*p < .05

⁺p < .10.

3.2. fMRI whole-brain analysis

The three-way interaction of sound condition, infant identity condition, and maternal ethnicity was not significant. We next tested a more parsimonious model in which four conditions of cry sounds were concatenated to create one factor with two levels: cry and white noise sounds. The two-way interaction of sound condition (cry, white noise) and maternal ethnicity (Latina, White) showed statistically significant clusters in the right posterior insula, left cerebellum, right premotor cortex, right dorsolateral prefrontal cortex, left somatosensory, left premotor cortex, left auditory, right medial prefrontal and right superior frontal cortex (see Table 3 for details). Pairwise comparisons in the right medial prefrontal and prefrontal cortices did not survive Bonferroni adjustments. Table 4 provides activation means in response to the cry and white noise conditions for the two racial/ethnic groups. White mothers showed steeper increases in BOLD activation for cry (versus white noise sounds) compared to Latina mothers in the right posterior insula (p = .001), left cerebellum (p = .02) and left auditory cortex (p = .03; see Figure 1). White mothers showed differential BOLD activation between cry and white noise sounds (compared to Latina mothers, who did not differentially activate) in the right dorsolateral superior frontal cortex (p = .01), left somatosensory (p = .04), right (p = .01) and left premotor cortices (p = .02), and the right posterior insula (p = .04; see Supplementary Appendix for Figures).

Table 3.

Brain regions showing statistically significant Maternal Ethnicity × Sound Condition interactions and post-hoc analyses.

Region	BA	Side	MNI Coordinates			Cluster size	F (1, 45)
			x	y	z
Posterior Insula	21	R	40.0	−6.0	−12.0	912	16.69***
Cerebellum	-	L	−17.0	−68	−45	189	13.19**
Precentral Gyrus	4	R	50	−9	53	111	13.83**
Dorsolateral Prefrontal Cortex	10	R	21	52	18	82	12.66**
Somatosensory Cortex	43	L	−56	−15	19	62	14.49***
Precentral Gyrus	6	L	−53	−3	36	42	11.59**
Auditory Cortex	13	L	−46	−20	−1	41	11.97**
Medial Prefrontal Cortex	9	R	5	48	12	41	9.81**
Superior Frontal Cortex	8	R	5	31	48	40	10.51**

Note: BA = Brodmann Area, MNI = Montreal Neurological Institute

^***p < .001

^**p < .01.

Table 4.

Mean activation patterns for white noise and cry sounds of Latina and White participants.

		White Noise			Cry
Regions		Mean	SE	95% CI	Mean	SE	95% CI
Right Posterior Insula	Latina	.07	.03	.02, .13	.10	.03	.03, .16
	White	.16	.03	.10, .23	.27	.04	.20, .34
Left Cerebellum	Latina	.04	.02	.003, .09	.09	.03	.03, .16
	White	.03	.02	−.02, .07	.21	.03	.14, .28
Right Premotor Cortex	Latina	−.03	.03	−.08, .02	−.18	.05	−.27, −.09
	White	−.02	.03	−.08, .03	.01	.05	−.09, .10
Right Dorsolateral Prefrontal Cortex	Latina	−.02	.02	−.05, .02	−.10	.02	−.10, −.01
	White	−.03	.02	−.06, .01	.03	.02	−.02, .07
Left Somatosensory Cortex	Latina	.04	.03	−.03, .10	−.01	.05	−.10, .10
	White	.01	.04	−.06, .08	.14	.05	.05, .24
Left Premotor Cortex	Latina	−.07	.03	−.13, −.01	−.19	.05	−.30, −.09
	White	−.05	.03	−.12, .01	.01	.06	−.11, .12
Left Auditory Cortex	Latina	.27	.04	.20, .35	.38	.05	.27, .50
	White	.31	.04	.23, .39	.57	.06	.45, .68
Right Medial Prefrontal Cortex	Latina	−.07	.03	−.13, −.02	−.12	.03	−.18, −.06
	White	−.11	.03	−.17, −.05	−.03	.03	−.10, .04
Right Superior Frontal Cortex	Latina	−.02	.03	−.07, .03	−.02	.03	−.08, .05
	White	.07	.03	−.12, −.01	.06	.03	−.001, −.13

Figure 1. Blood Oxygenation Level Dependent (BOLD) activation illustrating White Noise by Infant Cry contrast in White and Latina groups for three selected regions of interest.

