Human Brain in Standard MNI Space: A Comprehensive Pocket Atlas

Human Brain in Standard MNI Space

A Comprehensive Pocket Atlas

STRUCTURE AND FUNCTION

Jürgen K. Mai

Institute for Anatomy, Heinrich-Heine University Düsseldorf, Düsseldorf, Germany MR-X Brain GmbH, Düsseldorf, Germany

Milan Majtanik

Institute for Informatics, Heinrich-Heine University Düsseldorf, Düsseldorf, Germany MR-X Brain GmbH, Düsseldorf, Germany

Table of Contents

Cover

Title page

Copyright

Dedication

Preface

1: Introduction

Stereotaxic Spaces

Construction of this Atlas

Human Brain in Standard MNI Space

Layout of the Atlas

Licence for the MNI images

2: Surface Views

3: Coronal Atlas

4: Axial Atlas

5: Sagittal Atlas

Brodman areas BA 1-52

List of Structures

List of Structures (Latin)

List of Abbreviations

References

MNI Licence for MRI images

Brodman areas BA 1-52

Copyright

Academic Press is an imprint of Elsevier

125 London Wall, London EC2Y 5AS, United Kingdom

525 B Street, Suite 1800, San Diego, CA 92101-4495, United States

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

Copyright © 2017 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the Publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website:www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-12-811275-5

For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mara Conner

Acquisition Editor: Mara Conner

Editorial Project Manager: Kathy Padilla

Production Project Manager: Julia Haynes

Designer: Milan Majtanik

Printed in China

Dedication

We dedicate this atlas to Gunter Corsten-Gerhards.

Preface

The past years have seen an explosive growth of imaging the human brain both for research and clinical purposes. Technological advances now show the brain tissue with excellent detail. The problem engendered with this great advancement is the incremental demand of structural interpretation of the images. We present this Atlas of the Human Brain in standard MNI Space to further structural knowledge about the brain, to support decisions made from that information and to facilitate cross-study comparisons.

This Atlas provides a consistent resource for the interpretation of MR-images that are registered in the MNI-space. This space has been evolved as the gold standard of present time imaging technology and of MRI data analysis equally for clinical and scientific purposes. The Atlas shows MNI-images at original (in vivo) size of one hemisphere in the three cardinal planes. This way the user can directly compare measurements on the MR-images on the screen with those in the Atlas. The images are presented in consistent order and are supplemented by very detailed drawings of every MNI image. The drawings and annotations are derived from the 3D brain model generated from the data shown in our Atlas of the Human Brain (Mai et al. 2016). With this comprehensive material in background it shall be possible to localize and identify any relevant anatomical structure on the MNI images.

Placement of the brain in the extensively used MNI-space also allows easy comparison with other images or reports related to this standard. It thus serves as platform allowing for comparing spatial entities on the basis of voxels avoiding the use of conflicting terms.

An important aspect in comparing different brains is the inter-individual variability. This problem is addressed by 3D warping and registration between brain scans. In our Atlas of the Human Brain we have registered an extensively analyzed individual brain hemisphere to the MNI-space providing a template for the analysis based on the principles of histological brain organization. In the Pocket Atlas we present the probabilistic neuroanatomy of 152 individuals to which the meticulously analyzed Atlas of the Human Brain is registered. This Atlas thus implements structural information at a detail that cannot be recorded from the image data alone.

The authors received invaluable assistance in delineation of the temporal lobe by Ricardo Insausti. It is a pleasure to acknowledge the professional guidance of a talented Mara Conner, Kathy Padilla and Julia Haynes of Elsevier/Academic Press who supported this edition with dedication.

1

Introduction

Stereotaxic Spaces

Stereotaxic spaces have been introduced for more reliable guidance of stereotaxic neurosurgery procedures and for precise data localization (Brett et al. 2002).

