Maternal postnatal depression is associated with difficulties in mother–child interactions (Field, 1995
) and, in particular, with poor maternal responsiveness (Murray et al., 1993
; Stanley et al., 2004
). Research suggests that these difficulties might be due, in part, to the fact that depressed mothers perceive their infants’ behaviour more negatively than non-depressed mothers (Field et al., 1993
). As most early infant communication derives from facial expressions (Papousek and Papousek, 1977
), one possible explanation is that mothers with postnatal depression are differentially affected by, and are less sensitive to, infant facial expressions and thus appraise them more negatively.
Such a pattern of negative perceptions is believed to be a function of broader information processing biases that are commonly observed in depressed people (Mathews and MacLeod, 2005
; Donaldson et al., 2007
). For example, depressed mood has been shown to be associated with negative interpretations of ambiguous facial expressions (Hale, 1998
; Lee et al., 2008
). Also, depression has been associated with an attentional bias towards negative emotional cues in facial expressions (i.e. sad emotions) (Gotlib et al., 2004
), although evidence for attentional biases towards negative stimuli in depression has not been consistent (Bouhuys et al., 1996
; Yoon et al., 2009
). Interestingly, one study which has examined pregnant women’s responses to infant facial expressions (Pearson et al., 2010
) showed that whereas non-depressed pregnant women revealed an engagement bias towards distressed infant faces, those who were depressed tended to actually disengage quicker from the images. The authors suggest that an avoidance mechanism may therefore operate.
Two of our own recent studies focused on processing biases in postnatally depressed mothers. In the first study (Stein et al., 2010
), we found evidence of an appraisal bias, as mothers suffering from postnatal depression were more likely to rate sad infant faces as more negative than controls; in particular, if the faces were shown for a longer period of time (2000 ms). In the second study, using a morphed faces task, mothers with depression were less likely to accurately identify happy infant faces than controls, although there were no group differences in the processing of sad faces (Arteche et al., 2011
). These findings in relation to the processing of positive faces are consistent with studies using morphed adult faces (e.g. Joorman and Gotlib, 2006
), and with other recent evidence that indicates that depression is characterized by difficulties in processing positive affect perhaps even more than by biases in processing of negative affect (Deveney and Deldin, 2004
; Surguladze et al., 2004
Given the crucial role of early mother–child interactions in later child development (Yarrow et al., 1984
), it is vital to investigate potential interventions that might help to ameliorate the negative cognitive bias and, in particular, the negative effects of depression on the appraisal of facial expressions. To date, most studies using pharmacotherapy for depression have focused on the effects of antidepressants on negative mood. More recently, research using healthy volunteer models has demonstrated that both selective serotonin reuptake inhibitors (SSRIs) and selective noradrenaline reuptake inhibitors (SNRIs) have direct effects on the processing of emotional information, in particular reducing the processing of negative emotional material, such as fearful and angry faces, and increasing the perception of, and memory for, positively valenced emotional material. For example, seven days treatment with either the serotonergic antidepressant, citalopram, or the noradrenergic antidepressant, reboxetine, has been shown to positively bias the appraisal of ambiguous facial expressions, with decreased recognition of negative facial expressions such as fear and anger (Harmer et al., 2004
). In addition, SSRIs, but not SNRIs, have been shown to be useful in the treatment of anxiety, potentially by having an effect on emotional processing biases that are relevant to anxiety (Dhillon et al., 2006
). For example, Murphy et al. (2009)
have shown that the SSRI citalopram, but not the SNRI reboxetine, reduces attentional vigilance towards threat-related stimuli.
The current study aimed to examine the effects of two different antidepressants (reboxetine and citalopram) on the processing of infant-related emotional information in a group of healthy volunteers. In particular, we sought to explore the effects of these antidepressants on the appraisal of infant facial expressions of emotion using a faces rating task, and on attention to infant emotion using an attentional probe task. It was hypothesized that both drugs would be associated with a positive bias in the appraisal of infant faces of emotion (i.e. that positive faces would be rated more positively) and that citalopram, but not reboxetine, would be associated with a reduction in attentional bias to negative infant emotion.