Characterising mood instability: clinical, behavioural, and neural implications for mood disorders

Mood instability is a core characteristic of, and risk factor for, mood disorders including bipolar disorder (BD) and major depressive disorder (MDD). However, little is currently understood of the wider, clinical, behavioural, and cognitive correlates of mood instability, and the neural mechanisms that may underlie it. It is hoped that a more thorough understanding of this pathophysiological basis may aid in better distinguishing between disorders, in predicting the onset and severity of symptoms, and in developing better suited targets for therapeutic intervention. Therefore, the current thesis aims to explore the profile of mood instability using a translational neuroscience approach combining remote technology, wearable devices, and neuroimaging techniques, in order to investigate the temporal dynamics of this relationship. Throughout this thesis data is presented from the Cognition and Mood Evolution across Time (COMET) study. The COMET study is a prospective, between-subjects experimental cohort study conducted over 10 weeks. Seventy-four participants took part, 37 of whom screened in for risk of BD using the Mood Disorder Questionnaire (MDQ) (MDQ7; high MDQ group), and 37 who did not (MDQ5; low MDQ group). All participants underwent a battery of multi-modal investigations, including remote monitoring of mood and activity, as well as cognitive testing and functional magnetic resonance imaging (fMRI). Clinical assessments revealed that 9 out of 37 high MDQ participants met criteria for one or more DSM-IV axis I psychiatric disorder (chapter 2). Given the emergence of new technologies and the ubiquity of internet access, I first explored whether prospective monitoring of daily mood could capture longitudinal mood instability in the high MDQ group (chapter 3). High MDQ individuals showed greater variability in daily mood ratings compared to low MDQ individuals, and reported significantly greater absolute negative affect across time. They also had significantly greater clinical symptoms of mania, depression, and anxiety, and reported greater variability across these measures. Additionally, increased levels of negative affect were reported in high MDQ participants who met criteria for an axis I disorder, regardless of classification, compared to those without a diagnosis. Therefore, analyses showed that daily mood monitoring could capture dynamic mood instability in individuals who screen in for BD. Despite screening in on a tool that targets manic and hypomanic symptoms, mood instability seemed to be more strongly associated with experiences of negative affect. While confirming the presence of mood instability in the high MDQ group, these results also paved the way for a closer examination of the related behavioural dynamics, a common symptom of mood disorders often manifesting in the form of circadian rest-activity disruptions. The subsequent study aimed to explore whether longitudinal actigraphy monitoring could distinguish between high MDQ participants and low MDQ participants, and also examined the influence of concurrent mood instability (chapter 4). High MDQ individuals showed significantly lower circadian relative amplitude, a marker of differentiation between daytime and night-time activity, driven by reduced levels of daytime activity. Relative amplitude was also differentially associated with mood instability and absolute negative mood in high MDQ participants and low MDQ participants, who also showed significant association with both night-time and daytime activity and mood instability. Together, these findings suggest that disruptions in circadian rhythms in mood disorders persist outside of the effects of clinical diagnosis and functional impairment, and are significantly impacted by the experience of mood instability. Subsequently, the second part of my thesis aimed to determine whether an overarching disruption in neural homeostatic mechanisms may be responsible for both regulating mood and circadian rhythms. Thus, resting-state fMRI was carried out at two time points during the COMET study, and analyses explored both static and dynamic patterns of functional connectivity, so as to more accurately capture the temporally dynamic experience of mood instability. Firstly (chapter 5), whole-brain analyses using the ICA method did not demonstrate network variability between groups and across time. Similarly, and using a seed-based approach, there was no difference in specific connectivity with the emotion processing and regulation regions of the amygdala and insula between groups or across time. This paved the way for more nuanced analyses of the dynamic patterns of functional connectivity that may be present. Therefore, subsequent analyses (chapter 6) aimed to build on this by exploring whether mood instability is reflected in altered patterns of dynamic neural connectivity, representing an overarching disruption in neural homeostatic mechanisms. A sliding-window approach was employed in order to characterise dynamic functional connectivity. Results revealed a comparative level of activation within four distinct meta-states (namely, highly interconnected, DMN, sensorimotor, and low connectivity) across all participants and scan sessions. When assessing dynamic properties, high MDQ individuals were found to occupy a greater number of meta-states and change more frequently across states than those in the low MDQ group. This pattern of instability was found to persist across sessions in the high MDQ group, whilst the low MDQ group showed a pattern of habituation or regulation across these metrics over time. Analyses of the relationship between dynamic properties with mood measures revealed a pattern of lowered variability in functional connectivity as a function of mood instability, as well as a potentially protective association between dynamic properties and circadian rest-activity patterns. Together, these findings uncover more dynamically nuanced details of the pathophysiology of mood and related instabilities. Led by the need to better understand the wide-ranging correlates of fluctuating mood and to improve clinical outcomes for patients with mood disorders, the aim of this thesis was to build a profile of the translational neuroscientific nature of mood instability. As the first study to do this, I was able to uncover the beginnings of an overarching disruption in neural homeostatic mechanisms that are responsible for regulating mood and related clinical and behavioural outcomes. It is hoped that the doctoral research presented here will encourage continued investigation of this transdiagnostic marker employing further novel and sophisticated methods, in order to capture the temporally dynamic nature of mood instability. Ultimately, such research will prove valuable in transforming our understanding of mood disorders and in the search for better-targeted therapeutic interventions.