Argumentation involves collaboration and regulation of learning. In recent years, regulation of learning, which involves self-, co-, and socially shared regulations, has attracted much attention across a range of domains. To date, studies on these forms of regulation, especially on cognitive regulation, have explored the temporal and sequential patterns of regulatory processes using various sample sizes, especially in collaborative learning, whereas few studies were found on how different argumentation scaffolds affect the structure of these three types of regulation. With the maturity of cloud computing and 5G technology, many kinds of collaborative whiteboards can be adopted to support students' argumentation. In this study, Toulmin's Argument Pattern (TAP) and Nussbaum's Argument Vee Diagram (AVD) scaffolds were integrated into an online whiteboard to assist with collaborative argumentation. A total of 193 pre-service teachers of five intact classes were recruited and evenly assigned to the control cohort, the TAP cohort and the AVD cohort. This study conducted a quasi-experiment for 6 weeks, collecting pretests on prior instructional design knowledge & prior argumentation skills and individual metacognition as covariates, screen recording of online whiteboard-based argumentation activities, and group argumentative reports. Using MANCOVA and epistemic network analysis (ENA), this study drew the following conclusions: (1) Compared to the conventional discussion, TAP and AVD scaffolds respectively activated prominently more self-regulation and co-regulation. However, in socially shared regulation, there was no significant difference between the two scaffolds and the conventional discussion; (2) The TAP and AVD scaffolds significantly enhanced students' written argument skills compared to the conventional discussion, but no significant difference existed in the TAP and AVD scaffolds for improving students' written argument skills since the AVD scaffold evoked a great deal of ineffective co-regulation; (3) In the claim phase, the TAP scaffold activated a stronger connection between co-regulation and socially shared regulation than the other two scaffolds, while the AVD scaffold had a connection between self-regulation and socially shared regulation that was stronger than the conventional discussion and as strong as the TAP scaffold. In the justification phase, the AVD scaffold showed a far stronger connection between self-regulation and co-regulation, while the TAP scaffold illustrated a stronger connection between co-regulation and socially shared regulation. In the rebuttal phase, all three scaffolds showed the strongest connection between self-regulation and co-regulation. This study has implications for effectively using online whiteboards to support productive argumentation, in turn boosting argumentation skills along with regulation characteristics of productive argumentation.