Absolute plate motion (APM) estimates are key to understand the driving forces of plates, particularly the role of the sublithospheric mantle flow, which has recently gained renewed recognition as a dominant driver. Tectonic plates that lack a subducting boundary (e.g., South America and Nubia) are prime examples of dynamics governed by mantle flow. Both the aforementioned plates, however, lack the hotspot space/age coverage required for high-resolution, well-constrained APM estimates. Here we resort to highly-resolved data sets of relative plate motions (RPM) across a number of spreading ridges in order to extract information on APM changes through geological time. Our analyses involve first mitigating the impact of noise in RPM data sets via Bayesian inference. This allows us to identify time periods that feature a relatively high probability of staging RPM changes. By extending these analyses to several neighboring plates, we can assess whether any of them is likely to feature an APM change through geological time. We apply such a method to RPM data set in the Atlantic realm and identify three time-intervals for changes in the APMs of the Nubia and South America plates. Our analyses are complemented by a quantitative assessment of the forces required to generate such APM changes.