White matter integrity moderates the relation between experienced childhood maltreatment and fathers’ behavioral response to infant crying

Abstract The ability to provide appropriate responses to infant distress is vital to paternal care, but may be affected by fathers’ experiences of childhood maltreatment. Detrimental effects of childhood maltreatment have been found in the adult brain's white matter fibers, accompanied with impaired emotional and cognitive functioning. In the current study (N = 121), we examined new and expectant fathers’ childhood maltreatment experiences (i.e. emotional and physical abuse and neglect), current behavioral responses (i.e. handgrip force) to infant cry sounds, and white matter integrity using diffusion tensor imaging. First, more exposure to childhood maltreatment was associated with more use of excessive handgrip force in response to infant crying by fathers. Second, the association between experienced childhood maltreatment and white matter integrity was not significant in whole‐brain analyses. Lastly, we found that the association between maltreatment exposure and excessive handgrip force during infant crying was absent in fathers with high tract integrity in the bilateral uncinate fasciculus. These findings possibly point to insufficient behavioral inhibition or emotional dysregulation in fathers who experienced childhood maltreatment, but buffering for this effect in those with larger integrity in brain fibers connecting the amygdala and prefrontal cortex.


| INTRODUC TI ON
Paternal care greatly impacts the development of children; higher quality of care has been found to positively affect children's social, cognitive, and linguistic development (see Lamb, 2010, for reviews).
Several neural factors such as brain structure and function have been shown to play an important role for the quality of paternal care (see Feldman et al., 2019;Rogers & Bales, 2019 for reviews). In females, experiences of childhood maltreatment have been found to affect both later parenting (see Norman et al., 2012; and brain structure (e.g. see Teicher & Samson, 2016 for a review).
To date, studies investigating the influence of experiences of childhood maltreatment on paternal brain and behavior are rare, It is important to note that ERN findings in the anxiety and externalizing literatures have not been entirely consistent, such that symptom/disorder and ERN patterns do not always demonstrate the expected directional relationships, particularly when there is comorbid psychopathology. For example, some adult studies have found that ERN and anxiety disorder associations differ when individuals have comorbid depression (Weinberg, Klein, et al., 2012;Weinberg, Kotov, & Proudfit, 2015). In addition, Stieben et al. (2007) found that the association between externalizing problems and a blunted ERN was less apparent among children with comorbid internalizing problems.
Despite the high prevalence of anxiety disorders (Beesdo et al., 2009;Copeland et al., 2014) and frequent co-occurrence of externalizing problems in childhood (Angold, Costello, & Erkanli, 1999;Kendall et al., 2010;Verduin & Kendall, 2003), how anxiety and externalizing comorbidity patterns relate to ERN profiles has yet to be tested. Research that examines these relations could have the potential to resolve some of the inconsistencies observed in the ERN pediatric literature, increase precision in our understanding of neural correlates of anxiety comorbidity subtypes, and elucidate primary sources of impairment that could have translational implications.
Accordingly, the primary aim of this preliminary study was to be one of the first to investigate how ERN responses could differentiate anxious subtypes as defined by those with and without comorbid externalizing problems in a pediatric sample. Healthy youth, defined as those without lifetime histories of psychiatric disorders, were also included in order to examine psychiatric symptom and ERN relations dimensionally. The overall sample consisted of anxious youth and healthy controls (ages 7-19 years old) who were recruited for a study of the treatment of pediatric anxiety disorders. Based on prior work on ERN in anxiety disorders (Meyer, 2017;Meyer et al., 2013), we hypothesized that youth with greater anxiety symptoms who did not have co-occurring externalizing problems would demonstrate enhanced ERN (i.e., greater sensitivity to errors). Drawing on preliminary studies suggesting anxiety and ERN relations could be qualified by psychiatric comorbidity (Stieben et al., 2007), we predicted that youth with co-occurring anxiety symptoms and externalizing 2 | ALYOUSEFI-vAn DIJK Et AL. even though their participation in childcare has substantially grown in modern western societies .
Moreover, parenting researchers have recently been calling for a shift in attention toward biobehavioral models of fatherhood as the underlying mechanisms of paternal care remain relatively unknown and likely developed along different evolutionary pathways than those of mothers Saxbe, 2017).
In particular, studies examining responses to infant crying by fathers with varying degrees of experienced childhood maltreatment are needed because responses to infant signals are a crucial component of the parent-child relationship (Leerkes et al., 2011). Also, child maltreatment is still prevalent worldwide (Stoltenborgh et al., 2015) with emotional maltreatment being particularly prevalent in many countries (Gilbert et al., 2009;Stoltenborgh et al., 2012).
In the current study, we therefore examined the relation between fathers' childhood maltreatment experiences (i.e. emotional and physical abuse and neglect) and their current behavioral responses to infant cry sounds (i.e. handgrip force). Additionally, we explore if this association is mediated or moderated by fathers' brain structure (i.e. white matter integrity). Our focus lies on a particularly important but understudied area of parenting; the perinatal period, a period in which the foundation for postnatal parenting is build (e.g. Cabrera et al., 2008;Hechler et al., 2019;Lucassen et al., 2015;Witte et al., 2019).
