Well my answer for you question comes in a few parts.
First off I'm not too sure if you understand what you mean when you say that "heads up play is completely solved from the point of view of nash equilibria". Yes you can google for a couple of charts and get a good idea of some basic heads up strategy but saying "completely solved" is a bit misleading. Being in a Nash equilibria for a two player game means that player A cannot change his strategy unilaterally, in the knowledge of players B's strategy to make more profit and vice versa. One problem here is that player A needs to know player B's strategy and vice versa, in poker this is simply not the case. If both players are playing heads up play completely according to Nash equilibria then great but that isn't really the case and if it was then one player would surely adjust to make their play more explotive.
The thing to notice about the Nash equilibria is that it give exploitative play but not optimal play. So saying that heads up play is completely solved by Nash equilibria is maybe a bit misleading (there is a Nash solution but it's not optimal).
So I would say that my first point is that some people do seem to overestimate the importance of mathematics in poker. You can't just look up a few charts and have the game solved and maybe more importantly people sometimes fail to notice when mathematics is giving an explotive solution not an optimal one.
Bringing me onto my second point which is maybe a bit more in response to some of the answers here is that people also seem to underestimate the importance of mathematics in poker. In the above answers I can see maybe see some misconceptions already (maybe not, I could just be reading them wrong). In reference to Isuldur1's hyper agressive play being "mathematically poor" and non-optimal, I would say is maybe a bit naive. Looking at that sort of strategy with some pretty basic mathematical tools we can see some advantages of it.
(from phil gordon's little green book of poker)
If you consider a heads up match with a hyper aggresive player (HAP) and player A playing regular poker, consider the following stacks are $5000, preflop pot $500:
Player A (A, k)
HAP (7,7) (Ts, 8s) ( 8s,5h) (5h,9s)
HAP moves all in. What does player A do? Well, if you are in the knowledge that you are playing a hyper aggresive player with top pair on a flop such as this you may be thinking great, easy call but lets look at a rough table:
Hand A chances A equitiy
1 (Ts, 8s) 52 5,481
2 (8s,5h) 63 6,615
3 (7,7) 1.6 168
4.(5h,9s) 76.3 8012
Player A total equitiy vs HAP: -$1448
Now that is interesting. Playing very aggressively with these hand can actually give you a positive equity against a strong made hand.
Now I am not attempting to argue that this in any way clears the matter up or that this is an in depth analysis- it clearly isn't but what it does show is a particular example of where some (very basic) math can give us a good strategy and some idea of where Isuldur1 and other players hyper aggresive strategy comes from.
Of course if you want some more in depth analysis then you need to use some more advanced mathematics (it's easy to come up with a strategy that counters the above type of play) but as you start to look in more generality the math becomes more vague. This does not make it less useful and (I would argue contrary to some of the above answers) it does not make it less applicable in higher level games it just means you need some higher level maths.
For anyone that is interested in maths and poker I would reccomend Bill Chen's Mathematics of Poker as the best book by far.
I may need to come back and edit this post as it is a bit long. Also I would also like to say that we really should get latex (mathjax or something) for this forum as it would be extremely useful! (I might bring it up on meta as a suggestion)