Brain networks, including olfactory system, communicate internally and externally via neural oscillations. Specifically, two key nodes of olfactory system–the olfactory bulb and the piriform cortex –have been demonstrated in non-humans to communicate with each other via neural oscillations to form olfactory percept. Nevertheless, the communication of these two critical olfactory nodes is not well understood in humans. We used our recent developed and validated EEG-based method to assess the OB-PC communication in healthy human participants. With the help of source reconstruction method, the cross-spectrogram of the OB and PC source signals was evaluated, where we found that there is evaluation in the frequency of the OB-PC connection from early gamma to late theta/delta bands. Moreover, we further found that there is a bottom-up information flow from the OB to the PC in the beta and gamma frequency bands, while top-down information from the PC to the OB is facilitated by delta and theta oscillations. Critically, using machine learning methods, we demonstrate that there was enough information to decode odor identity above chance as early as 100ms after odor onset from the level of OB-PC connection in low gamma band. These data illuminate further the critical role of bidirectional information flow in human sensory systems to produce perception. Nevertheless, it cannot be yet fully determined what specific odor information is extracted and communicated in the information exchange and future studies are needed to address this matter adequately.