Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain.

Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion

Variations in neural circuitry, inherited or acquired, may underlie important individual differences in thought, feeling, and action patterns. Here, we used task-free connectivity analyses to isolate and characterize two distinct networks typically coactivated during functional MRI tasks. We identified a "salience network," anchored by dorsal anterior cingulate (dACC) and orbital frontoinsular cortices with robust connectivity to subcortical and limbic structures, and an "executive-control network" that links dorsolateral frontal and parietal neocortices. These intrinsic connectivity networks showed dissociable correlations with functions measured outside the scanner. Prescan anxiety ratings correlated with intrinsic functional connectivity of the dACC node of the salience network, but with no region in the executive-control network, whereas executive task performance correlated with lateral parietal nodes of the executive-control network, but with no region in the salience network. Our findings suggest that task-free analysis of intrinsic connectivity networks may help elucidate the neural architectures that support fundamental aspects of human behavior.

https://www.jneurosci.org/content/jneuro/27/9/2349.full.pdf

Dissociable Intrinsic Connectivity Networks for Salience Processing and Executive Control

Functional MRI (fMRI) can be applied to study the functional connectivity of the human brain. It has been suggested that fluctuations in the blood oxygenation level-dependent (BOLD) signal during rest reflect the neuronal baseline activity of the brain, representing the state of the human brain in the absence of goal-directed neuronal action and external input, and that these slow fluctuations correspond to functionally relevant resting-state networks. Several studies on resting fMRI have been conducted, reporting an apparent similarity between the identified patterns. The spatial consistency of these resting patterns, however, has not yet been evaluated and quantified. In this study, we apply a data analysis approach called tensor probabilistic independent component analysis to resting-state fMRI data to find coherencies that are consistent across subjects and sessions. We characterize and quantify the consistency of these effects by using a bootstrapping approach, and we estimate the BOLD amplitude modulation as well as the voxel-wise cross-subject variation. The analysis found 10 patterns with potential functional relevance, consisting of regions known to be involved in motor function, visual processing, executive functioning, auditory processing, memory, and the so-called default-mode network, each with BOLD signal changes up to 3%. In general, areas with a high mean percentage BOLD signal are consistent and show the least variation around the mean. These findings show that the baseline activity of the brain is consistent across subjects exhibiting significant temporal dynamics, with percentage BOLD signal change comparable with the signal changes found in task-related experiments.

/pdf/consistent-resting-state-networks-across-healthy-subjects-xd3o1l0nyp.pdf

Consistent resting-state networks across healthy subjects

In this paper, we introduce two efficient methods for second and higher-order correlation, or linear and nonlinear dependence, analysis of several multidimensional variables We show that both the second and higher-order correlation analysis can be cast into a specific joint diagonalization problem Compared with existing multiset canonical correlation analysis (MCCA) and independent vector analysis (IVA) algorithms, desired features of the new methods are that they can exploit the nonwhiteness of observations, they do not assume a specific density model, and they use simultaneous separation and thus are free of error accumulation arising in deflationary separation Simulation results are presented to show the performance gain of the new methods over MCCA and IVA approaches

Second and higher-order correlation analysis of multiple multidimensional variables by joint diagonalization

This paper addresses the quantification and segmentation in brain tissue analysis by using MR brain scan. It is shown that this problem can be solved by distribution learning and relaxation labeling, an efficient method that may be particularly useful in quantifying and segmenting abnormal brain cases where the distribution of each tissue type may heavily overlap. The new technique utilizes suitable statistical models for both pixel and context images. The analysis is then formulated as an optimization problem of model-histogram fitting and global consistency labeling. The quantification is solved by a probabilistic self-organizing map, and the segmentation is performed through local Bayesian decisions. The experimental results show the efficient and robust performance of the new algorithm and that it outperforms conventional classification and Bayesian based approaches.

http://ieeexplore.ieee.org/iel2/3543/10634/00493131.pdf

MR brain image analysis by distribution learning and relaxation labeling

On joint analysis of multi-task brain imaging data sets, a variety of multivariate methods have shown their
strengths and been applied to achieve different purposes based on their respective assumptions. In this paper, we provide
a comprehensive review on optimization assumptions of six data fusion models, including 1) four blind methods: joint
independent component analysis (jICA), multimodal canonical correlation analysis (mCCA), CCA on blind source
separation (sCCA) and partial least squares (PLS); 2) two semi-blind methods: parallel ICA and coefficient-constrained
ICA (CC-ICA). We also propose a novel model for joint blind source separation (BSS) of two datasets using a
combination of sCCA and jICA, i.e., 'CCA+ICA', which, compared with other joint BSS methods, can achieve higher
decomposition accuracy as well as the correct automatic source link. Applications of the proposed model to real multitask
fMRI data are compared to joint ICA and mCCA; CCA+ICA further shows its advantages in capturing both shared
and distinct information, differentiating groups, and interpreting duration of illness in schizophrenia patients, hence
promising applicability to a wide variety of medical imaging problems.

A review of multivariate methods in brain imaging data fusion

Graph-theoretical methods are being increasingly used in areas of interest within the IEEE and beyond. Graphs are mathematical abstractions that can be used to represent networks of various types: physical (e.g., the internet or electrical networks), biological (e.g., brain networks), or social (e.g., online social networks). Furthermore, graphs can provide tools for flexible representation of data sets in which data points have irregular positions with respect to each other. Common examples of this include data sets acquired by a sensor network, where uniform sensor placement may not be possible, or machine learning data sets, where training samples are not uniformly distributed in feature space. In some instances, a graph representation arises as a natural way to describe the problem, while in other areas, e.g., image processing, they are being used to develop powerful, content-dependent alternatives to conventional processing tools.

Applications of Graph Theory

The importance of learning and adaptation in statistical signal processing creates a symbiotic relationship with machine learning. However, the two disciplines possess different momentum and emphasis, which makes it attractive to periodically review trends and new developments in their overlapping spheres of influence. Looking at the recent trends in machine learning, we see increasing interest in kernel methods, Bayesian reasoning, causality, information theoretic learning, reinforcement learning, and nonnumeric data processing, just to name a few. While some of the machine-learning community trends are clearly visible in signal processing, such as the increased popularity of the Bayesian methods and graphical models, others such as kernel approaches are still less prominent. However, kernel methods offer a number of unique advantages for signal processing, and this special issue aims to review some of those. KerneL-BASed LeArnIng: BAcKground

Tulay Adali

Papers

Second and higher-order correlation analysis of multiple multidimensional variables by joint diagonalization

MR brain image analysis by distribution learning and relaxation labeling

A review of multivariate methods in brain imaging data fusion

Applications of Graph Theory

Special Issue on Advances in Kernel-Based Learning for Signal Processing