Social disconnection, covering loneliness and social isolation, might be associated with the development of paranoid thoughts. Differential effects of loneliness and social isolation on the occurrence of paranoia have not been tested so far. Moreover, the role of cognitive mechanisms in these associations remains unknown. This study aimed to investigate differential associations of loneliness and social isolation with paranoid thoughts in the general population, considering the role of cognitive mechanisms.

Altogether, 3,275 individuals, enrolled from the general population, completed baseline and follow-up assessments spanning 6–7 months. Cognitive biases (rejection sensitivity, attributional biases, and safety behaviors), social cognitive problems, and subjective cognitive problems were measured. The cross-lagged panel network (CLPN) analysis was performed, controlling for the effects of sociodemographic characteristics, psychiatric treatment, substance use, depressive, and anxiety symptoms. Additionally, mediation was tested for the CLPN paths linking social disconnection with paranoid ideation, with one intermediary node representing cognitive processes.

Loneliness was the most important node in terms of predicting other network variables. It was bidirectionally associated with paranoid thoughts. Cognitive processes mediated these associations (partial mediation for ideas of reference and full mediation for ideas of persecution). In turn, social isolation predicted paranoid thoughts through the effects on loneliness. It was also predicted by paranoid thoughts through attributional biases.

Social disconnection might be bidirectionally associated with paranoid thoughts. However, loneliness is more closely tied to paranoid thoughts compared to social isolation. Cognitive processes might mediate the association of social disconnection with paranoid thoughts.