Adrenal sensitivity was assessed in 16 non-depressed patients with NINCDS/ADRDA Alzheimer's disease (AD) and 18 control subjects by measuring cortisol response to low dose (0·05 μg/kg i.v.) exogenous adrenocorticotrophic hormone (ACTH). Controlling for sex and medication, both peak cortisol level (peak–baseline) and area under cortisol response curve (AUC above baseline) were significantly greater in AD subjects. This shows that HPA axis hyperactivity, as demonstrated by enhanced adrenal sensitivity to ACTH, occurs in AD. Similar findings have been reported to occur in depression. Among AD subjects, AUC cortisol response correlated with current age (r = 0·70, P = 0·001) and age at onset of dementia (r = 0·73, P = 0·001) and an inverse correlation was seen between cortisol AUC and cognitive test (CAMCOG) score (r = −0·51, P = 0·044). Our findings suggest that HPA axis hyperactivity in AD is associated with advancing age and cognitive dysfunction. Such changes may be cause, or consequence, of neuronal loss.

The neuropsychological performance and Magnetic Resonance Imaging (MRI) brain appearance of a consecutive series of 46 in-patients with anorexia nervosa (AN) was compared with hat of 41 normal-weight controls. The groups were matched for sex, age, estimated pre-morbid intelligence and education. AN patients who had gained at least 10% of their body weight were retested and rescanned. Controls were retested after a similar interval. The AN group performed significantly worse than the controls on tasks measuring attention, visuospatial ability and memory. On tasks assessing flexibility and learning, no group differences were evident although an examination of deficits in individuals revealed that more anorexics were impaired on both. Following treatment, the AN group improved relative to the control group only on those tasks assessing attention. Comparison of MRI measures showed a significant proportion of anorexics had enlarged lateral ventricles and dilated sulci on both cortical and cerebellar surfaces, but no dilatation was evident for the third and fourth ventricular measures. Improvements were found after treatment on some of the radiological measures but many differences remained. Relationships between morphological brain changes and cognitive impairments were weak. Lower weight, but not duration of illness, was associated with poorer performance on tasks assessing flexibility/inhibition and memory, and with greater MRI ventricular size.

The neural correlates of inner speech and of auditory verbal imagery were examined in normal volunteers, using positron emission tomography (PET). Subjects were shown single words which they used to generate short, stereotyped sentences without speaking. In an inner speech task, sentences were silently articulated, while in an auditory verbal imagery condition, subjects imagined sentences being spoken to them in an another person's voice. Inner speech was associated with increased activity in the left inferior frontal gyrus. Auditory verbal imagery was associated with increases in the same region, and in the left premotor cortex, the supplementary motor area and the left temporal cortex. The data suggest that the silent articulation of sentences involves activity in an area concerned with speech generation, while imagining speech is associated with additional activity in regions associated with speech perception.

To examine whether poor verbal fluency in schizophrenia represents a degraded semantic store or inefficient access to a normal semantic store, 25 normal volunteers and 50 DSM-III-R schizophrenic patients, matched for age, sex and IQ, were recruited. Although schizophrenic patients were impaired on both letter and category fluency, they showed a normal pattern of output in that category was superior to letter fluency, and an improvement in category fluency when a cueing technique was employed (Randolph et al. 1993). These results resemble those found in disorders of frontostriatal systems (Parkinson's and Huntington's disease) and suggest that poor verbal fluency in schizophrenia is because of inefficient access to semantic store. A measure of improvement with cueing was directly related to performance on the Stroop executive task. Of all symptom measures derived from SANS and Manchester Scales, only alogia was related to verbal fluency in that superior improvement correlated inversely with the degree of alogia. It is suggested that both alogia and poor verbal fluency are mediated by the same underlying cognitive abnormality that reflects frontostriatal dysfunction.

