Abstract
Bonding is described as a complex two-way process in which the child becomes emotionally attached to family members with central importance of the mother. It is a biological and adaptive process that enables the needs of the child for nurture and protection to be provided. It is a kind of a ‘control system’ in which signals sent by both mother and child maintain the proximity between mother and child. The later development of social relationships seems to be based on the bonding between mother and child [1–3].
Bonding as a two-way process is dependent on the contribution of the mother as well as of the child. Personality, physical and mental health of mothers can influence the bonding between mother and child [4] as well as the capacity of emotional expression of especially intense feelings towards the mother or of empathy with the feelings of the mothers or temperament of the child [5]. Kendler et al. [6] described the prediction of parenting by parental personality, psychopathology, educational level, religious fundamentalism, parental marital quality and child temperament as well as by neurotic/anxious traits in both parents and child. Parker et al. [7] summarized ‘that parent-child bonds would be broadly influenced by characteristics of the child (e.g. individual differences in attachment behaviour), characteristics of the parent or care-taking system (e.g. psychological and cultural influences), and by characteristics of the reciprocal, dynamic and evolving relationship between the child and the parent’.
Thus, the influence of childhood characteristics on bonding between mother and child seems to be important. Abnormalities in personality or social functioning may have consequences on maternal bonding behaviour. Preschizophrenic patients seem to be different from unaffected children [8], especially in their emotionality and interpersonal relationships [9] and especially from their same-sex, healthy siblings [10]. They often show poor premorbid social adjustment and premorbid schizotypal or schizoid personality traits [11–13]. McCreadie et al. [14] found a small, but consistent, correlation between both premorbid social functioning and personality, and perceived paternal rearing attitudes in schizophrenic patients. These patients who were described by their mothers as presenting more evidence of schizotypal or schizoid personality traits or poor social functioning rated their parents as treating them with less warmth.
Several studies reported less care and more overprotection in schizophrenic patients [15–17]. However, this type of parental bonding, called ‘affectionless control’, does not seem to be specific for patients with schizophrenia but for several patient groups like borderline personality disorder [17, 18], agoraphobia [19], panic disorder [20], drug-dependence [21], and depression [22, 23].
In the present study we have been interested in the perceptions of maternal bonding behaviour by schizophrenic and schizoaffective patients and their same-sex, unaffected siblings and in premorbid personality traits in the patients compared to their siblings. In order to look for relationships between maternal behaviour and premorbid personality traits and for correcting for possible significant associations we have reanalysed the ratings on maternal bonding behaviour, using the premorbid personality as covariates.
Methods
Subjects
Selection procedure for the samples of patients and siblings has been described in detail in a previous paper [10]. The inclusion criteria for the patients were diagnoses of schizophrenia or schizoaffective disorder, depressive type (core phenotypes within schizophrenia-specturm according to Cloninger [24]. The inclusion criteria for the siblings were, (i) same-sex, and (ii) no psychiatric disorder on axis I and II. Thirty-six patients, satisfying DSM-III-R [25] criteria for schizophrenia or schizoaffective disorder, depressive type and their same-sex siblings without a psychiatric disorder on axis I and axis II (consensus diagnosis) were included in the study. The patient sample consisted of 15 individuals (42%) with residual type of schizophrenia, 9 (25%) with schizoaffective disorder, depressive type, 8 (22%) with undifferentiated type of schizophrenia, 3 (8%) individuals with paranoid type of schizophrenia and 1 (3%) with disorganized type of schizophrenia. There was no significant difference between patients and their siblings in age (t-value = 1.55; p = 0.130). The mean age of patients when they entered the study was 31.9 years (SD = 9.9), the mean age of the siblings was 30.6 years (SD = 9.9). The number of female and male patients was equal. The mean age of onset (first medical treatment of psychotic symptoms) of male patients was 21.4 years (SD = 3.8) and of female patients 24.3 years (SD = 9.0). There was no significant difference between females and males in their age of onset (t-value = 1.2; p = 0.242). All patients had been hospitalized at least once, and all were on maintenance doses of antipsychotic medication. All subjects gave written informed consent prior to the study. The protocol was approved by the Institutional Ethics Committee.
