The majority of trials possess at least one characteristic that may compromise their ability to guide clinical practice.

C. elegans foraging efficiently approximate maximally informative search strategies that involve abrupt switching between different types of behaviors.

Animals work in a world full of surprises, where using energy to position sensors proportional to the location's expected information avoids the pitfalls of positioning them at the information maxima.

In paediatric concussions, there are a multiplicity of ways that brain damage gives rise to symptoms, most of which are missed by conventional statistical approaches and study designs.

A nonlinear deep generative model can represent fetal–neonatal resting-state functional magnetic resonance imaging better than conventional linear models.

The Reproducibility for Everyone initiative aims to provide researchers at all career stages and across many disciplines with training in reproducible research practices via customizable workshop material.

To track odors borne on turbulent airflow, mice strategically follow their nose, with sampling movements and sensory computations reminiscent of those demonstrated in nematodes and insects.

An analysis of 67,885 preprints on bioRxiv finds evidence for disparities in international participation that are similar to the disparities found in conventional journals.

Results from the Reproducibility Project: Cancer Biology provide a picture of the rate of false positives in cancer preclinical research, but it is still not clear if research in this field is less reproducible than it should be.