Infant crying is a highly salient stimulus aimed at motivating a caregiver to provide appropriate parental care and consequently stop the infant's distress (Soltis, 2004). However, infant crying also holds the potential for triggering child abuse and neglect because it can onset feelings of anxiety, aversion, or anger in parents (e.g. Out, Bakermans-Kranenburg, Van Pelt, & Van IJzendoorn, ;Reijneveld et al., 2004, Soltis, 2004. Having endured childhood maltreatment in one's past has been documented to be associated with emotional dysregulation throughout later life (e.g. Dvir et al., 2014;Pears & Capaldi, 2001). In turn, emotional dysregulation is thought to be detrimental for effectively managing emotional and distressing situations (e.g. see Mikulincer & Shaver, 2008;Reijman et al., 2016 for reviews), such as soothing a crying infant. Adults who experienced more childhood maltreatment show more negative behavioral responses to child signals and respond with harsher reactions when exposed to infant crying (Buisman et al., 2018(Buisman et al., , 2019. Commonly found deficits in inhibitory and affective control in maltreated individuals (e.g. see Cowell et al., 2015;Dvir et al., 2014 for reviews) might cause these individuals to interpret infant signals as relatively negative and call upon time and energy consuming cognitive strategies to provide appropriate responses. This cognitive control may break down in cases of unconscious decision making or when under time pressure, such as is the case in responding to infant crying. Although all adults experience some level of physiological arousal when exposed to infant crying (e.g. Groh & Roisman, 2009;, men tend to find infant crying particularly aversive (Zeifman, 2003).
Fathers have however been included in very few studies examining possible biological mechanisms underlying the relation between childhood maltreatment and responses to infant signals (but see Buisman et al., 2018; for a meta-analytic comparisons of male vs. female brain reactivity to infant crying, see Witteman et al., 2019). As men now participate more in childcare and an increasing number of studies confirm their influence on child development (e.g. see Lamb, 2010), a better understanding of (biological) factors contributing to their ability to provide appropriate care is needed.
The handgrip dynamometer paradigm is used in parenting research as a behavioral marker of the ability to modulate a behavioral response (i.e. handgrip force) while being exposed to infant crying.
Typically, females with insecure attachment representations (often resulting from experiencing suboptimal parenting) as well as parents at risk for perpetrating child abuse experience more hostile feelings and irritation while listening to infant crying and use more excessive handgrip force than females with secure attachment representations and parents at low risk for perpetrating child abuse (Crouch et al., 2008;Riem et al., 2012). Similarly, maltreating mothers used more excessive handgrip force than non-maltreating mothers while listening to both infant crying and laughter (Compier- de Block et al., 2015). Notably, males and females with experiences of childhood maltreatment (i.e. parental neglect) were found to have more difficulty modulating handgrip force when exposed to infant crying even though they did not rate the sound more negatively (Buisman et al., 2018). Combined, these studies call for a closer look at how fathers' experiences of childhood maltreatment relate to maladaptive responses to infant signals. In particular, fathers' functioning in the peripartum is of interest as this is a formative time that poses high demands on parents' mental and physical health but is also crucial for successful adjustment to parenthood (e.g. see Saxbe et al., 2018).
Moreover, exposure to infant crying, and the incidence of shaken baby syndrome, is especially frequent in the early postpartum (e.g. see Barr et al., 2006).
Connective white matter fibers are sensitive to adverse experiences throughout childhood and have, therefore, been a focal point in investigating the biological effects of childhood maltreatment (e.g. see Teicher & Samson, 2016 for a review). Since Teicher and Samson concluded in 2016 that childhood maltreatment likely affects white matter tracts of the dorsolateral prefrontal and orbitofrontal cortex, anterior cingulate, hippocampus, and corpus callosum (CC), several studies have added findings on the relation between childhood maltreatment exposure and white matter structure in young healthy individuals in particular. These studies confirmed the existence of a relation between childhood maltreatment and fibers connecting prefrontal and occipital areas in general (Ohashi et al., 2017), in the fronto-occipital fasciculi (FOF; Lim et al., 2019;McCarthy-Jones et al., 2018;Meinert et al., 2019), and in the longitudinal fasciculi specifically (LF; Lim et al., 2019;Meinert et al., 2019;Tendolkar et al., 2018). Findings of atypical structure in fibers connecting the orbitofrontal cortex and temporal regions (e.g. the uncinate fasciculus [UF]) were also confirmed in several studies (McCarthy-Jones et al., 2018;Meinert et al., 2019;Ohashi et al., 2017). Four recent studies confirmed structural changes in the cingulum or in fibers connecting the anywhere from 15% to 23% of anxious youth meet criteria for attention-deficit hyperactivity disorder (ADHD; Angold, Costello, & Erkanli, 1999;Kendall, Brady, & Verduin, 2001), and about 9% of anxious youth meet criteria for oppositional defiant disorder (ODD;Kendall et al., 2001). It is well established that childhood comorbidity is associated with significantly worse short-and long-term psychosocial impairments (Fraire & Ollendick, 2013;Franco, Saavedra, & Silverman, 2007). Despite moderate rates of anxiety and externalizing psychopathology comorbidity, their neurobiological origins are not well understood, which is significant given increasing clinical research emphasis on understanding the underlying pathophysiology of common psychiatric disorders and syndromes, and identifying the extent to which disorders and subtypes are associated with similar or unique neurophysiological characteristics (Insel et al., 2010;Shankman & Gorka, 2015).
