From Neural Dynamics to Psychological Profiles: A Comprehensive Study of Anxiety in Young and Adult Individuals

Anxiety, a common yet complex emotional state, ranges over a spectrum from transient feelings of worry to chronic disorders with significant social and economic impacts. While traditional models have focused on its psychological and clinical aspects, recent advances in neuroscience highlight the need for a more detailed understanding of anxiety as a brain-based phenomenon, exploiting multivariate approaches. This thesis indeed explores the neural underpinnings of trait and state anxiety using a multimodal approach, incorporating structural and functional neuroimaging techniques as well as machine learning methodologies, attempting to clarify the neural and psychological disruptions associated to anxiety. Across four studies, we thus aim to identify the brain networks involved in trait anxiety and uncover potential neural biomarkers. Study 1 investigates the structural neural correlates of trait anxiety in adults by applying Parallel ICA, a data fusion machine learning technique, to explore joint gray-white matter networks. We hypothesized that specific structural neural networks could predict trait anxiety and that some maladaptive strategies would correlate positively with anxiety levels. Indeed, we identified a parieto-temporal and a fronto-parietal network linked to trait anxiety, along with correlations between anxiety and maladaptive strategies like rumination. Study 2 extends this investigation to a larger sample of young individuals, hypothesizing that gray and white matter networks, along with a specific psychological profile characterized by elevated stress, could distinct higher from lower anxious young individuals. Findings revealed gray matter networks with lower concentration in higher anxious individuals, linked to areas such as the insula and the anterior cingulate, and white matter networks with greater concentration in lower anxious individuals, linked to the insula and the precuneus. Additionally, elevated stress and impulsivity were found in the higher anxiety group. Study 3 examines trait anxiety in adolescents, utilizing resting-state functional connectivity analysis hypothesizing altered connectivity in large-scale brain networks, particularly the Default Mode Network. Results showed a temporo-parietal and an occipito-cerebellar network, including regions ascribable to the Default Mode Network, to have an altered resting-state connectivity associated with trait anxiety. Study 4 (ongoing) shifts the focus to state anxiety, using the novel "Threat of Scream" paradigm and functional near-infrared spectroscopy (fNIRS) to capture real-time neural responses to unpredictable aversive stimuli. We hypothesized that fronto-temporal regions would characterize anxiety responses, and that scream stimuli would affect interpersonal distancing during a social task. Preliminary findings indicate that screams elicit anxiety and influence social decision-making. By integrating multivariate machine learning approaches and neuroimaging techniques, this research provides new insights into the brain’s structural and functional features of anxiety, advancing the potential for identifying neural biomarkers. Our findings contribute to the growing body of literature emphasizing the importance of network-based perspectives in understanding anxiety, offering a more comprehensive view of its neural architecture and manifestations.