Theses - MRC Cognition and Brain Sciences Unit

Permanent URI for this collection

https://www.repository.cam.ac.uk/handle/1810/263982

Browse

Now showing 1 - 20 of 65

Open Access
The Relationship Between Phonological Short-Term Memory (pSTM) and Categorical Perception of Speech
Smith, Harriet Jordan
In real-world listening situations, the speech signal is often perceptually ambiguous. To achieve successful speech identification, listeners must resolve this ambiguity by organising the continuously varying phonetic features into discrete phoneme categories (categorical perception). The cognitive demand associated with categorisation depends on the degree of ambiguity in the signal, and on the perceptual acuity of the individual listener. Prior research suggests that phonological short-term memory (pSTM) capacity is, in part, determined by the quality of the speech input. For example, recall is poorer for speech-in-noise, and individuals with poor pSTM may also show weak categorical perception. However, the relationship between categorical perception of phonemes and pSTM capacity remains under-studied. This thesis assesses whether the difficulty associated with phoneme categorisation has an impact on serial recall performance, using a variety of methods that advance the current evidence base. Firstly, speech-sound discrimination was experimentally disrupted in a pSTM task, by asking typical adult listeners to recall audio-morphed syllables with varying perceptual ambiguity (Chapter Two). Recall for high-ambiguity syllables from letter-letter (“B”-“P”) continua was poorer than for low-ambiguity equivalents under specific circumstances, but when top-down control could be engaged to disambiguate the sounds, the ambiguity effect was abolished. Secondly, sensitivity to phoneme contrasts was compared between adolescents (10-15 years) with phonological deficits and controls, using a measure of categorical perception with minimal extraneous cognitive demands (Chapter Three). No between-groups differences were identified on either speech or non-speech auditory perception. In an independent experiment (Chapter Four), individual differences in speech perception acuity could not predict pSTM capacity in typical adults (N = 75). Finally, a functional MRI study provided preliminary evidence that perceptual demands and pSTM demands do not load on a shared brain network (Chapter Five). Thus, by collating evidence from children and adults, and using behavioural and neuroimaging approaches, I conclude that categorical perception and pSTM are dissociable in individuals, and should be considered relatively independent functions.
Open Access
A New Eye On Semantics: Probing the Spatiotemporal Brain Dynamics Underlying Natural Reading
Magnabosco, Federica
In this thesis, I examined the influence of task and context on the semantic processing of single words in isolation and in a sentence context. In one study, I considered task effects on single-word semantic representations using multivariate pattern analysis. In two further studies, I tested whether sentence context affects the semantic processing of single words during natural reading. Overall, these studies aimed at testing predictions made by the Controlled Semantic Cognition (CSC) framework during natural reading, in order to better characterise the dynamic relationship between semantic control and representation mechanisms. To achieve this, I considered dynamic brain activity with good spatio-temporal resolution as recorded by EEG/MEG as well as behaviour as recorded by eye-tracking. In my multivariate pattern analysis in EEG/MEG source space data, I found that while semantic representations show some stability across tasks, even subtle task differences are associated with widespread differences. I found that stimulus and task effects are better assessed by different approaches: stimulus information is best investigated by considering the decoding performance in each individual ROI (by-ROI approach) while distributed, task effects are better understood by including all ROIs and assessing the back-projected model’s weights (across-ROI approach). In my behavioural study, I found additive effects of predictability and concreteness on early reading time measures, paralleling previous results on the relationship between predictability and frequency. An analysis on a larger and freely available corpus produced compatible results. For the first time, I performed a co-registration study combining EEG+MEG and eyetracking using a deconvolution approach to separate brain activity generated by overlapping fixations. Firstly, I tested some of the proposed procedures for ensuring a good data quality, i.e., removing artefacts generated by eye movements, and proposed some novel methodological improvements, such as the combination of the variance ratio and EOG correlation to detect artefactual components in independent component analysis. Analyses of the neural activity revealed an effect of predictability in sensor space and decoding analysis, which indicated that the left temporal lobe is crucial for integrating words into higher, sentence-level meaning. I found no effect of concreteness in the analyses conducted. Moreover, the behavioural results obtained from this co-registration analysis did not replicate the results obtained in the previous behavioural study. In summary, this thesis reveals novel insights into the dynamic interaction of control and representation mechanisms, by considering the effects of task and context. The methodological advancements contribute to the growing field of EEG/MEG and eyetracking co-registration.
Open Access
Neural Dynamics of Goal-Directed Action Selection: A Multivariate Approach to Evidence Accumulation and Cognitive Control in Response to Competing Stimuli
Minors, Dorian
This thesis sets out to explore how two preconditions for goal-directed behaviour might arise from neural activity—the capacity to integrate information about a stimulus to inform action-selection, and the capacity to resolve conflict in the presence of competing alternatives. I examine this over the course of three empirical studies incorporating behavioural measures, computational modelling, magneto- and electroencephalography (M/EEG), functional Magnetic Resonance Imaging (fMRI), as well as both univariate and multivariate analysis techniques. In the Introduction, I introduce the background to and motivation for the thesis, exploring the evidence accumulation framework as a tool to understand the latent cognitive processes involved in adaptive action-selection, and exploring the phenomenon of conflict resolution through the lens of the Stroop task. In Chapter 2, I asked how different kinds of task difficulty influence the neural process of accumulating evidence in service of an action. I used an Evidence Accumulation Model (EAM) and combined magneto- and electroencephalography (M/EEG) to explore the impact of manipulating perceptual difficulty and the difficulty of categorising the percept according to a rule. I used multivariate pattern analysis (MVPA) to quantify how information, latent in the neural activity, about the percept, rule, the application of the percept to the rule (decision), and the response, varied across the manipulations. Manipulating perceptual difficulty influenced the model parameter associated with, and hallmark neural signals of, evidence accumulation as well as information relating to the percept, decision, and response. In contrast, manipulating categorisation difficulty influenced the model parameter associated with the amount of evidence required to make a decision, had a subtler effect on ramping neural activity, and only influenced information relating to the decision and response. Together the results gesture at multiple underlying processes within the evidence accumulation process that are affected differently by a manipulation of perceptual evidence relative to manipulating the difficulty of evaluating the percept according to a rule. In Chapters 3 and 4, I asked how conflict in the arc from perception to action is resolved: though the inhibition of task-irrelevant information, the enhancement of relevant information, or a combination of both. In Chapter 3, I developed variations of the classic Stroop task suitable for a subsequent fMRI investigation to detect neural processes of inhibition and enhancement in the Stroop task. This comprised a novel baseline condition in which participants named the sizes of Stroop (word-colour) stimuli, and a baseline condition in which participants named the colours of non-words matched to Stroop stimuli for low-level visual features. In both cases, less interference should be produced by the irrelevant stimulus features relative to the Stroop, modulating the need for neural inhibition and/or enhancement relative to the more interfering Stroop task. I validated the baselines behaviourally in both manual and vocal versions of the tasks, demonstrating that interference was reduced relative to the Stroop task, and this effect was greater in the vocal than the manual versions of the tasks. In Chapter 4, I used the baseline tasks developed in Chapter 3 in an fMRI study exploring inhibition and enhancement in the Stroop task. I used MVPA to quantify the latent word-related information in word-specialised cortex and colour-related information in colour-specialised cortex, as well as the word- and colour- information available in frontoparietal “multiple demand” (MD) regions. I found initial evidence to support the notion that the brain inhibits task-irrelevant word information in word-sensitive cortex during the task, but no evidence to suggest that the brain enhances task-relevant colour information in colour-sensitive cortex. I replicated findings that frontoparietal MD regions preferentially maintain task-relevant information, but found no evidence to discriminate whether this was achieved by enhancement or inhibition of information in MD. The findings suggest that the brain prioritises task-relevant information in MD region, and this may be associated with inhibition of task-irrelevant information in specialised cortex when there is conflict in the arc from perception to action. The thesis concludes with a discussion of the limitations and future directions for research in the field of cognitive control, decision-making, and attentional processes, alongside the potential of using MVPA and EAMs to further our understanding of these complex neural processes.