Psychophysiological tools, such as scalp-recorded electroencephalogram (EEG) recordings in response to environmental cues of threat and error monitoring, have proven to be useful in identifying neural correlates of different forms of psychopathology across the life span (e.g., Shankman & Gorka, 2015). Over 50 studies have focused on the error-related negativity (ERN), an event-related potential (ERP) component typically measured at frontocentral electrodes 50-100 ms following commission of an error (Olvet & Hajcak, 2008;Weinberg, Riesel, & Hajcak, 2012). Source localization studies and investigations employing EEG and functional magnetic resonance imaging (fMRI;Debener et al., 2005;Fitzgerald et al., 2005;Mathalon, Whitfield, & Ford, 2003) have identified the anterior cingulate cortex (ACC) as the region of the brain that generates the ERN. The ACC is the primary brain mechanism involved in online monitoring for conflict between simultaneously active but incompatible streams of information (Shiels & Hawk, 2010). Greater ERN amplitudes are theorized to reflect processes of heightened conflict monitoring (Yeung, Botvinick, & Cohen, 2004), and sensitivity to threat (Weinberg et al., 2016) andpunishment (Shackman et al., 2011;Zambrano-Vazquez, & Allen, 2014). Research has indicated that the ERN represents a trait-like neural response to errors, as it demonstrates strong test-retest reliability and rank-order stability. Furthermore, ERN amplitudes appear to be multiply determined by genetic and environmental influences (Weinberg, Klein, & Hajcak et al., 2012).
Research has shown that the ERN is a neural correlate of psychopathology, particularly anxiety disorders. Specifically, enhanced (i.e., more negative) ERN in anxious individuals is frequently observed, and has been consistently replicated in both adult (Hajcak, Klawohn, & Meyer, 2019;Weinberg, Dieterich, & Riesel, 2015) and pediatric samples (Ladouceur, Dahl, Birmaher, Axelson, & Ryan, 2006;Meyer, 2017 cingulum with other brain regions McCarthy-Jones et al., 2018;Meinert et al., 2019;Tendolkar et al., 2018), and in the CC (Jensen et al., 2018;Lim et al., 2019;McCarthy-Jones et al., 2018;Meinert et al., 2019). Only one study so far confirmed a relation between childhood maltreatment and hippocampal projections (McCarthy-Jones et al., 2018). An overview of samples, maltreatment measures, imaging parameters, tract delineation, and performed analyses in these studies has been presented in Table 1. In summary, consensus seems to be emerging on which white matter tracts are affected by exposure to childhood maltreatment in healthy adults. However, no studies so far have looked at possible functional outcomes related to these changes in white matter integrity that might be relevant for parenting behaviors.
A growing body of literature suggests that atypical neural structure and function resulting from received parental care underlie parenting-related responses in mothers (e.g. Kim et al., 2010;Mielke et al., 2016). This is not surprising since atypical white matter structure has been found to play an important role in cognitive and emotional difficulties as well as psychopathology among maltreatment survivors (e.g. Riem et al., 2019). Although fathers have not yet been investigated directly, atypical brain structure associated with childhood maltreatment has been found to mediate detrimental functional outcomes such as psychopathology and vulnerability to later life stress in both young males and females (Gorka et al., 2014;Hanson et al., 2015). Speculatively, the documented risk of inter- Based on the available literature, we suggest that the association between experienced maltreatment and fathers' ability to modulate handgrip force in reaction to infant crying might be mediated by the brain's white matter structure in regions commonly associated with childhood maltreatment. Alternatively, brain structure might play a moderating role leaving individuals more or less susceptible to cognitive and emotional dysfunction after experiencing maltreatment depending on their biological make-up (e.g. see McCrory et al., 2010 for a review on the moderating role of genetics on the effects of maltreatment). Moreover, in infants, white matter structure has been found to moderate environmental influences on their negative emotional reactivity (Nolvi et al., 2020). Since few studies are available on a possible moderating role of white matter on behavioral outcomes relevant for parenting, we will merely explore (i.e. test without specific hypotheses) this moderation model. Additionally, participants were excluded when abnormalities were found during the medical 20-week ultrasound examination, or in case of known birth defects in the families of either parent that caused excessive worry for the current pregnancy. For new fathers, the infants had to be full-term (i.e. born after 37 week gestation), healthy, and around 2-4 months of age at time of inclusion. Additionally, new fathers were excluded if they used a baby carrier more than 5 hr a week due to an intervention taking place after the measurements reported here. One infant was born at 36 weeks and 6 days, but was considered healthy and therefore included. One father was not the biological father of the infant, but he had been cohabitating with the mother since mid-pregnancy. Three participants reported a past diagnosis of depression, and one participant reported a diagnosis of a past anxiety disorder. All diagnoses occurred between 2 and 10 years before inclusion into the study. They did not have a current psychiatric diagnosis and did not currently use psychotropic medications and were therefore included in the analyses here.
As a result, 121 first-time fathers with diffusion tensor imaging (DTI) data available were included into the analyses reported here, see Figure S1 for a flow chart. Of these 121 participants, handgrip data were missing for three participants due to time constraints during the lab session or technical difficulties, and information based on questionnaires was missing for five participants. Independent t- Kolmogorov-Smirnov test indicated that the education scores were not normally distributed (p = .00) and the distribution was skewed (i.e. skewness = 5.81). A chi-square test indicated that country of birth (i.e. the Netherlands or other) did not differ between new and expectant fathers. See Table 2 for more sample characteristics.
Participants received financial compensation for each laboratory visit; €25 (plus travel allowance) for the visit described here.
Participants received an extra €10 after the completion of the study if they completed at least 80% of the questionnaires. The study was approved by the Ethics Committees of the Leiden University Medical Centre and of the Department of Education and Child Studies at Leiden University. The study was carried out in accordance with the declaration of Helsinki and all participants gave informed consent.

