Executive Function in the OCD Spectrum: Phenomenology, Neurocognition, and the Development of a New Assessment Tool

This thesis investigated executive impairments in the OCD spectrum. Chapter 1 offers an overview of the general background of this work, focusing on OCD symptomatology, neurocognitive deficits in OCD, and the various forms of measuring both OCD and neurocognitive deficits in OCD. This narrative review highlights key domains of executive dysfunction in OCD that require further empirical investigation to bridge existing knowledge gaps. The first empirical chapter (Chapter 2) presents a systematic review and meta-analysis of three decades of research (112 studies), examining neurocognitive impairments of cognitive inflexibility, response inhibition, and decisional impulsivity (representing latent compulsivity and impulsivity, respectively) in patients diagnosed with OCD. This comprehensive analysis examines a total of 8,313 participants, comparing 4,289 patients with OCD and 4,024 healthy controls on various neurocognitive tasks. To maintain consistency with previously discussed meta-analytical paradigms, this meta-analysis focused on the WCST, IDED, SST, and IGT, given their relevance in capturing latent neurocognitive phenomena. The review found minimal publication bias and generally adequate methodological quality, although some limitations were noted, concerning sample size justification, inconsistent use of diagnostic screening tools, and discussion of limitations. This chapter revealed moderate effect sizes of deficient neurocognitive performance in patients with OCD that were comparable across neurocognitive compulsivity and impulsivity. These effect sizes were orthogonal to various clinical variables such as: depression and anxiety symptoms, effects of medication, comorbid conditions, and duration of illness. Importantly, findings revealed distinct relationships between compulsivity and impulsivity and symptoms of OCD. The severity of OCD symptoms did not moderate effect sizes of either neurocognitive compulsivity or impulsivity for between-group contrasts, however, within-group correlations revealed a significant relationship between OCD symptoms and compulsivity but not impulsivity. This work revealed implications for these impairments in OCD that requires substantiation, specifically, the manifestation of these latent cognitive phenotypes and how they are assessed. The second of the empirical chapters (Chapter 3) explores obsessive-compulsive phenomenology during the easing of lockdown restrictions within the general population of the UK. The pandemic offered an opportunity to look at salient pre-clinical obsessive and compulsive tendencies that may have developed during the lockdown (such as avoidance and excessive cleanliness) and the potential persistence of such tendencies once these restrictions had lifted. This work was a replication and extension of Fineberg et al (2021) and so the methodology largely reflected that of the original paradigm. Indeed, the mediation model was replicated; while the original work demonstrated that OCD symptoms predicted lockdown adjustment difficulties only indirectly via affective symptoms, the current study found both direct and indirect effects of OCD symptoms on post-pandemic adjustment difficulty. This suggested that ongoing nature of the pandemic may have strengthened the impact of obsessive-compulsive symptoms on adjustment over time. Findings also revealed contrasting accounts of compulsivity between assessment modalities. Longitudinal analyses revealed strong correlations between self-rated compulsivity (on the Compulsive Personality Assessment Scale (CPAS; Fineberg et al, 2007) and symptom severity across the three assessment points despite symptom severity declining and compulsivity remaining stable. As such, compulsivity may be a trait vulnerability for the development of OCD, though as this was not reflected on neurocognitive tasks (the Wisconsin Card Sorting Test: Milner et al, 1963) and the Intra-Extradimensional Set-Shifting Task: Owen et al, 1991) further work is required to unpick the manifestation of compulsivity at different junctures of OCD symptom severity. Chronologically speaking, the COVID-19 longitudinal study was conducted first and influenced the methodological choices of the meta-analysis presented in Chapter 2, particularly the decision to expand the assessment of executive functioning in OCD to include measures of impulsivity alongside compulsivity. It was logical to place the meta-analysis first in this thesis, as it provides a comprehensive overview of neurocognitive impairments in clinical patients diagnosed with OCD, forming the foundation of this work. The longitudinal analysis is then presented to build upon these findings by addressing the clinically relevant questions in a subclinical sample that were raised within the meta-analysis, further advancing the investigation into phenotypic manifestations of OCD. A synthesis of the findings of the preceding two chapters revealed three core themes to this thesis: (1) the exploration of latent cognitive phenotypes of compulsivity and impulsivity, (2) the two leading methods of assessing these phenotypes, self-reports and neurocognitive tasks, and (3) the distinct relationships that compulsivity and impulsivity each have with OCD symptom severity. This apparent orthogonality between symptom severity and executive impairments raises questions about the adequacy of the current diagnostic scope for OCD. As such, Chapter 4 reviews current screening tools for OCD, highlighting their respective strengths and limitations. Classical