Notes. BOLD activation in the a) right posterior insula, b) left premotor cortex, and c) left auditory cortex. In all three figures, the dashed line represents activation for Latina participants and the solid line represents White participants.

3.3. Group differences and correlational analyses

See Table 1 for t-test results. Maternal sensitivity and endorsed negative life events (count and valence scores, and their composite) did not differ between groups. A greater number of Latina mothers reported living with non-biological children in household compared to the number of White mothers; examined dichotomously, about a third (n = 8) of Latina mothers and a tenth (n = 3) of White mothers lived with at least one non-biological child. White mothers reported slightly higher depression symptoms compared to Latina mothers. No group differences were tested for acculturation because acculturation was only measured among Latinx participants.

3.3.1. Number of non-biological children in the household.

Across all participants, number of non-biological children was correlated with cry-white noise activation difference in right premotor (r = −.31, p = .03), right dorsolateral prefrontal cortex (r = −.35, p = .01), and right medial prefrontal (r = −.28, p = .05) areas such that more non-biological children in the household was related to less activation to infant cries.

3.3.2. Maternal depression.

Across all participants, higher BDI scores were correlated with larger cry-white noise activation difference in right premotor (r = .31, p = .03) and left premotor (r = .29, p = .05) regions.

3.3.3. Acculturation.

Latina mothers, on average, endorsed moderate Mexican Orientation and Anglo Orientation. Mexican Orientation was correlated to activation patterns of cry-white noise in the right posterior insula (r = −.43, p = .05) and right dorsolateral prefrontal cortex (r = −.64, p = .002) regions. Anglo Orientation was not correlated with brain activation.

Discussion

The current study sought to contextualize differences in neural activation patterns to parenting-relevant stimuli (i.e., infant cries) between Latina and White mothers experiencing a mix of low- and middle- income in the United States, situating findings and future directions within a cultural neuroscience framework (Kitayama & Uskul, 2011). We found increases in BOLD signal activation in response to cry sounds in the left somatosensory, left auditory, bilateral premotor cortex, left cerebellum, right posterior insula, right dorsolateral prefrontal cortex, right prefrontal, and medial prefrontal cortices relative to matched white noise conditions. These brain regions are involved in sensory processing, motor planning, movement, social information processing, affect regulation, and cognitive control and have been commonly reported in prior neuroimaging studies for infant stimuli (Barrett & Fleming, 2011; Bornstein et al., 2017; Lorberbaum et al., 2002; Rilling, 2013; Seifritz et al., 2003; Witteman et al., 2019). Further, we found significant differences in activation patterns between Latina and White mothers, such that BOLD activation to infant cries (but not white noise) in regions related to motor planning, movement, and sensory processing were higher among White mothers compared to Latina mothers. In the right posterior insula, an area thought to support social information processing, White mothers showed higher activation across sound conditions compared to Latina mothers. It remains unclear whether lower or lack of activation to infant cries represents more efficient processing in the postpartum period. Because previous studies are predominantly based on non-Latinx White samples, our finding might represent differential processing of infant cues among Latina mothers in the postpartum period, however, this interpretation is necessarily preliminary given the limited work inclusive of Latina mothers.

Results showed that Latina and White mothers differed in the reported number of non-biological children living in the household and severity of depression symptoms. We found that depression symptoms positively correlated with BOLD activation in response to sound condition suggesting that the higher activation patterns among White mothers might be related to higher reported depressive symptoms. Number of non-biological children in the household, on the other hand, was negatively associated with less activation in infant cries. This suggests that lower activation patterns among Latina mothers might be related to higher number of non-biological children they live with, and putatively to greater alloparenting experience.

Acculturation

Ethnicity was used as a proxy for cultural comparisons across the full sample, with acculturation to mainstream culture measured within Latina mothers. Endorsement by Latina mothers of higher Mexican Orientation (as indicated by higher affinity for the Spanish language and affiliation to Mexicans) was related to lower differential cry versus white noise activation in the right posterior insula and right dorsolateral prefrontal cortex regions, areas related to social information processing and cognitive control. We interpret this finding as supporting the idea that culture shapes parenting responses. It may be that differences in endorsed cultural values shape ideals of maternal or infant behavior, and correspondingly, which social cues or information are relevant. Emerging cultural neuroscience research has shown relations between endorsed collectivism and increased mean EEG amplitude when working toward a socially contextualized goal (i.e., winning points for a team) compared to individual goals for Latinx adolescents (Rapp et al., 2021). To our knowledge, no neuroimaging studies have examined effects of acculturation in the postpartum period, and our findings imply that acculturation and cultural values might be one mechanism through which parental brain is shaped in the postpartum period.