Measures
Maternal bonding behaviour
Information about maternal bonding behaviour during the first 16 years was assessed on the Parental Bonding Instrument (PBI) [7] by self-rating of patients and siblings. Care was measured by 12 items on a dimension with one pole defined by empathy, closeness, emotional warmth, affection and the other by neglect, indifference and emotional coldness. Overprotection was evaluated by 13 items, ranging from overprotection, intrusion, excessive contact, control, infantilization and prevention of independent behaviour to autonomy and allowance of independence. The validity and reliability of the PBI have been shown to be acceptable and to be independent of the parent's sex. In addition, both scales of the PBI have been shown to be stable over time [26].
Premorbid personality traits
The ‘Gießen-Test’ [27] was used to retrospectively assess mother's perception of premorbid (defined as the period from birth to the age of 16 years according to the ‘PBI’) personality traits of patients and their siblings. This personality inventory contains six scales (social resonance, permeability, social competence, basic mood, dominance, control). Social resonance is evaluated on a continuum ranging from attractive, charming, popular, liked at work, generally accepted, and interested in looking good to unattractive, disliked, ignored, disliked at work, unaccepted, and uninterested in looking good. Permeability is ranging from communicative, close to other people, rather very exposed to others, starved for affection, confining and to be able to experience love to reserved, distant, rarely exposed to others, reservation in showing affection, suspicious, and unable to experience love. Social competence has one pole defined by being sociable, natural in heterosexual contacts, able to compete, able to have lasting relationships, and being imaginative while the opposite pole is defined as being unsociable, shy in heterosexual contacts, unable to compete, unable to have lasting relationships, and being unimaginative. Basic mood is assessed on a dimension ranging from seldom depressed, seldom thinking about problems, seldom anxious, hardly self-critical, to able to be angry and rather independent, often depressed, often thinking about problems, often anxious, often self-critical, unable to be angry, and rather dependent. Dominance ranges from often quarrelling with others, being wilful, often impatient, to seldom quarrelling with others, being compliant and being patient. Control ranges from being untidy, being easy going and being generous to being very tidy, being very eager and being stingy.
Statistical analyses
Group differences between patients and siblings were tested for significance by paired t-tests and by analyses of covariances with that premorbid personality traits as covariates, which were significantly different between patients and siblings. The cut-off level for statistical significance was set at P < 0.05, 2-tailed. Data handling and analyses were carried out using SPSS for Windows, Version 6.1.
Results
Maternal bonding behaviour
Significant differences were found in care and protection between patients and siblings. The schizophrenic and schizoaffective patients rated their mothers to be less caring and more overprotective (conceptionalized as ‘affectionless control’) than their same-sex, healthy siblings (see Table 1). The scores on the care scale in the patients and the scores on the care and overprotection scale in the healthy siblings are corresponding to the scores (same scales) in dizygote twins, discordant for schizophrenia [16]. The scores on the care and overprotection scales in the healthy siblings also correspond to scores in a norm-population [7]. Only the scores on the overprotection in our patients are much higher than the scores (same scales) in the dizygote, schizophrenic twins [16].
Maternal bonding behaviour and premorbid personality in patients and siblings. Mean scores, standard deviations in brackets
Premorbid personality
Patients and their same-sex, unaffected siblings were described as significantly different by their mothers in social resonance, permeability, social competence and basic mood. Mothers perceived their schizophrenic or schizoaffective children to be rather disliked and unaccepted (‘social resonance’), rather reserved, distant and suspicious (‘permeability’), rather unsociable and unable to have lasting relationships (‘social competence’) as well as often depressed, often selfcritical, and often anxious (‘basic mood’). The patients did not differ from their siblings in dominance and control (see Table 1).