Open Access
Social-semantic knowledge and behavioural changes in frontotemporal dementia
Rouse, Matthew; Rouse, Matthew [0000-0002-3157-4301]
Key insights into the organisation of semantic memory in the brain have been derived from semantic dementia, a subtype of frontotemporal dementia (FTD) associated with a degradation of conceptual knowledge following atrophy in the anterior temporal lobes (ATLs). Recent research in FTD has begun to focus on social-semantic knowledge, and its relationship to the known behavioural changes in this disorder. The aims of the thesis were to (i) determine how social-semantic knowledge is affected in FTD in relation to general conceptual knowledge, (ii) explore the roles of the left and right ATLs in social-semantic knowledge, and (iii) investigate the relation between social-semantic knowledge and behavioural changes in FTD. An additional theme that ran throughout the thesis was the use of a transdiagnostic approach to FTD. First, I introduced a model of impaired social behaviour in FTD – controlled social-semantic cognition. In a formal review chapter, I proposed that the impaired social behaviour in semantic dementia results predominantly from a degradation of social-semantic knowledge, and that social-semantic knowledge is part of a broader conceptual system supported by a transmodal hub in the bilateral ATLs. In Chapter 3, I assembled a neuropsychological battery comprising social- and non-social-semantic tasks. Both batteries were run in a large cohort of people with FTD, as well as people with unilateral ATL resection for temporal lobe epilepsy (TLE). I found that social-semantic knowledge was degraded in parallel with general conceptual knowledge in FTD following bilateral ATL atrophy. In contrast unilateral ATL damage yielded a mild impairment for social and non-social concepts, with no left vs right differences. Chapter 4 focussed on a specific type of social concept – face/person knowledge. I compared the impact of bilateral ATL damage vs unilateral ATL damage on face recognition, person knowledge and general semantic memory. Bilateral ATL damage generated a severe impairment in both person knowledge and general semantic memory, whereas only a mild impairment was found after unilateral ATL damage. Strikingly this bilateral vs unilateral difference for semantic performance occurred despite similar levels of overall ATL volume. In Chapter 5, I used a battery of informant questionnaires to explore the range and nature of behavioural changes in FTD, and tested the hypothesis that social-semantic knowledge is related to impaired social behaviour in FTD. By recruiting the same FTD cohort as Chapter 3, I was able to explore the relation between behavioural changes and (i) neuropsychological performance, and (ii) grey matter volume. Behavioural changes were prevalent in both FTD subtypes and differed quantitatively (more frequent/severe in bvFTD) but not qualitatively. There was no association found between impaired social-semantic knowledge and behavioural changes, which may reflect informant questionnaires being more sensitive to prefrontal cortical-based behavioural changes rather than semantically driven behaviours. In conclusion, my findings are in line with a theoretical framework whereby the bilateral ATLs underpin a transmodal representational hub critical for all aspects of conceptual knowledge, including social-semantic knowledge. My findings have important implications for FTD and for theories regarding the roles of the ATL in semantic memory.
Open Access
Notice, Observe, Step Back and Experience: The assessment and potentiation of psychological decentering for adolescent mental health
Knight, Rachel; Knight, Rachel [0000-0002-8315-2864]
Half of all mental health disorders begin by the age of 14, with significant effects both for the individual and for wider society. As such, adolescence is a critical period for the prevention of and intervention in mental health difficulties. In hopes of developing new treatments and improving pre-existing therapies, there has been much investigation of potential “active ingredients” by which therapeutic change may be driven. One of these active ingredients may be psychological decentering, a self-reflective capacity featuring across a spectrum of psychological therapies. Psychological decentering is a skill through which distressing internal events, including difficult thoughts, feelings, and memories, are re-perceived from an observer perspective. This typically results in a reduction in identification with these experiences, and attenuates emotional reactivity. This skill has been hypothesised to be protective against mental health difficulties. The work presented in this thesis attempted to answer four core research questions: (1) what are the characteristics of decentering in an adolescent sample? (Chapters 2, 3 & 4) (2) How is decentering connected to adolescent mental health? (Chapters 2, 3 & 4) (3) What cognitive processes are associated with decentering? (Chapters 3 & 4) And (4) can decentering be trained and does this impact mental health in adolescence? (Chapters 5 & 6). This was achieved using a narrative review, regression models calculated using data from two cross-sectional samples (N = 553 and N = 904), and a preliminary randomised controlled trial comparing low-intensity remote decentering skills training against an active control condition (N = 115). There was conflicting evidence for the association between age and decentering across adolescence, with one sample suggesting a positive association, and one a negative. There was consistent strong evidence for relationships between decentering and mental health symptoms, in which higher self-reported decentering was associated with fewer symptoms, and better wellbeing. There was also evidence of association between decentering and emotion regulation strategies, with higher self-reported decentering predicting broader use of emotion regulation strategies, and fewer difficulties with emotion regulation. In terms of cognitive processes and their association with decentering, there was consistent evidence of a lack of association between decentering and cognitive control on a range of tasks. However, there was evidence of association between decentering and a self-referential processing task, in which participants with higher self-reported decentering committed fewer errors on a condition in which they were required to manipulate self-perspective. Finally, a randomised controlled trial conducted with adolescents with elevated depression symptoms showed good preliminary evidence of improving individuals self-reported decentering, reducing depression symptoms, and increasing psychological wellbeing. In conclusion, the work presented in this thesis provides further evidence of decentering as a predictor of broader mental health and emotion regulation in adolescence. It is currently unclear how this skill may develop across this period, with further longitudinal studies needed. Most importantly, a fully-powered randomized-controlled trial shows promising evidence that, compared with an active control, decentering can be trained, using a remote, low-intensity intervention. This improvement in decentering coincides with a significant reduction in depression symptoms. However, a larger effectiveness trial is needed to confirm the benefits of this intervention, and to investigate possible moderators and mediators of the observed effects. This thesis offers promising possibilities for advancing interventions for young people, and uncovering how these treatments function.
