arXiv subjects

Joshua Lukemire

Publications and source records attributed to Joshua Lukemire.

8 recordsLinked to original sources

arXiv2022

BrainGB: A Benchmark for Brain Network Analysis with Graph Neural Networks

Mapping the connectome of the human brain using structural or functional connectivity has become one of the most pervasive paradigms for neuroimaging analysis. Recently, Graph Neural Networks (GNNs) motivated from geometric deep learning have attracted broad interest due to their established power for modeling complex networked data. Despite their superior performance in many fields, there has not yet been a systematic study of how to design effective GNNs for brain network analysis. To bridge this gap, we present BrainGB, a benchmark for brain network analysis with GNNs. BrainGB standardizes the process by (1) summarizing brain network construction pipelines for both functional and structural neuroimaging modalities and (2) modularizing the implementation of GNN designs. We conduct extensive experiments on datasets across cohorts and modalities and recommend a set of general recipes for effective GNN designs on brain networks. To support open and reproducible research on GNN-based brain network analysis, we host the BrainGB website at https://braingb.us with models, tutorials, examples, as well as an out-of-box Python package. We hope that this work will provide useful empirical evidence and offer insights for future research in this novel and promising direction.

q-bio.NC↗

arXiv2022

Flexible Bayesian Product Mixture Models for Vector Autoregressions

Bayesian non-parametric methods based on Dirichlet process mixtures have seen tremendous success in various domains and are appealing in being able to borrow information by clustering samples that share identical parameters. However, such methods can face hurdles in heterogeneous settings where objects are expected to cluster only along a subset of axes or where clusters of samples share only a subset of identical parameters. We overcome such limitations by developing a novel class of product of Dirichlet process location-scale mixtures that enable independent clustering at multiple scales, which result in varying levels of information sharing across samples. First, we develop the approach for independent multivariate data. Subsequently we generalize it to multivariate time-series data under the framework of multi-subject Vector Autoregressive (VAR) models that is our primary focus, which go beyond parametric single-subject VAR models. We establish posterior consistency and develop efficient posterior computation for implementation. Extensive numerical studies involving VAR models show distinct advantages over competing methods, in terms of estimation, clustering, and feature selection accuracy. Our resting state fMRI analysis from the Human Connectome Project reveals biologically interpretable connectivity differences between distinct intelligence groups, while another air pollution application illustrates the superior forecasting accuracy compared to alternate methods.

stat.ME↗

arXiv2022

FBNETGEN: Task-aware GNN-based fMRI Analysis via Functional Brain Network Generation

Functional magnetic resonance imaging (fMRI) is one of the most common imaging modalities to investigate brain functions. Recent studies in neuroscience stress the great potential of functional brain networks constructed from fMRI data for clinical predictions. Traditional functional brain networks, however, are noisy and unaware of downstream prediction tasks, while also incompatible with the deep graph neural network (GNN) models. In order to fully unleash the power of GNNs in network-based fMRI analysis, we develop FBNETGEN, a task-aware and interpretable fMRI analysis framework via deep brain network generation. In particular, we formulate (1) prominent region of interest (ROI) features extraction, (2) brain networks generation, and (3) clinical predictions with GNNs, in an end-to-end trainable model under the guidance of particular prediction tasks. Along with the process, the key novel component is the graph generator which learns to transform raw time-series features into task-oriented brain networks. Our learnable graphs also provide unique interpretations by highlighting prediction-related brain regions. Comprehensive experiments on two datasets, i.e., the recently released and currently largest publicly available fMRI dataset Adolescent Brain Cognitive Development (ABCD), and the widely-used fMRI dataset PNC, prove the superior effectiveness and interpretability of FBNETGEN. The implementation is available at https://github.com/Wayfear/FBNETGEN.

cs.LG↗

arXiv2021

Statistical Analysis of Complex Computer Models in Astronomy

We introduce statistical techniques required to handle complex computer models with potential applications to astronomy. Computer experiments play a critical role in almost all fields of scientific research and engineering. These computer experiments, or simulators, are often computationally expensive, leading to the use of emulators for rapidly approximating the outcome of the experiment. Gaussian process models, also known as Kriging, are the most common choice of emulator. While emulators offer significant improvements in computation over computer simulators, they require a selection of inputs along with the corresponding outputs of the computer experiment to function well. Thus, it is important to select inputs judiciously for the full computer simulation to construct an accurate emulator. Space-filling designs are efficient when the general response surface of the outcome is unknown, and thus they are a popular choice when selecting simulator inputs for building an emulator. In this tutorial we discuss how to construct these space filling designs, perform the subsequent fitting of the Gaussian process surrogates, and briefly indicate their potential applications to astronomy research.

