سطوح تحت بالینی افسردگی مادر و حساسیت نوزاد به اقتضای اجتماعی

Abstract The aim of the study was to investigate how young infants respond to contingent and non-contingent interaction in relation to maternal level of depressive symptoms in a non-clinical sample of mothers and infants. Two groups of three-month-olds interacted with their mother who was assessed as either non-depressed or sub-clinically depressed, based on self-reported scores on the Edinburgh Postnatal Depression Scale (EPDS). The infants were presented with a continuous image and voice of their mother in a closed circuit computer system, using the double video procedure. The experiment comprised five sequences, alternating between contingent (Live) and non-contingent (Replay) maternal behaviur in a fixed Live1–Replay1–Live2–Replay2–Live3 sequence. The infants of the sub-clinically depressed mothers showed a high gaze focus at their mother independently of the quality of interaction, while the infants of the non-depressed mothers showed a preference for looking at the mother only when the interaction with their mother was contingent. Further, the infants of the sub-clinically depressed mothers showed no differentiation in affective expression between contingent and non-contingent interactions, while the infants of the non-depressed mothers expressed more positive affect than negative affect only when the interaction with their mother was contingent. Finally, there was a significant relation between the infant's preference for looking at the mother and the infant's amount of positive affect, but this was only found for the infants of the non-depressed. These results indicate that young infants’ sensitivity to social contingency is related to maternal level of depression, even in a non-clinical sample. This expands the implications of earlier findings on the impact of maternal depression on infant sensitivity to social contingency, demonstrating that even sub-clinical levels of maternal depression may effect early interaction and child development.

. Introduction Young infants’ capacity for intersubjective awareness is fundamental for early human interaction (Trevarthen, 2001 and Trevarthen and Aitken, 2001). This capacity is shaped by the infants early experiences of sensitive, contingent responses from their caregivers, which comprises the core in the attachment between infants and caregivers (Ainsworth, 1991). It is well established that maternal depression is associated with impairment in the mother's ability or motivation to synchronize with the infant's emotional state (Field, Healy, Goldstein, & Guthertz, 1990). Microanalytic studies of depressed mothers have shown that they respond more slowly and less consequently to their infants than non-depressed mothers and that they have less positive and more negative facial expressions during interaction with their infants (Cohn et al., 1986 and Field, 1995). Interestingly, studies using continuous scales to measure depression have shown that even sub-clinical levels of maternal depression is associated with impaired parenting (Lovejoy, Graczyk, O‘Hare, & Neuman, 2000). Postnatal depression is prevalent, estimated to affect around 10% of all mothers (Berle et al., 2003 and O‘Hara and Swain, 1996) and coincides with the period where the infant establish a pattern of joint interaction, often termed the period of primary intersubjectivity (Trevarthen, 1979). It is demonstrated that infants respond to the mother's depression with less attention of gaze and less emotional expressions when interacting with their mother (Cohn et al., 1986 and Field, 1995), close to the concept of social withdrawal (Guedeney, 2007). For these reasons, it is important to understand the mechanisms through which maternal depression contributes to increased psychosocial risk in infants and children. One explanation is that postnatal depression affects infants’ capacity to detect and expect social contingency (Field et al., 2005 and Nadel et al., 2005). Two experimental perturbation procedures have been used to examine infant sensitivity to social contingency. In the still-face procedure (Tronick, Als, Adamson, Wise, & Brazelton, 1978) the mother interrupts an on-going face-to-face interaction with the infant by holding a still face while she continues to gaze at the infant. Infants react to the mothers’ still face with gaze aversion and less positive affect. The still-face effect is thought to indicate infant sensitivity to changes in adults’ behavior and the tendency to react with distress because they expect adults to produce contingent responses and positive affect during the interaction. When presented with a blank face in the still face procedure, infants of depressed mothers show less negative reactions compared to infants of non-depressed mothers. A limitation with the still face procedure is that the infants’ responses to their mother's still face could be due to the distress witnessing their mother become unresponsive and expressionless rather than the shift in contingency. The double video procedure represents a more subtle and less stressful manipulation, where the infant interact face-to-face with their mother via a closed-circuit computer system (Murray & Trevarthen, 1985). Following an initial live sequence, where the mother and infant interact in real time, the infant is presented with a replay of the mother from the first live sequence. In the replay sequence, the mother's facial expressions and vocalizations are exactly the same as in the initial live sequence, except for the fact that the mother's behavior is unrelated to the infant's current behavior. Thus, in the double video procedure all parameters of communication are unchanged during the replay sequence, except for the interruption of the contingency in the interaction between the mother and infant. In their classical report, Murray and Trevarthen (1985) found that 6–12 weeks olds showed a decline in gaze and positive affect in the replay sequence compared to the initial live sequence. Subsequent studies failed to replicate this and attributed the original finding to various other mechanisms, such as increased fussiness over the course of the experiment (Rochat, Neisser, & Marian, 1998), familiarity or boredom with the adult (Bigelow and Birch, 1999 and Hains and Muir, 1996) memory of the mother's previous behavior (Hains & Muir, 1996) or differences in the degree to which the mothers mirror the infants’ affect (Legerstee & Varghese, 2001). However, when contingent interaction between mother and infant was established during the first live sequence, both Nadel, Carchon, Kervella, Marcelli, and Reserbat-Plantey (1999), Stormark and Braarud (2004) and Braarud and Stormark (2006) found that 2-to-3 month olds discriminated between contingent and non-contingent interaction with their mother, thus replicating Murray and Trevarthen's (1985) original finding. Furthermore, the infants were also able to re-engage with their mother in a subsequent contingent live sequence, after the interaction was set out of phase in the replay sequence. Using the double video procedure, both Field et al. (2005) and Nadel et al. (2005) found that infants of depressed mothers, like infants of non-depressed mothers were sensitive to the shift from contingent to non-contingent interaction, by showing reduced gaze at their mother during the non-contingent interaction. However, only the infants in Field et al. (2005) study reacted to the non-contingent interaction with a change in affect. In addition, while the infants of depressed mothers in Field et al. (2005) reengaged with their mothers and returned back to low levels of negative affect in the subsequent, contingent interaction, the infants of depressed mothers in Nadel et al. (2005) continued with gaze aversion and actually showed an increase in negative affect in the second contingent live interaction. Thus, the finding that infants of depressed mothers also detect shifts from contingent to contingent interaction could both reflect the same sensitivity to social contingency observed in infants of non-depressed mothers (Field et al., 2005) or be part of a persistent reaction to non-contingency (Nadel et al., 2005), related to the reaction of social withdrawal. The purpose of this study was to examine the extent to which young infants’ sensitivity to social contingency in the double video paradigm is related to their mothers’ level of depression, as an indicator of the infants’ past experience with contingent and responsive parenting. Building on the work by Field et al. (2005) and Nadel et al. (2005), we made several important modifications in the design and the recruitment of the subjects. First, in order to control for a possible memory bias where the infant recognize the maternal behaviour that is being replayed for them, we introduced a second replay sequence where the infants witnessed their mothers’ responses to a replay of non-contingent infant behavior, in accordance with the design developed by Stormark and Braarud (2004; see also Braarud & Stormark, 2006). If infants of mothers with higher depressive symptoms are sensitive to shifts in social contingency, we would expect that the infants would respond equally to both replay sequences. Second, the mothers were naive with respect to the experimental manipulation, unlike the mothers in both Field et al. (2005) and Nadel et al. (2005), who were informed in advance that the infants would be presented with a replay of their behavior during the experiment. We would like to argue that not disclosing the experimental manipulation would provide valuable information on how infants respond to shifts between contingent and non-contingent interaction that resembles naturally occurring interaction and processes of repair in a more ecologically valid manner. Third, the participants were recruited from a population-based sample where the mothers were screened for postpartum depression. As pointed out in their review, Lovejoy et al. (2000) suggest that deficits in parenting behaviour may not be related to depressive disorders per se, but the level of psychological distress. This warrants a more dimensional approach, targeting population-based, rather than comparing infants of clinically depressed with non-depressed mothers. If infants of depressed mothers are characterized by social withdrawal during regular interaction with their mothers we would expect low discrimination between contingent and non-contingent interaction sequences, expressed as an increase in gaze aversion and blunted emotional reactions in infants of mothers with elevated levels of depressive symptoms, also in a population-based sample. Alternatively, it is possible that gaze aversion and blunted emotional reactions would be more strongly related to the delimitating effect of a clinical diagnosis of depression and that infants of mothers with sub-clinical levels of depression would discriminate between contingent and non-contingent interaction sequences like infants of non-depressed mothers, expressed as a higher gaze focus at their mother and more positive and less negative affect when the interaction with the mother is contingent than non-contingent.

