Background The disruption of brain functional connections has been linked to cognitive impairment in Alzheimer’s disease. The functional architecture of the brain can be modeled using graph theory, which provides metrics to describe both local and global network properties. Autosomal dominant Alzheimer’s disease (ADAD) provides a unique opportunity to study brain changes in preclinical AD since carriers of ADAD mutations are destined to develop dementia at young age and have a well‐characterized disease trajectory. Here, we applied a graph theory approach to examine the functional network topology in Presenilin‐1 (PSEN1) E280A carriers and non‐carriers. Methods A total of 43 cognitively‐unimpaired individuals (18 PSEN1 carriers, 25 non‐carriers; mean age = 36.6±5.2) from the Colombia‐Boston biomarker study were included. Participants underwent resting‐state fMRI in a 3T scanner, and a comprehensive multi‐domain neuropsychological battery. fMRI data were preprocessed using C‐PAC. Global functional connectivity (FC) was derived from the nodal strength of the Pearson correlation matrix between time‐series of 400 ROIs Shaeffer‐2016 within 6 different functional networks Yeo‐2011. The strength of within‐ (z) and between‐network (h) functional integration was derived from a consensus modularity analysis Dwyer‐2014. Differences in cognitive variables were evaluated using Mann‐Whitney tests. Differences in FC were assessed using permutation methods: global FC using FSL‐randomize and integration parameters were based on a consensus analysis applied to random controlled matrices. Both methods were repeated 1000 times, with a p<0.05 significance threshold. Results There were no group differences in cognitive variables. Compared to non‐carriers, carriers exhibited decreased local FC in higher‐order association areas, and increased connectivity in somatosensory areas. There were no group differences in the within‐network integration. Carriers showed decreased between‐network integration, especially in higher‐order association networks, mainly the default mode network (Fig.1). Conclusions Findings support the hypothesis that alterations of functional network topology are evident in preclinical ADAD, primarily in the higher‐order cognitive networks. Early identification of abnormalities within these networks may be useful in understanding how FC changes may lead to subsequent cognitive decline in individuals at increased risk for AD. Future work with larger samples should investigate the relationships between FC changes and the neurobiological factors underlying AD.