Arnaud Cachia's home page

I am interested in brain imaging in general, brain development and cortex morphometry in particular. My research has a particular focus on the use of image analysis methods for studying the brain anatomy in developmental psychiatric disorders, such as schizophrenia and autism (detailled overview (in french) ).

1. Cortical pattern variability and brain development

The most striking, interesting, yet poorly understood gross morphological features of the human cerebral cortex are the diverse and complex arrangements of cortical folds (gyri and sulci). Many methodological issues in brain imaging - as brain matching or brain morphometry -  stem from  their high inter-individual variability. Cortical folds are formed during fetal and early childhood age and their investigation could provide a new way for understanding the normal and pathological brain anatomy variability.

Some recent work have indicated that the first cortical folds that appear on the fetal brain (called the "sulcal roots") are especially stable across individual, and might indicate the existence of a stable cortical organization, underlying the apparent variable pattern.

Thanks to recent technological progresses, the quality of antenatal imaging has improved dramatically and enables to study in vivo the earliest stages of the brain morphogenesis (collaboration with Necker Hospital (Paris, France), Institute H San Raffaele (Milano, Italia), and La Timone hospital (Marseille, France). 

Central sulcus (CS) morphogenesis. 
(Left) 3D reconstruction of a fetal brain from  prenatal MRI.
At this developpement stage, the two CS sulcal roots are still visible.
(Right) During the brain maturation, the two sulcal roots merge leading to the CS.
In child brain, the presence of a burried gyrus, splitting the two roots, is still visible.
In adults, only a modification in the sulcus wall curvature gives a cue of the sulcal root limits.
Cachia et al. 2004 Mangin et al. 2004

In parallel, we try to recover a trace of these putative primal cortical folds in the adult cortex. A structural multi-scale model of the cortical surface curvature - the primal sketch of the cortex mean curvature - have been introduced to address this issue. Its main goal is to segment the sulcal roots from the depth of the adult cortical folds (collaboration with LSIS, Marseille, France).

Extraction of sulcal roots burried in the depth of adult brain
A] Extraction of the central sulcus sulcal roots (in red and yellow). 
B] Extraction of  the superior temporal sulcus sulcal roots.
Two different sulcal root organization lead to two different shapes for the temporal sulcus.
Cachia et al. 2004 Mangin et al. 2004 Coulon et al. 2000

2. Cortical fold morphometry

A standard approach in computer-based brain morphometry consists in analyzing the anatomy at the voxel level (brain tissue density, cortical ribbon thickness, white matter fiber orientation...). This approach requires a spatial normalization process to spatially match the brains to be compared. The complementary approach we develop consists in studying the brain anatomy from the analysis of the cortical folds shape and spatial organization. This approach does not use any spatial normalisation pre-processing and can then take into account the individual cortex morphology specificities.

Extraction and automatic labelisation of the main sulci and gyri. 
(From left to right) Representation of some cortical gyri and sulci with 3D brain (cortex / CSF interface) reconstruction and
3D white matter ( cortex  / white matter interface) reconstruction
Cachia et al. 2004 Mangin et al. 2004

 

Morphological variability of superior temporal, central and intraparietal sulci
Turner patients (left) and control subjects (right)
Molko et al. 2004