The study investigated which client, therapist, and therapy characteristics were associated with the quality of the therapeutic relationship in Cognitive-Behavioural therapy (CBT) for psychosis. The development of a good therapeutic relationship is essential to the work of CBT for psychosis but is often more difficult to achieve due to the nature of psychotic symptoms. Despite this, there has been little research investigating what affects the development of the therapeutic relationship within CBT for psychosis. The study had a cross-sectional, correlational design with measures taken at around the sixth session of therapy. Clients and therapists completed questionnaires measuring client, therapist, and therapy factors, and their perceptions of the therapeutic relationship. On average, both clients and therapists rated the therapeutic relationship as good. Although there was a trend towards agreement on the quality of the therapeutic relationship, on average clients rated it higher than therapists. There were few significant effects of client or therapist factors on the therapeutic relationship, including some unexpected negative results, e.g. no effect for psychotic symptoms. There were some significant effects for therapy factors and clients reports of the therapeutic relationship, e.g. the presentation of a case formulation.

Bipolar disorder (BD) is a common highly heritable disorder. The calcium channel gene family has been widely implicated in BD aetiology and these genes include CACNA1C, and CACNG4. The association signal for CACNA1C with BD is located in the middle of the third intron of the gene. CACNG4 encodes a transmembrane AMPA receptor regulator that is involved in trafficking AMPA receptors to the neuronal post-synaptic density. High-resolution melting curve (HRM) analysis and whole genome sequencing (WGS) methods were used to identify functional variants in calcium channels genes in the UCL BD cohort. Variants that were predicted to impact gene regulation, transcription or to be damaging to protein structure were genotyped in the larger UCL BD and control cohort. HRM analysis identified 26 calcium channel gene variants. These included two non-synonymous CACNG4 variants that were associated with mental illness (rs371128228, p=1.05x10-4, OR=4.39 and 17:65026851 (C/T), p=5x10-4, OR=9.52). Fluorescent activated cell sorting analysis was used to determine the effect of rs371128228 on trafficking of GluR1 and GluR2 to the cell surface. This analysis demonstrated that the risk allele of rs371128228 significantly decreased cell surface trafficking of GluR1 (p=0.026) but no effect was observed on GluR2 trafficking. WGS analysis of CACNA1C intron 3 identified two BD associated (p=0.015, OR=1.15) variants 105bp apart that were in complete LD. Both variants are predicted to create YY1 transcription factor binding sites. Luciferase reporter assays show a significant decrease in gene expression in the presence of both variants (p=0.004). Protein-DNA complex assays of the CACNA1C variants demonstrate increased nuclear proteins binding affinity for the variant alleles. If these calcium channel variants are confirmed to be important risk factors for BD they could be used as markers for personalised treatment or in the identification of genetic subtypes of BD or other psychiatric illness.

