Search

This study investigates self-paced actions in freely foraging macaques. Findings highlight continuously evolving neural components that capture beliefs about latent reward dynamics, which are crucial for informed decision-making in a natural setting.

Neural Decomposition (NEURD) is a software package that decomposes neuronal data from high-resolution electron microscopy volumes into feature-rich graph representations to facilitate analysis for neuroscience research.

A foundation model trained on neural activity of visual cortex from multiple mice accurately predicts responses to video stimuli and cell types, dendritic features and connectivity within the MICrONS functional connectomics dataset.

The MICrONS mouse visual cortex dataset shows that neurons with similar response properties preferentially connect, a pattern that emerges within and across brain areas and layers, and independently emerges in artificial neural networks where these ‘like-to-like’ connections prove important for task performance.

Natural behaviors induce changes to hidden states of the world that may be vital to track. Here, in monkeys navigating virtually to hidden goals, the authors show that neural interactions in the posterior parietal cortex play a role in tracking displacement from an unobservable goal.

Dense calcium imaging combined with co-registered high-resolution electron microscopy reconstruction of the brain of the same mouse provide a functional connectomics map of tens of thousands of neurons of a region of the primary cortex and higher visual areas.

Understanding how rewards can override our initial sensory judgments and influence decision-making is crucial. Here the authors show that the anterior cingulate cortex plays an important role in mediating reward biasing.

Sensory data about most natural task-relevant variables are entangled with task-irrelevant nuisance variables. Here, the authors present a theoretical framework for quantifying how the brain uses or decodes its nonlinear information which indicates near-optimal nonlinear decoding.

The authors develop a deep learning approach that enables an efficient search of the input space to find the best stimuli for modeled neurons. When tested, these stimuli are most effective at driving their matching cells in the brain.

Correlations of noise in neural population activity are thought to limit the amount of information contained in such population activity, whereas decorrelation is suggested to increase information content. Here the authors show that decorrelation does not imply an increase in information, and only certain types of correlations limit information content.

It has been proposed that the center-surround receptive fields encountered in the early visual system serve to reduce the redundancy that is always present in natural scenes. The authors test this idea by recording from salamander retinal ganglion cells. They find strong decorrelation that is primarily a result of non-linear processing in the retina, rather than center-surround interactions. These nonlinearities serve to enhance efficient coding in the presence of noise.