Associations between retrospective ratings of parental behaviour and adult affective symptoms were investigated in a British national sample. Symptom scores at ages 36 and 43 years showed low but significant correlations with care (negative) and control (positive), as measured by the Parental Bonding Instrument. Prevalence of high symptom scores was much greater in respondents with low care–high control (affectionless control) parents than in those with high care–low control parents, but there was no synergistic effect of combined care and control. Degree of affectionless control was progressively related to risk of depression. No significant gender differences were found in these associations. Findings could not be explained as spurious relationships resulting from association with other features of childhood adversity, and there was evidence that distorted recall arising from contemporaneous depressed mood was not responsible. Work is needed to establish the causal mechanisms underlying observed associations, including inter-relationships between parental style and other early adversity, and factors mediating or moderating the long-term effects of parental behaviour.

Potential mediators of the modest association between retrospectively rated parental behaviour and adult affective symptoms in offspring were investigated in a national longitudinal study of a cohort followed to the age of 43. Personality measures from adolescence could account for a small part of this association. Personal relationships in adulthood were more strongly associated with both parental behaviour and symptoms: marital history, emotional support, social network and availability of help in a crisis. Poor parenting did not lead to a general vulnerability to later life events, and socio-economic status and financial hardship were not implicated in the link between parental behaviour and adult symptoms. However, parental affectionless control was associated with certain types of life stressors in adulthood, i.e. interpersonal as opposed to non-interpersonal life events. Collectively, aspects of personal relationships accounted for much of the elevated symptom levels in those rating parents as low on care or high on control. Findings were consistent with the notion that interpersonal competence is important in the continuity between childhood experience and adult mental health, but other possible interpretations are discussed.

Childhood parental loss may be an important risk factor for psychiatric illness in adulthood. While this association has been carefully examined for depression, little is known about the role of parental loss in predisposing to alcoholism. We examined an epidemiological sample of female twin pairs with the same history of continuity or disruption in parent–child relationships (N = 1018 pairs; mean age 30 years), using a range of definitions of alcoholism. Childhood parental loss through separation, but not death, substantially increased the risk in adulthood for all definitions of alcoholism. Furthermore, both paternal and maternal alcoholism substantially increased the probability of parental separation from their children. Proposing a structural equation twin-family model that incorporates childhood parental loss as a specified environmental risk factor, we examined how much of the association between childhood parental loss and alcoholism was causal (i.e. mediated by environmental factors) v. non-causal (mediated by genetic factors, with parental loss serving as an index of parental genetic susceptibility to alcoholism). Both the causal and non-causal paths were significant for all definitions of alcoholism. However, the causal–environmental pathway consistently accounted for most of the association. While a significant proportion of the association is due to non-causal genetic mechanisms, childhood parental loss (or the familial discord that precedes or follows it) is probably a direct and significant environmental risk factor for the development of alcoholism in women.

From 1983–1994 a community mental health programme was set up in Guinea-Bissau. The first part of the programme concentrated on epidemiological aspects: rural and urban study areas were selected on socio-economic level and participation in the liberation war. A two-stage design was used to screen 351 adult consecutive general health care attenders and 100 children in a rural and an urban area for mental disorder. Psychiatric disorders have a morbidity of 12% among adults seen in Primary Health Care. Disorders were mainly neuroses (74%), but more psychoses were found than in other countries. No statistically significant difference in morbidity was found between rural-urban areas or between previous war and non-war zones. The diagnostic sensitivity of the Primary Health Care workers was 31%, their diagnostic specificity 88%. Thirteen per cent of the children showed neuropsychiatric disturbances. There were no sociocultural impediments to this public mental health approach. During the following intervention programme 850 Primary Health Care workers were trained and supervised nationwide. The diagnostic sensitivity of major mental disorders and epilepsy increased from 31% to an average of 85%. Before the training, their knowledge of the treatment of these disorders was nil whereas after training 82% of the patients received appropriate treatment. Moreover, this model programme shows a profitable cost/benefit ratio and a high sustainability over the last 10 years.

