Recent advances in neuroimaging technologies have generated large-scale multimodal datasets that capture complementary aspects of brain organization. These developments have fueled growing interest in identifying and quantifying associations between spatially distributed brain features across modalities, referred to as brain map correspondence. To meet this need, several permutation tests, including the spin test, Brain Surrogate Maps with Autocorrelated Spatial Heterogeneity (BrainSMASH), and the simple permutation-based intermodal correspondence (SPICE) test, have been developed to assess brain map correspondence while accounting for spatial autocorrelation. Despite their widespread use, these methods often yield inconsistent conclusions when applied to the same data, leading to long-standing confusion in the field regarding their interpretation and relationships.

We address this challenge by introducing a unified two-factor mixed-effects model that explicitly parameterizes inter-subject variability and spatial variability across brain locations. This formulation provides a principled statistical framework for decomposing sources of variation in brain maps and for formally linking each permutation test to the specific component of variability it probes. Within this framework, we derive analytical expressions for the null distributions of the spin test, BrainSMASH, and SPICE, thereby clarifying their implicit assumptions and fundamental distinctions.

Using extensive simulations and empirical analyses of both structural and functional brain maps, we demonstrate that the proposed modeling-based approach achieves well-calibrated type I error, substantially improved statistical power, and a more comprehensive and interpretable characterization of brain map correspondence.

I will review the two main kinds of electrode recordings made within the brain by neurophysiologists: spike trains from individual neurons (often modeled as point processes) and local field potentials, which represent activity from many dozens of neurons (they are time series). I will focus on the Allen Brain Observatory’s Neuropixels Dataset in the context of identifying interactions across distinct parts of the brain. My goal is to explain, briefly, the way electrodes produce spike trains and local field potentials; the information these measurements can provide; and both the opportunities and statistical challenges in identifying neural communication across the brain using this kind of relatively large dataset. I will also mention a hybrid AI-statistical method based on what we have called a GLM-Transformer model, and I will touch on the ways multiple investigators have been developing and deploying foundation models (typically based on Transformers) to learn from neural data.

Alzheimer’s disease (AD) and related dementia (ADRD) research has become increasingly big data-driven, relying on large, heterogeneous, and longitudinal datasets that span imaging, biomarkers, genetics, and clinical phenotyping. Experimental data from model systems and clinical trials now often include rich high dimensional data along with key outcomes. Although researchers have known about plaques and tangles, composed of amyloid beta and tau proteins for over a century, the underlying causes of most forms of AD/ADRD remain unknown. This lecture will discuss foundational questions and challenges in ADRD research including how data sciences/AI can help advance precision diagnostic and therapeutic strategies for ADRD and will introduce the growing ensemble of accessible data resources available for analysis in an open science framework. Major NIA-funded observational data resources such as the National Alzheimer’s Coordinating Center (NACC) and Alzheimer’s Disease Neuroimaging Initiative (ADNI) and affiliated neuroimaging, genetics, and multi-omics data repositories will be described, along with efforts at data harmonization. Future directions include increasing use of large-scale electronic health records and national databases providing a window to diagnoses, treatment and outcomes in real-world clinical healthcare settings. Finally, bottlenecks in advancing knowledge of causal mechanisms, early detection, and development of novel therapeutic strategies will be discussed.

Estimation of biomarkers related to disease classification and modeling of its progression is essential for treatment development for Alzheimer’s Disease (AD). The task is more daunting for characterizing relatively rare AD subtypes such as the early-onset AD. In this talk, I will describe the Longitudinal Early-onset Alzheimer’s Disease Study (LEADS) intending to collect and publicly distribute clinical, imaging, genetic, and other types of data from people with EOAD, as well as cognitively normal (CN) controls and people with early-onset non-amyloid positive (EOnonAD) dementias. I will discuss manifold estimation methods for estimation of surfaces of shapes in the brain using data clouds, factor-analytic methods for estimation of clinical biomarkers of AD and their use for modeling differences in longitudinal trajectories of clinical deterioration between CN, EOAD, and EOnonAD groups.

In this talk we consider the fascinating possibility of organoid intelligence (OI). Organoid intelligence using human derived pluripotent stem cells to create neural clusters with measurable neuronal activity. We discuss a recent JHU effort in OI from a team of neuroscientists, engineers, signal processors and statisticians. The goal is to use organoids to perform complex computing tasks through stimulation and response. Measurement is obtained from an electrode shell custom designed for three dimensional measurements. Apart from OI, organoids electrophysiology experiments are useful for studying human genetic disorders and toxicity through another phenotype in vitro. Here, we will focus on the statistical challenges underlying this novel form of measurement. Time permitting, we will discuss other efforts in biocomputing.

It is now widely accepted that openness and transparency are keys to improving the reproducibility of scientific research, but many challenges remain to adoption of these practices. I will discuss the growth of an ecosystem for open science within the field of neuroimaging, focusing on platforms for open data sharing and open source tools for reproducible data analysis. I will also discuss the role of the Brain Imaging Data Structure (BIDS), a community standard for data organization, in enabling this open science ecosystem, and will outline the scientific impacts of these resources.

Diffusion-weighted imaging (DWI) is a powerful neuroimaging technique that captures the microscopic movement of water molecules in brain tissue, providing unique insights into white matter organization and structural connectivity. In this talk, I will introduce statistical and computational approaches for analyzing DWI data with a focus on both microstructure and connectome-level inference. We will begin with a discussion of what DWI measures and how acquisition strategies can be optimized to improve data quality. Next, I will review statistical methods for analyzing diffusion-derived microstructure measures, such as fractional anisotropy and mean diffusivity, and extend the discussion to structural connectome construction and analysis. Finally, I will highlight exciting new directions that combine advanced statistical frameworks and AI-based methods to tackle challenges in structural connectivity research. The goal of the talk is to provide researchers with an accessible overview of DWI data characteristics, analysis strategies, and emerging opportunities, and to encourage statisticians to contribute to novel methodological developments in this important field.

