Loneliness is a complex and subjective experience deeply intertwined with one’s internal perceptions, interpersonal relationships, and neural characteristics. This dissertation investigated these dimensions through three interconnected studies, with a focus on younger adults. Chapter 2 examined how self-perception, reciprocal social connectedness, and social network structures collectively explain loneliness. Utilizing structural equation modeling with round-robin data from campus social groups, we found that higher self-liking and larger egocentric networks were associated with lower loneliness, above and beyond the effects of peer connections. Chapter 3 applied a machine-learning framework to predict loneliness levels from whole-brain functional connectivity networks, measured while participants viewed their peers’ faces. The results showed that functional connectivity patterns throughout the whole brain predicted loneliness, with greater connectivity between unimodal and transmodal neurocognitive systems associated with higher loneliness. In contrast, increased connectivity within unimodal sensory systems and between the default mode system and other transmodal systems predicted lower loneliness. Chapter 4 extended this analysis longitudinally, examining whether pre-pandemic measurements of functional brain networks could predict loneliness levels during the pandemic. This study confirmed the predictive power of functional brain networks, demonstrating that the consistent patterns of connectivity between unimodal and transmodal neurocognitive systems, as observed in Chapter 3, also predicted different levels of loneliness during enforced social isolation. Synthesizing findings from all chapters, the analysis highlighted that while psychological and social factors contribute to loneliness, neural markers provide a more direct pathway to understanding its foundations. Particularly, enhanced task-evoked connectivity between the frontoparietal control and visual systems, as well as between the ventral attention and somatomotor systems, was consistently predictive of loneliness levels over time, surpassing the statistical significance of self-reported psychosocial variables. Overall, this dissertation underscores the critical role of neurobiological factors in loneliness and suggests the potential for developing targeted interventions that modulate specific markers within brain networks. These findings advocate for a multidisciplinary approach to tackling loneliness, combining insights from psychology, social network theory, and neurobiology to devise effective strategies for prevention and intervention.