Asymptotic normality of a maximum likelihood estimator

In a regular $p$-parameter model with one-observation Fisher information $I({\theta }_0)$,