When exploring a location in space of an intracerebral structure one faces two fundamental questions: the space and the template question. The space question refers to an unequivocal description of a position in the brain. To answer this question one must define a coordinate system for the space. To define a coordinate system the origin of the space (the point with the coordinates = [0,0,0]) must be specified along with identification of three perpendicular directions in that space (left-right, anterior-posterior and ventral-dorsal). The template question relates to the interpretation of the position in the brain, i.e. what structure am I currently in? A brain template is a representation of the brain depicting anatomical information (e.g., nuclei, cortical areas). Different templates in the same space may assign different structures to the same position. The templates used in stereotaxic navigation come from various sources:

• histological preparations and drawings of individual brains: Brodmann, 1909; Schaltenbrand and Bailey, 1959; Talairach and Tournoux, 1988; Mai et al. 2016, as examples,

• single MRI brain scans: Holmes et al. 1998,

• average of many MRI scans: Evans et al., 1992; Mazziotta et al. 1995; Lee et al. 2005; Fonov et al. 2009, as examples.

The MNI standard space was introduced as a neuroimaging approximation of a widely used stereotaxic space, the Talairach space. The first MNI space consisted of large series of MRI scans on normal young adults transformed into the Talairach space.

Talairach space

Talairach and Tournoux (1988) developed their atlas from a single 60 year old female brain. Their original atlas (T&T) was a printed paper atlas developed for stereotaxic procedures (e.g. pallidotomy or thalamotomy). It has fairly good precision in the central region of the brain but the accuracy in the cortex is low (Sütfels, 2011). The T&T atlas brain does not use histological sections. The Talairach space is based on the bicommissural line (BCL). This line connects the upper border of the anterior commissure (AC) with the lower border of the posterior commissure (PC). The vertical line of the anterior commissure (VAC) orients perpendicular to the BCL from the posterior margin of the AC. Consequently, the zero point (intersection of the bicommissural line and VAC) is located at the posterior margin of AC.

Mapping to the Talairach space involves manual identification of the BCL by the named landmarks followed by the application of three piecewise linear scaling factors to fit the brain into the Talairach space (Evans, 1992b). The manual mapping of neuroimaging data to the Talairach space is time consuming and prone to errors and biases.

Average250_T1 Template

The average250_T1 template was the first attempt to approximation of an average brain shape (Evans et al., 1992a). It attempted to reduce the bias originating from the use of a single brain. In the 1980th a large library of MRI and PET scans from young, normal healthy subjects were created at the Montreal Neurological Institute (MNI). These scans were mapped to the Talairach space by manual identification of anatomical landmarks, BCL line and brain extension. The visually identified landmarks from each of the subjects were adjusted together by means of least squares linear regression that was designed to match the resulting AC-PC line to the T&T atlas. A volume based average of a subset (250 subjects) of the scans from the library was used to generate an approximation of the Talairach stereotaxic space: the average250_T1. The subcortical structures in this template are well visible due to relatively good adjustment of anatomical variability by linear registration. The high variability of the cortex, however, provokes intense blurring of cortical convolutions. Most importantly, the least squares AC-PC line identification method was lightly biased in z direction and induced a shift of the AC center in the vertical direction (Petrides, 2012). Hence the AC is not center of the coordinate system, but is located at [0,0,-2 mm]. It was T1 only (Collins et al., 1994).

MNI305_T1 Template

The next generation of the average brain approximation of the population used a subject independent automatic mapping routine that registered the individual brain scans to the stereotaxic space (Evans et al., 1992, 1993; Collins et al., 1994). The co-registration applied 9 parameter linear transformation with maximization of cross-correlation of intensities between the average250_T1 and the individual MRI. The number of subjects used for this average increased by 55 to 305 subjects (239 males and 66 females, age: 23.4 +/- 4.1 years). The automatic mapping procedure and the larger number of subjects resulted in improved tissue contrast but also lacked anatomical detail at the cortex. Due to the AC-PC line unconstrained mapping procedure the displacement of the AC center is also present in this average. The location of the AC is at [0,1,-3.5 mm]. The MNI305_T1 template was the first public MNI template available and it was T1 only.

