Functional MRI often requires general anesthesia for noncompliant individuals. However, it remains unclear how anesthesia affects brain network organization and whether individuals’ idiosyncratic functional fingerprints remain identifiable. Here, we performed individualized analyses on prolonged fMRI data (96 minutes) collected from 14 participants in both awake and anesthetized states. We found that functional connectivity in the primary sensorimotor, visual and auditory cortices was particularly susceptible to anesthesia. Within- and between-network connectivity was generally reduced under anesthesia, indicating weakened integration. Individual differences in connectivity profile were diminished, however, subject-specific functional fingerprints were well-preserved and highly identifiable, especially in the association cortices. A model was developed to reconstruct individuals’ awake connectomes using anesthesia data. The model's ability to identify autism-related dysfunctions in social network was demonstrated using data from 29 anesthetized, autistic children. This work revealed anesthetic effects on functional connectivity, and established a reconstructive model that may bridge conscious and unconscious states.