Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
The authors show that neural activity and synaptic plasticity in the orbitofrontal cortex mediate multiple timescales of reinforcement learning (RL) for meta-RL, which parallels a form of meta-RL in artificial intelligence.
Nelson et al. report that the APOE-R136S mutation protects against APOE4-promoted Alzheimer’s disease pathologies, including phosphorylated Tau accumulation, neuroinflammation and neurodegeneration, in mouse and human neuron models.
The authors describe the connectivity, response profile and behavioral roles of two transcriptionally defined amygdala populations from separate embryonic lineages and show how responses of one population change with social experience.
This study applies topological analysis to hippocampal ripple waveforms, uncovering a low-dimensional continuum that encodes layer-specific synaptic input information. It also reveals how ripple waveforms vary during wakefulness, sleep and learning.
Experiments in human cortical organoid and mouse models of SYNGAP1 haploinsufficiency, which is associated with autism spectrum disorder (ASD), reveal altered cortical neurogenesis, suggesting that a non-synaptic mechanism contributes to the disorder.
McGinty and Lupkin show that value-based choices in monkeys are explained by multi-neuron activity patterns in the orbitofrontal cortex (OFC) that are not evident in single cells. Identifying this neural–behavioral link sheds light on the OFC’s role in decision-making.
Radulescu et al. show that homeostatic mechanisms that reduce cortical activity following overstimulation are dysregulated later in life, such that overstimulation results in synaptic strengthening, elevated activity and cognitive impairment.
The study by Pallucchi et al. links the molecular identity of motoneuron and V2a interneuron subtypes to their function and uncovers orthogonal transcriptomic rules for their assembly into separate circuit modules controlling locomotor speed.
Chen et al. find that cerebellar Purkinje cells directly inhibit neurons in parabrachial nuclei that in turn influence many forebrain regions. This alternative output pathway could enable the cerebellum to regulate emotions, anxiety, aggression and affect.
What neurons encode when animals face a dangerous situation is unclear. Here, the authors show that the prefrontal cortex encodes both threat-specific information and a more general representation of the presence of danger.
Using direct intracranial recordings and modern speech AI models, Li and colleagues show representational and computational similarities between deep neural networks for self-supervised speech learning and the human auditory pathway.
Electrical deep brain stimulation therapy is limited by the risks of inserting electrodes into the brain. Here the authors report non-invasive deep brain stimulation in the human hippocampus using temporal interference of kHz electric fields.
This study examined the role of rat frontal and parietal cortices in choosing whether to gamble versus play it safe. A combination of perturbations, electrophysiology and quantitative modeling establishes that the frontal cortex is important for representing the expected utility of options in the service of economic choice.
The Hummel lab demonstrated that the striatum can be successfully and focally reached noninvasively via transcranial electrical temporal interference stimulation in humans, which resulted in improvements of motor learning in older adults.
The authors test artificial neural networks with stimuli whose activations are matched to those of a natural stimulus. These ‘model metamers’ are often unrecognizable to humans, demonstrating a discrepancy between human and model sensory systems.
Spatial transcriptomics reveals distinct composition and organization of cells and circuits in the mouse prefrontal cortex (PFC) relative to adjacent cortices, which concur with PFC’s diverse functions, and also help detect neurons involved in chronic pain.
We typically assume that we lose the ability to react to the outside world when sleeping. Oudiette et al. show that, in most sleep stages, humans can use their facial muscles to respond to spoken words during transient ‘connected’ periods.
In primates, activity in the visual cortex is not driven by spontaneous body movements. These results confirm the functional specialization of primate visual processing, in contrast with findings in mice, and highlight the importance of cross-species comparisons.
Neural networks must balance associative plasticity with rapid compensatory processes to maintain stable activity patterns. Andrei et al. provide in vivo evidence of a rapid homeostatic process that decreases network connectivity when excitatory neurons are synchronously activated.