Intermediate frontal sulcusFrontalis medius

Introduction Materials & Methods

A quantitative comparison of the brain and the inner surface of the cranium

Jean Dumoncel1, Gérard Subsol2, Stanley Durrleman3, Anna Oettlé4,5, Zarina Lockhat6, Farhana E Suleman6, Edwin de Jager4, Amélie Beaudet4,7

(i) Segmentation

(iii) Sulci detection

(i) Sulci identification

(ii) Image registration

(i) Surface comparison

Conclusion and perspectives

1Laboratoire d’Anthropologie Moléculaire et Imagerie de Synthèse, UMR 5288 CNRS, Université Toulouse 3 Paul Sabatier, France; 2Laboratoire d’Informatique, de Robotique et de Microélectronique de Montpellier, UMR 5506 CNRS, Université de Montpellier II, France; 3Aramis team, INRIA Paris, Sorbonne Universités, UPMC Université Paris 06 UMR S 1127, Inserm U 1127, CNRS UMR 7225, Institut du Cerveau et de la Moelle épinière, France; 4Department of Anatomy, University of Pretoria, South Africa; 5Department of Anatomy and Histology, School of Medicine, Sefako Makgatho Health Sciences University, South Africa; 6Department of Radiology,

Faculty of Health Sciences, University of Pretoria, South Africa; 7School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, South Africa

References: Beaudet A. et al. 2016. J. Hum. Evol. 101; Beaudet A. et al. 2017. J. Anat. 232; Boule M. et Anthony R. 1917. J. Anat. 51; Durrleman S. Et al. 2015. Neur. 101; Fournier M. et al. 2011. SPIE Med. Ima. 7962; Rivière D. et al. 2009. 15th HBM; Subsol G. Et al. 2010. Am. J. Phys. Anth. 141; Yoshizawa S. Et al. 2008. Comp. Aid. Geom. 25; Zollikofer C. P. E. Et al. 2000. Firenze University Press.

Acknowledgements: Model-based deformation computation was granted access to the HPC resources of CALMIP (Grant 2018-P1440). This research has been supported the National Center for Scientific Research (CNRS), the Centre of Excellence in Palaeosciences and the Claude Leon Foundation. We are grateful to J. Braga for scientific discussion.

While our results suggest a close relationship between the sulcal pattern detected on the cerebral and endocranial surfaces, the correspondence in terms of morphology between the superior part of the brain and the endocast is more questionable. The lack of correspondance in this area might be due to potential geometrical distortions between CT scans and MRIs as well as differences in patient positioning during scans.In this context, we plan to refine the deformation process between the brain and the endocast by testing different parameters, which may improve the quality of the registration. Moreover, we will include additional specimens to test our methodological approach. By providing a quantitative comparison of the brain and the endocast, our study constitutes a usefull reference for paleoneurological studies.

MRI

CT SCAN Endocast segmentation

Sulci identification

Brain segmentation

Sulci detection

Sulci detection

Brain hull surface

Surface comparison

Sulci comparison

Sulci transfert Sulci identification

Segmentation, registration Shape analysis Sulci analysis

Occipitotemporal

Lunate sulcusParietal inferior

Paramedialis

Intraparietal sulcus

Postcentral sulcus

Central sulcus Precentral sulcus

Inferior frontal sulcus

Ascending ramusHorizontal ramus

OrbitalOlfactory sulcus

Central sulcus

Intraparietal sulcus

Lunate sulcus

Inferior frontal sulcus

Orbital

Superior temporal sulcus

Inferior temporal sulcus

Sylvian/lateral fissure

Frontomarginal sulcus

Intermediate frontal sulcusFrontalis medius

Superior frontal sulcus

Lateral/inferior occipital sulcus

Lateral calcarine sulcus

Transverse occipital sulcus

Lateral/inferior occipital sulcus

Lateral calcarine sulcus

Transverse occipital sulcus

Sylvian/lateral fissure

Superior frontal sulcus

Frontomarginal sulcus

Superior temporal sulcus

Inferior temporal sulcus

(ii) Sulci comparison

The brain (1) has been automatically extracted from MRIs with Brainvisa [Rivière 2009] and the corresponding brain hull (2) has been subsequently extracted with Endex [Subsol 2010]. The endocast (3) has been automatically extracted from the CT scans with Endex and cleaned with Avizo.

MRIs (1) and CT scans (2) have been registered with a similarity measure using a normalized mutual information metric with Avizo.

The correspondence of the brain and the endocranial shapes, as well as the coincidence of the gyral and sulcal pattern in the brain with the bulges and furrows imprinted on the braincase, have been the focus of historical debates [Boule 1917, Zollikofer 2000]. Because the brain rarely fossilizes, the endocast, i.e., a replica of the internal table of the bony braincase, is the only direct evidence of brain evolution in the human lineage. However, for now, only few publications have attempted to quantitatively compare the brain and the endocast [Fournier 2011].

1 Brain surface 2 Brain hull surface 3 Endocast surface

In this study, we provide a morphological analysis of the brain and the corresponding endocast of the same individual by using different modalities (MRI and CT scan) and a combination of “state of the art” software coupled with custom-written program in MATLAB.

1 Slice of MRI with delimitation of brain (red), brain hull (blue) and endocast (green)

1 Brain surface with sulci

2 Brain hull surface with sulci

3 Endocast surface with sulci

Brain sulci (1) have been detected with Brainvisa, and subsequently projected onto the brain hull surface (2). Endocranial sulci (3) have been detected using a C++ program [Yoshizawa 2008].

1 Result for the identification of the brain hull sulci

2 Result for the identification of the endocranial sulci

3 Displacement map (mm) of the deformation from the source to the target

Shape analysis has been performed using Deformetrica, a C++ software for the statistical analysis of 3D shape data [Durrleman et al 2015, Beaudet et al 2016, Beaudet et al 2017]. After superimposition of the two surfaces, the Brain hull surface (1) has been deformed to the endocast surface (2).The displacement map (3) points out significant morphological changes on the superior part of the brain /endocast.

Brain and endocranial sulci have been manually identified on both the brain (1) and the endocast (2) using a MATLAB program.

Deformation obtained from the surface comparison has been applied to the brain hull sulci in order to compare with the endocranial sulci (1). We computed the distance between the sulci detected on the endocast and the corresponding curves on the deformed brain hull (2). The best correspondances correspond to the lateral aspects of the brain, where the brain is relatively close in terms of distance to the inner surface of the cranium.

1 Brain hull sulci (red) after deformation compared to the endocast sulci (blue)

2 Distances (mm) from endocast sulci to the

corresponding brain hull sulci

1 Source object (brain hull surface)

2 Target object (endocast surface)

We collected the MRIs and CT scans of a total of seven individuals from the clinical record of the Steve Biko Hospital in Pretoria (South Africa). Individuals suffering from cerebral pathologies or disorders affecting substantially the inner table were systematically excluded. Accordingly, we finally selected one individual.

2 Slice of CT scan with delimitation of brain (red), brain hull (blue) and endocast (green)

Brain

Endocast