Russell Taki Shinohara

ComBatLS improves multi-site image harmonization by preserving biological variance in features, outperforming existing methods in sex-imbalanced site scenarios and reducing sex biases.

The paper identifies robust sex differences in personalized functional brain network topography in youth, linked to X-linked genes and brain regions involved in psychiatric disorders.

This paper introduces a self-supervised deep learning framework to compute individualized multi-scale hierarchical functional networks from fMRI data and shows their association with biological phenotypes.

This paper introduces ConnReg, a regression framework for analyzing brain functional connectivity networks, and applies it to the Human Connectome Project to study how connectivity varies with covariates.

DeepCombat removes batch effects from neuroimaging data using a deep learning approach, outperforming existing methods in preserving biological variation.

This paper introduces a method to analyze functional brain networks by treating network diagnostics as functions of parameters, enabling regression and alignment with cognitive performance measures.

The paper shows thalamocortical connections develop along a sensorimotor-association gradient in youth, linking structural connectivity to cortical plasticity and environmental sensitivity.

The paper introduces NEST, a framework for testing brain-phenotype associations within functional networks, applied to neurodevelopmental imaging data.

pNet is an open-source toolbox for creating personalized functional networks from fMRI data, offering multiple interfaces and methods to enhance reliability and reproducibility.

The paper examines how functional connectivity develops along the sensorimotor-association axis during childhood to adolescence, finding that sensorimotor connectivity increases while association connectivity decreases, reinforcing cortical hierarchy.

The paper investigates how an individual's exposome influences functional brain network organization and cognitive functioning in youth, finding that exposome factors predict cognitive performance better than neuroimaging data.

XCP-D provides a standardized, reproducible post-processing pipeline for fMRI data, supporting multiple pre-processing pipelines and BIDS best practices.

The paper evaluates resampling methods for neuroimage topological features, finding that wild bootstrap and permutation perform well in larger samples, while GRF-based methods have inflated error rates.

This study identifies white matter fiber covariance networks in youth, showing their development aligns with executive function and hierarchical cortical maturation patterns.

This paper introduces a distance-based regression method for multimodal data that can handle different data types and dimensionalities, showing its effectiveness through simulations and real-world studies.

This study shows that myelination and excitation-inhibition balance influence structure-function coupling in the human cortex, with myelination becoming more important in agranular regions.

This paper introduces a framework for applying network control theory to the human structural connectome to study neural dynamics and control energy, controllability, and their correlations with behavior and mental health.

The paper investigates how functional connectivity patterns in brain networks predict symptoms of borderline personality disorder in adolescents and young adults using fMRI data.

This paper integrates multiscale cortical biology by generating spatially dense gene expression maps from postmortem data, revealing transcriptional zones linked to microstructure and neuroimaging features.

This study evaluates six compressed sensing diffusion spectrum imaging (CS-DSI) schemes, finding they provide accurate and reliable white matter structure measures with up to 80% reduced scan time compared to full DSI.

This study tracks cortical activity propagations in youth, finding increased top-down propagations with cognitive control demands and development.

The paper develops a multivariate residualization method for medical imaging data to better control for nuisance variables and applies it to ADNI data.

The paper introduces the PARE framework to remove confounders from dimension reduction methods, showing it can eliminate batch effects in single-cell data and separate clinical from technical variability in neuroimaging.

The paper uses network control theory to show that bottom-up signaling across the sensory-fugal axis of the human connectome requires less energy than top-down signaling.

The paper introduces CLEAN-R, a method to test and localize intermodal correspondence in neuroimaging data, demonstrating high statistical power and nominal type I error levels.

The paper defines personalized functional networks in children to predict cognitive differences, finding that fronto-parietal network representation correlates with general cognition.

CuBIDS streamlines reproducible curation of large BIDS datasets by validating metadata, managing variability, and preparing data for preprocessing pipelines.

The paper shows that intrinsic brain activity decreases along a sensorimotor-association cortical axis during human development, with reduced activity in disadvantaged youth.

The paper introduces CoCoA, a statistical framework to estimate conditional correlations and association sizes, demonstrating its effectiveness in neurocognitive data analysis.

Simulations and neuroimaging analyses argue that unreliable clinical and cognitive phenotypes, not only noisy fMRI, cap biomarker discovery; aggregating across raters or repeated measures beats enlarging samples.