**Meet the editor**

Dr Peter Bright was educated at the Universities of Surrey (BSc, 1991), Reading (MSc, 1993) and Cambridge (PhD, 1999). His research in the fields of memory and conceptual knowledge are well known. He has held research positions at the MRC Cognition and Brain Sciences Unit in Cambridge (1994-1995), King's College London (1998-2001), and the University of Cambridge

(2001-2005). He currently holds the position of Reader at Anglia Ruskin University in Cambridge (since 2005).

Contents

**Preface IX** 

Chapter 1 **Functional Neuroimaging:** 

**A Historical Perspective 1**  Stefano Zago, Lorenzo Lorusso, Roberta Ferrucci and Alberto Priori

Chapter 4 **Measurement of Brain Function Using** 

Chapter 5 **Towards Model-Based Brain Imaging with Multi-Scale Modeling 99**  Lars Schwabe and Youwei Zheng

Chapter 6 **Functional Brain Imaging** 

Irene Karanasiou

Linda J. Lanyon

Chapter 8 **A Triangulation-Based MRI-Guided** 

Chapter 2 **fMRI for the Assessment of Functional Connectivity 29** 

**Principles and Neuroscientific Applications 47**  José León-Carrión and Umberto León-Domínguez

Chapter 3 **Functional Near-Infrared Spectroscopy (fNIRS):** 

**Near-Infrared Spectroscopy (NIRS) 75** 

**Using Non-Invasive Non-Ionizing Methods:** 

Chapter 7 **Diffusion Tensor Imaging: Structural Connectivity** 

**Method for TMS Coil Positioning 163**  Jamila Andoh and Jean-Luc Martinot

**Towards Multimodal and Multiscale Imaging 115** 

**Insights, Limitations and Future Directions 137** 

Till Nierhaus, Daniel Margulies, Xiangyu Long and Arno Villringer

Hitoshi Tsunashima, Kazuki Yanagisawa and Masako Iwadate

## Contents

### **Preface XI**


X Contents


## Preface

Neuroimaging methodologies continue to develop at a remarkable rate, providing ever more sophisticated techniques for investigating brain structure and function. The scope of this book is not to provide a comprehensive overview of methods and applications but to provide a "snapshot" of current approaches using well established and newly emerging techniques. Taken together, these chapters provide a broad sense of how the limits of what is achievable with neuroimaging methods are being stretched. In cognitive neuroscience research, however, it is increasingly recognised that key theoretical debates about brain function are only likely to be resolved with reference to converging evidence from a range of methods. All neuroimaging techniques have important limitations which should always be acknowledged. For example, functional magnetic resonance imaging (fMRI) is a correlational, indirect method of measuring brain activation and interpretation of signal should always reflect this fact. Spatial resolution and sensitivity is improving with the commercial availability of ultra-high field human scanners, but a single voxel (the smallest unit of measurement) still corresponds to many thousands of individual neurons. Haemodynamic response to input is slow (in the order of seconds) and the relationship between this function and neural activity remains incompletely understood. Furthermore, choice of image preprocessing parameters can appear somewhat arbitrary and an obvious rationale for selection of statistical thresholds, correction for multiple corrections, etc. at the analysis stage can likewise be lacking. Therefore, to advance our knowledge about the neural bases of cognition, rigorous methodological control, well developed theory with testable predictions, and inferences drawn on the basis of a range of methods is likely to be required.