Open Access
The Effect of Mindfulness-Based Programmes on Work Performance
Vainre, Maris; Vainre, Maris [0000-0001-9570-3726]
Mindfulness-based programmes (MBPs), suggested to improve work and academic performance, are increasingly used in occupational and educational settings. This thesis advances the field by synthesising the evidence, optimising the operationalisation of work performance, providing preliminary data on MBPs’ effectiveness and mechanisms for work performance, and testing acceptability and feasibility for future trials. First, I led a systematic review and meta-analysis which aimed to map how work performance has been operationalised in randomised controlled trials (RCTs) of MBPs and to assess the impact of MBPs on adults’ academic and work performance. The pre-registered primary outcome was task performance, a key aspect of work performance, up to 4 weeks after the intervention (PROSPERO: CRD42020191756). Secondary outcomes were the remaining aspects: contextual performance, adaptive performance, and counterproductive work behaviours. Pairwise random-effects meta-analyses were used to calculate Hedges’ g. A total of 47 RCTs with 5041 participants were included. Adaptive performance outcomes were collected most frequently. There was no support for MBPs significantly improving task performance (7 RCTs, 454 participants, Hedges’ g = 0.52, 95% CI -0.03 to 1.07, p = 0.06) compared with passive control groups. However, MBPs statistically significantly improved adaptive performance and contextual performance. There were an insufficient number of RCTs to allow meta-analysis comparing MBP to active control interventions. All bar one RCT demonstrated high risk of bias. Confidence in the review results was ‘low’ to ‘very low’, according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria. Second, I led a feasibility RCT which evaluated whether improved cognitive control and/or enhanced mental health may be mechanisms underpinning the effect of MBPs on work performance. Two hundred and forty-one employees were recruited from eight employers. The participants were randomised on 1:1 basis to the offer of a four-week, self-guided, digitally delivered intervention of either the Be Mindful MBP or a light physical exercise programme (control intervention). We assessed the acceptability of interventions and trial procedures, and estimated effect sizes to inform sample size calculations for a later-stage trial. The primary effectiveness outcome was self-reported work performance at post-intervention measured using the Work Role Functioning Questionnaire. Secondary outcomes included depression, anxiety, stress, and cognitive processes hypothesised to be targeted by mindfulness, including decentering and executive functions. All outcomes were assessed at pre-intervention, post-intervention, and 12-week follow-up. The trial protocol was pre-registered (NCT04631302) and published. We concluded that a full-scale trial would be feasible and acceptable, based on recruitment and retention rates. Yet, a full-scale trial may not be warranted: the MBP, compared to light physical exercise, offered negligible benefits for work performance at post-intervention (Cohen’s d = 0.06) and 12 weeks later (Cohen’s d = 0.02). For the potential mechanisms, we observed similarly small effect sizes for the differences between the MBP and the alternative intervention on mental health and cognitive control outcomes. Both interventions improved mental health outcomes compared to baseline. In conclusion, while MBPs may have some potential in enhancing some aspects of work performance when compared to passive control groups, the results of this thesis suggest the evidence is of low quality. Furthermore, the effectiveness of MBPs compared to alternative interventions, like physical exercise, remains uncertain. We found no evidence to suggest cognitive control could be a mechanism underlying MBPs effects on work performance, when compared to light physical exercise. These findings underscore the importance carefully operationalising work performance and conducting high-quality trials to establish the impact of MBPs on work performance in occupational settings.
Open Access
Reality Resists Classification: A Transdiagnostic, Network-Based Approach to Behavioural and Neural Variation in Childhood
Zdorovtsova, Natalia
Childhood development is shaped, and characterised, by a variety of interacting processes that encompass biological and environmental phenomena. Key developmental outcomes, such as cognitive ability and behaviour, are closely associated with the structure and function of the brain. These features of neurobiology are shaped concurrently by genetic factors, individuals’ interactions with the environment, and stochastic effects that instantiate divergent trajectories of development over time. Some behavioural and cognitive profiles are recognised as being particularly divergent from the population-norm, such that they are included in clinical taxonomies of neurodevelopmental conditions. Research in developmental cognitive neuroscience, until relatively recently, has assumed that different neurodevelopmental conditions—such as ADHD, autism, and dyslexia—constitute fundamentally separate developmental trajectories, each with their own genetic, neurological, cognitive, and behavioural distinctions. However, there is considerable heterogeneity within, and overlap between, neurodevelopmental conditions, suggesting that diagnostic categories are not robust predictors of behavioural and cognitive differences between individuals. A growing number of researchers are therefore choosing to study neurodevelopmental heterogeneity in a manner that is agnostic to the presence, or absence, of formal diagnoses. This thesis builds on the current literature in developmental cognitive neuroscience by taking a *transdiagnostic* approach to studying the associations between neurology, cognition, and behaviour. We endeavoured to address these three questions: 1. Do the topological features of structural brain network connectivity differentiate children with elevated inattention and hyperactivity? 2. Do patterns of functional brain network connectivity differentiate the brains of children with elevated inattention and hyperactivity? 3. How are resting-state neural dynamics related to individual differences in behaviour and cognition? We addressed the first two questions by analysing the relationships between structural MRI, functional MRI, cognitive, and behavioural data from the Centre for Attention, Learning, and Memory, which included participants aged 6-17 (CALM; n=383). To address our third question, we analysed the spontaneous, transient dynamics of resting-state MEG data from a sample of children aged 8-13 (n = 46). Additionally, we worked with a multidisciplinary team of academic researchers, charity leaders, educators, policymakers, and neurodivergent community partners to develop a set of freely-available online resources that help schools create inclusive educational frameworks. Exploratory factor analysis indicated that inattention and hyperactivity are best represented as one latent factor in the CALM sample. No single component of structural brain organisation predicted linear differences in inattention and hyperactivity in our sample. However, a further analysis that combined multidimensional scaling with k-means clustering revealed two structural neural subtypes in children with elevated levels of inattention and hyperactivity (n = 232), differentiated primarily by communicability—a measure which demarcates the extent to which neural signals propagate through specific brain regions. These different clusters had similar behavioural profiles, which included high levels of inattention and hyperactivity. However, one of the clusters scored higher on multiple cognitive assessment measures of executive function. Further analyses that compared measures of localised functional connectivity between these clusters revealed no significant differences; however, between-cluster differences were found on measures of intra- and inter-network connectivity between global brain networks. In our third empirical study, we inferred a Hidden Markov Model using resting-state MEG data to investigate the relationships between neurodevelopmental features of interest and transient states of neural activity. The complexity of participants’ MEG time-courses was positively related to their cognitive ability. Higher probabilities of transitioning into certain states, particularly those involving the default-mode network, fronto-parietal networks, and sensory processing regions, also predicted individual differences in cognitive ability. Finally, we completed a large public engagement project centred around neurodiversity and inclusive practices in schools. After running a multi-stakeholder workshop about barriers to learning and wellbeing in the UK, we collected evidence-based recommendations for educational and social care policy change, which informed our comprehensive set of inclusion resources for schools. This thesis represents an important advance towards the transdiagnostic understanding of neurodevelopmental differences. This work builds on previous research in cognitive neuroscience while placing brain network complexity and neural dynamics in a developmental context. We believe that the research and practical endeavours described here will help guide future efforts in the scientific study of neurodiversity, in addition to the creation of more equitable, evidence-based educational frameworks.
Open Access
Social interaction, social status, and mental health and how they were impacted by the COVID-19 pandemic as a natural experiment of social isolation.
