dimensionality reduction

Brightness demixing for simultaneous multi-target imaging in 3D single-molecule localization microscopy

This paper introduces Brightness Demixing, a method for distinguishing multiple fluorophores in SMLM by using brightness instead of spectral separation, enabling multi-target imaging in 2D and 3D.

Tumor extracellular vesicle RNA profiling predicts treatment response in pediatric diffuse midline glioma

This paper shows that analyzing RNA in tumor-derived extracellular vesicles from blood can predict treatment response in pediatric diffuse midline glioma.

Foveated Light-Field Compound Imager

The paper introduces FOLIC, a foveated light-field imaging system that combines compound and chambered eye principles to enable multiscale, depth-extended visualization across biomedical specimens.

Non-saturating dimensionality, contextual dependence, and the limits of unsupervised decoding in motor cortex

The paper shows that unsupervised dimensionality in motor cortex does not reflect movement complexity but instead scales with neuron count and is context-dependent, challenging assumptions about neural representation.

sEEG-Suite: An Interactive Pipeline for Semi-Automated Contact Localization and Anatomical Labeling with Brainstorm

sEEG-Suite provides a semi-automated pipeline for sEEG contact localization and anatomical labeling within Brainstorm, improving accuracy and efficiency in epileptic network mapping.

CIPHER: An end-to-end framework for designing optimized aggregated spatial transcriptomics experiments

CIPHER optimizes aggregate spatial transcriptomics experiments by jointly designing gene signatures and cell embeddings, improving decoding accuracy and experimental reliability.

Depression Severity Modulates the Cognitive Process of Facial Affect Representation: Evidence from a Genetic-Algorithm Face Synthesis Task

This study uses a genetic-algorithm face synthesis task to show that depression severity affects the cognitive process of facial affect representation, with At Risk participants showing faster convergence on embarrassment and greater instability in anger representations.

Topology-aware reconstruction of cellular state landscapes from microscopy using self-supervised learning

The paper introduces SI-SimCLR, a self-supervised learning framework for reconstructing cellular phenotypic landscapes from microscopy data without segmentation or annotation, revealing substate organization and connections in astrocytes.

Category selectivity observed in the human brain is distinct from category selectivity observed in artificial neural networks

The paper compares category selectivity in human brain regions with that in artificial neural networks, finding significant differences in response tuning and stability.

Alignment-free integration of single-nucleus ATAC-seq across species with sPYce

sPYce integrates single-nucleus ATAC-seq data across species using k-mer histograms, revealing conserved and divergent regulatory programs without requiring genome alignments.

ATI_Box: A Simple tool for convolutional neural network-based image semantic segmentation

ATI_Box simplifies CNN-based image segmentation for researchers without coding skills by integrating annotation, training, and analysis into a single platform.

Ultra-efficient High Resolution 3D Reconstruction of Spatial Omics Data with Neural Transcriptomic Field

This paper introduces NTF, a deep learning framework for high-resolution 3D reconstruction of spatial omics data, achieving fast and scalable results across multiple species and datasets.

sBOSC: A method for source-level identification of neural oscillations in electromagnetic brain signals

sBOSC identifies neural oscillations in electromagnetic brain signals by detecting oscillatory episodes at their sources with high accuracy.

Instant Prior-Free Resolution Enhancement for Cross-Modality Microscopy

This paper introduces NFR, a prior-free algorithm that enhances microscopy resolution without requiring PSF estimation, achieving results beyond the Sparrow limit and improving super-resolution techniques.

A Growth-Based Framework for Scaling Infant Anthropometry in Motion Capture Studies

This paper develops a growth-adjusted anthropometric scaling framework to improve biomechanical analysis of infants by integrating CDC growth charts with motion capture data.

Robust High-Throughput Imaging Analysis with Wasserstein Geodesic Transformations

This paper introduces Wasserstein Geodesic Transformations to improve analysis of high-throughput cell imaging data by addressing long-tailed distributions and outliers.

CycleVI: Isolating cell cycle variation with an interpretable deep generative model

CycleVI isolates cell cycle variation in scRNA-seq data using a deep generative model, improving the resolution of biological processes and drug response signals.

Learning to See Like a Child: Why Viewpoint Diversity is Fundamental for Human-Aligned Object Recognition

This paper shows that increasing viewpoint diversity in training improves generalization to novel views, aligning models more closely with human object recognition strategies.

Reorganization of Human Brain Waves Across Diverse States of Consciousness

This study identifies four wave propagation motifs in the human brain across different states of consciousness and shows how they reconfigure in diminished and psychedelic states.

PEPE: Scalable extraction of multi-modal protein language model representations

PEPE efficiently extracts multimodal protein embeddings from language models, enabling scalable and memory-efficient generation of large embedding datasets for downstream biological tasks.

Development positions malignant cellular states but does notexplain their diversification

The paper develops a framework to analyze how malignant cellular states in pancreatic ductal adenocarcinoma relate to normal developmental trajectories, finding that while developmental programs help locate cancer states, they do not explain their diversification.

The hippocampus as a small-world cognitive map

The paper uses a geometry-aware autoencoder to analyze hippocampal activity in mice, revealing a small-world network structure that enables both local fidelity and global searchability in cognitive maps.

Luminance-corrected Pupillometry for Reliable Effort-tracking in Dynamic Environments

This paper introduces a luminance-corrected pupillometry method to improve effort tracking in dynamic environments, showing it enhances discriminability and classification performance when combined with other measures.

HARP: Hierarchical Anatomical Refinement of Pathways for Whole-Brain Tractography

HARP improves diffusion MRI tractography by integrating hierarchical anatomical constraints, reducing implausible streamlines by up to 9%.

FlowTransOP: Distributional Translation of Omics Signatures via Constrained Deep Flow Matching

FlowTransOP translates omics signatures between domains without paired samples or feature mappings, outperforming existing methods when data are sparse or features are weakly correlated.

FASTIMAGES: Validating replay detection methods in human Neuroimaging using a combined MEG and fMRI dataset

The paper introduces FASTIMAGES, a benchmark dataset combining MEG and fMRI to evaluate replay detection methods in humans, showing method-specific strengths and cross-modality challenges.

bMAE: Masked Autoencoder Latent Representations for Bulk RNA-seq Tissues

The paper introduces a masked autoencoder for bulk RNA-seq data to learn latent representations that outperform existing methods in tissue discrimination and structure preservation.

SQANTI-browser: visualization and curation of SQANTI3-classified long-read transcriptomes within the UCSC Genome Browser

SQANTI-browser converts SQANTI3 classifications into UCSC Genome Browser Track Hubs for visualizing and curating long-read transcriptomes with metadata and filtering capabilities.

Superpixel-ComBat multi-site harmonization of unpaired T1W MRI data in Huntingtons disease

This paper adapts Superpixel-ComBat to harmonize unpaired multi-site T1W MRI data in HD, improving image quality and enabling unified VBM analyses across scanners.

TopOmics: Topic Modelling for All Omics

TopOmics introduces a flexible topic modeling framework for multi-omics data, demonstrating competitive performance and interpretability across diverse datasets.