Parenting behaviors

Behavioral coding of maternal sensitivity across a naturalistic, parent-infant interaction task did not show any group differences between Latina and White mothers. Additionally, brain activation across regions of interest were not associated with maternal sensitivity. This suggests that although Latina and White mothers recruit differential brain regions to support responding to infant cries, these do not correspond to differences in quality of maternal sensitivity. It may be important for future studies to be sufficiently powered to capture the complicated interplay between stress exposure, neural activation to infant cry, and sensitive parenting behaviors, as another study in this sample found that differential lower brain activation to parenting cues by cumulative risk exposure accounted for variability in maternal behavioral sensitivity (Kim et al., 2020).

Number of non-biological children

Number of non-biological children in the household significantly differed between groups, such that Latina mothers reported more children living in the household. We use this variable as a proxy for prior caregiving experience, as our sample consisted of first-time biological mothers and any additional children in the household would thus be non-biological children. In our sample, number of non-biological children in the household was correlated with reduced differential activation to cry sounds versus white noise in the premotor, dorsolateral prefrontal cortex, and medial prefrontal regions. This aligns with prior studies indicating that increased exposure to non-biological children and childrearing responsibilities may shape brain adaptations in the postpartum infant caregiving period (Abraham et al., 2018). Although we did not directly measure prior caregiving experience, studies indicate that Latinx parenting tends to be more communal (Barker et al., 2010). More non-biological children in the household may also be more likely within Latinx communities, particularly for those who are foreign-born and among those experiencing low-income, due to discriminatory practices such as predatory lending and assumptions of legal status (Ayón, 2015). These practices can lead to a higher number of people sharing the same household and being accustomed to relatively higher noise levels. Thus, in addition to the alloparenting hypothesis we have offered, these results could also potentially result from desensitization to cry sounds due to high noise exposure in the house. Lower activation within parenting-relevant regions for Latina compared to White mothers might be also related to parenting practices in Latinx cultures. These findings are also consistent with an event-related potential study in which first-time mothers showed greater neural activation to infant cues compared to more experienced mothers (Maupin et al., 2019). However, that study also documented similar patterns in response to infant cry among primiparous and multiparous mothers. Therefore, these findings should be carefully interpreted and as generally supportive of more studies investigating potential cultural effects on neurobiology. Our study provides preliminary evidence that brain activation patterns might be different for Latinx parents compared to White parents. Although we did not have power to test it, research signals differences between these two ethnic groups due to differential experience with children or experiences of acculturation (Calzada, 2010; García Coll, 1990). Indeed, there were more mothers who were living with additional children at home in the Latina group compared to the White mothers. However, these differential patterns of brain activation were not reflected in observed parenting behaviors.

Strengths, limitations, and future directions

To our knowledge, few parental neuroimaging studies have incorporated attention to context as shaped by race/ethnicity or culture. This study uses an existing dataset to demonstrate differences in brain activation by one racial/ethnic group contrast (White and Latina), accounting for several potential sources of systemic variation between the two groups, and that cultural context may be important to attend to in the neuroimaging literature. Although we examined potential sociodemographic differences between Latina and White mothers, we acknowledge that other meaningful constructs may also differentially shape parenting experiences of Latina and White mothers. For example, a third of Latinx individuals in the United States were born outside of the country (Noe-Bustamante & Flores, 2019) and experiences of immigration and acculturation are a relevant context for many Latinx families in the United States. Acculturative stress is associated with elevated depression and anxiety symptoms perinatally (Luis Sanchez et al., 2020; Preciado & D’Anna-Hernandez, 2017). While we did not measure acculturative stress, in our sample, Latina mothers had slightly lower depressive symptoms than White mothers and the two groups did not differ in the number of negative life events endorsed. Therefore, research focusing on these constructs along with acculturation and discrimination experiences is needed but may not be the most relevant explanation for the racial/ethnic group differences that we report here.