Gormley, Siobhan
Affective difficulties are some of the most prevalent health problems worldwide and can lead to severe disruptions in daily functions. Socio-evolutionary theories stress the importance of social groups and status for mental well-being. The work presented in this thesis aims to further our understanding of the link between social interaction, status and mental health. Social isolation due to the COVID-19 pandemic provided a natural experiment of what can happen in cases of extreme reductions in socialisation. We attempted to investigate this by asking three major questions. First, is perceived status associated with, and predictive of, depressive and anxious symptomology? (Chapters 3, 4, 5). Second are affective symptoms and perceptions of involuntary subordination associated with individual differences in perception of others and behaviour? (Chapters 3 & 4). Finally, has the COVID- 19 pandemic lead to changes in mental health symptoms (Chapter 5), and are there any predictors of individual differences in COVID-19-specific behaviour? (Chapter 6). To do this I used a mixture of survey and experimental designs, along with cross-sectional and longitudinal research. We found strong consistent evidence for the association between low perceptions of status, measured by involuntary subordination and depressive and anxious symptoms (Chapters 3, 4, 5). There was also evidence for the predictive nature of involuntary subordination providing some support for a causal relationship (Chapter 5). Affective symptoms were associated with individual differences in general perceptions of social status (Chapter 3) and showed indirect effects on behavioural responses to socially defeating situations (Chapter 4). Sustained increases in anxiety, involuntary subordination, and interpersonal hyper-sensitivity were observed during the COVID-19 pandemic when compared to pre-pandemic levels, and involuntary subordination showed consistent predictive power (Chapter 5). Behaviourally executive function and depression were not predictors of social distance behaviours during the pandemic in adolescents. We did see a significant effect of anxiety when depression symptoms were controlled for; however, this finding was not robust. Future research should aim to investigate the causal relationship of social status and groups on affect, utilising experimental designs, neuroimaging, and new statistical techniques. Cross-cultural research is also needed if we are to establish the evolutionary nature of these constructs.
Open Access
Genetic and Environmental Influences on Cognitive and Neural Development
Smith, Tess
The brain self-organises slowly over time, being shaped by both endogenous and exogenous factors. This process results in a unique system for each person, with neuronal circuits shaped by an individual’s genetic background and experience. This thesis aims to better understand these multifactorial gene-environment-outcome relationships, and how they relate to human development. Using multiple data types from the Avon Longitudinal Study of Parents and Children (ALSPAC) combined with a diverse methodological approach, I highlight compounding environmental factors which longitudinally interact to predict developmental outcomes, generate meaningful and valid measures of polygenic propensity, and consider the ways the early environment and polygenic heritability influence and interact to influence specific features of the structural connectome, features that are essential for coherent neural communication and brain function. In the first empirical chapter, I use structural equation modelling to assess whether maternal mental health longitudinally mediates associations between the early socioeconomic status (SES) and key developmental outcomes (i.e., cognitive ability and child mental health). I show that maternal mental health mediates these relationships, and primarily during the first year of life. In the second empirical chapter, I employ polygenic score (PGS) analysis to generate valid and meaningful PGSs for cognitive ability and educational attainment, with the PGSs for cognitive ability ultimately taken forward in the proceeding empirical chapters. The third empirical chapter incorporates measures of socioeconomic status, PGSs for cognitive ability, and diffusion tract imaging data to explore how these data types are associated with global and local features of the connectome. This was made possible by employing graph theory metrics and partial least squares regression analysis. I demonstrate the early child environment and the genome influence the structure of the brain across at least three local metrics of network connectivity (i.e., node strength, degree, and clustering coefficient), with node strength playing a particularly significant role. Drawing on local node strength and generalised linear modelling, the fourth and final empirical chapter considers how measures of SES and polygenic propensity interact to influence developmental outcome and specific features of the connectome that fall under the rich-club framework (i.e., rich-club, feeder, and peripheral nodes, and rich-rich, rich-feeder, and peripheral connection types). Here I show both SES and PGSs to interact to influence developmental outcome, as well as node and connection type. In particular, PGS and the SES-by-PGS interaction appear to relate to the connectivity of a rich-club of highly interconnected nodes most strongly, but these variables do this by shaping so-called ‘feeder’ connections. Finally, the links between polygenic propensity and connection strength are patterned by gene expression, being strongest across connections that span regions with moderate levels of expression similarity. This thesis provides new insights made possible by employing a diverse methodological approach in combination with a rich prospective longitudinal dataset. The findings show how it is possible to span multiple levels of analysis within a contemporary developmental science framework, to consider how factors operate at a population level in terms of behaviour, environment, heritability, and brain organisation.
Open Access
A cross-species network science approach to childhood adversity
Carozza, Sofia
The brain is an adaptive organ whose development is responsive to its early-life environment. This means that early experiences shape the biological substrates of lifelong health and well-being. Indeed, exposure to significant adversity in childhood increases the risk of a range of cognitive and behavioural difficulties later in life. Previous work has sought to understand this pathway by identifying neural correlates of exposure to poverty and violence, focusing particularly on the morphology of several key cortical and sub-cortical structures. However, region-specific correlations alone are not enough to generate a coherent account of the impact of child adversity. This thesis aims to address this gap by conducting a global, causal, and mechanistic investigation of the impact of adverse experiences in childhood, leveraging network science methods and computational modelling to accommodate more of the complexity inherent in the nature of childhood experiences and development. In Chapter 2, I aim to ascertain whether early-life adversity alters the development of rodent structural brain organisation. Using imaging data obtained from collaborators at Yale University, I first reconstruct the structural connectomes of a sample of mice, half of which were exposed to a paradigm of unpredictable postnatal stress. I then simulate the development of each connectome using generative network modelling, a computational approach that generates complex networks probabilistically based on a trade-off between the cost and value of connections. After validating the quality of the simulations, I conduct a case-control comparison of the models that best replicate the brain networks of adversity-exposed and unexposed mice, and explore potential implications of an observed difference in model parameters. I conclude that early adversity may cause an increase in stochasticity in the formation of the structural connectome. In Chapter 3, I aim to establish a global understanding of the nature of early adverse experiences as they occur in the general population. The analysis tests a prominent theory within the field, the Dimensional Model of Early Adversity (McLaughlin, Sheridan, & Lambert, 2014). Using data from a longitudinal cohort study, I apply a data-driven network and clustering approach to identify dimensions of adversity, or groups of adverse experiences that impact children in similar ways. Contrary to my pre-registered hypotheses—and prior work on the dimensional model—I show that a range of deprivation-related experiences are closely related to later cognitive and socioemotional difficulties. I conclude that deprivation is a broad and robust predictor of later psychological outcomes, and that theories of early adversity would benefit from triangulation of findings through diverse methodologies. In Chapter 4, I aim to replicate findings from Chapter 2 in humans, and to extend them by determining which early experiences are most predictive of brain wiring in young adulthood and by assessing the relationship between model parameters and cognitive and socioemotional difficulties. I first reconstruct and simulate the development of the structural connectomes of a subsample of participants from the longitudinal cohort used in Chapter 3. A data-driven partial least squares (PLS) regression shows that no early experiences of adversity predict brain wiring parameters in young adulthood. Sensitivity analyses confirm the robustness of this null finding, for which I find some previous support in the literature. Finally, in the General Discussion I delineate broader conclusions that can be drawn from this work, acknowledge its overarching limitations, and suggest promising next steps in the study of early adversity. I conclude with an addendum on the application of neuroscience research to public education and policy, in which I warn against neuroscience overreach and make specific proposals for improving the integrity, quality, and utility of the scientific study of childhood adversity.
