Reproducibility of data has never been more at the forefront of scientific endeavour than now. In their latest contribution, the International Brain Laboratory (a virtual network of neuroscience researchers from 22 laboratories established in 2017) have made a significant contribution in answering how it is possible to perform repeatable experiments across different laboratories in different countries using animals from different vendors. Implanted mice were head-fixed and exposed to a decision-making task based on visual contrast using a steering wheel. When images were high contrast, the mice could rely on their vision, when contrasts were low, they had to decide based on memory of the most frequent direction. A total of 140 mice in 7 laboratories were trained in 5 million choices in a standardised protocol, identical equipment and pre-specified data analysis. Although learning curves (trials to achieve a pre-specified criterion) varied between laboratories, once training was completed, performance was indistinguishable between sites. Intriguingly, a probabilistic model further revealed the implementation of a common performance strategy in all participating laboratories. This confirms the utility of harmonisation of study factors to achieve identical research outcomes in cognitively complex tasks. The study provides evidence for means to design reproducible studies in the future and to handle within and between laboratory variability.
Leave A Comment