Analyses of covariances
Analyses of covariances with the significant premorbid personality traits (‘social resonance’, ‘permeability’, ‘social competence’, and ‘basic mood’) as covariates showed that the significant univariate difference in care disappeared (F = 2.096; p = 0.154). Patients and their same-sex, unaffected siblings showed no more different perceptions of maternal care. The significant univariate difference in overprotection was still left (F = 4.247; p = 0.044). Permeability and social competence were the only significant influencing variables on care and overprotection. Permeability had a significant influence on both scales, care and overprotection; social competence had a significant influence on care, too. Being reserved, distant, unable to experience love and being rarely exposed to others in the first 16 years of life seemed to be connected with high maternal overprotection and less maternal care. The later seemed also be influenced by being unsociable, shy in heterosexual contacts and being unable to have lasting relationships (see Table 2).
Multivariate difference between patients and siblings in maternal care and overprotection, using premorbid personality traits as covariates
Discussion
In the current study we compared schizophrenic and schizoaffective patients, depressive type (core phenotypes within schizophrenia-spectrum according to Cloninger [24]), with their same-sex, healthy siblings with respect to maternal bonding behaviour and premorbid personality traits. We decided to take same-sex, healthy siblings as controls because, despite of studies pointing at differential experiences of siblings in the same family [28] patients and their same-sex siblings shared more of the same familial experiences than do patients and controls stemming from different families. The information about premorbid personality traits was obtained from a single informant, their mothers. Information about paternal bonding behaviour was restricted to maternal bonding behaviour because mothers were most frequently selected as key relatives [29].
In the present study preschizophrenic and preschizoaffective patients were described by their mothers to be more depressed, to be less socially resonant, to be more permeable, to be less socially competent, and to be more anxious than their same-sex, unaffected children. These results fit into earlier findings about premorbid characteristics in schizophrenic patients. Parents considered their schizophrenic children as showing significantly more childhood behaviour problems (e.g. ‘not liked by peers’) than their siblings with healthy adult outcomes [30]. Jones et al. [13] described associations between later schizophrenia with solitary play preferences at ages 4 and 6, self-reported anxiety in social situations at age 13, and teacher ratings regarding severe anxiety at school at age 15. McCreadie et al. [12] reported that patients with schizophrenia presented more schizotypal or schizoid personality traits in youth than their unaffected siblings.
In addition, schizophrenic and schizoaffective patients reported less maternal care and more maternal overprotection than their same-sex, unaffected siblings. Several studies found this ‘affectionless control’ used by schizophrenic patients to describe their mothers [15–17].
Our multivariate analyses showed significant associations between premorbid personality traits and maternal bonding behaviour. Being reserved, distant, unable to experience love and being rarely exposed to others in the first 16 years of life seemed to be connected with high maternal overprotection and less maternal care. The latter seemed to be also influenced by being unsociable, shy in heterosexual contacts and being unable to have lasting relationships. After correcting for the influence of the premorbid personality traits on perceived maternal behaviour the univariate significant difference in care disappeared, but the significant difference in overprotection still remained, presumably caused by high scores in overprotection in the patients.
There are at least six possible explanations for these associations between maternal bonding behaviour and premorbid personality traits:
1. Actual premorbid personality of subjects may lead to different maternal bonding behaviour. Maternal bonding behaviour could be influenced by child characteristics [6] and may be moderated by variables such as maternal personality [31]. According to Mangelsdorf et al. [32] the security of attachment could be predicted by an interaction between maternal personality and the child's proneness-to-distress. A difficult or demanding child may produce negative feelings in his/her mother possibly causing more negativity in the child. Both may then be caught in a vicious circle of negative feelings [16]. Children who were more difficult in childhood were more liable to experience rejection and overprotection. Otherwise, no statistical, but weak association of premorbid personality and social functioning with rejection and overprotection was found [14].