The question addressed to in this paper is whether severely mentally ill patients are treated differently in a community mental health service without the back-up of a mental hospital (South-Verona, Italy) compared with an institution-based system in which mental hospitals, although highly modernized, are still predominant (Groningen, The Netherlands). Using the psychiatric case-registers in both areas, the patterns of care in 2 years of follow-up of schizophrenic patients were constructed. Survival analysis was used to analyse in-, day- and out-patient episodes of care. Three-quarters of the Groningen and half of the South-Verona patients experienced at least one episode of hospitalization; 20% of the Groningen and 5% of the South-Verona patients were long-stay patients at the end of the observation period. The South-Verona patients had more episodes of in-patient and especially of day-patient and out-patient care. Cox's regression showed that the duration of episodes controlled for the history of events and sociodemographic characteristics, was significantly shorter in South-Verona. One of the main conclusion was that hospitalizations for the severely mental ill are also needed in a community-based system of care, supporting the assumption of a ‘bed-rock’ of mental illness. However, the South-Verona community mental health service seems to be able to reduce the duration of hospitalizations considerably.

The effects of life events, social support and marital relationships on outcome were examined in a predominantly recurrent in-patient sample of depressives studied longitudinally every 3 months to remission and up to 15 months thereafter. Outcomes examined were length of time to remission, presence of residual symptoms at remission, and subsequent relapse. There were few associations between these outcomes and the social variables. These findings add to other recent evidence that psychosocial factors are relatively unimportant in the subsequent course of severe and recurrent depressions, in contrast to their contribution to onset of such depressions and subsequent outcome of milder depressions.

The point prevalence of depression measured on one occasion between 2 and 5 months after childbirth was estimated to be 13·1% in 352 mothers living in urban Portugal; the criterion was a score of 13 or more on a translated version of the Edinburgh Post-natal Depression Scale (EPDS). More detailed comparisons were made between a subgroup of 118 mothers and 118 matched controls who had not borne a child in the previous 2 years. Post-natal women were twice as likely as non-childbearing controls to meet the EPDS criterion for depression. In comparison with controls, they were also more severely depressed as judged by their total scores on another questionnaire, the Zung Scale. Comparisons of individual symptom scores (Zung Scale) showed that childbearing women, as a whole, reported more somatic symptoms than controls, but when only those women judged to be depressed or dysphoric by the EPDS were compared, this difference disappeared. Stepwise logistic analyses of symptoms contributing to the categorization of a ‘case’ of post-natal versus non-post-natal depression did not reveal any very clear divergences in selfreported psychopathology. In childbearing women, two factors were found significantly to contribute to higher depression scores; women with more children and those from lower socioeconomic groups were most at risk.

The birth distribution of 399 cases of Alzheimer's disease (AD) identified in the region of Saguenay-Lac-St-Jean (Québec) was compared with that of: (a) the population currently living in the area; and (b) the population born during the same period in the area. AD cases have been recruited since 1986 by the IMAGE Project. Cases and controls were grouped according to the month of birth and according to the day of birth using density estimation. Analyses showed a significant deficit of births in the month of May. We believe these preliminary results deserve further attention and we suggest two possible explanations that could lead to a deficit of AD births at specific periods during the year.

This study reports the results of a comparison of DSM-III-R and ICD-10 personality disorder criteria by application of both sets of criteria to the same group of patients. Despite the clinical relevance of these disorders and the need for reliable diagnostic criteria, such a comparison has not previously been reported. DSM-III-R and ICD-10 have converged in their classification of personality disorders, but some important differences between the two systems remain. Personality disorder diagnoses from both systems were obtained in 52 out-patients, using the Standardized Assessment of Personality (SAP), a brief, informant-based interview which yields diagnoses in both DSM-III-R and ICD-10. For individual personality disorder diagnoses, agreement between systems was limited. Thirty-four subjects received a personality disorder diagnosis that had an equivalent form in both systems, but only 10 subjects (29%) received the same primary diagnosis in each system. There was a difference in rate of diagnosis, with ICD-10 making significantly more personality disorder diagnoses. The lower diagnostic threshold of the ICD-10 contributed most of this effect. Further modifications in ICD-10 Diagnostic Criteria for Research (DCR) and DSM-IV to the personality disorder category have been considered. The omission in DSM-IV of three categories unique to that system and the raising of the threshold in ICD-10 DCR, do seem to have been helpful in promoting convergence.

