The influence of maternal anxiety and depression symptoms on fNIRS brain responses to emotional faces in 5- and 7-month-old infants

Highlights

• fNIRS asymmetry of facial emotional processing was investigated in infants.

• fNIRS responses to emotional faces were not lateralized in 5- and 7-month-olds.

• Maternal anxiety and depression were not associated with infants’ fNIRS asymmetry.

• Infants of mothers with higher depression symptoms had greater oxyHb in the left IFG.

• Maternal depression and early experiences may influence infants’ emotion processing.

Abstract

Greater relative right (versus left) frontal cortical activation to emotional faces as measured with alpha power in the electroencephalogram (EEG), has been considered a promising neural marker of increased vulnerability to psychopathology and emotional disorders. We set out to explore multichannel fNIRS as a tool to investigate infants’ frontal asymmetry responses (hypothesizing greater right versus left frontal cortex activation) to emotional faces as influenced by maternal anxiety and depression symptoms during the postnatal period. We also explored activation differences in fronto-temporal regions associated with facial emotion processing. Ninety-one typically developing 5- and 7-month-old infants were shown photographs of women portraying happy, fearful and angry expressions. Hemodynamic brain responses were analyzed over two frontopolar and seven bilateral cortical regions subdivided into frontal, temporal and parietal areas, defined by age-appropriate MRI templates. Infants of mothers reporting higher negative affect had greater oxyhemoglobin (oxyHb) activation across all emotions over the left inferior frontal gyrus, a region implicated in emotional communication. Follow-up analyses indicated that associations were driven by maternal depression, but not anxiety symptoms. Overall, we found no support for greater right versus left frontal cortex activation in association with maternal negative affect. Findings point to the potential utility of fNIRS as a method for identifying altered neural substrates associated with exposure to maternal depression in infancy.

Introduction

Offspring of anxious and depressed mothers are at increased risk for developing behavioral and emotional problems in infancy (Feldman et al., 2009), which may later develop into clinically significant internalizing disorders (Keenan, 2006). Infants’ exposure to maternal anxiety and depression may contribute to experience-dependent fine-tuning of emotional brain systems in the first years of life (Leppänen & Nelson, 2009; Leppänen, 2011). Evidence for neurodevelopmental embedding of maternal depression and anxiety is therefore a growing area of inquiry (Aktar & Bögels, 2017; Porto, Nunes, & Nelson, 2016), that may elucidate mechanisms linking parental risk with maladaptive outcomes in children, and offer unique biomarkers that can aid in early detection and intervention. To contribute to this line of work, we explored connections between maternal depression and anxiety symptoms and infants’ neural responses to emotional facial expressions using functional near infrared spectroscopy (fNIRS). Based on substantive literature demonstrating associations between maternal depression and infants’ EEG-based frontal asymmetry patterns, we specifically tested whether patterns of frontal asymmetry would be associated with exposure to maternal depression and anxiety. We also explored connections between maternal depression and anxiety and activation in key frontal and temporal regions of interest, previously implicated in facial and emotion processing. Thus, the present study aimed to contribute to the current body of knowledge about by considering the influence of maternal depression and anxiety exposure on neurodevelopmental underpinnings of emotional face processing in infancy.

Substantive work has demonstrated links between maternal depression and anxiety and altered neural function in infants. Most studies to date have used EEG, which is non-invasive, and given tolerance for motion, is well suited to assess neural function in infants. Prior work in children and adults has demonstrated greater relative right versus left frontal EEG activation, as indicated by greater left versus right alpha power, referred to as frontal asymmetry, in association with depression exposure (Davidson, 1988). Greater relative right versus left frontal cortex activation is theorized to underlie greater tendencies to show negative emotion, negative affect and avoidance (Fox, 1991), which may increase susceptibility to clinically significant mood and anxiety disorders. Relative to infants of non-depressed mothers, infants of depressed mothers have shown greater relative right versus left frontal alpha asymmetry at rest in numerous studies (see Field & Diego, 2008, for a review). Effects are shown to appear as early as the first month of life, and persist until at least early childhood (Diego, Jones, & Field, 2010; Field et al., 2004; Thibodeau, Jorgensen, & Kim, 2006).

Most work has investigated frontal alpha asymmetry in the context of a “resting state” measure, in which EEG is recorded while children are not explicitly engaged in a cognitive task. A limited number of studies have also tested for frontal alpha asymmetry while infants view salient emotional stimuli, such as to emotional facial expressions. Building off work in resting EEG studies, it has been theorized that individual differences in frontal asymmetry to emotional faces may signal individual differences in emotion processing. Specifically, greater right versus left activation to facial expressions may indicate greater avoidance or negative affect experienced during its processing, which may serve as a neural marker for later affective risk.

As a means for potentially identifying intergeneration pathways to risk for affective problems, a number of studies have examined associations between maternal depression exposure and frontal asymmetry patterns when infants view salient or emotional facial expressions. In one study, infants of depressed mothers with high levels of intrusive behavior have shown greater relative right frontal asymmetry when viewing sad versus happy expressions to strangers’ facial expressions (Diego et al., 2002). Greater relative right versus left frontal activation has also been observed when infants of depressed mothers view images of their mothers face showing various emotional expressions (Diego et al., 2004) or engage in playful interactions with their mothers (Jones, Field, Fox, Davalos, & Gomez, 2001).