Psychotic disorders, including schizophrenia and bipolar disorder, are amongst the most severe and enduring mental illnesses. Recent research has identified several genetic variants associated with an increased risk of developing psychosis; however, it remains largely unknown how these lead to the illness. This is where endophenotypes – heritable traits associated with the illness and observed in unaffected family members of patients – could be valuable. Endophenotypes are linked to the genetic underpinnings of disorders, and can help elucidate the functional effects of genetic risk variants. This thesis investigates endophenotypes for psychosis, with the overall aim of identify such biological markers, as well as to examine the relationships between different endophenotypes and their associations with genetic risk for psychosis. A family design has been used throughout, including patients with psychosis, their unaffected first-degree relatives, as well as healthy controls. In chapter 1, I review the endophenotype approach and those markers proposed for psychosis genetic research. Chapters 2 and 3 investigate whether different neurophysiological measures are potential endophenotypes for psychosis. In chapter 2, resting state EEG was studied and it was shown that risk groups, including unaffected relatives and people with an at-risk mental state, presented no abnormalities. This suggests that – rather than endophenotypes – the low frequency electrophysiological abnormalities seen in chronic patients in this study might be related to illness progression or long-term medication effects, and be more useful as biomarkers in non-genetic research. In chapter 3, I used dynamic causal modelling to investigate effective connectivity – the influence that one neuronal system exerts over another – underlying the mismatch negativity evoked potential, a marker of pre-attentive auditory perception. Results indicate that, compared to controls, both patients and their relatives show abnormalities of the excitability of superficial pyramidal cells in prefrontal cortex. Hence, this appears to be linked to the genetic aetiology of psychosis, and constitutes a potential endophenotype. Chapters 4 and 5 investigate several pre-identified endophenotypes for psychosis: Electrophysiological (the P300 event related potential), cognitive (working memory, spatial visualisation, and verbal memory), and neuroanatomical (lateral ventricular volume). In chapter 4, the associations between these endophenotypes were examined. Results showed that the P300 amplitude and latency are independent measures; the former indexing attention and working memory and the latter possibly a correlate of basic speed of processing. Importantly, individuals with psychosis, their unaffected relatives, and healthy controls all showed similar patterns of associations between all pairs of endophenotypes, supporting the notion of a continuum of psychosis across the population. Lastly, in chapter 5, polygenic risk scores – a measure of the combined effect of a large number of common genetic risk variants – were used to investigate the relationships between genetic risk for schizophrenia and bipolar disorder, and the endophenotypes studied in the previous chapter. Results showed that higher polygenic score for schizophrenia nominally predicts poorer performance on a spatial visualisation task; providing some evidence that the two traits share genetic risk variants as hypothesised. No other associations approached significance, possibly due to insufficient statistical power. However, as discovery samples grow, the use of polygenic scores is promising. This thesis has thus contributed to the field of mental health research by investigating key electrophysiological, cognitive and imaging endophenotypes for psychosis, as well as their genetic influences. Well defined and reliably measured endophenotypes are valuable in mental health research by clarifying the functional effects of identified genetic risk factors, and by providing ways of identifying groups of people with similar abnormalities, both within and between current diagnostic categories.

BACKGROUND: Improving the quality of life (QOL), of people with dementia living in care homes is a priority. People living in care homes with more severe dementia are often unable to self-report QOL, so proxy ratings are needed. However, we do not know if, or how, paid and family caregiver proxy reports differ in care homes. Aim: To compare paid and family caregiver DEMQOL-Proxy-reports of care home residents with dementia and investigate any differences. METHODS: This study is nested in an epidemiological study of 86 care homes. I compared DEMQOL-Proxy total and global ratings of 1,056 pairs of staff and family carers and explored the associated factors using multilevel modelling. I interviewed 12 staff and 12 relatives about their proxy ratings and analysed their answers thematically. RESULTS: Proxy ratings were weakly correlated (rs = 0.35, p <0.001); paid carers’ median total scores were higher than family carers (104 vs 101; Z = -7.15, p < 0.001). Family carer global ratings were also more negative (X2 = 20.69 (N = 1,016), p < 0.001). Staff and family rated QOL as better when residents had fewer neuropsychiatric symptoms. Staff who were native English speakers rated residents QOL as better. Staff rated residents with severe dementia more highly on QOL. Resident’s QOL was rated more highly in homes with lower staff:resident ratio by staff. Family also rated QOL as higher when the resident spoke English as a first language, had lived for longer in the home and had no recent hospital admission. Resident’s spouses rated residents’ QOL higher than children. Qualitative results suggest differences arise because staff felt good care gave high quality of life but families experience loss and sadness at dementia and care home placement. CONCLUSION: Proxy reports are influenced by the rater and differ systematically between family and paid carers.

In 2014, the British Psychological Society and Division of Clinical Psychology (BPS/DCP) published a report outlining the multiple ways of thinking about psychosis or “schizophrenia” beyond a dominant bio-medical framing. They highlighted that psychosis is a contested area. One area of contention has been the role of dissociation as a response to traumatic experiences leading to psychosis. The current literature review aimed to evaluate the role for dissociation in the relationship between trauma and psychosis. Eighteen, peer-reviewed journal articles were included. There was evidence that dissociation is an important aspect of the relationship between trauma and psychosis. The type, frequency and chronicity were found to be important factors. The mechanisms of dissociation remain to be elucidated. The current research study aimed to engage psychiatrists in narrative accounts of their understanding of psychosis as a contested area considering the BPS/DCP report. Semi-structured interviews were conducted with 12 participants at both trainee and qualified level. Critical Narrative Analysis was used which comprises six stages. The results highlighted the role of power, the barriers to opportunities for exposure to alternative views, and the role of dialogue and reflexivity. The Critical Appraisal considers, using a reflective account, the strengths and limitations of the current research, along with reflections on the whole research process.