Open Access
The Relation Among Thought Suppression, Forgetting, and Mental Health
Mamat, Zulkayda
Things are known by their opposites. If we want to dive deeper into how we remember, we must not forget about forgetting. Even after more than a decade of research on forgetting having an adaptive and active function in our lives beyond a passive role as a failed attempt to remember, it is still not widely studied in its own right. Not only can forgetting be an adaptive gateway for learning new things and regulating our emotions, but also be a critical opening for dealing with the traumas that haunt us, the worries that consume us, or the thoughts that just never seem to stop. Our ability to forget should not only be studied as a reaction – a failure to remember, a symptom in patients, or a response to unprocessed trauma. Rather, forgetting can and has been increasingly studied through its acts. At its core, this thesis is intended to put forth an appeal to urge and inspire research efforts to move beyond seeing the water in the half-full glass only after half of it is gone, but to take hold of the glass and drink from it! For instance, we passively experience decay through observing the changing colours of leaves in nature; however, this does not preclude our knowledge of a very active force behind the cycle of decay, indeed for the purposes of nothing other than renewal. Similarly, forgetting can feel passive experientially, but there could be very active forces that bring about this seeming decay, perhaps to bring out the light that shines on life. In this way, this thesis will examine one aspect of this mighty force – motivated forgetting – our capacity to actively block out distressful thoughts and render them less memorable or intrusive. Just as discipline is the root of building virtuous character, controlling our thoughts is perhaps the beginning of disciplining our minds in this arduous walk of life. Active forgetting can be a tool, which in times of difficulty can allow us to get up and step back into walking in life, rather than letting life walk all over us. Indeed, when the COVID-19 pandemic happened, it felt like life was getting out of control, and anxiety started to walk over everyone, especially the most vulnerable of us. Although decades of research have “hinted” and “suggested” that suppression “may” promote resilience in facing difficulties of life, we were all too afraid to truly affirm this with conviction because of an even longer line of work going back to Freud that engraved suppressing of thoughts as fundamentally maladaptive into the public mind. Thus, the pandemic actually presented an opportunity to put these competing theories to test and for the first time causally investigate the impact of suppression training on mental health. This research study forms the heart of this thesis and will be discussed in detail in Chapter 3. Of course, our conviction in the beneficial impact of suppression would not have been as strongly present if not supported by existing research as well as personal experiences of everyday people using suppression to overcome life’s challenges. With this in mind, Chapter 1 will highlight existing research that gave us the extra confidence to grab hold of that glass of water, so to say. It will also walk us through how the pandemic specially brewed so much anxiety amongst the population that in a way made everyone parched for some sort of relief. However, we must also recognize that research into motivated forgetting has been around for some time now, the resistance of its applicability is stickier than one may think. Hence, Chapter 2 is dedicated to a detailed discussion of the theories and experiments that have been proposed to showcase an ironic heightening of memory for the supposedly suppressed information, and will offer alternative explanations for why such efforts may be misdirected. Then, we will swiftly move into the suppression training study in Chapter 3 and conclude with some heart-warming reflections from many people who learned to suppress their fears and intuitively integrated the technique into their own lives in the midst of the pandemic. It seems that we have not only found water, but sweet fragrant rosewater! Having tasted its sweetness, we wanted everyone to enjoy it, moving us into Chapter 4 where we present our translation of the suppression training from a remote testing procedure into an accessible app, hoping for a future where we all taste the sweetness of relief from intrusive thoughts. This then prompts us to think: how can we find more people who know what this drink tastes like? Such a question shall lead us into Chapter 5 in which we provide a novel tool for investigating the phenomenon of selective forgetting and its relevant phenomena such as distinct suppression strategies and prevalence of recovered memories. With that, we conclude this writing, having equipped wayfarers with signposts to find out more about the quality of this precious glass of rosewater that we seem to all possess but many do not yet know how to find or drink from.
Open Access
Exploring the functional organization of cerebral cortex using data from the Human Connectome Project
Rajimehr, Reza
Neuroimaging techniques such as MRI and fMRI provide a non-invasive window to look at the macroscale organization of cerebral cortex. Despite monumental efforts to map the structure and function of cerebral cortex, our understanding has been limited due to insufficient sample sizes, low-resolution imaging, and suboptimal volume-based analyses. In recent years, the Human Connectome Project (HCP) has released high-quality imaging datasets from an unprecedented number of subjects. The surface-based analysis of HCP data provides a unique opportunity to investigate detailed topography of structural features and functional activation maps in cerebral cortex. In my thesis, I will describe three studies that were conducted using the HCP database. In the first study, I tested genetic influences on the organization of category-selective areas of visual cortex. Using fMRI data from monozygotic (MZ) and dizygotic (DZ) twins, I found that category-selective maps for faces, bodies, and places were more identical in MZ than DZ twins. Within each category-selective area, distinct subregions showed significant genetic influences. Structural MRI analysis revealed that the cortical curvature, thickness, and myelination in genetic and non-genetic subregions were different. The results suggest a link between genetic influences and cortical morphology during cortical development. In the second study, I analyzed functional activation maps in language and social tasks, and found an important link between language and social processing in the human brain. While certain cortical areas in the left hemisphere were specialized in language processing, their homologues in the right hemisphere were specialized in processing of social information. This complementary hemispheric lateralization provides novel insights into the origin of language selectivity in the brain, and it could have implications in understanding neurocognitive mechanisms of social disorders such as autism. In the third study, I comprehensively investigated the large-scale functional architecture of cerebral cortex during naturalistic movie-watching. A data-driven clustering approach revealed a map of 24 functional areas/networks, each related to a specific aspect of sensory or cognitive processing. The map included three distinct executive control (domain-general) networks which showed a strong push-pull interaction with domain-specific regions in visual, auditory, and language cortex. The cortical parcellation scheme presented here provides a comprehensive and unified map of functionally defined areas, which could replace a large set of functional localizer maps. At the end of the thesis, I outline additional ongoing projects using the HCP database, addressing a number of further questions concerning the organization of human cerebral cortex.
Open Access
The neurocognitive mechanisms underlying dissociative amnesia
Marsh, Laura
Dissociative amnesia refers to the loss of autobiographical memory with a presumed psychological cause. This can involve memory loss for traumatic experiences, such as in PTSD. It can also involve more extensive amnesia for several years, or, in its most extreme form, a fugue state involving total loss of memory and sense of identity. These latter, ‘generalised’ forms of dissociative amnesia are rare, and the mechanisms underlying the memory loss are poorly understood. It has been theorised that the amnesia results from a prefrontally-mediated inhibition of memory systems, which is presumed to occur sub or semi-consciously (Kopelman, 2000, 2002). A parallel body of experimental work has defined a prefrontally-mediated ‘memory control' network, which can be engaged to inhibit memory retrieval voluntarily. These findings offer a candidate neurobiological mechanism for the inhibition of retrieval in dissociative amnesia, and provide a set of specific, testable predictions regarding the putative role of this mechanism in the memory loss. In the current thesis, we systematically reviewed the literature on the neurophysiological correlates of dissociative amnesia, evaluating the existing evidence in support of this proposed neurobiological mechanism. Chapter 3 reports a reanalysis of existing fMRI data from two patients with dissociative amnesia (Kikuchi et al. 2010), in which we formally tested whether the patterns of neural activation associated with dissociative amnesia aligned with those observed during voluntary retrieval inhibition. We found strong evidence for the hypothesised prefrontally-medicated inhibition of memory systems. Thus, we attempted to replicate and extend these findings in an ongoing prospective case series. By combining qualitative, quantitative and neuroimaging methods, we aimed to develop a more cohesive understanding of the bio-psycho-(social) mechanisms underlying dissociative amnesia. Chapters 4-7 report the results from the first 3 patients, recruited to the study over a period of 2 ½ years, with neuroimaging completed in one case. Chapter 8 provides a discussion of these findings in relation to the broader clinical understanding dissociative amnesia, and in relation to current understanding of dissociative symptoms more broadly.
