To better understand outcomes after early brain injuries, studies must address multiple variables including age at injury, the mechanisms and severity of injury, environmental factors (before and after injury) and developmental factors. Animal models are helpful for elucidating these different aspects. First, this paper describes a new model of shaken baby syndrome (SBS) in mice, without impact or hypoxia. Mortality was 27%; 75% of survivors had focal brain lesions consisting of haemorrhagic or cystic lesions of the white matter, corpus callosum and cerebellum. All shaken animals, with and without focal lesions, showed delayed white matter atrophy. White matter damage and atrophy were reduced by pre-treatment with an NMDA receptor antagonist, indicating that excess glutamate release contributed to the pathophysiology of the lesions. Secondly, it discusses data on neuroprotection after early brain injuries; drugs targeting the NMDA receptors cannot be used in clinical practice but indirect neuroprotection strategies including anti-NO, anti-free radicals and trophic factors hold promise for limiting the excitotoxic white matter damage induced by early injury, in particular caused by shaking, during brain development. Thirdly, it describes two experimental models in which SBS outcomes are determined when the trauma is combined with environmental influences, namely medications during the acute phase, most notably anti-epileptic drugs and rearing conditions.