Previous research has shown that patients are reluctant to use medical artificial intelligence (AI). Cadario et al. find that this reluctance is due to people perceiving algorithms as a ‘black box’, coupled with an illusory sense of understanding medical decisions made by humans. Brief interventions that target subjective understanding of medical AI increase people’s willingness to use it.

See Cadario et al.

A preregistered systematic review and meta-analysis finds that, when self-reported media use is compared with digital logs of media use, subjective judgements are often inaccurate. This suggests caution when self-reports are used to test associations between media use and other outcomes.

See Parry et al.

Zeng and colleagues quantify the freshness of a scientific team by the absence of prior collaboration among members. Papers of fresher teams are associated with higher originality and more multidisciplinary impact. These effects are stronger in larger teams.

See Zeng et al.

How does information percolate through real- world social networks? Xie et al. directly observe phase transitions in large-scale social media data and find that the positive-feedback coevolution between users’ activity level and network structure greatly increases spreading power.

See Xie et al.

Using summary statistics from large-scale GWAS, Marees et al. examine the extent to which genetic overlap with socio-economic status (SES) influences genetic variance in and genetic overlap across 16 mental health phenotypes. The authors show that removing the variance of the latent SES factor significantly changes the pattern of genetic relationships between mental health traits.

See Marees et al.

Is the use of spice a response to local variations in the risk of foodborne infection? Bromham et al. examined the relationship between use of spice and factors such as local temperature and risk of infection across 70 cuisines. Contrary to predictions, use of spice was not strongly associated with these variables, but varied with socio-economic factors.

SeeBromham et al.

Are aesthetic preferences for visual art entirely idiosyncratic or predictable? Igaya et al. show that aesthetic preferences can be predicted by a mixture of low- and high-level image features, and that a convolutional neural network trained only on object recognition naturally encodes many of these features.

See Igaya et al. See also News & Views by Zhang and Kreiman

Can non-monetary interventions help combat climate change? Based on data from two large-scale field studies in Switzerland, Liebe et al. examine the acceptance of green energy defaults in the household and business sectors. They find large and temporally stable green default effects in both sectors, demonstrating that such defaults can be highly effective.

See Liebe et al.

See also News & Views by Sunstein

How do gendered economic roles structure space use? Wood et al. examine gender differences in Hadza hunter-gatherer spatial behaviour using 2,078 days of GPS-recorded travel. As predicted from principles of foraging ecology, Hadza men walked further per day, explored more land, followed more sinuous paths and were much more likely to travel alone.

See Wood et al.

Timbre is an auditory attribute that conveys important information about the identity of a sound source, especially for music. Thoret et al. re-analyse past research to identify the multiple acoustical facets of musical instrument perception, capitalizing on spectrotemporal modulation models and metric learning.

Controlled use of fire is one of the most important discoveries of early hominins. Agam et al. use Raman spectroscopy and machine learning to estimate the heating temperatures of flint tools fabricated by hominins over 300,000 years ago at Qesem Cave, providing insights into both advanced behaviours and the cognitive evolution of our species.

See Agam et al.

How long does the average person sleep? And are changes in sleep quality and quantity predictive of age-related cognitive decline? Two articles in this issue address these questions. Kocevska et al. conducted a meta-analysis including over 1.1 million people to produce age- and sex-specific population reference charts for sleep duration and efficiency. Djonlagic et al. identify 23 objective sleep metrics that predict cognitive performance and processing speed in older adults.

See Kocevska et al. and Djonlagic et al. .