1795 papers
Neuroscience explores the intricate machinery of the brain and nervous system, seeking to understand how we think, feel, and move. From the microscopic dance of individual neurons to the complex networks that shape our memories and behaviors, this field peels back the layers of our biological selves to reveal the origins of consciousness and disease.
At Gist.Science, we bring these discoveries directly from bioRxiv, the leading preprint server for biological sciences, to a broader audience. We process every new neuroscience preprint as it is uploaded, transforming dense academic manuscripts into clear, plain-language explanations alongside detailed technical summaries. This ensures that both curious readers and specialists can stay current with the latest breakthroughs before they are formally published.
Below are the latest neuroscience papers we have processed from bioRxiv, offering fresh insights into the workings of the mind.
This study demonstrates that a transformer neural network can decode specific episodic memories from human intracranial neuronal spikes, revealing a dynamic shift in memory representation from the medial temporal lobe to the frontal cortex across wake-sleep cycles, which suggests new avenues for treating memory disorders via brain-computer interfaces.
By integrating whole-brain -synuclein pathology mapping with network diffusion modeling and a spatial gene expression atlas, this study identifies group II p21-activated kinases (PAKs) as key molecular regulators of regional vulnerability in Parkinson's disease and demonstrates that their inhibition effectively reduces aggregation and neuron loss both in vitro and in vivo.
This study presents the first systematic benchmark of four structural MRI foundation models, demonstrating that while performance varies across tasks and datasets, models like 3D-Neuro-SimCLR and AnatCL generally outperform traditional FreeSurfer-based baselines in predicting age, sex, and body-mass index, thereby establishing a new standard for evaluating learned representations in neuroimaging via the BrainFMBench platform.
This study demonstrates that the α2δ-2 subunit is essential for coupling presynaptic calcium entry to vesicle release in hippocampal parvalbumin-expressing interneurons, and its absence impairs synaptic inhibition, thereby contributing to the spontaneous seizures observed in knockout mice.
This paper demonstrates that recurrent excitatory-inhibitory networks with purely local plasticity, particularly BCM-like rules, can achieve predictive inference and learn sparse spatiotemporal features by maintaining weights on a consistency manifold, thereby linking predictive coding to biologically plausible circuit mechanisms without requiring explicit error representations.
This study demonstrates that excitatory theta burst stimulation of the dorsolateral prefrontal cortex causally improves motor performance in Parkinson's disease, particularly under high cognitive load, by reshaping functional connectivity within the somato-cognitive action network.
This study validates a recently proposed seven-protein blood-based Alzheimer's disease staging panel by demonstrating its constituent biomarkers consistently correlate with brain proteomic remodeling across multiple independent post-mortem cohorts, achieved through the harmonized re-analysis of legacy mass spectrometry data.
This paper proposes an Echo State Network architecture incorporating a dedicated decision-confidence readout neuron that successfully monitors classification quality in real-time, demonstrating that confidence levels correlate with decision accuracy and discriminability in EEG-based motor imagery tasks.
This study demonstrates that the human cortex actively encodes speech timing variability as meaningful information across multiple timescales, specifically representing the deviation of syllable and phoneme durations from expectations rather than treating such variability as noise to be discarded.
The paper introduces PURELIGHT, a unified hardware and software framework that eliminates photon pile-up distortions in time-correlated single-photon counting, enabling video-rate, quantitative fluorescence microscopy with accurate intensity and lifetime measurements across diverse detector technologies.