The Demerara plateau, located offshore French Guiana and Suriname, is part of a passive transform continental margin particularly prone to develop slope instabilities, probably in relation to the presence of a free distal border along its steep continental slope. Slope failure occurred at different periods (Cretaceous to Neogene) and shows an overall retrogressive evolution through time. Upslope these failure headscarp, an enigmatic regional MioPliocene unconformity has been discovered through the interpretation of new academic and industrial datasets. The aim of this work is to describe and understand the origin of this surface. Our analysis shows that this unconformity is made of a series of valleys that cross-cut sedimentary strata. Each one of these valleys has a short lateral extent and is closed along two perpendicular directions, which suggests that it could correspond to a highly meandering system, or to some sub-circular depressions. The infill of these features is equivalent to the regional stratigraphic strata found outside the structures, but in a subdued position. This seems to imply that the structures have originated by a local loss of sediments at their base or by sliding processes. Furthermore, these depressions intersect each other through time, while migrating progressively downslope. We discuss a series of hypotheses that try to explain the onset and evolution of these depressions forming the Mio-Pliocene unconformity (Canyons? Slope failures? Contourite moats? Hydrate pockmarks?). Having established that these structures are depressions formed by collapse, and have many similarities with structures recently described in the literature as pockmarks associated with gas hydrate dissolution, we favor this hypothesis. We propose that these hydrate pockmarks form with a mass failure that was triggered by fluid-overpressure development at the base of the hydrate stability zone.