1523 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.
By modeling large-scale brain-state dynamics against a cognitively normal aging reference, this study demonstrates that Alzheimer's disease and mild cognitive impairment exhibit progressively greater deviations in transition entropy, supporting a single absolute-deviation score as a compact and interpretable biomarker for distinguishing pathological neurodegeneration from healthy aging.
This study identifies the ESCRT-0 protein Hrs as a dynamic, bidirectional regulator of synaptic strength that senses neuronal activity to modulate AMPA receptor currents and Ca2+-dependent signaling, thereby influencing synaptic plasticity and preventing neurodegenerative protein accumulation.
This study demonstrates that exposing fixed pigeon brains to gadolinium-based contrast agents during the rehydration phase alone is sufficient to achieve stable relaxation parameters and reliable diffusion-weighted imaging, offering a simplified and reproducible protocol for ex vivo avian neuroimaging.
Using the complete *Drosophila* adult brain connectome, this study reveals that mixed-polarity synapses are widespread and highly structured, with their distribution predictable by neuronal complexity and playing a critical role in shaping non-canonical and reciprocal circuit motifs.
This study presents an automated framework that combines deep-learning-based segmentation of digital holographic microscopy images with graph-theoretical analysis to quantitatively characterize the organization and maturation of neuronal networks, achieving high accuracy in classifying developmental stages through label-free imaging.
SpikeCleaner is a semi-automated algorithm that standardizes the labeling of unit quality (Good, Multi-Unit Activity, or Noise) following automated spike sorting by integrating physiological and timing metrics, achieving 97% agreement with expert curators to enable scalable, high-quality curation of large-scale neuronal datasets.
This study presents a novel, automated behavioral paradigm that utilizes cell-type-specific optogenetic stimulation of TRPV1-expressing corneal nociceptors combined with DeepLabCut-based machine learning to achieve high-accuracy, observer-independent detection of blink reflexes in transgenic mice, offering a precise alternative to conventional air puff methods for studying corneal pain.
This paper demonstrates that a multi-timescale sensory model utilizing a fractional power-law adaptation mechanism () effectively stabilizes primary sensory encoding against natural environmental fluctuations by preventing refractory saturation and maintaining homeostatic firing rates, a function that can be efficiently approximated by a simple three-pole system.
This study identifies ventral pallidal GABAergic neurons as a distinct neural population that selectively drives hedonic feeding and diet-induced obesity in mice while remaining largely insensitive to homeostatic hunger signals, offering a potential therapeutic target to block overeating without affecting normal caloric intake.
ComCat is a novel neuroimaging harmonization framework that extends the ComBat method to simultaneously remove unwanted variability from both categorical site indicators and continuous nuisance variables (such as image quality and motion) using B-spline basis expansion, thereby improving prediction accuracy and preserving biologically relevant signals across diverse multi-site datasets.