I hope it's OK to post an answer to my own question since it's community-wiki. Here are a couple of things I found down this rabbit-hole.

This first paper is crap, but ranks high in entertainment value:

Kanazawa, S. “Why productivity fades with age: The crime-genius connection”, Journal of Research in Personality 37(2003), 257-272.

[http://personal.lse.ac.uk/kanazawa/pdfs/JRP2003.pdf][1]

> There is another uncanny resemblance between crime and scientific productivity. Coles (1979) study of a representative sample of contemporary mathematicians in the United States demonstrates that, while the career trajectories of a majority of mathematicians follow what I call the "age–genius curve," where their productivity, measured both by the quality and quantity of their publications, peaks very early in their careers and gradually declines thereafter, there is a small minority of mathematicians who produce a large quantity of high-quality work throughout their careers. This dichotomy of mathematicians is reminiscent of Moffitts (1993) taxonomy of "adolescence-limiteds" and "life-course persistents" among criminals.

More seriously, Dean Simonton at UC Davis has done some work claiming that there *is* a slow age-related decline in quality and quantity of creative output, but the relevant variable is *career age*, not biological age. He also makes it clear that although he believes there is a clear aggregate trend, the individual variability is much greater than the aggregate variability.

Simonton, D. K. (1997). Creative productivity: A predictive and explanatory model of career trajectories and landmarks. Psychological Review, 104, 66-89.

This paper is behind a subscription paywall, so instead I'm posting this link to the PowerPoint (sorry) of his 2005 talk at the Max Planck International Research Network on Aging:

[http://psychology.ucdavis.edu/Simonton/MxAgCrProd.ppt][2]


  [1]: http://personal.lse.ac.uk/kanazawa/pdfs/JRP2003.pdf
  [2]: http://psychology.ucdavis.edu/Simonton/MxAgCrProd.ppt