This work focuses on the layout analysis of historical handwritten registers, in which local religious ceremonies were recorded. The aim of this work is to delimit each record in these registers. To this end, two approaches are proposed. Firstly, object detection networks are explored, as three state-of-the-art architectures are compared. Further experiments are then conducted on Mask R-CNN, as it yields the best performance. Secondly, we introduce and investigate Deep Syntax, a hybrid system that takes advantages of recurrent patterns to delimit each record, by combining ushaped networks and logical rules. Finally, these two approaches are evaluated on 3708 French records (16-18th centuries), as well as on the Esposalles public database, containing 253 Spanish records (17th century). While both systems perform well on homogeneous documents, we observe a significant drop in performance with Mask R-CNN on heterogeneous documents, especially when trained on a non-representative subset. By contrast, Deep Syntax relies on steady patterns, and is therefore able to process a wider range of documents with less training data. Not only Deep Syntax produces 15% more match configurations and reduces the ZoneMap surface error metric by 30% when both systems are trained on 120 images, but it also outperforms Mask R-CNN when trained on a database three times smaller. As Deep Syntax generalizes better, we believe it can be used in the context of massive document processing, as collecting and annotating a sufficiently large and representative set of training data is not always achievable.