Volume 8 Issue 4, April 2024

Much well-designed and preregistered research is conducted but never published. The reasons for these studies ending up in the ‘file drawer’ are varied. Making this research public would help us all to do better science.

Effectively engaging with large language models is becoming increasingly vital as they proliferate across research landscapes. This Comment presents a practical guide for understanding their capabilities and limitations, along with strategies for crafting well-structured queries, to extract maximum utility from these artificial intelligence tools.

Being able to deliver a persuasive and informative talk is an essential skill for academics, whether speaking to students, experts, grant funders or the public. Yet formal training on how to structure and deliver an effective talk is rare. In this Comment, we give practical tips to help academics to give great talks to a range of different audiences.

Belonging is an essential part of human identity. But with belonging comes ‘otherness’ — the tendency to label ‘others’ on the basis of gender, race, ethnicity, religion, ability or some other dimension. To advance science, we need to recognize how otherness affects research and implement interventions to overcome the biases that it creates.

Generative AI tools can quickly translate or summarize large volumes of complex information. This technology could revolutionize the way that we communicate science, but there are many reasons for caution. We asked six experts about the potential and pitfalls of generative AI for science communication.

The past 35 years have seen Bayesian models applied to many areas of the cognitive and brain sciences, which suggests that reasoning and decision-making may be rational. Wishful thinking provides a serious challenge, as it questions a core assumption of Bayesian belief updating. Melnikoff and Strohminger develop a Bayesian model that uses affective prediction errors and meets this challenge.

Fast neuromodulator release was measured in patients undergoing awake brain surgery while they played an economic game with human and computer players. The findings show that dopamine and serotonin track social context and value statistics.

Time is running out to prevent our closest living relatives—the great apes—from being driven to extinction at our own hands. Mitani et al. advocate for substantial changes to policy and research practices before it is too late.

The authors document wide variation in information density and speed of communication across the world’s languages. They find that higher-density languages communicate information more quickly but with more sustained focus than lower-density languages.

In 2021, the United States provided an unconditional child allowance to most families with children. Using anonymized mobile-location and debit/credit card data, the authors find that the benefits increased spending at childcare centres, health- and personal-care establishments, and grocery stores.

Zhao et al. find that children exhibit greater honesty after having been trusted by adults. The findings validate philosophical conjectures and offer practical strategies to foster honesty in children by nurturing adult trust.

Thomas et al. examine how dataset effects may limit the generalizability of machine learning-based models of human choice.

People often believe what they want to believe rather than what the evidence implies. Here Melnikoff and Strohminger find that this seemingly irrational tendency may emerge from fully rational Bayesian calculations.

Here Shoham and colleagues use deep learning algorithms to disentangle the contributions of visual, visual–semantic and semantic information in human face and object representations. Visual–semantic and semantic algorithms improve prediction of human representations.

In this article, Batten and colleagues measure fast neurotransmitter release in patients undergoing awake brain surgery. As volunteers play an economic game with human and computer partners, dopamine and serotonin track social context and value statistics.

Although action and motor imagery share similar population-wide neural responses in motor cortex, a subset of those responses exists in orthogonal action-unique and imagery-unique subspaces.

Brus et al. show that modulation of slow oscillatory neural activity with non-invasive electrical stimulation of the prefrontal cortex can be used to modulate top-down control and behavioural performance in non-spatial attention.

Using intracranial EEG in human participants, the authors identify a functionally distinct set of brain regions which exhibited characteristic signatures of decision formation independently of the motor action associated with the choice.

Using a unique high-quality dataset of 37,000 parent–offspring trios, the authors probe the mechanisms of the so-called indirect genetic effects on educational attainment. Surprisingly, they find that these effects cannot be explained by processes that operate exclusively within the nuclear family and instead are consistent with dynastic social effects.

A genome-wide association study of the human hypothalamus discovers 23 unique loci and examines genetic associations with neuropsychiatric behaviours and disorders.