While magnetic resonance imaging (MRI) studies are critical for the diagnosis, monitoring, and study of a wide variety of diseases, their use in quantitative analysis can be complex. An increasingly recognized issue involves the differences between MRI scanners that are used in large multi-center studies. To address this, the current state of the art is to “regress out” or “adjust for” scanner differences. The field has found these methods to be insufficient and has advocated for the adaptation of methods pioneered in genomics to help mitigate inter-scanner differences, which can vary across the brain and result in both mean and variance shifts. We further study the implications of differences in correlation structures across and between images, and how this affects downstream inference.

Multisite imaging studies increase statistical power and enable the generalization of research outcomes; however, due to the variety of imaging acquisition, different PET tracer properties and inter-scanner variability hinders the direct comparability of multi-scanner PET data. The PET imaging field is lacking behind in terms of harmonization methods due to the complexity associated with the combination of different tracers and different scanners. Similarly, MRI present similar challenges, but mainly due to scanner differences. In this study we investigate samples of cognitively normal participants, mild cognitive impaired and Alzheimer’s disease subjects in two major multisite studies of Alzheimer’s disease.

We present challenges and solutions associated with different MRI scanners, PET tracers, as well as analysis and harmonization techniques including simple imaging standardization, Combat and deep learning methods. We show regions of interest differences in PET outcome measures before and after the harmonization in multisite studies of Alzheimer’s Disease as well as voxel level harmonization along with summary measures before and after harmonization in MRI studies.

Modeling dependence between nodes in a brain network is central to understanding underlying neural mechanisms such as perception, action, and memory. In this talk, we present a broad range of statistical methods for characterizing dependence in a brain network. We first review a general framework which decomposes each signal into various frequency components and then characterize the dependence properties through these oscillatory activities. The unifying theme across the talk is to explore the strength of dependence and possible lead-lag dynamics through filtering. The proposed framework is capable of representing both linear and non-linear dependencies that could occur instantaneously or after some delay (lagged dependence).

In the first part of the talk, some of the most prominent frequency domain measures such as coherence, partial coherence, and dual-frequency coherence can be derived as special cases under this general framework. Coherence can be viewed in the proposed framework as the maximal squared correlation between oscillations from a pair of nodes. It gives a more specific information that correlation because it identifies the frequency bands that drive the dependence between nodes.

In this second part of the talk, we will discuss methods for characterizing dependence between a pair of communities (of nodes) under the high dimensional setting. Here, we will introduce our proposed Kendall’s Canonical Coherence (KenCoh). The classical method of Canonical Correlation Analysis (CCA) is limited to capturing linear associations, cross-sectional studies, and is sensitive to heavy-tailed observations. The proposed KenCoh method mitigates the limitations by using a rank-based approach under elliptic symmetry. It captures non-linear associations and is robust against outliers. The work presented here is limited to multiple independent trials of multivariate time-series, although this is the usual set-up of experimental studies. Future direction of this work will look into potential Markovian dependence of trials which is applicable when trials have sequencing.

In the third and final part of the talk, we will briefly describe current work on information-theoretic measures of connectivity. This class of measures are based on joint (and conditional) distributions rather than just the first two moments. The proposed spectral transfer entropy (STE) quantifies the magnitude and direction of information flow from a certain frequency-band oscillation of a channel to an oscillation of another channel. The main advantage of our proposed approach is that it allows adjustments for multiple comparisons to control family-wise error rate. Another novel contribution is a simple, yet efficient estimation method based on vine copula theory that enables estimates to capture zero (boundary point) without the need for bias adjustments. With the vine copula representation, a null copula model, which exhibits zero STE, is defined, making significance testing for STE straightforward through a standard resampling approach. Lastly, we illustrate the advantage of our proposed measure through some numerical experiments and provide interesting and novel findings on the analysis of EEG recordings linked to a visual task.

This is joint work with the Mara Talento, Sarbojit Roy and Paolo Redondo of the Biostatistics Group at KAUST.

Neuroimaging technology has played a central role in advancing scientific understanding of normal brain function in humans, as well as in investigating the neural basis of psychiatric and neurological disorders. Among the various neuroimaging techniques, functional magnetic resonance imaging (fMRI) is one of the most commonly used tools in current neuroscience studies. This talk aims to introduce analytical techniques and statistical methods for investigating brain function and connectivity using fMRI. We will begin with a discussion of the background and key characteristics of fMRI data that are relevant to statistical modeling. This will be followed by an overview of fMRI analysis, with a focus on brain network and connectome analysis using statistical and AI-based approaches. The goal of the talk is to help familiarize researchers with data considerations, analysis procedures, and challenges in fMRI studies, thereby enabling more statisticians to contribute advanced and well-suited analytical tools for the field.

The field of neuroimaging, particularly functional magnetic resonance imaging (fMRI), is rapidly growing, with over 1,000 new publications annually and widespread use across diverse academic disciplines. As fMRI becomes a core tool for psychologists, neuroscientists, physicians, economists, lawyers, and engineers, understanding its intricacies has never been more critical. However, the interdisciplinary nature of neuroimaging presents challenges due to its combination of necessary biological, computational, statistical, and psychological expertise. This talk seeks to provide an accessible overview of how fMRI works, focusing on its underlying principles, and explain the variety of techniques used to analyze fMRI data. We will focus on essential conceptual tools to critically evaluate fMRI studies, avoid common pitfalls, and deepen the understanding of this powerful imaging technique. Our aim is to empower researchers to navigate and assess the complexities of neuroimaging in a rapidly evolving field.