Relevant to the current study, epidemiological studies indicate that Latinx Americans, on average, experience higher levels of stressors (Myers, 2008) due to historical marginalization and systemic racism (Bailey et al., 2017). One strength of this study is that it includes some variables indicating exposure to stressors to begin disentangling race/ethnicity and stressors in relation to patterns of brain activation. Prior studies in this sample have examined the role of socioeconomic disadvantage, finding that current perceived stress mediated the link between lower income-to-needs ratio and lower neural responses to infant cry sounds in prefrontal regions among first-time mothers in the first six months of infancy (Kim et al., 2016). In another fMRI paradigm where mothers viewed emotional infant faces, socioeconomic disadvantage was associated with lower amygdala activation to positive infant faces but increased activation to negative infant expressions (Kim et al., 2017). Therefore, we carefully examined sociodemographic measures (e.g., household income, relationship status, highest level of education) between the Latina and White groups in our sample. Groups were similar across relationship status, employment, child age, and endorsed negative life events. While income and stress in prior work were important variables, in the current study neither differed between our racial/ethnic groups nor explained activation pattern differences. Latina mothers were slightly younger and had fewer years of formal education compared to White mothers in the sample. Although these demographic variables were included as covariates in our analysis, they might have limited our ability to explain differences between two groups.

Maternal depressive symptoms showed some small, positive associations with cry- white noise activation difference in bilateral premotor regions. However, it should be noted that depression symptoms or negative life events may not be the best measures for stress among Latina mothers and sociocultural stressors, such as discrimination, immigration status and acculturative stress should be considered in future research.

The findings of the current study should be interpreted in the context of several limitations. First, we had a small sample size for investigating the nuance of culture, and replications with larger sample sizes will be critical to validate the main findings and to further examine possible correlates and mediators (particularly stress). Next, we included individuals from different backgrounds in one broad Latinx term because our sample did not allow differentiating among specific Latinx groups. Although there is behavioral evidence that both Dominican and Puerto Rican mothers self-report engaging in similar caregiving behaviors (Guilamo-Ramos et al., 2007) and that Mexican immigrants and Mexican Americans do not significantly differ in their caregiving behaviors (Varela et al., 2004), these similarities might not translate at the neural level. We used number of non-biological children in the household as a proxy for prior caregiving experience⁴ and future studies take into account measures of alloparenting and childcare experiences as well as the number of years living with children. We did not measure cultural beliefs, values, culturally informed parenting practices, or child socialization goals, nor did we measure acculturation across the entire sample (i.e., White mothers were not asked to complete the acculturation measure) to capture the full variability in endorsed culture. Most Latina mothers within our sample were born in the US and foreign-born experiences are not as represented in this study; however, this may be an important dimension to consider in future studies as epidemiological studies support a “Latinx Health Paradox” such that foreign-born Latinx individuals show better physical health compared to their US-born counterparts (Fox et al., 2015; Ruiz et al., 2016). Lastly, measuring parental values and practices with infant caregiving would allow for greater insight into conceptualizations of motherhood, which have been indicated to vary across cultures (Harkness et al., 2007). Future cultural neuroscience studies would strongly benefit from more detailed measurement of cultural constructs. Strengths of the current study include investigating patterns of brain activation by ethnicity instead of covarying for ethnicity, as well as examining endorsed acculturation within Latina mothers in relation to brain activation to infant stimuli. The study sample also equally represents Latina and White mothers, in contrast to most parenting neuroimaging studies that recruit predominantly White samples and includes a sample of Latina and White mothers with similar levels of stress and demonstrated parental sensitivity.

Conclusion

Our study contributes to the neuroimaging literature by documenting differences in neural activation patterns in the postpartum period between Latina and White first-time mothers in the United States. Although results should be interpreted with caution and our data do not provide a clear indication of why we see group differences in neural activation, we provide initial support for potentially culture-related influences in brain activation occurring in the postpartum period. Results illustrate compelling reasons to increase representation of racial/ethnic minorities in neuroimaging research, including with other types of infant stimuli beyond cry (Raghunath et al., 2022). In the fMRI task used in the current study, the participants were not asked to rate each sound. This design is to let the participants focus on their thoughts naturally occurred in response to each sound. The design is used in most of the fMRI infant cry tasks (Kim et al., 2011; 2020; Laurent & Ablow, 2012; 2013; Seifritz et al., 2003). However, the active rating task of infant cry sounds may shed light on attentional or emotion regulation processes used for infant stimuli (Capistrano et al., 2022; Mutschler et al., 2016; Witte et al., 2022). Although lower activation patterns to infant cry signals are usually thought to implicate unfavorable outcomes and have been associated with difficulties such as adverse caregiving history, current depression, substance use disorders and lower income to needs ratio, the lack of mediation by these measured factors suggest that lower activation in a racial/ethnic group may not be interpreted as a negative finding and there is need for future studies to explore this phenomenon.

Availability of data and materials

Data reported here are not available, as the study predated NIH implementation of open access norms.