Let $X$ be a symmetric random variable, that is $X$ and $-X$ have the same distribution function $F$. Suppose that $F$ is continuous and strictly increasing in a neighborhood of $0$. Then prove that the median $m$ of $F$ is equal to $0$, where we define $m:=\inf\{x\in \mathbb{R}|F(x)\ge \frac{1}{2}\}$.
This definition of the median kind of annoys me. I could easily show that $\mathbb{P}(X\le 0)\ge \frac{1}{2}$ and $\mathbb{P}(X\ge 0)\ge \frac{1}{2}$ and by the usual definition of the median I would be done, but I don't know how to deal with that $\inf$. I could only observe that my first equality implies that $m<0$. I think that the point of the qustion is to use that $F$ is invertible on that neighborhood, but I can't make any progress.
If you know how to prove that $P(X\le 0)=1/2$, then you can prove that $0$ satisfies the "inf" definition of median by contradiction. If you assume that $\textrm{inf}\, \{x, F(x)\ge 1/2\}>0$ then, by definition of infimum, there will be some $x_0>0$ such that $F(x_0)=P(X\le x_0)<1/2$. But then $P(X\le 0)\le P(X\le x_0)<1/2$, contradicting the fact that $P(X\le 0)=1/2$. A similar contradiction arises if you assume that the infimum is negative. So it has to be zero.