Functional connectivity networks derived from resting-state functional MRI (rsfMRI) have received increasing interest to further our understanding of brain function. The anesthesia in rodent models may influence the interpretation and comparison of results from functional connectivity MRI (fcMRI). More research is required on this aspect. In this study, we investigated rat brain connectivity networks under 1.5% isoflurane anesthesia in comparison with medetomidine sedation. rsfMRI data were acquired under both anesthesia conditions within one imaging session. Male Wistar rats (n = 17) were scanned at 11.7 T with focus on the sensorimotor system. The data underwent a per-subject independent component analysis (ICA), after which individual components were grouped using hierarchical clustering. Consistent and reliable networks were identified under medetomidine in sensorimotor cortex (three networks) and striatum (two networks). The incidence of these networks was drastically reduced under isoflurane. Seed correlation analysis confirmed these results and revealed globally elevated correlations with low topical specificity under isoflurane, stemming from low-frequency global signal fluctuations. Global signal removal thus enhanced slightly regional specificity under isoflurane and showed anti-correlations of cortico-striatal connections in both anesthesia regimes. Functional connectivity networks are thus reliably detected in medetomidine-sedated animals on an individual basis using ICA. Their occurrence, however, is heavily compromised under isoflurane as a result of global signal fluctuations potentially stemming from burst-suppression-like neural activity. Anesthesia and pharmacologically induced modulations may provide insight into network mechanisms in the future. As an agent for fcMRI in brain disease studies, light sedation using medetomidine preserves connectivity networks in a greater level of detail, and may therefore be considered superior to standard isoflurane anesthesia.