Background: Theory associates shame with violence but research is inconsistent. The Compassion Focused Therapy shame concept distinguishes internal shame, other shame and social rank, offering a novel research approach. Adverse and traumatic experiences have been associated with violence in adulthood. Aims: This study aimed to distinguish between internal, other and social rank shame with the intention of introducing a relational and social understanding of shame and violence. Secondly, it aimed to explore developmental psychopathology theories of violence by profiling the central and traumatic features of male offenders’ shame memories. Method: Drawing on a pragmatist philosophy, this study adopted a cross sectional, quantitative approach. Male offenders (N = 121) in a young offenders prison were recruited via the healthcare suite. Participants were invited to complete a series of established self-report questionnaires via one to one interview. Two questionnaires required responses with reference to a strong shame memory. Results: Multiple regression analysis found proactive aggression was predicted by other shame, social rank and shame memory avoidance. Only other shame and participant age were independent predictors of proactive aggression. Reactive aggression was predicted by internal shame, other shame, shame memory avoidance and hyperarousal, however only age independently predicted reactive aggression. MANCOVA found no differences between groups with and without physical violence risk alerts in terms of shame when controlling for age. Structural Equation Modelling identified social rank and other shame as mediators of proactive aggression. Black and Asian/Other ethnic groups had significantly higher levels of social rank but not aggression. Conclusion: Although physically violent and nonviolent groups did not differ in terms of shame, different shame variables predicted proactive and reactive aggression in the whole population. The structural equation model is a novel analysis of proactive aggression. Ethnic differences in social rank are discussed in terms of BME overrepresentation in the criminal justice system.

Background: Though we know that psychological interventions for adults with chronic pain are effective, many unanswered questions remain. In particular, we have a poor understanding of how some aspects of treatment, such as treatment ‘dose’, relate to outcome. We also know little about the contribution of specific components included in broad treatment packages, such as CBT. Objectives: I replicated Williams, Eccleston and Morley’s (2012) Cochrane review of psychological interventions for adults with non-cancer chronic pain (excluding headache) but extended this in two ways: I investigated the relationship between dose and outcome, and developed a system of categorising treatment content across trials. Results: 64 randomised controlled trials (N = 7,840 participants) were included. Effect sizes(standardised mean differences) were calculated and used in meta-analysis to explore five outcome domains: pain experience, pain behaviour, emotional functioning, and coping and cognitive appraisal. My findings converge with previous reviews; psychological interventions were effective, but effect sizes were relatively small. Meta-regression analyses found limited evidence for the moderating role of treatment ‘dose’ on outcome. There was some evidence that the period over which treatment was delivered(i.e. ‘dose in weeks’), may be related to outcome for two out of five outcome domains. In terms of treatment content, trials that appeared to utilise the same broad treatment package (e.g. CBT), often appeared to feature different treatment components. Moreover, treatments described as being distinct (e.g. ACT and CBT) often used similar treatment components. Conclusions: The relationship between treatment dose and outcome is complex and dependent on the outcome being assessed. Broad labels of treatment types, such as CBT, are vague and do not represent homogenous groups. This suggests that all treatment components are not created equal. I propose single-case design and patient-level data analysis as tools to help further explore treatment components, including treatment dose, and outcome.

