This Perspective overviews the sources of prediction uncertainty in machine learning for applications in healthcare, and discusses how to implement suitable prediction-uncertainty metrics.

This Review highlights technological advances in exoskeleton technology from the past two decades, and challenges and opportunities in the exoskeleton-assisted rehabilitation of gait.

This Perspective discusses the applicability of integrated whole-body PET/MRI for the study of immune-mediated phenomena associated with haematopoietic activity and cardiovascular disease.

Comparisons of neural recordings across time, across subsets of neurons and across individuals requires the alignment of low-dimensional latent representations.

A subcutaneous hydrogel electrostatically entrapping live Zika virus recruited immune cells at the injection site and provided mice with effective protection against a lethal challenge.

A library of renally cleared fluorescence probes that are activated only in the presence of both tumour and leukocyte biomarkers allows for the tracking of specific populations of tumour-infiltrating leukocytes in vivo and in urine.

A Cas9-based gene therapy that replaces expanded CAG repeats in the mutant HTT allele causing Huntington’s disease by a normal CAG repeat led to substantial reductions in neurological symptoms in a pig model of the disease.

A fluorophore with peak excitation and emission wavelengths in the near-infrared window allows for the microscopic imaging of amyloid-β and tau aggregates through the intact skull in mouse models of Alzheimer’s disease.

The sensitivity of conventional lateral-flow assays can be improved by orders of magnitude by using plasmonically active fluorescent gold nanorods as antibody-conjugated labels.

The spatial distribution and placement of antigens in nanoparticle-based cancer vaccines can substantially affect antigen processing, cytokine production, the induction of immune memory and tumour growth.

New strategies for enhancing the sensitivity and resolution of contrast in magnetic resonance imaging will broaden its applicability.

The consequences of their use will be far reaching.

At first glance, some research articles published in the journal may appear to lack substantial novelty.

Graph neural networks and transformers taking advantage of contextual information and large unannotated multimodal datasets are redefining what is possible in computational medicine.

Comparisons of neural recordings across time, across subsets of neurons and across individuals requires the alignment of low-dimensional latent representations.

Everyone benefits from the filters for rigour, excellence and trust that scholarly journals should provide.