| Procedure
Participation in the study started with the visit to the laboratory described here. Participants were instructed to abstain from alcohol and excessive physical activity during the 24 hr before the visit, and from caffeine on the day of the visit. The visit started with instructions about the visit, after which saliva collection (for hormone measurements) took place. Next, participants took care of an infant simulator Note: Information on education and country of birth is missing for one participant as he dropped out of study between the laboratory session and filling out online questionnaires.
anywhere from 15% to 23% of anxious youth meet criteria for at- or their own infant for 10 min for the observation of their parenting behavior, during which the doll was programmed to cry uncontrollably for 5 min. This was followed by another saliva sample collection, after which participants underwent a brief training of the upcoming functional magnetic resonance imaging (fMRI) tasks on a laptop, which familiarized them with the tasks. A resting-state scan, and a fMRI paradigm (using cry stimuli), as well as a fMRI task using video vignettes preceded the DTI scan described here. The neural measurements were followed by the handgrip cry paradigm. In the week following the visits, participants completed an online questionnaire. This questionnaire included basic demographics, the Love Withdrawal subscale of the Children's Report of Parental Behavior Inventory (CRPBI, Beyers & Goossens, 2003;Schludermann & Schludermann, 1983), the Conflict Tactics Scale-Parent Child (CTS; Straus et al., 1998), and the Edinburgh Postnatal Depression Scale (EPDS; Cox et al., 1987).