The first chapter (Zago, Lorusso and Priori) provides an informative (and sometimes surprising) historical overview of functional neuroimaging techniques, drawing a direct line of influence from cerebral thermometry and brain "pulsation" recordings through Roy and Sherrington's late eighteenth century studies linking neural activity with energy consumption and blood flow to the development of photon emission tomography (PET), computed tomography (CT), and measurement of the bloodoxygen-level dependent effect with magnetic resonance imaging (MRI). Chapter 2 (Nierhaus, Margulies, Long and Villringer) provides a brief but excellent introduction to the BOLD effect and a more comprehensive consideration of functional connectivity

### X Preface

with specific reference to measurement of baseline or so called "resting state" networks. Principles and applications of functional near-infrared spectroscopy (fNIRS) are presented by León-Carrión and León-Domínguez (Chapter 3). These authors provide a strong case for more widespread application of fNIRS in a range of clinical populations and conditions. The size and portability of fNIRS devices provides opportunities for enhancing ecological validity of research investigations (in comparison to the restrictive conditions of fMRI), an argument also proposed by Tsunashima, Yanagisawa and Iwadate (Chapter 4). Tsunashima et al. provide a detailed consideration of NIRS signal analysis and offer a direct comparison of NIRS and fMRI data associated with systematic variations in cognitive demand.

Preface XI

connectivity under MRI. Their research provides compelling evidence for the suitability of newly synthesized imaging probes for detecting and quantifying distributed tracer uptake across the cortex. Puri and Treasaden (Chapter 10) describe limitations of using conventional 31-phosphorus neurospectroscopy for quantifying membrane phospholipid metabolism (estimated indirectly by the ratio of phosphomonoesters to phosphodiesters). Rather than focusing on signals associated with narrow resonances, the authors argue that the broad component in the 31 phosphorous neurospectroscopy signal offers a more direct measure of brain cell

Modern cranial ultrasound provides a number of advantages over neuroimaging methods such as MRI and CT, particularly with respect to costs and clinical practicalities of assessing intracranial pathology in gravely ill infants. Fickenscher, Bailey, Saettele, Dahl, and Lowe (Chapter 11) present a highly detailed overview of the approach which benefits greatly from the beautifully presented figures outlining

Level of recovery in survivors of ischemic stroke is highly variable and contingent upon success and speed with which blood flow can be restored to the affected area. Recovery can be gradual (months or even years), consistent with cerebral plasticity, i.e., the capacity for the structure and function of unaffected brain regions (e.g., the homologous contralesional site) to adapt to the behavioural or motor consequences of the stroke. Lindenberg and Seitz (Chapter 12) describe methodological issues associated with using DTI to assess post-stroke white matter integrity, and discuss evidence for structural changes possibly occurring in response to rehabilitative therapy. Their review indicates the clinical importance of DTI as a tool for predicting whether (and the extent to which) recovery is possible and the potential for understanding behavioural response to therapy in the context of underlying structural

Identification of the extent of penumbral tissue (and differentiation from irreversibly infarcted tissue) following stroke is critical for predicting optimal intervention (if any) and tissue outcome. However, while best practice guidelines are typically available, there is considerable debate in the literature concerning important factors (including time-window for successful outcome, statistical thresholds for differentiating viable from unrecoverable areas, etc.). Scalzo, Hu and Liebeskind (Chapter 13) evaluate different flow parameters from perfusion weighted images during the acute stroke phase in terms of their predictive power for estimating tissue fate four days post onset. Using nonlinear regression models, their results indicate that tissue outcome is best predicted by a regional rather than single-voxel based approach. Recovery from stroke is also considered in a targeted review by Fabien et al. (Chapter 14). Specifically, these authors address possible explanations for why clinical trials of anti-inflammatory treatments have failed despite overwhelming evidence for an inflammatory response after stroke. They argue for the use of MRI to enable selection of patients most likely to

membrane motion-restricted phospholipids.

normal and abnormal structure.

changes in the brain.

Neuroinformatics is concerned with advancing neuroscience through a process of sharing and integrating data and techniques across all levels and scales of investigation. Schwabe (Chapter 5) argues convincingly for the importance of neuroinformatics tools, able to accommodate a range of spatial and temporal scales, for the development of detailed computational models of complex cognitive functions. Karanasiou (Chapter 6) discusses strengths and limitations associated with multimodal data acquisition (including simultaneous fMRI and electroencephalography (EEG), fNIRS with EEG, magnetoencephalography (MEG) with fNIRS, and fMRI with fNIRS) and also considers the viability and potential for integrating optogenetics with fMRI (ofMRI).

