In ischaemic stroke the two major potential therapeutic strategies are aimed at either improving cerebral blood flow or directly interacting with the cytotoxic cascade--a large body of evidence gained from animal studies is in support of them. In clinical trials direct neuroprotection by blocking the neurotoxic cascade remained ineffective, although there are several clinical trials still in progress. We summarize the experimental data and present the results of clinical trials and also discuss why so many drugs, which were effective in animal studies, failed in human trials. It is emphasized, that 1. in most animal studies the reduction of infarct size, i.e. the amount of saved penumbral tissue, was the outcome measure, whereas neurological function remained unassessed; 2. the recovery of intellectual performance and higher cortical functions are of major importance in the future quality of life in stroke victims; however, it is impossible to examine these parameters appropriately in animal studies; 3. in many clinical trials the patient population was rather heterogenous and low in number, the study protocol was not optimal and the critical analysis of the subacute and chronic phase was lacking or insufficient. We present the major experimental stroke models, discuss their similarities, differences and limitations as compared to the human pathophysiological processes. The pitfalls of extrapolating data from animal studies to clinical practice are also summarized. The complex network of functional and morphological intercellular connections, the long timescale of neurotoxic and reparative events and the lessons learned from clinical trials suggest, that the use of drug combinations (therapeutic cocktails) targeting multiple steps of the neurotoxic cascade would hopefully result in more effective treatment of ischaemic stroke. Strategies to facilitate brain plasticity and regeneration is an additional promising tool to enhance recovery in brain ischaemia.