Open Access
Characterising hippocampal replay in the APP^NL-G-F mouse model of Alzheimer’s disease
Shipley, Sarah
Memories which are initially dependent upon the hippocampus can gradually be incorporated into cortical networks, a process known as system-level consolidation. Replay, the rapid recapitulation of neural sequences experienced during behaviour, is thought to be a key mechanism supporting consolidation and predominantly occurs during rest and sleep. More generally replay is also held to play a role in memory retrieval, navigation, and likely planning. Consistent with these multiple roles, replay disruption negatively impacts performance on memory related tasks. The hippocampus and associated structures are known to be among the first to exhibit signs of pathology in Alzheimer’s disease; a disorder characterized by progressive memory dysfunction. It is therefore pertinent to understand how functional changes in the hippocampus, including those related to replay, relate to the pathological and symptomatic changes that accumulate of the course of the disease. In this thesis I describe the results of two experiments in which the relationship between hippocampal replay and Alzheimer’s disease are investigated. In the first experiment, I present a characterization of hippocampal replay in the APP^NL-G-F knock-in rodent model of Alzheimer’s disease. Place cells were recorded from the CA1 region of the hippocampus as mice performed a radial-arm maze task, and during subsequent sleep. The task was designed to separably identify errors in reference and working memory. APP^NL-G-F mice were found to make greater numbers of reference memory errors than control animals. This reduction in memory performance was related to an overall reduction in the rate at which replay occurred, both during sleep and during brief periods of inactivity in the maze. During waking periods, this deficit appeared to result from a reduction in the number of population activity bursts; some of which incorporated bona fide replay sequences. In sleep, this loss was further exacerbated by a reduction in the coherence of replay sequences; fewer of the candidate population bursts carried decodable sequences. It seems plausible that a deeper understanding of how Alzheimer’s pathology produces the deficits in replay described here, may lead to the identification of new therapeutic targets. In the second experiment, I describe the interaction between cholinergic tone in the hippocampus and replay in both wild-type and APP^NL-G-F mice. Acetylcholine levels are reduced in Alzheimer’s disease and acetylcholinesterase inhibitors, which increase the availability of acetylcholine in the brain, form one of the few approved therapeutic treatments. However, acetylcholine has been hypothesised to modulate the balance between encoding and consolidation, with falling cholinergic levels being necessary for replay to occur after new memories are encoded. To assess the role of acetylcholine in replay, I used a viral vector to exclusively express Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) in medial septal cholinergic neurons of ChAT-cre mice. DREADDs were activated using systemic injection of an actuator (CNO), stimulating cholinergic release in the hippocampus and other brain regions. CNO was administered immediately prior to sleep, when Alzheimer’s sufferers are typically recommended to take acetylcholinesterase inhibitors. I found that increased cholinergic tone during sleep reduced the rate at which replay sequences occurred in APP^NL-G-F mice. However, this changed was not associated with any further reduction in memory performance. Nevertheless, these results do indicate that that timing of therapeutic treatments targeting the cholinergic system should be considered relative to diurnal changes in neuromodulatory levels, and more generally the sleep-wake cycle.
Open Access
Concepts and Schemas: Representational Format for Structured Knowledge
Bokeria, Levan; Bokeria, Levan [0000-0001-5982-4520]
One of the central challenges in cognitive neuroscience has been the study of internal mental representations of the external objects, events and relations that allow us to predict and interact with the world. Recently, researchers have uncovered parallels between the neural processing of physical space and of abstract knowledge, such that the established neural mechanisms for spatial navigation may also shed light on how we represent conceptual knowledge. In this thesis, we present a set of behavioural experiments examining the representational format of knowledge structures such as concepts and schemas, and develop learning paradigms that test algorithmic-level theories of spatial and non-spatial processing. We start by discussing classical geometric models of knowledge representation, which view concepts as regions in abstract, multidimensional spaces organised by metric principles. These models have been supported by recent neuroimaging studies that suggest shared neural representations for spatial and non-spatial reasoning. We consider an older set of behavioural results that uncovered violations of the metric axioms of such representations, and discuss augmented geometric models that have been developed in response. One such model – the distance-density model – is examined in Chapter 2, using similarity judgments on a novel one-dimensional stimulus space. We did not find support for the basic prediction that psychological density affects similarity. In Chapter 3, we adapted the conceptual stimulus spaces used in the recent neuroimaging studies, and found that violations of metric requirements depend on the nature of the dimensions defining the stimuli. Nonetheless, using simulations and considering the prior psychological literature, we argue that another type of augmented model – the attention-weighted geometric model – is unlikely to account for such violations. These chapters therefore cast doubt on geometric models as adequate algorithmic-level theories for human knowledge representation. The next two chapters develop schema learning tasks that lay the foundation for continued study of parallels between spatial and non-spatial reasoning. In Chapter 4, we examined how a non-spatial schema acquired in one conceptual space can influence learning in a different conceptual space. Across two experiments, we found effects consistent with generalisation of knowledge, but only for certain counterbalancing conditions. We argue for the importance of further refining our task and stimuli to develop a fast and flexible knowledge-transfer paradigm for studying relations between spatial and non-spatial processing, which could also be extended to analogical reasoning, categorisation and schemas. In Chapter 5, we examined the nature of representational elements constituting spatial schemas. The prior literature has defined such schemas as networks of stimulus-location associative elements that can benefit learning. An unexamined possibility is that, instead of forming a cohesive network, such elements act independently to influence acquisition of new knowledge only within their local neighbourhood. Across two experiments involving learning of image-location associations on 2D boards, we find evidence consistent with this interpretation, and we outline how our paradigm can be adapted to address analogous questions for non-spatial schemas. Taken together, our results question spatial representation of knowledge at the algorithmic level, as well as the nature of spatial schema, and emphasize the importance of continued research for elucidating commonalities and differences between spatial and non-spatial reasoning.