Diffusion tensor imaging is an MRI technique used to characterise white matter, specifically the directionality of pathways (or "tracts"). Visualisation of these tracts therefore provides the opportunity for examining structural connectivity in vivo. Lanyon (Chapter 7) presents an excellent overview of this technique, including consideration of its application for clinical diagnostic purposes. Transcranial magnetic stimulation (TMS) offers a number of important potential advantages over the classical neuropsychological (lesion-behaviour) approach to understanding neural basis of cognition. Through the creation of a "virtual lesion" in normal participants, the potential issues of cortical reorganisation, additional pathological substrates, distal pressure effects, psychiatric factors etc. in patient populations are avoided. Additionally, there is a great advantage in being able to study the same participants in control (i.e., pre-"lesion") and experimental (post-"lesion") conditions. Clinically, repetitive transcranial magnetic stimulation (rTMS) has proved successful (although not in all cases) in the treatment of a range of neurological or psychiatric conditions through the excitation or inhibition of target areas. However, among a number of methodological and interpretative challenges, perhaps the most critical is placement precision of the TMS coil. Andoh and Martinot (Chapter 8) present a validated and freely available triangulation-based MRI-guided manual method for ensuring accurate coil placement.

Mishra et al. (Chapter 9) describe in vivo neuronal tract tracing in the rat brain using biocytin-based tracers which clearly indicated their suitability for visualising cortical connectivity under MRI. Their research provides compelling evidence for the suitability of newly synthesized imaging probes for detecting and quantifying distributed tracer uptake across the cortex. Puri and Treasaden (Chapter 10) describe limitations of using conventional 31-phosphorus neurospectroscopy for quantifying membrane phospholipid metabolism (estimated indirectly by the ratio of phosphomonoesters to phosphodiesters). Rather than focusing on signals associated with narrow resonances, the authors argue that the broad component in the 31 phosphorous neurospectroscopy signal offers a more direct measure of brain cell membrane motion-restricted phospholipids.

X Preface

with specific reference to measurement of baseline or so called "resting state" networks. Principles and applications of functional near-infrared spectroscopy (fNIRS) are presented by León-Carrión and León-Domínguez (Chapter 3). These authors provide a strong case for more widespread application of fNIRS in a range of clinical populations and conditions. The size and portability of fNIRS devices provides opportunities for enhancing ecological validity of research investigations (in comparison to the restrictive conditions of fMRI), an argument also proposed by Tsunashima, Yanagisawa and Iwadate (Chapter 4). Tsunashima et al. provide a detailed consideration of NIRS signal analysis and offer a direct comparison of NIRS

Neuroinformatics is concerned with advancing neuroscience through a process of sharing and integrating data and techniques across all levels and scales of investigation. Schwabe (Chapter 5) argues convincingly for the importance of neuroinformatics tools, able to accommodate a range of spatial and temporal scales, for the development of detailed computational models of complex cognitive functions. Karanasiou (Chapter 6) discusses strengths and limitations associated with multimodal data acquisition (including simultaneous fMRI and electroencephalography (EEG), fNIRS with EEG, magnetoencephalography (MEG) with fNIRS, and fMRI with fNIRS) and also considers the viability and potential for

Diffusion tensor imaging is an MRI technique used to characterise white matter, specifically the directionality of pathways (or "tracts"). Visualisation of these tracts therefore provides the opportunity for examining structural connectivity in vivo. Lanyon (Chapter 7) presents an excellent overview of this technique, including consideration of its application for clinical diagnostic purposes. Transcranial magnetic stimulation (TMS) offers a number of important potential advantages over the classical neuropsychological (lesion-behaviour) approach to understanding neural basis of cognition. Through the creation of a "virtual lesion" in normal participants, the potential issues of cortical reorganisation, additional pathological substrates, distal pressure effects, psychiatric factors etc. in patient populations are avoided. Additionally, there is a great advantage in being able to study the same participants in control (i.e., pre-"lesion") and experimental (post-"lesion") conditions. Clinically, repetitive transcranial magnetic stimulation (rTMS) has proved successful (although not in all cases) in the treatment of a range of neurological or psychiatric conditions through the excitation or inhibition of target areas. However, among a number of methodological and interpretative challenges, perhaps the most critical is placement precision of the TMS coil. Andoh and Martinot (Chapter 8) present a validated and freely available triangulation-based MRI-guided manual method for ensuring accurate