Although EEG provides excellent temporal resolution, spatial precision is limited due to volume conduction of electrophysiology activity throughout the brain, surrounding tissue, and skull. In recent years there is growing usage of fNIRS in infant samples, given its greater tolerance of motion (as compared with fMRI), and greater spatial specificity (as compared with EEG), such that neural activation patterns can be localized to specific cortical regions (Vanderwert & Nelson, 2014). fNIRS is an optical imaging technique that measures changes in concentrations of oxygenated, deoxygenated, and total hemoglobin (oxyHb, deoxyHb, and totalHb) in cortical areas, providing an indirect measurement of neural activity (Gervain et al., 2011).

A growing number of studies have applied fNIRS to understand the development of circuitries supporting facial emotion processing (see Maria et al., 2018, for a review). Some have examined infants’ cortical activation patterns in response to mother’s versus stranger faces. One study involving 6- to 9-month-old infants, demonstrated significant right frontal and temporal activation when infants viewed their own mother’s face versus that of an unfamiliar woman (Carlsson, Lagercrantz, Olson, Printz, & Bartocci, 2008). Another study of 7- to 8-month-old infants showed bilateral temporal activation in response to mothers’ faces versus a baseline non-face condition (Nakato, Otsuka, Kanazawa, Yamaguchi, Honda et al., 2011).

A separate set of studies have examined cortical activation when infants view positive and negative facial emotion expressions. The results on the extent of fNIRS based frontal activation patterns to emotional facial images are mixed. In one study, 7-month-old infants indicated greater right frontal activation when infants were viewing videos of their mothers’ smiling facial expressions in contrast to their mothers’ neutral faces (Fox, Wagner, Shrock, Tager-Flusberg, & Nelson, 2013), In another study, 7-month-olds characterized as having a low negative emotionality temperament showed increased left prefrontal cortex activation in response to happy faces versus a neutral baseline video (Ravicz, Perdue, Westerlund, Vanderwert, & Nelson, 2015). In a separate study, 9- to 13-month-old infants showed greater medial prefrontal activation when viewing video clips of the own mother versus an unfamiliar woman smiling (Minagawa-Kawai et al., 2009).

A related question concerns the involvement of key temporal regions in infants’ processing of facial emotions. Data on temporal lobe involvement is also mixed. For example, infants between 7 and 8 months of age showed greater left temporal superior temporal sulcus (STS) activation in response to happy faces and right STS activation in response to angry faces versus a non-face condition (Nakato, Otsuka, Kanazawa, Yamaguchi, & Kakigi, 2011). However, in another study of younger, 5-month-old infants, only weak activation for temporal regions was found, with much stronger activation over occipital cortical regions (Di Lorenzo et al., 2019). This suggests that specialization of key temporal circuitries associated with face processing of positive and negative facial emotions may develop over the first year of life.

Therefore, data thus far suggests that fNIRS is a promising tool for understanding the neurodevelopmental processes underlying emotion processing in infants. However, there are outstanding questions on whether similar frontal asymmetry described in EEG work (greater right versus left frontal cortex activation to negative versus positive emotions) are also registered with fNIRS. Given enhanced spatial localization to frontal brain regions, fNIRS based patterns of frontal asymmetry would bolster theory that right versus left frontal activation varies according to early social input (maternal emotion exposure) and may also signal affective risk.

An additional outstanding question concerns the involvement of key temporal cortical regions when infants process emotional faces. Data thus far indicate mixed results, and significant variation across age. Part of the problem may be the lack of consideration of early environmental risk factors that may contribute to individual differences in activation. Indeed, the majority of work has focused on identifying normative neurodevelopmental patterns elicited with fNIRS.

Our goal was to use fNIRS to further address these questions, to better understand how maternal negative affect, including depression and anxiety symptoms, influence neural correlates of facial emotion processing during infancy. We specifically tested for influence on frontal asymmetry patterns, and in activation patterns of key frontal and temporal cortical regions of interest known to subserve salient emotional and facial input. We had several specific hypotheses. First, we expected that infants of mothers who reported higher levels of negative affect would show greater right versus left frontal activation, especially in ventral (more orbitofrontal) regions. Following findings from EEG work, we expected that frontal asymmetry would be most pronounced when infants viewed negative versus positive facial emotional expressions. Next, we expected that there would be significant associations between maternal depression and anxiety and key temporal regions known to subserve facial and emotional processing.

Depression and anxiety symptoms are highly comorbid (Pawluski, Lonstein, & Fleming, 2017). Therefore, we focused on the influence of maternal negative affect in our primary hypotheses. However, on an exploratory basis, we questioned whether there may be differential effects in the influence of maternal depression and anxiety on infants’ neural responses to facial emotions. This was motivated by prior work showing mothers with depression versus anxiety exhibited variations in their emotional expression to infants. For example, maternal depression has been associated with lower positive affect, greater emotional withdrawal and/or higher levels of anger than non-depressed mothers (Aktar, Colonnesi, de Vente, Majdandžić, & Bögels, 2017; Beck, 1998; Murray, Halligan, & Cooper, 2010; Stanley, Murray, & Stein, 2004), whereas maternal anxiety has been associated with greater displays of fear but not necessary reduced positive affect (Weinberg, Beeghly, Olson, & Tronick, 2008). Guided by this behavioral literature, we hypothesized that maternal depression (which may manifest as reduced positive affect and greater anger) might be more correlated with infants’ processing of positive and angry emotional faces. In contrast, we expected that maternal anxiety might be more strongly correlated with infants’ neural activation patterns when processing fearful emotional expressions.