Recent immediate early gene evidence suggests that proximal CA3 (proxCA3, close to dentate gyrus) and distal CA1 (distCA1, close to subiculum) form a specialized non-spatial hippocampal subnetwork (nakamura et al, JON, 2013; Beer and Vavra, Plos Biology, 2018) while distal CA3 (distCA3) and proximal CA1 (proxCA1) are more specialized in spatial information processing (Flashbeck et al, 2018). However, direct in-vivo evidence for such functional networks are still missing. Here, we used chronically implanted multi-tetrode recording technique to simultaneously record along the proximodistal axis of the two CA-fields while rats performed a high-demanding delayed non-match to odor memory task. In this task, rats smelled 10 (old) odors during the study phase, and after a 20-minute delay memory for the studied odors was tested by exposing rats to the same odors intermixed with 10 new odors. We recorded 193 CA3- and 367 CA1-neurons in 5 animals who could perfom above threshold (75%). Using Support Vector Machine (SVM) we tested whether proxCA3-distCA1 neurons (non-spatial network) can differentiate the old from new odors better than distCA3-proxCA1 neurons (spatial network). We found that activity in the proxCA3-distCA1 network was relevant for the discrimination between old from new odors and similar to behavior; in contrast, the activity of the distCA3-proxCA1 network was not. Further, we found a gradient in the distribution of task-relevant neurons along the transverse axis of CA1 as well as CA3. Overall, we provide clear in vivo electrophysiological evidence that supports the role of proxCA3-distCA1 network in non-spatial memory processing.