Mishra et al. (Chapter 9) describe in vivo neuronal tract tracing in the rat brain using biocytin-based tracers which clearly indicated their suitability for visualising cortical

and fMRI data associated with systematic variations in cognitive demand.

integrating optogenetics with fMRI (ofMRI).

coil placement.

Modern cranial ultrasound provides a number of advantages over neuroimaging methods such as MRI and CT, particularly with respect to costs and clinical practicalities of assessing intracranial pathology in gravely ill infants. Fickenscher, Bailey, Saettele, Dahl, and Lowe (Chapter 11) present a highly detailed overview of the approach which benefits greatly from the beautifully presented figures outlining normal and abnormal structure.

Level of recovery in survivors of ischemic stroke is highly variable and contingent upon success and speed with which blood flow can be restored to the affected area. Recovery can be gradual (months or even years), consistent with cerebral plasticity, i.e., the capacity for the structure and function of unaffected brain regions (e.g., the homologous contralesional site) to adapt to the behavioural or motor consequences of the stroke. Lindenberg and Seitz (Chapter 12) describe methodological issues associated with using DTI to assess post-stroke white matter integrity, and discuss evidence for structural changes possibly occurring in response to rehabilitative therapy. Their review indicates the clinical importance of DTI as a tool for predicting whether (and the extent to which) recovery is possible and the potential for understanding behavioural response to therapy in the context of underlying structural changes in the brain.

Identification of the extent of penumbral tissue (and differentiation from irreversibly infarcted tissue) following stroke is critical for predicting optimal intervention (if any) and tissue outcome. However, while best practice guidelines are typically available, there is considerable debate in the literature concerning important factors (including time-window for successful outcome, statistical thresholds for differentiating viable from unrecoverable areas, etc.). Scalzo, Hu and Liebeskind (Chapter 13) evaluate different flow parameters from perfusion weighted images during the acute stroke phase in terms of their predictive power for estimating tissue fate four days post onset. Using nonlinear regression models, their results indicate that tissue outcome is best predicted by a regional rather than single-voxel based approach. Recovery from stroke is also considered in a targeted review by Fabien et al. (Chapter 14). Specifically, these authors address possible explanations for why clinical trials of anti-inflammatory treatments have failed despite overwhelming evidence for an inflammatory response after stroke. They argue for the use of MRI to enable selection of patients most likely to benefit from anti-inflammatory treatment, to identify the therapeutic time-window, and to promote the development of interventions targeting specific pathophysiological processes. Arboix and Grive (Chapter 15) consider seven sites of intracerebral haemorrhage from the perspective of typical clinical spectrum and early outcome.

With the exception of the familial early onset form, the genetic risk factors of Alzheimer's disease remain poorly understood. Reitz (Chapter 16) discusses the potential of imaging genetics for exploring specific candidate genes on brain imaging phenotypes. Dementia with Lewy bodies (DLB), like some other forms of progressive dementing illness (e.g., semantic dementia), tends to be under-diagnosed. Baglio, Preti and Farina (Chapter 17) provide a valuable overview of DLB, with particular focus on neuroimaging findings. They convincingly argue for the employment of a range of MRI-based methods to improve differential diagnosis of DLB from other forms of dementia. In the final chapter, Jimenez-Vazquez provides an overview of endoscopic intracranial imaging for diagnosis and treatment of lesions within or adjacent to fluidfilled cavities.