astro-ph.IM↗

arXiv2020

HINT: A Hierarchical Independent Component Analysis Toolbox for Investigating Brain Functional Networks using Neuroimaging Data

Independent component analysis (ICA) is a popular tool for investigating brain organization in neuroscience research. In fMRI studies, an important goal is to study how brain networks are modulated by subjects' clinical and demographic variables. Existing ICA methods and toolboxes don't incorporate subjects' covariates effects in ICA estimation of brain networks, which potentially leads to loss in accuracy and statistical power in detecting brain network differences between subjects' groups. We introduce a Matlab toolbox, HINT (Hierarchical INdependent component analysis Toolbox), that provides a hierarchical covariate-adjusted ICA (hc-ICA) for modeling and testing covariate effects and generates model-based estimates of brain networks on both the population- and individual-level. HINT provides a user-friendly Matlab GUI that allows users to easily load images, specify covariate effects, monitor model estimation via an EM algorithm, specify hypothesis tests, and visualize results. HINT also has a command line interface which allows users to conveniently run and reproduce the analysis with a script. HINT implements a new multi-level probabilistic ICA model for group ICA. It provides a statistically principled ICA modeling framework for investigating covariate effects on brain networks. HINT can also generate and visualize model-based network estimates for user-specified subject groups, which greatly facilitates group comparisons.

stat.CO↗

arXiv2019

A-ComVar: A Flexible Extension of Common Variance Design

We consider nonregular fractions of factorial experiments for a class of linear models. These models have a common general mean and main effects, however they may have different 2-factor interactions. Here we assume for simplicity that 3-factor and higher order interactions are negligible. In the absence of a priori knowledge about which interactions are important, it is reasonable to prefer a design that results in equal variance for the estimates of all interaction effects to aid in model discrimination. Such designs are called common variance designs and can be quite challenging to identify without performing an exhaustive search of possible designs. In this work, we introduce an extension of common variance designs called approximate common variance, or A-ComVar designs. We develop a numerical approach to finding A-ComVar designs that is much more efficient than an exhaustive search. We present the types of A-ComVar designs that can be found for different number of factors, runs, and interactions. We further demonstrate the competitive performance of both common variance and A-ComVar designs using several comparisons to other popular designs in the literature.

stat.CO↗

arXiv2019

Bayesian Joint Modeling of Multiple Brain Functional Networks

Brain function is organized in coordinated modes of spatio-temporal activity (functional networks) exhibiting an intrinsic baseline structure with variations under different experimental conditions. Existing approaches for uncovering such network structures typically do not explicitly model shared and differential patterns across networks, thus potentially reducing the detection power. We develop an integrative modeling approach for jointly modeling multiple brain networks across experimental conditions. The proposed Bayesian Joint Network Learning approach develops flexible priors on the edge probabilities involving a common intrinsic baseline structure and differential effects specific to individual networks. Conditional on these edge probabilities, connection strengths are modeled under a Bayesian spike and slab prior on the off-diagonal elements of the inverse covariance matrix. The model is fit under a posterior computation scheme based on Markov chain Monte Carlo. Numerical simulations illustrate that the proposed joint modeling approach has increased power to detect true differential edges while providing adequate control on false positives and achieving greater accuracy in the estimation of edge strengths compared to existing methods. An application of the method to fMRI Stroop task data provides unique insights into brain network alterations between cognitive conditions which existing graphical modeling techniques failed to reveal.

stat.ME↗

arXiv2016

Using particle swarm optimization to search for locally $D$-optimal designs for mixed factor experiments with binary response

Identifying optimal designs for generalized linear models with a binary response can be a challenging task, especially when there are both continuous and discrete independent factors in the model. Theoretical results rarely exist for such models, and the handful that do exist come with restrictive assumptions. This paper investigates the use of particle swarm optimization (PSO) to search for locally $D$-optimal designs for generalized linear models with discrete and continuous factors and a binary outcome and demonstrates that PSO can be an effective method. We provide two real applications using PSO to identify designs for experiments with mixed factors: one to redesign an odor removal study and the second to find an optimal design for an electrostatic discharge study. In both cases we show that the $D$-efficiencies of the designs found by PSO are much better than the implemented designs. In addition, we show PSO can efficiently find $D$-optimal designs on a prototype or an irregularly shaped design space, provide insights on the existence of minimally supported optimal designs, and evaluate sensitivity of the $D$-optimal design to mis-specifications in the link function.

stat.AP↗