I think proving something like this is unfeasible. What is shown in studies of expert performance is that there's a very strong correlation between the hours of deliberate practice and expertness in some field. That is, pretty much without any exceptions, experts in any field have been practicing for thousands of hours, the number typically quoted is in the order of 10000 hours. This doesn't prove anything particular about the necessity of talent, but it proves that at least hard work is necessary.
But, on the other hand, nothing in these studies suggest that some talent-like property is necessary.