Results 3.1. Gaze focus There was a main effect of focus (F(1, 37) = 22.60, p < .01), independent of group and sequence, reflecting that the infants looked more at their mother than away from their mother. The planned comparison also yielded a marginal significant three-way group × focus × sequence interaction (F(1, 37) = 3.86, p = .057), reflecting that infants of non-depressed mothers had a higher gaze focus at their mothers than away from mother during the Live than the Replay sequences (F(1, 37) = 4.23, p < .05), while there were no such difference for the infants of sub-clinically depressed mothers, who showed a high gaze focus at their mother during both the Live and Replay sequences. Further, infants of non-depressed mothers looked more at than away from their mother during the Live (p < .05), but not during the Replay sequences. The infants of the sub-clinically depressed mothers looked more at than away from their mother in all of the sequences (all comparisons p < .05) (see Fig. 2). Mean percent of time and standard errors of infant amount of gaze at mother and ... Fig. 2. Mean percent of time and standard errors of infant amount of gaze at mother and away from mother during the Live1, Replay1, Live2, Replay2 and Live3 sequences in the sub-clinically and the non-depressed group. Figure options 3.2. Facial expression of affect There were no main effects, but the planned comparison revealed a three-way group × affect × sequence interaction (F(1, 37) = 7.83, p < .01), reflecting that the infants of the non-depressed mothers showed more smile than frown during the Live compared to the Replay sequences (F(1, 37) = 4.43, p < 05), while there was a tendency for the infants of the sub-clinically depressed mothers to show more frown than smile in the Live compared to the Replay sequence (F = 3.71, p = .06). Further, the infants of the non-depressed mothers showed more smile than frown in the live sequences (F(1, 37) = 7.75, p < .01), but not during the replay sequences. There were no differences between smile and frown during the live sequences, nor during the replay sequence for the infants of the sub-clinically depressed mothers, but there was an increase in frown from Replay1 to Live2 (F(1, 37) = 11.42, p < .01) (see Fig. 3). Mean percent of time and standard errors of infant amount of positive (smile) ... Fig. 3. Mean percent of time and standard errors of infant amount of positive (smile) and negative (frown) affect during the Live1, Replay2, Live2, Replay2 and Live3 sequences in the sub-clinically and the non-depressed group. Figure options In addition, there was a significant correlation between the infants’ gaze at (compared to away from) their mother and amount of smile (compared to frown) during the Live (compared to the Replay) sequences, but only for the infants of the non-depressed mothers (r = .41, p < 05).