I hope that, cumulatively, these chapters convey a broad sense of what can be achieved with neuroimaging methods. Rate of technological progress is encouraging increasingly sophisticated lines of enquiry in cognitive and clinical neuroscience and shows no sign of slowing down in the foreseeable future.

> **Dr. Peter Bright**  Anglia Ruskin University, Cambridge Campus, East Road, Cambridge United Kingdom

XII Preface

filled cavities.

benefit from anti-inflammatory treatment, to identify the therapeutic time-window, and to promote the development of interventions targeting specific pathophysiological processes. Arboix and Grive (Chapter 15) consider seven sites of intracerebral haemorrhage from the perspective of typical clinical spectrum and early outcome.

With the exception of the familial early onset form, the genetic risk factors of Alzheimer's disease remain poorly understood. Reitz (Chapter 16) discusses the potential of imaging genetics for exploring specific candidate genes on brain imaging phenotypes. Dementia with Lewy bodies (DLB), like some other forms of progressive dementing illness (e.g., semantic dementia), tends to be under-diagnosed. Baglio, Preti and Farina (Chapter 17) provide a valuable overview of DLB, with particular focus on neuroimaging findings. They convincingly argue for the employment of a range of MRI-based methods to improve differential diagnosis of DLB from other forms of dementia. In the final chapter, Jimenez-Vazquez provides an overview of endoscopic intracranial imaging for diagnosis and treatment of lesions within or adjacent to fluid-

I hope that, cumulatively, these chapters convey a broad sense of what can be achieved with neuroimaging methods. Rate of technological progress is encouraging increasingly sophisticated lines of enquiry in cognitive and clinical neuroscience and

**Dr. Peter Bright** 

Anglia Ruskin University, Cambridge Campus, East Road, Cambridge United Kingdom

shows no sign of slowing down in the foreseeable future.

**1** 

*Italy* 

**Functional Neuroimaging:** 

*2Unità Operativa di Neurologia, Azienda Ospedaliera* 

*Dipartimento di Neuroscienze ed organi di Senso,* 

*1Dipartimento di Neuroscienze ed Organi di Senso, Università degli Studi di Milano, U.O.C. di Neurologia, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico,* 

*3Centro Clinico per la Neurostimolazione, le Neurotecnologie e i Disordini del Movimento,* 

Modern in *vivo* functional neuroimaging techniques produce colorful computer images of the pattern of metabolic neuronal activity in living humans while engaged in performing cognitive and/or emotional tasks, and provide an unprecedented opportunity to examine how brain function supports activities in normal and abnormal conditions. The idea of the relation between blood flow and brain activity dates back to the second half of the 19th century, a time when some scientists made important contributions to this subject. In this period, functional brain activity began to be studied in the intact human brain by thermal recording from the scalp or by using techniques providing indirect graphical measurements of changes in cerebral flow when different mental tasks were performed. This relationship began to be quantified by measuring whole brain blood in animals and humans. The development of technologies that could measure these changes safely in normal human subjects would take another 60 years. Detailed mapping of regional flow changes during various mental and motor activities was achieved in 1970s and 1980s using new techniques (CT, PET and MRI) with further new valuable information about brain function in normal and in pathological conditions. This chapter will focus on a review of the historical

**1. Introduction** 

background of functional brain imaging techniques.

**2. The first evidence: Changes in scalp and cortex temperatures** 

Today, we know that brain temperature is a physiological parameter determined primarily by neural metabolism, regulated by cerebral blood flow, and affected by various environmental factors and drugs (Kiyatin, 2007). This aspect was already conjectured by some scientist in the mid-19th century. For example, the French physiologist Claude Bernard (1813-1878) in 1872 observed: '*If we now try to understand the relationship that one* 

**A Historical Perspective** 

Stefano Zago1, Lorenzo Lorusso2, Roberta Ferrucci3 and Alberto Priori3

*'M. Mellini' Chiari, Brescia,* 

*Università degli Studi di Milano,* 