and Modern theoretical frameworks are also explored for their unique provisions for scale development. The review highlighted that most symptom scales have only undergone limited psychometric validation and are very restricted in their respective acknowledgement of the more nuanced neurocognitive aspects of the disorder. The current shortcomings of OCD assessment, as revealed in the review, highlighted the need for a psychometrically sound measure that integrate the strengths of both classical and modern test frameworks to assess the subtle neurocognitive dimensions of OCD for early identification and novel target interventions. Study 4.1 outlines the initial validation of the Obsessive-Compulsive Phenomenological Assessment Scale (OCPAS). As patients with OCD are one singular point on a spectrum, developing a scale that is too close to the diagnosable illness may have eliminated salient pre-clinical features that may elucidate predisposing and precipitant factors at particular junctures of severity. As such, the OCPAS was initially validated in a non-clinical sample so as to retain features relevant to the development of OCD that might be overlooked at the severe end of the spectrum. An initial corpus of 59 items measuring phenotypic variance in OCD reduced to 45 items under classical test theory (exploratory factor analysis); and further reduced to the final 26-item scale under modern test theory (Rasch analysis) - to provide a novel psychometrically robust questionnaire for the self-assessment of compulsive and impulsive features in OCD phenomenology. The OCPAS consist of three factors. Cognitive Control, Behavioural Control, and Anankastic Goals. Cognitive Control measures attentional inflexibility and deficient executive functioning. The Behavioural Control factor describes a lack of control over behavioural and emotional output and subsequent ceasing of the behaviour once started, Finally, Anankastic Goals pertain to rigid beliefs, a need for control, and a motivational drive to achieve perfectionist standards. Study 4.2 assesses the construct and convergent validity of the OCPAS in a sample with diagnosable OCD. Scores on the OCPAS are contrasted with self-ratings of OCD symptom severity, compulsivity, and impulsivity (using the OCI-R, CPAS, and BIS-11, respectively) as well as with performance on neurocognitive tasks assessing cognitive flexibility and response inhibition (the IDED and the SST, respectively). These self-rating measures and neurocognitive tasks were selected for consistency with preceding chapters and prior research that has informed this thesis. This chapter shows that the OCPAS effectively distinguishes between those that do and do not have OCD, and correlates strongly with neurocognitive measures of compulsivity and impulsivity. Interestingly, a relationship was revealed between the OCPAS and the Stop Signal Task (Aron, Robbins, and Poldrack, 2004) suggesting that the OCPAS is sensitive to the nuanced neurocognitive features of the disorder. Again, self-rated compulsivity correlated with symptom severity, which was not reflected on the IDED. However, both self-ratings on the Barratt Impulsiveness Scale-11 (Patton, Stanford, and Barratt, 1995) and task performances (SST) of response inhibition correlated with symptom severity and with each other, which may seem contrary to the orthogonality between assessment modalities revealed in Chapters 2 and 3 in addition to the wider literature (e.g., Evenden, 1999; Frydman et al, 2020), challenging the previously conceived independence of impulsivity from symptom severity. These findings highlight the OCPAS as a robust scale for measuring neurocognitive dimensions of OCD, and offers insights into the relationship between compulsivity, impulsivity, and OCD symptomatology. Chapter 5 consolidates the findings from the previous empirical chapters by synthesising phenotypic variance across distinct levels of symptom severity, ranging from non- and sub-clinical presentations (in Study 4.1 and Chapter 2) to OCD-likely and formally diagnosed groups (in Study 4.2 and Chapter 1). This synthesis showed that compulsivity shares a similar trajectory to symptom severity across the chapters, given their moderate to strong correlations in Chapters 2, 3, and 4. In contrasting the inconsistencies in the relationship between impulsivity and symptom severity in Chapters 2 and 4, it appears that impulsivity at first correlates and contributes the precipitation of OCD and then remains stable for the duration of the illness given the orthogonality recorded in Chapter 2. An exploratory mediation model is presented, offering a potential mechanism by which these phenotypes influence symptom severity in OCD-likely groups. The model suggests that both compulsivity and impulsivity contribute to the transition from symptoms to syndrome to illness which may be the difference between managing and not-managing worsening symptomatology. In conclusion, this thesis reveals the respective manifestations of compulsivity and impulsivity in OCD phenomenology. Both phenotypes display unique associations with symptom severity and contribute to the precipitation of OCD proving to be potential trait markers. However, the current diagnostic assessment of OCD fails to capture such phenotypic variance and so, the OCPAS was developed to provide a psychometrically robust self-report measure of deficient cognitive and behavioural control. It is suggested that future research focuses effort on naturalistic and longitudinal analyses of these phenotypic manifestations to substantiate benchmarks for early identification of at-risk cohorts.