Research suggests that self-harming behaviour has increased and rates of self-harm are consistently higher in adolescence compared to adulthood. Understanding why adolescents engage in self-harming behaviour is important. Adolescents who self-harm are at higher risk of a repeated episode and self-harm is a key risk factor in completed suicide. Only a small number of studies have directly explored adolescents’ views of their self-harm using first-hand accounts. Data was gathered via a secondary analysis of video-recorded Family Therapy sessions collected as part of the Self-harm intervention: Family Therapy. Session recordings of 22 participants, approximately 170 hours of footage, formed the dataset. Data was only transcribed data for later analysis if the adolescent was directly involved in the conversation. Using thematic analysis to analyse the data, five core themes were developed; (1) Distress can be difficult to convey (2) Self-harm and suicidal ideation; a complex relationship (3) Self-harm as a form of communication (4) Self-harm to manage emotions and (5) Moving forward. Prominent social discourses around self-harm resulted in some adolescents attempting to manage alone and fearing the consequences if they talked about self-harm. Accounts highlighted the complex interplay between self-harm and suicidal intent; for some this fluctuated across episodes of self-harm. Self-harm was a means of communicating distress as well as managing emotions. Encouragingly, many participants described being able to resist self-harm, often mirroring why some adolescents harmed themselves in the first place. Findings from the analysis are discussed in relation to the literature along with strengths, limitations, clinical implications and future research.

Introduction: Serious childhood illness has a significant impact on the child and their family, but it is known that strong supportive relationships can act as a buffer. Due to limited NHS resources, volunteers are becoming integral to the support of such families. Despite this, little is known about the experiences of volunteers in such roles, and more specifically, those choosing to volunteer in emotionally challenging areas such as childhood cancer. Method: A sample of seven volunteers working for Candlelighters childhood cancer charity were interviewed using semi structured interviews; transcripts were analysed using interpretative phenomenological analysis. Results: Volunteer’s experiences of working in a childhood cancer charity were described in terms of five superordinate themes: ‘Motivation’, ‘Identity and Roles’, ‘Coping’, ‘Community’, and ‘Family’. Overall, the themes captured what motivates individuals to volunteer and in particular why they chose the area of childhood cancer. The themes also reflected the impact that the role has on their sense of identity and self, and how the participants seemed to cope with that role. It is speculated that this is in part due to the protective experience of belonging to a community as described by the participants, with frequent allusions to an even deeper link in terms of feeling part of a family. Discussion: The findings of this study are related to the wider literature with consideration of role identity theory and the significance of community and relationships. The strengths and limitations are discussed to assess the quality of the study. Implications for our current understanding of volunteers and suggestions for future research are proposed.

There is evidence that life events experienced during the perinatal period, such as relationship breakdown, increase the risk of depression. In addition, clinical comorbidity with anxiety and somatic symptoms is common during the perinatal period and is associated with more negative outcomes for the mother and child. However, few studies have investigated the association between life events and depression symptoms whilst controlling for comorbidity with anxiety and somatic symptoms. This study aimed to explore the associations between the number and type of life events and depression symptoms across the perinatal period, whilst controlling for age, deprivation levels, comorbid anxiety and somatic symptoms. Cross-sectional associations were explored at three time-points during the perinatal period (26 weeks pregnancy, 8 weeks postnatal and 1 year postnatal) and prospective associations were explored between life events during pregnancy and postnatal depression symptoms. Data from 917 women who took part in the Born and Bred in Yorkshire population cohort study between 2011-2015 were analysed. Depression, anxiety and somatic symptoms were assessed by the PHQ-8, GAD-7 and PHQ-15 respectively. Life events were assessed using the LTEQ which is a 12-item checklist measure. Results showed that specific types of life events were stronger predictors of perinatal depression symptoms than the number of life events alone. In particular, life events such as relationship breakdown and serious financial problems in the preceding six months were associated with concurrent perinatal depression symptoms in the cross-sectional analyses. The prospective analyses showed that serious problems in a woman’s social network (close friends and relatives) that occurred during pregnancy and soon after childbirth increased the likelihood of experiencing postnatal depression by approximately three times (OR 2.8-3.7). Moreover, problems in a woman’s social network during pregnancy were also shown to increase the risk of postnatal depression independently to depression, anxiety and somatic symptoms during pregnancy. Significant interpersonal and financial life events increase the risk of depression in women during the perinatal period. The results of this study are relevant to public health policymakers and suggestions for clinical implications and further research are discussed.