Computer viruses spread through online social networks taking advantage of the susceptibility arising from individuals' cognitive mechanisms. Many models that simulate the spread of cyber threats are based on time-aggregated networks and memoryless users, which limit the reliability of results and forecasts. This thesis aims to integrate the temporal dynamics of social networks with the decision-making processes of individuals by implementing a cognitive model that describes users' susceptibility to phishing and its impact on the spread of cyber threats. The dynamics of the networks are modeled using an activity-driven network; each node has its own activity, which measures its propensity to create connections.