Data acquisition
Diffusion tensor imaging data were collected on a Philips 3.0 T Achieva

Data preprocessing
The Oxford Centre for Functional MRI of the Brain software library (FMRIB's Software Library version 6.0; Smith et al., 2004) was used to preprocess and analyze DTI data. First, topup was run because data were collected with reversed phase-encode blips, resulting in pairs of images with distortions going in opposite directions. From these pairs, the susceptibility-induced off-resonance field was estimated using a method similar to that described in Andersson et al. (2003).
To run topup, the number of slices was reduced to 56 by removing the most inferior slice. Next, topup combined two images (i.e. the first b = 0 volume from the DWI dataset and the reversed b = 0 scan) into a single file which was used to estimate the susceptibility field. Then, all non-brain material was extracted using BET (Brain Extraction Tool, Smith, 2002). Second, diffusion data were corrected for eddy currentinduced distortions and subject movements using eddy . Outliers were replaced within eddy using the repol option . We visually examined corrected diffusion images, and also obtained quantitative metrics relating to image quality using FSL's QUAD (QUality Assessment for DMRI) and SQUAD (Study-wise QUality Assessment for DMRI) tools (Bastiani et al., 2019). Third, fractional anisotropy (FA) images were created by fitting a tensor model to the raw diffusion data using dtifit from the FMRIB's Diffusion Toolbox. From these maps, FA was calculated.
Fourth, to reduce partial volume effects and registration misalignments, TBSS (Tract-Based Spatial Statistics; Smith et al., 2006) was used. All subjects' FA data were aligned into a 1 × 1 × 1 mm standard space (i.e. FMRIB58_FA template) using the nonlinear registration tool FNIRT (Andersson et al., 2007;Smith et al., 2004), which uses a b-spline representation of the registration warp field (Rueckert et al., 1999).
Next, the mean FA images were created and thinned to create a mean FA skeleton which represents the centers of all tracts common to the group. After visual inspection of our data, we set a threshold of 0.3 for the average of all aligned FA images to reduce inter-subject variability when creating a white matter skeleton. Each subject's aligned FA data were then projected onto this skeleton and the resulting data fed into voxel-wise cross-subject statistics. Additionally, mean diffusivity (MD) images were also created and fed into TBSS.

DTI data analysis
To test for statistically significant associations between maltreatment scores and FA or MD values, we employed nonparametric permutation testing using maltreatment scores as a covariate.

| Behavioral assessment-handgrip paradigm
Procedure Participants were exposed to infant crying and images representing either their own or an unknown infant while they were asked to anywhere from 15% to 23% of anxious youth meet criteria for at-

Task images
To create suitable own infant images for this task, new fathers either The distinction between own and unknown infants was expected to be nonrelevant for the analyses reported here.

