Methods, systems, apparatuses, and non-transitory computer-readable media are provided for incremental convergence in an extended reality environment during virtual object selection. In one implementation, the computer-readable medium may contain instructions that when executed by at least one processor cause the at least one processor to display a plurality of dispersed virtual objects across a plurality of virtual regions, the plurality of virtual regions including at least a first virtual region and a second virtual region that differs from the first virtual region; receive an initial kinesics input tending toward the first virtual region; highlight a group of adjacent virtual objects in the first virtual region based on the initial kinesics input; receive a refined kinesics input tending toward a particular virtual object from among the highlighted group of adjacent virtual objects; and trigger a functionality associated with the particular virtual object based on the refined kinesics input.