As the other posters have explained, different neutral atoms have different lifetimes. For example, there are Helium-4 atoms (two neutrons and two protons) throughout the universe that have been around since shortly after the Big Bang--they're extremely stable. But if we built the another isotope of Helium, Helium-5 (three neutrons and two protons), it would shoot out its extra neutron within 10
-20 seconds--unlike Helium-4, it is unstable.
If we were to write down all the possible isotopes of elements in terms of their proton number Z and neutron number N, then each pair of numbers (Z,N) gives us a different isotope of some element. We could associate each pair (Z,N) with a spot in a two-dimensional grid, with Z on one axis and N on the other axis. Then, for each pair (Z,N), we could either measure the stability of the atom or try to theoretically predict it--then we could color that atom's square a dark red for very stable atoms and a blue for unstable atoms, which yields the following plot:
The bottom left stable atoms are what appear in our periodic table. The gap before the "peninsula" is the "band of instability" snorkak points out. The upper-right region of stability corresponds to elements that we have not been able to create yet but are predicted to have a relatively long lifetime, and it is called the "island of stability" because it does not "connect" to the "continent" via other stable atoms.
One example that I like to remember as a sort of opposite phenomenon is that the relatively light element Technetium (Tc, Z=43) actually lies in a "pond of instability" right among a bunch of other stable elements. It is the lightest radioactive "element" on the periodic table!