Task sounds
To create suitable sounds for this task, a total of six cry sounds were recorded from six infants (three males, three females), using a TasCam DR-05 solid state recorder with at a 44.1 kHz sampling rate and 16 bit. All sounds were recorded between 2 days and 5.5 months postnatally. All cry sounds were scaled, the intensity is normalized to the same mean intensity (74 Db) and sounds are edited to last for 10 s using PRAAT software (version 6.0.37; Boersma & Weenink, 2018). For each cry sound, a neutral auditory control stimulus was created by calculating the average spectral density over the entire duration of the original sound. A continuous sound of equal duration was re-synthesized from the average spectral density and amplitude modulated by the amplitude envelope, extracted from the original sound. After this procedure, all auditory stimuli and control stimuli were intensity matched. The neutral auditory control stimuli were identical to the original auditory stimuli in terms of duration, intensity, spectral content, and amplitude envelope.

Task design
The procedure used here was identical to that used in a similar (but not After an interval of 2s, participants were prompted to squeeze at half strength (instructions were displayed for 1s). A fixation cross was shown for 3s between each trial.

| Self-reported child maltreatment
Identical to the procedure in a similar (but not the same) sample the resulting scale). The resulting scale contained 11 items such as "My mother/father is a person who, when I disappoint her/him, tells me how sad I make her/him." Participants rated how well each of the statements described their mother's and father's behavior separately on a 5-point scale ranging from 1 = "not at all" to 5 = "very well." Two participants reported that one parent had not been present throughout their childhood and scores for this particular parent where therefore not taken into account. A mean parental love-withdrawal score was computed by averaging the highest scores per item, being either that reported about the participant's father or mother. Scores per item were roughly 50% of the times highest for father and for mother, except for item 7 (i.e. "When I disappointed my parent he/she told me how sad I made him/her"), where the score for mother was more often the highest. Additionally, participants completed the Conflict Tactics Scale-Parent Child (CTS, Straus et al., 1998). We used items from the subscales Psychological aggression, Minor physical assault, Severe physical assault, and Neglect, resulting in a total of 18 items. Items were answered on a 7-point scale (0 ="never," 1 = "once," 2 = "twice," 3 = "3-5 times," 4 = "6-10 times," 5 = "11-20 times," 6 = "more than 20 times"). Averaging scores on the minor and severe physical assault scales resulted in a Physical assault score, which combined with the scores on Psychological aggression formed an Abuse score. An overall CTS score was computed by averaging the Abuse and Neglect scales. Both maltreatment questionnaires were found to be reliable (α = 0.85 for CTS and α = 0.90 for CRPBI). Average maltreatment scores (n = 116) were 0.75 (SD = 0.75) for the CTS and 2.02 (SD = 0.78) for the CRPBI, range CTS = 0-3.48 and range CRPBI = 1-4.18. Additionally, we found good reliability for the combined CTS and CRPBI item scores (α = 0.89), and therefore combined the standardized total scores of both questionnaires into one average maltreatment score, which was used in all further analyses.
A Kolmogorov-Smirnov test indicated that the combined standardized maltreatment average was somewhat skewed (i.e. skewness = 5.47). Data were not transformed as maltreatment scores were the predictor in our model. Average maltreatment scores did not differ between new and expectant fathers (t(114) = 0.51, p = .61).

| Self-reported perinatal depression
Participants filled out all 10 items of the Postnatal Depression Scale (EPDS; Cox et al., 1987), where a total score represents the sum of all symptoms present. The EPDS is a self-report screening tool, devised to detect mild perinatal depression in the community. Previously, the EPDS has been shown to be suitable for the assessment of depression in the prenatal period (e.g. see Cox et al., 2014) and in fathers (e.g. Edmondson et al., 2010). The clinically relevant cutoff for EPDS scores in fathers is currently under debate, but it is believed to be lower than that for mothers, typically between 7 and 10 (e.g. Edmondson et al., 2010) but possibly even lower (Matthey et al., 2001). The EPDS questionnaire was found to be reliable in the sample described here (α = 0.74).
A Kolmogorov-Smirnov test indicated that the total EPDS score was somewhat skewed (skewness = 4.68). Data were not transformed as perinatal depression scores were a covariate in our model.   Table 3) and in the non-imputed dataset (see Table S1). Further analyses were conducted in SPSS version 25.