Colin27_T1 Template

To achieve improved anatomical detail of the cortex in the average250 and the MNI305 spaces a high definition single subject template was created (Holmes et al., 1998; Collins et al., 1998). A lab member of the MNI, was scanned 27 times and the scans were averaged in stereotaxic space. The resulting template was registered to the MNI305_T1. It showed high anatomical tissue definition for both the cortex and the deep brain structures. This template was used as the standard template in SPM96 and many other software packages. In the original release, only a T1 template was available. The AC centre displacement in the Colin27_T1 is [0,3.2,-5.5 mm] in the stereotaxic space. In 2008 the MNI also released a high-resolution (0.5×0.5×0.5 mm) version of the Colin27 T1, T2 and PD templates (Aubert-Broche et al., 2006).

ICBM/MNI152_T1, T2 and PD templates

In 2001 the International Consortium for Brain Mapping (ICBM) created a database of high resolution full-brain MRI scans (T1, T2 and PD) to build normative anatomical models of the human brain. The T1, T2 and PD MRI scans of 152 subjects acquired at the MNI were used to create the MNI152_T1, MNI152_T2 and MNI152_PD linear templates (Mazziotta et al., 1995; Evans et al., 2001; Mazziotta et al., 2001). The scans of each individual were linearly co-registered (9-parameters) to the MNI305_T1 space and averaged voxel by voxel. While these templates had better contrast and better tissue definition at the top of the brain and at the bottom of the cerebellum, the sulci and gyri remained blurred. These templates again continued to have the AC shift as compared to the original Talairach space centre located. The AC has the location [0,2.0,-3.5 mm].

Figure 1.1 Examples of the MNI Brain templates.

To improve the definition of sulci and gyri further refinement steps were realized in 2002 by using non-linear registration techniques to align cortical structures between subjects (Grabner et al. 2006). Each of the 152 individual scans was iteratively mapped to their own evolving average by non-linear registration. This technique allowed for significant reduction of anatomic variance within the group average. The nonlinear MNI152 templates have much higher cortical definition than the linear MNI templates.

Recent MNI/ICBM2009 templates

In 2009, the MNI released an updated version of the average of the 152 acquisitions. The new MNI/ICBM2009 spaces are a set of unbiased non-linear averages of the MNI152 database with both high-spatial resolution and high signal-to-noise ratio. These templates have the highest resolution and anatomical definition to date and is not biased by the deviations of any single brain (Fonov et al. 2009; Fonov et al., 2011). Furthermore, a specific age-based subpopulation of subjects became created. The scans within each group were automatically registered to the stereotaxic space target template with an iterative nonlinear co-registration algorithm applied to obtain the group averages (Fonov et al., 2011). The anatomy in these templates is slightly different in direct comparison to the 6th generation templates (MNI152 nonlinear). The ICBM/MNI2009 space comes in three versions that each has a symmetric and asymmetric subversion. The ICBM/MNI2009a template consists of T1, T2, PD and T2-relaxometry versions and has 1×1×1 mm resolution. Both, a symmetric and an asymmetric version is available. The asymmetric versions does not feature T2-relaxometry. The ICBM/MNI2009b is the only version of the novel templates available in 0.5×0.5×0.5 mm resolution. It comes in a T1, T2 and PD version. The ICBM/MNI2009c is a 1×1x1mm template which includes T1, T2, PD modalities, and tissue probability maps. Intensity inhomogeneity was performed using N3 version 1.11 and the sampling is different from 2009a template. The version used in this Atlas is the ICBM/MNI2009b symmetric template.

Space in neuroscience

The history of the MNI templates indicates that there is no single MNI space that can be directly used for the interpretation of imaging results. Rather the concept of the MNI space refers to a position in the brain, expressed in millimeters, from a selected origin. Depending on which MNI template is used, a selected position may point to different anatomical structures. For example, small subcortical structures such as the anteroventral thalamic nucleus may show a significant displacements depending