Background: It is known that both severe mental illness and violence have genetic components. Multiple genes play a role in the cause of violent behaviour. Violence is one of the leading causes of death for young people in South Africa and yet little is known about its prevalence in state patients and their family members. Aim: This study aimed to investigate the prevalence of violence and mental illness in the families of state patients, to what extend these coincide and to compare schizophrenia and mood psychosis in that context. Setting: The study included 60 state patients' folders, all of whom were diagnosed with a psychotic disorder, in accordance with DSM5 criteria. The subjects were divided into two groups: those who have committed violence and those who have not committed violence. Method: Patients' folders were selected by purposive sampling. These folders were then reviewed by the researcher and a questionnaire was completed. Results: Violent patients had more first-generation relatives with violent convictions (68.2%), compared to 36.4% of non-violent patients. Only 3.3% of non-violent patients witnessed domestic violence, whereas 13.3% of violent patients witnessed domestic violence. A significantly higher proportion of patients with bipolar disorder had been convicted of physical assault (p=0.035). 17.6% of violent schizophrenia patients had a family history of violence and mental illness, compared to 18.2% of violent patients with mood psychoses which is not statistically significant. Conclusion: It was found that violence runs in families and that mental illness and violence was prevalent in the described group. Of further concern was that more violent patients witnessed domestic violence compared to non-violent patients, emphasising the idea that the cause of violence is multifactorial (genetic, environmental), and that identification of not only high-risk patients but also high-risk families need to be implemented.

Thesis (M.Cur.) -- University of Limpopo, 2015 / The purpose of the study was to investigate the experiences of families towards psychiatric state patients during LOA. A qualitative, phenomenological, exploratory, descriptive and contextual design was used. Non-probability sampling of the purposive type was used to explore and describe the experiences of families towards psychiatric state patients during LOA. Data were collected from ten (10) families by using semi-structured face-to-face interviews. Data were audio recorded and field notes were also written. Trustworthiness was ensured by applying credibility, transferability, confirmability and dependability. Transferability was ensured by utilizing purposive sampling to include participants. Confirmability was ensured by collecting data from the participants who have experience on the problem studied. Data were analysed qualitatively by using Tech’s open coding method. Results from the study shows that families experiences uncontrolled and disruptive behaviour from the psychiatric state patients during LOA. Guidelines for the study include: Families of psychiatric state patients should receive counselling and be involved in family therapy prior and after the psychiatric state patient’s LOA. Family members of psychiatric state patient experience emotional pain, therefore they need support from multidisciplinary team members to assist them to facilitate promotion, maintenance and restoration of their mental health as an integral part of their health. Recommendations include improvement of aspects in nursing practice, nursing education and nursing research. Key words: Experiences, family, psychiatric state patients, Leave of absence. DEFINITION

Nous n’avons pas conscience de l’ensemble des processus réalisés par notre cerveau à chaque instant. Cette dissociation entre l’expérience subjective et l’activité neuronale présente un défi majeur à la fois pour les neurosciences fondamentales, mais également pour la pratique clinique. En effet, non seulement les mécanismes neuronaux de la prise de conscience sont mal compris, mais il reste extrêmement difficile de déterminer si des patients en état végétatif – éveillés mais non-communicants – perçoivent leur environnement consciemment. Ces questions théorique et clinique constituent les deux axes principaux de cette thèse. Dans un premier temps, je développe, à partir des récentes avancées aussi bien empiriques que théoriques, une série d’outils permettant de caractériser les mécanismes neuronaux et computationnels de la perception consciente. En particulier, je montre dans une première étude comment les analyses de classification multivariée permettent de décoder les signaux magnéto- et électro-encéphalographiques à l’échelle de l’essai unique. De plus, dans trois études successives, je propose de nouvelles méthodes de traitement du signal permettant de i) caractériser la structure dynamique des processus évoqués par une stimulation sensorielle ii) de quantifier la quantité d’information échangées entre différentes régions corticales et iii) d’estimer la complexité des réponses cérébrales. Enfin, je montre comment un modèle mathématique utilisant les principes d’inférence bayésienne permet de rendre compte d’un grand nombre de résultats observés dans les études de la perception consciente et inconsciente. Dans un second temps, j’applique ces méthodes aux EEG d’une large cohorte de patients végétatifs, minimalement conscients et conscients. Les résultats montrent que les patients végétatifs présentent i) une altération des réponses corticales tardives évoquées par une stimulation auditive, ii) une diminution de l’échange d’information entre régions cérébrales, iii) des rythmes EEG moyens et lents (< 13Hz) anormaux et iv) une réduction de la complexité de l’activité EEG. A l’avenir, ces différentes signatures neurales de la conscience pourraient être utilisées en synergie pour décoder le contenu conscient et aider au diagnostic, au pronostic et au monitoring des patients non-communicants. / We are not aware of everything our brain does. This dissociation between subjective experience and objective neural activity challenges both theoretical neuroscience and clinical practice. Indeed, not only are the neuronal mechanisms of conscious perception poorly understood, but it remains extremely difficult to deter-mine whether vegetative state patients – who are thus awake but non-communicating – perceive their envi-ronment consciously. These theoretical and clinical questions constitute the two main axes of this thesis. In a first part, I develop, from the recent empirical and theoretical advances, a series of methods to characterize the neural and computational mechanisms of conscious perception. In particular, I show in a first study how multivariate pattern classifiers can decode magneto- and electroencephalographic recordings at the single trial level. In three successive studies, I then propose new signal processing methods to i) characterize the dynamical structure of stimulus-evoked processes ii) quantify the amount of information exchanged across cortical regions and iii) estimate the complexity of cerebral responses. At last, I show how a mathematical model based on Bayesian inference principles, can account for a large number of empirical findings observed in studies of conscious and unconscious perception. In a second part, I apply these methods on EEG recordings acquired from a large cohort of vegetative, minimally conscious and conscious patients. The results show that vegetative state patients present i) impaired late and sustained sound-evoked brain responses, ii) a reduction of the exchange of information across cortical regions iii) abnormal slow and medium EEG rhythms (<13Hz) and iv) a decrease of the EEG complexity. Ultimately, these various neural signatures of consciousness could be used in synergy to decode conscious contents and help to diagnose, predict and monitor the state of consciousness of non-communicating patients.