It is unlikely that this neutral gene when coupled with another mutation would create such a big difference as to overcome genetic drift.
I tend to agree. But apparently I did not explain myself clearly enough.
In the real environment, mutations can occur which are relatively neutral. The mutated genes can become part of the gene pool, where they are subject to thorough testing in combination with various other genes (at different loci). Thus a species can build up variability that has been well tested in the original environment.
At some later time, there might be a change in environment. What resulted from neutral mutations in the original environment might now become beneficial in the altered environment. Moreover, the gene pool might contain several well tested variants that were all near neutral in the original environment. It is possible that these, in combination, could result in a so-called IC change within the altered environment.
If we look at lab tests under radiation, at least as I understand those experiments, the resulting mutation occurs at the same time as the changed environment (i.e. the artificially applied selection), so there is no time for the thorough testing that would occur in real life situations. I would expect the lab conditions to produce evolutionary changes that are less robust than occurs in real life. And I would expect there to be less creativity (less opportunity for so-called IC changes).
Again, I'll remind you that I am not a biologist. I suppose I am a bit of a theoretician, and I am giving my theoretical analysis. But it is not informed by any detailed knowledge of the lab experiments with drosophila. I had hoped that somebody familiar with that lab work would comment.