Open Access
The contributions of posterior lateral temporal cortex to language and control networks
Hodgson, Victoria
Within cognitive neuroscience, understanding how the brain supports the flexible use of language and meaning remains an ongoing challenge. In recent years, the Controlled Semantic Cognition framework has offered an account of the neural correlates of representation and control of meaningful information. The context-independent storage of semantic representations is subserved by the bilateral anterior temporal lobe hub and its connections to modality-specific spokes distributed throughout the cortex, while semantic control – the flexible, goal-oriented access and manipulation of this meaningful information – relies upon a left-lateralised set of frontal and temporal brain regions. However, we do not yet understand precisely how these two networks may integrate. Further, the set of regions supporting semantics overlaps somewhat with phonological processing, and the semantic control network is distinct, but overlapping with, the multiple demand network for domain-general executive control. This raises the question of how semantic representation and control might relate to other domains. The exact extent of this overlap, and the implications this has both for role of individual regions, is still uncertain. One such region is the posterior lateral temporal cortex (pLTC). Many disparate literatures have made claims about the function of the pLTC or subregions within, suggesting involvement in domains such as phonology, semantics, social processing, and both semantic and domain-general control. However, it remains unclear how this area is organised, including which subregions might be recruited by the different language, semantic and control networks in which the pLTC has been implicated, the domain-specificity of the control processes supported by these subregions, and whether the pLTC is a site of integration between networks. To address these questions, Chapters 2 and 3 employ activation likelihood estimation meta-analysis at the whole brain and region of interest (ROI) levels to compare the regions of consistent activation across domains. Then, Chapters 4 and 5 use task-based functional MRI with a factorial design in neurotypical participants, to explore the independent and interacting effects of stimulus domain and task process on the activation and connectivity of control regions. In Chapter 2, whole brain meta-analyses are used to delineate dimensions of functional organisation in the language network, highlighting the importance of language subdomain (semantic versus phonology), as well as the distinction between representation and control processes as key factors in determining the functional organisation of the language network. In Chapter 3, a large cross-domain meta-analysis within a bilateral pLTC ROI compares activation convergence across seven cognitive domains simultaneously (semantics, semantic control, phonology, tool processing, face processing, theory of mind and biological motion processing), in order to map out subregions within the pLTC that are shared and distinct across domains, and generate hypotheses about the cognitive processes that may be supported by these regions. From the meta-analyses, several hypotheses emerged regarding the domain-specificity of control regions, in particular the left pLTC, that warranted further testing. In Chapter 4, univariate whole brain analyses indicate that both task process and stimulus domain affect the activation of control networks and their constituent regions, separating these factors for the first time to give a more nuanced understanding of what is “special” about semantic control. ROI analyses also demonstrate that regions for semantic and non-semantic control are not homogenous in their profile of activity. Chapter 5 uses task-based psycho-physiological interaction analyses in an attempt to assess how the functional connectivity of key seeds in the left pLTC changes with task process and stimulus modality. Though these results are minimal and provide no insight, resting-state functional connectivity analyses on an independent fMRI dataset reveal different intrinsic connectivity for the left pMTG/STS versus pITG, giving a better understanding of the precise contributions of these pLTC subregions to semantic and control networks.
Open Access
Development of novel multidimensional pattern-based EEG/MEG connectivity methods and their application to investigate the semantic brain network
Rahimi Ghazikalayeh, Masoomeh
Functional and effective connectivity methods are key to understanding how brain regions interplay to perform complex cognitive processes. Yet, most current connectivity methods summarise activity within brain regions to unidimensional measures, resulting in a loss of information. Only recently, new functional connectivity methods have been introduced that exploit multidimensional information, i.e. pattern-to-pattern relationships across regions. To date, these methods have mostly been applied to functional Magnetic Resonance Imaging (fMRI) data, and no method allows the estimation of vertex-to-vertex transformations with the temporal specificity of electro-/magnetoencephalography (EEG/MEG) data. The current thesis introduces novel multidimensional pattern-based connectivity methods for event-related EEG/MEG applications. I introduced time-lagged multidimensional pattern connectivity (TL-MDPC), nonlinear TL-MDPC (nTL-MDPC), and multivariate TL-MDPC (mvTL-MDPC), as novel bivariate and multivariate functional connectivity metrics for EEG/MEG research. They detect linear as well as nonlinear dependencies between patterns, through estimation vertex-to-vertex transformations, for pairs of brain regions and latencies. I evaluated my methods on simulated data as well as on an existing EEG/MEG dataset. Particularly, I asked: 1) whether multidimensional connectivity methods capture more information than unidimensional ones, 2) whether a nonlinear multidimensional connectivity method captures more and different information than its linear counterpart, and 3) whether moving from a bivariate multidimensional connectivity method towards its multivariate version yields more realistic results. In numerical simulations I could show that: 1) none of my novel methods are prone to producing false positives for independent random patterns, 2) TL-MDPC captures both unidimensional and multidimensional connectivity, and performs better than its unidimensional version in the case of multidimensional effects, 3) nTL-MDPC captures both linear and nonlinear dependencies, and performs slightly better than its linear version when nonlinear dependencies exist, 4) mvTL-MDPC produces more realistic results than its bivariate version, TL-MDPC. I used my new methods to gain novel insights into the brain semantic network. In order to do so, I compared two semantic visual word processing tasks, varying the depth of semantic processing of words by contrasting a semantic decision (SD) and a lexical decision (LD) task. I used both conventional unidimensional approaches, including evoked responses and coherence, as well as my novel multidimensional connectivity methods, and found 1) my multidimensional methods provide a more complete picture of the brain dynamics compared to coherence and another unidimensional connectivity method, by producing richer connectivity among ROIs, 2) four semantic regions, lATL, rATL, PTC, and IFG, showed rich connectivity with each other, 3) lATL and rATL showed strong connectivity throughout using all unidimensional and multidimensional connectivity methods, confirming the essential role of a bilateral ATL hub for semantic representation, and finally 4) I did not find a key semantic role for AG. Importantly, to deal with the limited spatial resolution issue of EEG/MEG, I proposed a novel spatial subsampling method to select the most informative vertices within each ROI’s patterns, based on a k-means clustering algorithm. I also provided a quantitative assessment of homogeneous and non-homogeneous leakage among my ROIs. In summary, this thesis proposes novel connectivity methods for exploiting the multidimensional information of EEG/MEG patterns and demonstrates their usefulness for the investigation of the brain semantic network. My key findings support the principles of the Controlled Semantic Cognition (CSC) framework, and highlight the usefulness and advantages of multidimensional connectivity methods for cognitive neuroscience.
Open Access
Computational Principles of Brain Network Development
Akarca, Danyal; Akarca, Danyal [0000-0002-5931-0295]
Brain development can be viewed through many lenses and studied at many scales. However, multiple theoretical perspectives have argued that brain organisation develops via competitive interactions between its constituent units, dynamically over time. In this thesis, I focus on modelling these interactions. In Chapter 1, after providing an historical backdrop to the field of developmental systems neuroscience, I introduce generative network models. These relatively new family of models are capable of simulating probabilistic network development. The basis for these models includes simple sets of wiring rules, existing within various imposed biophysical constraints, which steer the developmental trajectory of the network. In Chapter 2, I show how the applications of these models can reveal simple principles that may contribute to our understanding of neurodiversity. In particular, small iterative updates in networks can lead to constrained variability in a child’s macroscopic structural brain organisation inferred via in vivo diffusion imaging. I highlight how decompositions of networks into the generative components used to construct them in this way can be useful lowerdimensional representations of developmental ingredients. This is particularly relevant when aiming to bridge associations between genomics, cognition and the brain for answering developmental questions. Generative network models emphasise the evolving economic context of dynamic interactive negotiations between brain regions. These regions can be defined at any scale. In Chapter 3 I pivot from studying cross-sectional macroscopic connectomes, to modelling the microstructural longitudinal development of in vitro neuronal networks at the cellular scale. I show that current instantiations of a homophily generative model are an effective growth model of in vitro neuronal network development. This simple model can recapitulate observable local topological organisation of functional networks across species, time, plating densities, cell-types and experimental conditions. Together, Chapter 2 and 3 can be considered as a test of whether generative network models can simulate biological brain topologies, in an unsupervised fashion, according to intrinsic wiring rules. The nervous system has evolved, in part, to sustain and ensure survival of the organism. Therefore, the structural organisation of the brain must be considered with respect to how it directly supports function in order to achieve behavioural goals. However, many current frameworks posit only associations between neural structure and function, rather than direct bidirectional influences. In Chapter 4, I aim to model how previously aforementioned economic negotiations may facilitate direct structure-function interactions. I introduce an extension of artificial neural networks for which I term spatially-embedded recurrent neural networks (seRNNs). seRNNs add simple biophysical constraints into the model within a regularisation term, that change the nature of how connections change during optimisation. I show that adding local spatial and communication constraints to this neural network points towards a convergent solution whereby optimal functional trade-offs are attained where sparsity, homophily generative mechanisms, small-worldness, functional configuration in space and energetic efficiency together coalesce. In Chapter 5 I summarise key take-aways and provide what I believe to be promising future avenues for the applications of computational modelling to developmental systems neuroscience.
