Attachment to the mother at 1 year of age was studied in 46 "high-risk" offspring of index mothers with a history of nonorganic psychosis and in 80 demographically similar control offspring. Attachment was measured in the home in a standardized manner by a modified version of Ainsworth's Strange Situation procedure. Three different patterns, representing "secure attachment" (Type B), "anxiously avoidant attachment" (Type A) and "anxiously ambivalent attachment" (Type C), were defined. While no significant difference on attachment type was found between the total index versus control group, a significantly increased rate of anxious attachment (A + C) was found for offspring of schizophrenics but no other diagnostic group. Attachment type was unrelated to index mothers' psychiatric hospitalization, psychotic status and known mental disturbance during the infants' first year of life, as well as to the infants' sex.

Mother-infant interaction during feeding and in an unstructured play situation was studied in the home at 3 weeks and 6 weeks of age in index mother-infant pairs in which the mother had a history of nonorganic psychosis (n = 42 and 51 at 3 and 6 weeks, respectively) and in demographically similar control pairs (n = 60 and 78). At both ages, interaction was significantly more negative in index than control cases, index mothers showing increased tension and a lack of harmony, decreased social contact, and reduced sensitivity to the infant's needs. Fewer significant differences were found between index and control infants. Mothers in the Schizophrenic, Cycloid and Nonendogenous groups evidenced more negative interaction characteristics than did their matched controls, but the Affective group was not in any way more negative than its controls.

Mother-infant interaction during feeding and in an unstructured play situation was studied in the home at 1 year of age in 46 index mother-infant pairs in which the mother had a history of nonorganic psychosis and in 80 demographically similar control pairs. As was true at the five previous observation ages, some aspects of the interaction were significantly more negative in index than control pairs. Index mothers showed increased tension and uncertainty regarding the infant's needs, increased physical contact during feeding and a discrepancy between the intonation and content of the mother's verbal contact with the infant. Index infants did not differ from controls in behavior in interaction. Across the entire first year, the Cycloid and Schizophrenic mothers deviated most frequently from controls, while the Affectives' interaction was more negative than controls' for the first time at the 1-year observation.

Mother-infant interaction during a feeding was studied at about 3 days of age in 51 index mother-infant pairs in which the mother had a history of nonorganic psychosis and in 73 demographically similar control pairs. Interaction was significantly more negative and deviant in index than control cases, index mothers establishing a significantly more negative emotional climate and less harmony in feeding, evidencing more tension and uncertainty, and showing less social contact toward the infant. Index infants showed significantly less social contact toward the mother and index pairs less reciprocal visual contact than did controls. While the Schizophrenic, Cycloid and Nonendogenous Psychoses groups each differed from their matched controls on a number of variables, the Affective group was not in any way more negative than its controls on interaction.

Mother-infant interaction during feeding and in an unstructured play situation was studied in the home at 3.5 and 6 months of age in index mother-infant pairs in which the mother had a history of nonorganic psychosis (n = 48 and 52 at 3.5 and 6 months, respectively) and in demographically similar control pairs (n = 80 and 79). Interaction was significantly more negative in index than control cases at both ages, index cases showing decreased maternal and infant social contact and reduced maternal sensitivity to the infant's needs. Schizophrenic and Cycloid groups evidenced more negative interaction characteristics than did their matched controls, while the Affective group was not generally more negative than its controls.