2. Actual premorbid personality of subjects may influence the ratings on maternal bonding behaviour. Few or no social contacts in childhood and youth may result in ‘no experience in social competence’. Furthermore patients could have a special form of perception of social stimuli [33]. Both may lead to extreme ratings concerning maternal bonding behaviour.
3. Actual maternal bonding behaviour may influence actual premorbid personality of subjects. Bowlby [1–3] discussed the importance of bonding between mother and child for later social relationships. Ainsworth et al. [34] distinguished three different patterns of attachment (secure attachment, anxious/resistant attachment, anxious/avoidant attachment) and their effect on later social relationships, especially disorder of social relationships and anxiety disorders. Parker et al. [7] reported four types of parental bonding: high care and low overprotection (conceptualized as ‘optimal bonding’), low care and low overprotection (conceptualized as ‘absent or weak bonding’), high care and high overprotection, (conceptualized as ‘affectionate constraint’), and low care and high overprotection (conceptualized as ‘affectionless control’). Each type of anxious attachment or each maternal bonding style, with the exception of optimal maternal bonding, may cause social anxiety and social incompetence in children. Canetti et al. [35] found associations of parenting and social functioning with distress in adolescents. Thus, low care and high control (‘affectionless control’) were linked with distress and isolation; high care and low control (‘optimal bonding’) with less distress, better social support and better general wellbeing. Latas et al. [36] pointed out that findings of their study suggest some specificity for the association between parental overprotection in childhood and personality disturbance in patients with panic disorder and agoraphobia.
4. Perceived maternal bonding behaviour may influence perceived premorbid personality of subjects. This explanation is rather impossible because the ratings by mothers and by children were performed independently.
5. The association between maternal bonding behaviour and premorbid personality traits may be a random one. This explanation also seems to be rather unlikely due to significant associations between maternal bonding behaviour and premorbid personality traits and due to the fact that the significant difference in the rating on maternal care behaviour between patients and siblings disappeared when considering premorbid characteristics as covariates.
6. The association between maternal bonding behaviour and premorbid personality traits is caused by a third factor. Retrospective ratings of the mothers could be influenced by the diagnosis in their children and subsequent events. Mothers may be looking for an explanation for the illness of their child and recall worse childhood behaviour for the child later affected than for the children who remained unaffected. Otherwise there might also be ‘idealization’ of childhood of the affected child. So, possible differences between the affected and the unaffected child could be underestimated [37]. There are also studies comparing two severe patient groups (patients with schizophrenic psychoses vs patients with affective psychoses) concerning their premorbid social functioning and personality traits, assessed retrospectively by mothers. In these studies schizophrenic patients were also described differently from patients with affective psychoses [11, 38].
The retrospective ratings on maternal bonding behaviour may be influenced by the presence of schizophrenia. Patients who show current symptoms of suspicions and social and emotional withdrawal may be more likely to recall their childhood as experiencing rejection and less warmth [14]. Otherwise Parker [39] could not find any effect of psychopathological symptoms on ratings on paternal bonding behaviour. Furthermore, ‘affectionless control’ (low care and more overprotection) has been reported for several psychiatric patient samples [40].
Summing up, we can confirm less social resonance, more permeability, less social competence and a more depressed and anxious mood in schizophrenic and schizoaffective patients as compared with their same-sex, unaffected siblings in their first 16 years of life. Also, the patients rated their mothers to be less caring and to more overprotective (conceptualized as ‘affectionless control’) than their siblings. But there were significant associations between maternal bonding behaviour and premorbid personality traits being emphasized by the disappearance of the significant difference in ratings on maternal care behaviour when considering premorbid personality traits as covariates. These results may suggest that premorbid personality traits should be taken into account in analyses of paternal bonding behaviour in schizophrenic and schizoaffective patients. Furthermore, it would be interesting if ratings on paternal bonding behaviour are different or changing in other psychiatric patient groups considering premorbid characteristics. Nevertheless, high maternal overprotection, perceived by patients with schizophrenia and schizoaffective disorder, still remained after correcting for the influence of premorbid personality traits.