Open Access
Developmental burden of childhood adversity: Insight from longitudinal perspectives
Nweze, Tochukwu
Childhood adversity has been implicated in poorer developmental outcomes such as behavioral problems, poorer mental health and cognitive deficits. Studies have also linked adversity to alterations in cortical brain structures. To date however, almost all knowledge of the effects of adversity on outcomes has come from cross-sectional studies or longitudinal studies that used cross-sectional data analysis method. In an attempt to bridge this gap, across three empirical studies, this thesis sets out to implement series of longitudinal data modelling aimed at disentangling the intricacies of the effects of childhood adversity on mental health, cognitive abilities and brain development. In study 1, I analyzed a large sample (N=13,287) of 5 wave longitudinal data obtained from the Millennium Cohort Study in an attempt to understand how early-life adversity, mental health and cognition affect one another or how the effects unfold over time. To achieve this, I used focused longitudinal mediation model via path model approach. Results showed that early-life adversity was associated with poorer performance in spatial working memory and vocabulary performance. Notably, current and previous mental health mediated a substantial proportion (working memory: 59%; vocabulary: ¬70%), of these effects. Findings also showed that adversity has an enduring adverse effect on mental health, and that poorer mental health is associated with poorer cognitive performance later on in development. Moreover, the adverse effects of mental health were cumulative: poor mental health early on is associated with poorer cognitive scores up to 11 years later, above and beyond contemporaneous mental health. Based on this evidence, I suggested that the academic and cognitive competence of vulnerable children may be enhanced if their early mental health conditions are given deliberate clinical attention. In a follow-up study 2, I attempted to provide empirical support for dimensional model of adversity which argues that childhood adversity can be classified into subgroups, known as dimensions. For this purpose, I analyzed rich set of adverse childhood experiences obtained from a subset of ALSPAC cohort sample (N = 2,965) using latent class analysis. Findings showed evidence of five distinct adversity subgroups, namely, low adversity, dysfunctional family, parental deprivation, family poverty and global adversity. To establish a pathway to cognitive functioning among the adversity subgroups, a further analysis using latent class regression revealed that family poverty subgroup performed poorest in working memory and inhibition tasks. A separate analysis revealed that the effects of each individual adversity types on cognitive outcomes were mostly consistent with the observed class performance in which they co-occurred. Regardless, sensitive periods (timing of adversity exposure) explained more variability in these observed effects compared to accumulation hypothesis. In study 3, I analysed a subset of IMAGEN cohort sample (N = 502) using latent change score model and complete longitudinal mediation model via autoregressive path approach, aimed to understand the long-term interrelations between adverse life events, cortical development and cognitive functioning. Results of latent change score model showed that greater baseline adverse life events predicted a marginal reduction in the right anterior cingulate surface area. In addition, baseline right orbitofrontal cortical thickness predicted a decrease change in adverse life events. I found no evidence of association between adverse life events and volumes of cortical structures or cognitive outcomes. In separate longitudinal analyses, I found no evidence of indirect effects in the two neurocognitive pathways that link adverse life events in adolescence to brain and cognitive outcomes. Although the results of latent change score model appear to support the robust cross-sectional studies which have implicated adverse events in brain alterations, especially in the prefrontal, however, the magnitude of effects observed in this study 3 are smaller than have been reported in the cross-sectional studies, suggesting that potential long-term impact of adverse life events on brain structures may likely be more modest than previously noted. I end the thesis by articulating the implications of these findings across the 3 empirical studies, indicating the strengths and limitations, and suggesting areas for future directions. Generally, it is my hope that new insight drawn from these longitudinal studies will inform the right policies in the society. Such policies may include but not limited to increase clinical intervention for the vulnerable and most underprivileged children as well greater financial aids to families living in poverty, given recent reports that such aid package can alter the trajectories of developmental outcomes of children in a positive way.
Open Access
Early adversity and individual differences in adolescent mental health
Uh, Stepheni
Adolescence is a dynamic period of profound changes during which individuals are particularly susceptible to poor mental health. These biological, cognitive, and psychosocial changes – alongside the increased influence of environmental factors –interact and impact both immediate and long-term wellbeing. Early adversity, in particular, confers significant risks to adolescent mental health. Thus, adolescence is an important window for researchers to investigate processes underlying different outcomes. However, the heterogeneity across the various factors that influence adolescent mental health presents many empirical challenges. This thesis, therefore, applies novel multivariate approaches to explore the heterogeneity of individual differences, risk factors, and outcomes outlined in theoretical and empirical literature of adolescent mental health and development. The first empirical chapter entails a prospective longitudinal study that investigates distinct developmental trajectories and associated risk factors to self-harm. I first used unsupervised machine learning algorithms to identify distinct psychological subgroups of young people who self-harm. This was followed by a supervised machine learning algorithm to identify significant risk factors across a nine-year period for the different subgroups – ultimately distinguishing two different pathways to self-harm. The second empirical chapter provides a novel analytical pipeline to investigate the heterogeneity of overlapping types of adversity, and risks for internalising problems, in a large sample of adolescents. This pipeline involved three steps: constructing a topological representation (“map”) of co-occurring adversities experienced by adolescents; identifying whether poor mental health outcomes are co-localised with specific features of this adversity map; and applying a precision stratification method to investigate differences in corticolimbic connectivity associated with differential outcomes in participants matched for adversity profiles. This approach revealed an adversity profile highly linked with internalising problems; importantly, differential mental health outcomes in participants sharing this adversity profile were linked with differences in corticolimbic connectivity. The third and final empirical chapter targets a specific mechanism strongly implicated in adolescent mental health and development: implicit emotion regulation. I used a modified emotional Go/NoGo fMRI task, in addition to reproducible preprocessing pipelines and a nonparametric group analysis, to explore neural correlates of implicit emotion regulation and individual differences in childhood. The results showed multiple significant response inhibition effects (i.e., larger NoGo vs Go activation in the IFG, insula, and ACC) and valence effects in the putamen and pallidum. Though I did not find significant relationships between these neural responses and individual differences in mental health. This thesis offers advances towards capturing the heterogeneity across multiple levels of factors that interact and influence adolescent mental health. The multivariate and targeted approaches targeting the bio-psycho-social processes that influence mental health outcomes provide important avenues for future research and policies to promote mental wellbeing in adolescents.