Development of a forward model of the rat’s cortex based on an empirically derived lead-field

Group members involved in the project

George Dimitriadis, Anne Fransen, Eric Maris

Background and aim of the project

The brain is an organ of highly complicated organization at all levels of life, from the molecular all the way to the systems level. Given the surprising homogeneity of the of neocortex with respect to neuronal types, the cognitive functions it implements must depend on the neuronal network architecture. Crucially, information processing at this network level depends on electrical signalling between neurons. Thus, understanding brain function requires the understanding of its electric behaviour and of the theory behind the electrical signals that we record in behaving animals.

Our lab has developed an invasive electrophysiology setup targeting the somatosensory system of rats. We have developed a method of implanting electrode micro-grids based on polyimide technology which allows to follow the electrophysiology of a large part of the cortex with high temporal and spatial precision. Our main focus is the elucidation of top-down effects on the sensory system of mammals as in, for example, attention and expectation.

Our original experiments have shown that our electrodes capture activity from deeper structures of the brain (e.g., thalamus) together with the local activity under the electrodes (i.c., the neocortex). For the elucidation of the signal we require the construction of a model of how the rat's brain conducts current originating from these different sources. For the human scalp EEG, the volume conduction model (forward model) is typically derived from Maxwell's equations under the assumption of isotropic conductance. This involves assumptions about the conductance of the brain, the cerebrospinal fluid (CSF), the skull and the skin, plus 3D anatomical information about geometry of these four tissues.

In our case, there is no CSF, skull nor skin between the electrodes and the brain, which calls for a different approach than the existing methods for scalp EEG. At the same time, we are in a position to conduct experiments not possible in humans, which will allow an empirical validation of the forward model as derived from Maxwell's equations under the assumption of isotropic conductance.

Internship

We have an internship available for a student who would be interested in (1) developing the apparatus required for such experiments and (2) to calculate the forward model for the rat brain based on these data. This will involve the development of a device that allow the electrical stimulation of parts of the rat's cortex (while anesthetized) and the simultaneous measurement of the electrical activity using our existing recording setup. After data collection, the student will be involved in deriving a standard forward model on the basis of Maxwell's equations, the isotropy assumption, and 3D anatomical information about the rat's brain. The theoretically derived forward model will then be compared with the empirical (data-based) one, allowing the former to be validated.

The student will come in contact with current methodologies in animal electrophysiology, will learn to understand current models of the brain electrical behaviour, will further develop them, and will finally provide the scientific community with a much required sanity check of the assumptions currently used in the construction of forward models.

Requirements

The prospective student should have a basic understanding of electromagnetism, linear algebra, and Matlab. He or she should interact with our electronics and engineering workshops to construct an experimental apparatus for electrical stimulation in the neuropil, which will subsequently be used to conduct a series of experiments.

References

Yvert B., Crouzeix-Cheylus A., Pernier J. (2001) Fast realistic modelling in bioelectromagnetism using lead-field interpolation, Human Brain Mapping, V.14, Issue 1, pp. 48-63

Fuchs M., Kastner J., Wagner M., Hawes S., Ebersole J.S. (2002) A standardized boundary element method volume conductor model, Clinical Neurophysiology, V.113, Issue 5, p. 702-71

Contact Information

George Dimitriadis (g.dimitriadis at donders.ru.nl)