See the other answers about increasing pressure until the density of the gas is equivalent to that of the object.
Additionally, why is it that objects tend to spread to the edges of the container, rather than the center?
That's just entropy.
The number of states (e.g., position) of the marble that would count as "near the edges" is vastly, vastly larger than the amount of states that we would consider "at the center". Now, divide the volume into an arbitrarily fine grid of voxels (like 3D-pixels) and label each one either "near the edge" or "near the center" by whatever definition you please. You will have hugely more voxels labeled "near the edge".
Unless you actively stabilize the marble in the center, any wayward motion of the marble will put it at a more or less random other place. If we assume that nothing influences it after getting a little nudge in a random direction with a random amount of energy, then it will bounce around like a 3D billiard ball and eventually come to rest somewhere after some of the energy is bled off through collisions with the box (large effect); and the rest by the constant bombardment of gas molecules (minuscule effect, but it is still there).
Simply counting the voxels, it is much more likely that it will come to rest in one that is labeled "near the edge" because there simply are more than those around than the few "in the center" voxels.
(N.B. to address a technicality mentioned in the comments: the above is slightly simplified; in practice you would need 3 categories ("close to the edge", "at the center", "in between"). If the box gets large, a large amount of those voxels would be "in between". Still, the amount "at the center" would be basically constant and very small, while both other categories grow. So you see why it is at least highly improbable that wherever the marble goes, it will not be at the center, or at any other specific place.)