Pillai JJ, ed. Mukherji SK, consulting ed. Neuroimaging Clinics of North America: Functional Connectivity. Elsevier; 2017;27(4):547–724; $365.00
This textbook is divided into 2 sections; the first section covers the basic principles and technical aspects of various methods of resting-state functional connectivity analysis, which would be very valuable to basic scientists/neuroradiologists working in this field. The second section covers various clinical applications of the methods described in the first, which would be of great importance to neuroradiologists. This book is a comprehensive guide for using functional magnetic resonance imaging (fMRI) in various clinical settings with all the technical caveats that one needs to be aware of when performing resting-state functional imaging (rs-fMRI).
Section 1: The first chapter describes fundamental aspects of various methods that are used for analyzing functional connectivity. It discusses several approaches to analyzing dynamic functional connectivity (eg, sliding window analysis that examines fMRI time frames of seconds to minutes; time frequency analysis; the application of a brief stimulus and use of point process analysis; and temporal clustering).
Chapter 2 gives an overview of 10 applications of independent component analysis (ICA) that have been used to look at spatial, temporal, and subject variations in fMRI data. It describes how ICA is robust in capturing spatiotemporal patterns in fMRI data.
Chapter 3 reviews rs-fMRI data analysis methods; in the temporal domain, it describes different model-driven methods, such as seed-based, multiple regression, or GLM, and data-driven methods, such as ICA, principle component analysis, and regional homogeneity. Similarly, it gives an overview of the frequency and time-frequency domains using coherence models and wavelet transform coherence methods, respectively, and how they can be applied for patients with Alzheimer disease and schizophrenia.
Chapter 4 reviews approaches to graph theoretic analysis of rs-fMRI data using the following approaches: microscale (which refers to the organization of nodes and the edges in the network), mesoscale (which analyzes the arrangement of nodes into modules), and macroscale (where networks can be with single values representing a property of the network).
Chapter 5 gives an overview of the application of machine learning for rs-fMRI analysis using support vector machines, random forests, and artificial neural networks. It also provides the advantages and disadvantages of each of the methodologies. There are in-depth discussions of the uses of machine learning methods in clinical studies, such as traumatic brain injury, epilepsy, schizophrenia, bipolar disorder, depression, social anxiety, ADHD, autism, addiction, and aging.
Section 2: Chapter 1 evaluates the clinical application of task-based (eg, finger tapping) and rs-fMRI for sensorimotor localization in presurgical planning; it shows that rs-fMRI revealed a full sensorimotor network when compared with a task-based method.
Chapter 2 describes the use of rs-fMRI in language mapping; this is very useful in instances when paradigms are not available in patients’ languages or when patients have large enough lesions that they cannot perform a language task. It is also useful in pediatric patients. It shows that there is good agreement of language region localization between resting-state and task-based fMRI.
Chapter 3 describes the limitations of resting-state functional imaging of focal brain lesions involving motor and language areas. It discusses image processing challenges in the settings of neurovascular uncoupling and susceptibility artifacts at 3T and 7T.
Chapter 4 discusses the use of rs-fMRI in neurodegenerative disease, Alzheimer disease, frontotemporal dementia, and mild cognitive impairment. It shows that rs-fMRI can be used to characterize degrees of functional connectivity and describes interrelationships among neural signals between different brain regions.
Chapter 5 describes in great detail the use of rs-fMRI in traumatic brain injury; combining functional imaging and machine-based learning shows great promise. Also, augmenting rs-fMRI with resting-state magnetoencephalography will augment fMRI findings.
Chapter 6 discusses the use of rs-fMRI in epilepsy, which is very promising in mapping language networks and characterizing the integrity of these networks.
Chapter 7 describes the use of rs-fMRI in neuropsychiatric disease, indicating that it is a very convenient method, but also mentioning that care needs to be taken in acquiring and analyzing the data, as there are sources of noise that interfere when interpreting the results.
