One factor may well be that when you start studying a subject a large part of what contributes to new developments may well be that you bring a new angle, and
lack of understanding of what is accepted.
E.g. I remember my first CS classes, and how I during my algorithm classes "invented" thing after thing just to read on to the next chapter and see my "inventions" covered. But now and again someone will come up with a genuine invention to a problem that is novel and important in part because they've not yet come to learn what the "right solution" to that problem is.
As you learn more, you tend to focus on the perceived narrow gaps of uncharted territories or at adding more at the peaks, where making progress is a massive slog.
From a different perspective, having done multiple startups, I think one of the advantages I had in terms of willingness to jump into something new when I was younger, was lack of knowledge.
I went head first into starting and ISP without knowing either the business side, or the technical side of it - I had enough knowledge to know where to start figuring it out, but not enough to know why it would be hard. And so I did it. We weren't tremendously successful, but we learned a hell of a lot, and in the end it kickstarted my career.
Today, it's far easier for me to look at an idea and think up a hundred reasons why it will probably fail because I know more. While that probably saves me a lot of misery, I've also been through enough startups to know that a lot of the objections people raise that looks fatal can be easily overcome; it's very hard to know which flaws will actually be a problem and which will not, and some even turns out to be advantages (e.g. you find a way past them but your competitors don't). In a sense it is frustrating, and I sometimes wish I could look at an idea and ignore the obvious flaws the way I would have at 20...
Another advantage was being less risk averse even when I did see the risks and flaws. I had far less to lose:
No career to ruin; no house to lose; no child I was responsible for, more time to do something else if I got nowhere.
So when we then look at breakthroughs, to determine if it is down to cognitive differences, we also need to consider how large a proportion at different ages are willing to take those risks and put in that level of effort.
My first startup was planned in a couple of weeks; we got seed funding from some random person one of us met in a bar (...); the following week I moved out from parents into a room in our new office.
I'd never be willing to move that fast today, nor live in the office, because my priorities are different. And while I strongly believe I am more likely to be successful at the things I try today, I am also a lot more cautious about the attempts I make. In the end I'm probably less likely to end up with a stratospheric success due to less willingness to take huge risks and fewer attempts, but I'm also less likely (I think, at least) to end up crashing and burning. Startups are obviously different from e.g. success in science, but I'd suggest similar mechanisms are at play - you build a reputation that depends on certain things that are easier if you stay "safe" in a niche you know you can keep publishing in than if you go for something of totally uncertain payoff; you get comfortable in a working pattern that fits your life etc.
Of course that does not rule out cognitive decline as part of the explanation as well.