[Ieee_vis_open_positions] PhD position at Inria Saclay, France: Visualization of structural and functional connectivity in the brain

Tobias Isenberg tobias.isenberg at inria.fr
Fri Jul 4 14:55:34 CEST 2014

*PhD position at Inria Saclay, France: Visualization of structural and
functional connectivity in the brain*

PDF version:

application deadline: September 15, 2014

starting date: end of 2014 at the latest

The study of brain connectivity is one of the fundamental ways to
investigate the complex functions of the (human) brain. For this
purpose, neuroanatomists capture and investigate two different types of
connectivity: anatomical connectivity arising from diffusion-weighted
MRI measurements and functional connectivity based on fMRI scans. Both
types of data have advantages and disadvantages, but ultimately it is
essential to study them in concert.

The research project:
The research in this PhD project will employ and combine approaches from
two sub-fields of visualization: the visualization of spatial
relationships (SciVis, for anatomical connectivity) and the
visualization of abstract data (InfoVis, for functional connectivity
data). The goal is to be able to start the interactive investigation
with either type of data, being able to interactively and freely switch
between the different representations as it is needed for the data
exploration. The ultimate vision is two-fold: The first and foremost
aspect is to get to a more in-depth understanding on how to support
interactive data exploration using various new and state-of-the-art
visualization techniques in the neurosciences. The second aspect is that
to more generally push the boundaries of multimodal visualization to be
able to generalize the findings of this research to other fields that
work on a daily basis with data that has both spatial and abstract

For this purpose the project extends past results in the visualization
of dense line data as well as the visualization of weighted graphs. For
the first aspect of anatomical connectivity, the project will use
methods from illustrative visualization to deal with the dense
fibertract datasets that are generated from diffusion-weighted MRI. This
aspect of the visualization will provide an important visual reference
and landmark for the exploration of functional connectivity, for which
the project will rely on the visualization of weighted graphs. A general
challenge in this context is the question on how to create a
visualization that combines both data types, either in separate views or
in a combined view. Separate linked views are common in abstract data
visualization and we will thus explore their application for our
application. A view that integrates both could combine fibertracts
inside the brain for anatomical connectivity with a bundled view of
functional links on its outside. This approach has the potential benefit
of not requiring a mental integration of separate points of reference.
On the other hand, this approach may lead to a cluttered and overloaded
depiction. The project will therefore explore new ways of controlling
the abstraction in the data depiction to deal with this issue to be able
to show the realistic large and complex datasets. This work will thus
also require research to understand how to apply illustrative
visualization to abstract data.

Such visualizations of the fibertracts in a more or less realistic way
is convenient for the neuroscientist, but we have to consider other
complementary linked views. These views often do not match the realistic
physical appearance of fibertracts but instead focus on the task at hand
the neuroscientist wants to perform with the data. So other
representations than graphs will be explored as part of this project
such as scatter plots or space-filling or pixel-based designs enabling
to provide a mapping of the information space more efficient to solve a
specific visual analytic task. Machine learrning techniques such as
generative graphs will be used to automatically extract summaries of the
anatomical and functional connectivity data. These geometrical and
topological summaries will be used as a backbone structure for the
visualization of the information space to be used for visual analysis tasks.

Moreover, an integral aspect of our approach is to combine the
visualization techniques in an interactive exploration tool that
supports analists in adjusting their exploration strategy as needed. An
integral part of the neuroanatomists' data exploration is the comparison
of different datasets, either derived from different people or captured
at different points in time. Therefore, the comparison of different
datasets and the temporal exploration will be an essential aspect of the
project. To be successful in this project, the PhD student will work
closely with domain experts in the neurosciences from the Université
Pierre et Marie Curie, both to develop the integrated interactive
visualization techniques using a participatory design approach as well
as to evaluate the new techniques in controlled experiements.

By closely working with the domain experts, the PhD student will work
toward an interactive tool for neuroanatomical data exploration that
integrates the new visualization techniques. The goal for this tool is
that it can be used in a realistic context for the everyday analysis
tasks of the neuroanatomists and that it will be provided to the public
as open-source software. Beyond this implementation, the project will
result in a deeper understanding of how to combine spatially explicit
data with connected abstract data aspects to benefit visualization in
the sciences in general.

The PhD research will be conducted under the supervision of Tobias
Isenberg and within the AVIZ research team at INRIA Saclay—Île-de-France
which concentrates on the visualization of complex data. AVIZ is one of
the most respected research labs in information visualization and visual
analytics worldwide. The PhD student will closely collaborate, in
particular, with Cédric Gouy-Pailler from the  Laboratoire Analyse de
Données et Intelligence des Systèmes at CEA whose expertise in machine
learning will be essential for the work. In addition, we will work with
domain experts in the neurosciences from the Université Pierre et Marie

Required applicants skills:
* highly motivated student
* degree (M.Sc., M. Eng, or equivalent) in computer science or closely
related fields
* education background in one or more of the following fields:
visualization, human-computer interaction, computer graphics, and
machine learning
* interest in applications in neuroimaging or in knowledge discovery
* previous experience in these fields (in particular, neuroimaging)
would be highly beneficial
* experience in modern computer graphics (GPU) programming
* fluent in written and spoken English (French language skills are not
required but would be beneficial for living in France and interacting
with people outside of the lab)
* previous experience in research and publication of research results

Application package:
* detailed CV
* motivation letter
* summary of the master thesis
* transcript of the grades
* contact details for two academic references
* prepare all application documents electronically and in English
* application deadline: applications are reviewed as they are received;
however, for full consideration please submit your application by
September 15

Dr. Tobias Isenberg <tobias.isenberg at inria.fr>

AVIZ team, INRIA Saclay

More information about the ieee_vis_open_positions mailing list