<p>Resting-state functional connectivity (RSFC) contains participant-specific patterns that can support differentiation among individuals within a cohort. This paper proposed computational approaches for determining the differential identifiability of nodes within the RSFC from the Human Connectome Project (HCP) and Leipzig Mind-Brain-Body (LEMON) datasets, and defined fingerprint-like nodes to explore individual brain functional fingerprint recognition and cognitive prediction. In addition, the predictions of fluid intelligence and working memory were explored. Experiments revealed that fingerprint-like nodes achieved higher recognition rates than whole-brain functional connectivity nodes, with individual recognition rates of at least 95.6% and 99.5% in the HCP and LEMON datasets, respectively. These nodes showed variability across brain atlases and datasets but consistent distribution within functional networks. Predictive models using these nodes also improved fluid intelligence and working memory predictions. Compared with existing methods, the proposed approaches define fingerprint-like nodes and systematically reveal their spatial distribution characteristics and contribution weights to individual functional fingerprint recognition at the node level. Although the VAN and DAN contribute less to individual identification of functional fingerprints compared to the FPN and DMN, our experiments find there are nodes within these networks that significantly impact individual identification. The fingerprint-like nodes performed well in both individual brain functional fingerprint identification and cognitive prediction. These approaches would contribute to a deeper understanding of the neural mechanisms of RSFC organization and function, as well as the mechanisms of RSFC in individual recognition and cognitive prediction.</p>