The aim of this study was to identify underlying dimensions of psychopathology in a cohort of patients with functional psychosis of recent onset, and to examine their prognostic value. Factor analysis of the psychopathological features of 166 consecutively admitted patients with functional psychosis of recent onset revealed seven psychopathological dimensions, which explained 63% of the variance. Five of these seven syndromes bore differential associations with subsequent treatment and illness course, independent of: (i) associations with DSM-III-R diagnosis; (ii) associations with other prognostic factors; and (iii) associations with the baseline values of outcome variables. The most striking associations were shown for an early and insidious onset syndrome with affective flattening, which predicted a more disabled course of illness on three of four outcome dimensions, and which was more common in males and unmarried individuals. A second syndrome, characterized by bizarre behaviour, inappropriate affect, catatonia, and poor rapport showed similar, slightly less striking, associations with illness course, as well as with poor pre-morbid social functioning. A third syndrome, characterized by positive psychotic symptoms was to a lesser degree associated with poorer outcome, whereas a fourth syndrome distinguished by manic symptomatology predicted a more benign illness course. A fifth syndrome identified by lack of insight predicted more time in hospital and admission under a section of the Mental Health Act during the follow-up period.

A further finding was that dimensional representations of psychopathological features were considerably more useful than categorical representations (DSM-III-R and ICD-10) as predictors of illness course and treatment decisions.

A comprehensive semi-structured questionnaire was administered to 100 psychotic patients who had experienced auditory hallucinations. The aim was to extend the phenomenology of the hallucination into areas of both form and content and also to guide future theoretical development. All subjects heard ‘voices’ talking to or about them. The location of the voice, its characteristics and the nature of address were described. Precipitants and alleviating factors plus the effect of the hallucinations on the sufferer were identified. Other hallucinatory experiences, thought insertion and insight were examined for their inter-relationships. A pattern emerged of increasing complexity of the auditory–verbal hallucination over time by a process of accretion, with the addition of more voices and extended dialogues, and more intimacy between subject and voice. Such evolution seemed to relate to the lessening of distress and improved coping. These findings should inform both neurological and cognitive accounts of the pathogenesis of auditory hallucinations in psychotic disorders.

Case studies of patients with familial schizophrenia may help to define the pathophysiology of this illness and indicate potential candidate genes for genetic linkage studies. In this regard, the clinical, radiological and pathological assessments of a 39-year-old affected man from a pedigree with familial schizophrenia are presented. Brain imaging with CT indicated moderate cortical atrophy, particularly of the temporal lobes. Neuropathological examination revealed granular ependymitis, indicating possible past ventricular pathology. Granular ependymitis was reported to occur in genetic developmental disorders with neuronal migration abnormalities. In the present case, heterotopic clusters of neurons were visualized in the entorhinal cortex, suggesting that temporal lobe development was not entirely normal. This case study suggests that genetic factors could be investigated further as one possible aetiology of certain neurodevelopmental abnormalities observed in schizophrenia.

In 15 women with Late Luteal Phase Dysphoric Disorder (LLPDD) and in 15 normal control subjects, personality traits were assessed using the Millon Clinical Multiaxial Inventory (MCMI) during follicular and luteal menstrual-cycle phases. Compared with controls, LLPDD subjects had less compulsive but more passive/aggressive and borderline/cycloid traits, and more depression and hypomania. Menstrual-cycle phase did not significantly affect personality variables in either group. In particular, depression and hypomania in LLPDD subjects suggests a relationship with affective disorders.

Age and sex-adjusted first admission rates for operationally-defined schizophrenia and other non-affective psychosis in different ethnic groups were calculated over the period 1988–1992 in a defined catchment area in South London. Standardized rates for schizophrenia, corrected for age- and gender-related under-reporting in the 1991 census and a 20% underestimate of the size of the ethnic minority populations in the area, were not only higher in the Afro-Caribbean group (SMR: 3·1; 95% Cl: 2·0–4·7), but also in the African group (SMR: 4·2; 95% Cl: 2·8–6·2). It was further found that higher rates were not specific to schizophrenia. These findings suggest that some common factor associated with ethnic minority membership is important in producing an excess of psychotic illness.