Ventilation of the deep ocean is thought to dominate glacial-interglacial (G-IG) variations of atmospheric CO₂ concentration. Much of our current understanding regarding the changes of deep ocean ventilation during the last G-IG transition comes from records of the radiocarbon content of benthic foraminifera and deep-sea corals. However, most studies focus on few records or records from a limited number of sites, and controversies between studies are not rare. A global perspective based on overall consideration of existing data can reduce the uncertainties associated with individual records. Here we present some preliminary work on reconstructing changes in basin-scale ventilation using an updated global data synthesis and inverse modeling approaches. A box model is fitted to the paleo-¹⁴C data using the sequential methods of optimal estimation theory. Our results suggest that the spatial pattern of ventilation changes may have been complex and the most significant feature is a more isolated circulation cell in the Atlantic basin during the last glacial interval than today. The inferred ventilation changes can explain 74-89% of all data, with the range reflecting the different assumptions about the errors in the air-sea ¹⁴CO₂ exchange and in the random walk used to model the temporal evolution of flow rates.