OBJECTIVES: Despite decades of research using animals to develop pharmaceutical treatments for patients who have had a stroke, few therapeutic options exist. The vast majority of interventions successful in preclinical animal studies have turned out to have no efficacy in humans or to be harmful to humans. In view of this, we explore whether there is evidence of a move away from animal models in this field. METHODS: We used an innovative methodology, the analysis of opinion papers. Although we took a systematic approach to literature searching and data extraction, this is not a systematic review because the study involves the synthesis of opinions, not research evidence. Data were extracted from retrieved papers in chronological order and analysed qualitatively and descriptively. RESULTS: Eighty eligible papers, published between 1979 and 2018, were identified. Most authors were from academic departments of neurology, neuroscience or stroke research. Authors agreed that translational stroke research was in crisis. They held diverse views about the causes of this crisis, most of which did not fundamentally challenge the use of animal models. Some, however, attributed the translational crisis to animal-human species differences and one to a lack of human in vitro models. Most of the proposed solutions involved fine-tuning animal models, but authors disagreed about whether such modifications would improve translation. A minority suggested using human in vitro methods alongside animal models. One proposed focusing only on human in vitro methods. CONCLUSION: Despite recognising that animal models have been unsuccessful in the field of stroke, most researchers exhibited a strong resistance to relinquishing them. Nevertheless, there is an emerging challenge to the use of animal models, in the form of human-focused in vitro approaches. For the sake of stroke patients there is an urgent need to revitalise translational stroke research and explore the evidence for these new approaches.