| Statistical analyses
A mediation analysis was run on the imputed dataset to test individual FA or MD values in any significant clusters (i.e. clusters significantly associated with experienced maltreatment scores in whole-brain analyses) as mediators for an association between experienced maltreatment scores and handgrip force during infant crying. Next, an exploratory moderation analysis was run on the imputed dataset to test moderation by individual FA values for the anatomical structures mentioned in the hypotheses on an association between experienced maltreatment and handgrip force. Perinatal depression, age, and education levels were used as covariates.

| Association between childhood maltreatment and handgrip force
Childhood maltreatment scores showed a significant positive correlation with handgrip force in response to infant crying (i.e. r(119) = 0.22, p = .02; see Table 3), indicating that higher levels of maltreatment were related to more use of excessive force during infant crying in new and expectant fathers.

| Association between childhood maltreatment and white matter integrity
No white matter clusters were found to relate significantly to childhood maltreatment scores in whole-brain analyses (i.e. p > .05 corrected) for either FA or MD; thus, no mediation analysis was run.
Likewise, there were no significant correlations between extracted FA values in the regions of interest (ROI) and experienced maltreatment, see Table 3.  anywhere from 15% to 23% of anxious youth meet criteria for attention-deficit hyperactivity disorder (ADHD; Angold, Costello, & Erkanli, 1999;Kendall, Brady, & Verduin, 2001), and about 9% of anxious youth meet criteria for oppositional defiant disorder (ODD;Kendall et al., 2001). It is well established that childhood comorbidity is associated with significantly worse short-and long-term psychosocial impairments (Fraire & Ollendick, 2013;Franco, Saavedra, & Silverman, 2007).   No significant moderation effects were found for the CC, bilateral cingulum, bilateral sagittal stratum, bilateral superior FOF, or bilateral superior LF. Findings of the moderation analyses were comparable with and without multiple imputation (see Table S2). Also, results were comparable in new and expectant fathers (i.e. confidence intervals in both groups overlapped with the overall results), see Tables S3 and S4.

