P-value of two-sided t-test

Let $x_1, x_2, x_3$ and $x_4$ be observed values of a random sample from an $N(\theta, \sigma^2)$ distribution, where $\theta\in\mathbb{R}$ and $\sigma > 0$ are unknown parameters. Suppose that $\bar{x}=\frac{1}{4}\sum_{i=1}^4 x_i = 3.6$ and $\frac{1}{3}\sum_{i=1}^4 (x_i - \bar{x})^2 = 20.25$. For testing the null hypothesis $H_0: \theta=0$ against $H_1: \theta\neq 0$, the $p$-value of the likelihood ratio test equals