| D ISCUSS I ON
In this study, we investigated the relation between childhood maltreatment experiences, handgrip force in reaction to infant crying, and white matter tract integrity in new and expectant fathers. As expected, we found that childhood maltreatment experiences were related to paternal behavioral responses; fathers with higher levels of experienced maltreatment used more excessive handgrip force while listening to infant crying. No significant relation was found between reported maltreatment and white matter integrity in the whole-brain analysis. Therefore, white matter integrity did not mediate the relation between maltreatment and handgrip force. However, this relation was moderated by white matter integrity in the bilateral UF such that childhood maltreatment experiences were only predictive of handgrip force during infant cry exposure in fathers with low levels of white matter integrity in the bilateral UF.
In line with earlier findings in mothers and fathers (Buisman et al., 2018), we found that fathers' experiences of childhood maltreatment were predictive of using more handgrip force during infant cry exposure. Speculatively, this finding might indicate that maltreated fathers experience more aversion and anxiety when presented with negative infant signals (e.g. Out, Pieper, Bakermans-Kranenburg, Zeskind, et al., 2010), and are therefore more likely to respond harshly. Emotional dysregulation or deficient inhibitory control may underlie this phenomenon. Indeed, childhood maltreatment is known to be related to an attentional bias toward negative or threatening stimuli in both childhood and adulthood, possible serving as an adaptive mechanism preparing the individual for frequent exposure to threatening situations (e.g. Gibb et al., 2009). Also, neural hyperactivation in regions involved in threat-detection in response to various emotional facial expressions (e.g., Van Harmelen et al., 2013) points to a mechanism of interpreting emotional expressions as highly salient and potentially dangerous in maltreated individuals (see also Oosterman et al., 2019;Pollak et al., 2000).
Parenting research supports this idea as mothers exposed to childhood sexual and/or physical abuse have been found to provide lower quality of care with their own child, and this parental quality has been found to relate to the anatomy of brain structures known to underlie cognitive (and not emotional) empathy (e.g. Mielke et al., 2016). Specifically, when given the instruction to inhibit prepotent motor responses, maltreated individuals show larger reaction times despite increased brain activation in areas responsible for inhibitory and response control, as compared to individuals who were not exposed to early life adversity (Mueller et al., 2010;Navalta et al., 2006). Childhood maltreatment has been found to impact on the effective connectivity of the neural inhibitory control network during a task in which prepotent motor responses had to be inhibited (Elton et al., 2014). Taken together, it is plausible that fathers with higher levels of experienced childhood maltreatment in our sample struggled to modulate handgrip force in response to infant crying because the stimulus was processed as particularly aversive and inhibitory control over their behavioral response was inadequate.
Contrary to our expectations, we did not find a direct relation between fathers' childhood maltreatment experiences and white matter integrity in a whole-brain approach. Previously, several stud-  not find correlations between self-reported childhood maltreatment and commonly reported ROIs, indicating that the absence of findings in our whole-brain analysis is probably not merely due to stringent corrections for multiple testing in this analysis. Only one other study reported whole-brain white matter integrity in a similar sample size of healthy young men (N = 114), and no significant clusters were found when testing for an association with overall maltreatment severity (Tendolkar et al., 2018), although a significant reduction in white matter integrity for several areas (including the UF) was found when looking at physical neglect rather than a combined maltreatment score. Additionally, a recent large study examining whole-brain white matter integrity in adult participants with or without major depressive disorder (N = 396) found several clusters to be significantly associated with maltreatment scores (i.e. emotional and physical neglect and abuse, as well as sexual abuse), irrespective of psychiatric diagnosis (Meinert et al., 2019). These authors concluded that the exact type of abuse is probably irrelevant and that an "overall contribution of early negative life events" seems to underlie these effects.
Taken together, these findings raise an important question. Why do sufficiently powered studies such as Tendolkar et al. (2018) and the current study finds no significant association between DTI measures and overall maltreatment scores in whole-brain analyses such as in Exploratively, we found that white matter integrity of the UF moderated the association between experienced maltreatment and fathers' modulation of handgrip force during infant cry exposure, where the association between maltreatment and handgrip force was absent in fathers with higher tract integrity in the UF.
Interestingly, the UF is known to support adequate inhibitory regulation of emotional reactivity or motor responses that depend on effective communication between the frontal and temporal structures (i.e. prefrontal cortex and amygdala; Depue et al., 2016;Versace et al., 2015). As such, the UF plays an important role in "valence-based biasing of decisions" and in the incorporation of reward and punishment history into memory-dependent processes (Von der Heide et al., 2013). Likewise, higher structural integrity of the UF has previously been shown to protect against the well-documented relation between childhood maltreatment and later anxiety and stress resilience . Additionally, activity and connectivity of the prefrontal cortex has recently been found to be an important differential susceptibility marker for the effects of environmental influences on child development (Crone et al., 2020). Although further studies are needed to replicate and elaborate on this finding, we speculate that part of the dysfunctional effect of maltreatment on later parenting is dependent on structural integrity of the UF which is needed to effectively downregulate emotional hyperreactivity as well as support adequate inhibitory control over behavioral responses, particularly in reaction to negative stimuli. Hypothetically, maltreated fathers with high tract integrity in the UF might be protected against effects of childhood maltreatment on emotional hyperreactivity and impaired behavioral inhibition. If so, tract integrity in the uncinate fasciculus might be one of the underlying factors contributing to resilience in the intergenerational transmission of maltreatment.
Several limitations of the current study should be noted. First, the field of neurobiological aspects of paternal care is relatively unexplored territory and replication studies and meta-analyses will be highly valuable in approximating valid conclusions. Although the current study was relatively well-powered, more studies are needed to uncover the mechanisms and contributing elements of paternal care.
Specifically, the moderation effect reported here was part of an exploratory analysis and therefore not corrected for multiple testing.
Future studies could use these findings as hypotheses. Also, as our design was cross-sectional, any conclusion about the origins of the relation between maltreatment history and later parenting remain speculative. Although there is a large body of literature supporting the mechanism proposed here, it should be noted that other explanations for our findings (e.g. genetic factors contributing to both maltreatment and excessive force in the handgrip task) cannot be not ruled out. Second, there is little agreement between prospective and retrospective reports of maltreatment (Baldwin et al., 2019).