link to homepage

Institute of Neuroscience and Medicine

Navigation and service

Justin Domhof

M.Sc. Justin Domhof

Ph.D. student

Research Topics:

  • Derivation of mechanisms underlying brain functioning through computational modelling
  • Investigation of the effects of brain parcellations on modelling results

Methodological Topics:

  • Non-linear dynamical systems to approximate whole-brain dynamics
  • Parameter space exploration to uncover mechanistic underpinnings

As a computational neuroscientist, I aim to develop models of the brain in order to explain a variety of experimentally found phenomena. Usually, a model is first constructed so that it replicates the phenomena or the data reported in the empirical literature. Subsequently, when the model indeed sufficiently reproduces the experimental findings, we may explore the many properties of the model. These explorations may not only provide new insights to the phenomena at hand, but also aid experimentalists with information on which they can base their next series of experiments.

In our group, we, for example, aim to fit data derived from experimentally obtained MRI images to whole-brain models. My role as a PhD student within this project is to investigate how the choice for a particular parcellation scheme influences these fits. Since such a parcellation scheme describes how the numerous signals we receive from MRI measurements should be grouped together into brain regions, one may expect that different parcellation schemes can have drastic consequences for both data analysis and model fitting. Specifically, I would eventually like to be able to cluster parcellation schemes with similar effects together and explain why they influence the results in the way they do. 

First, I will do this using the data from healthy participants, but afterwards I would like to answer similar questions for data originating from clinical patients. By comparing the disease-based with the health-based models, we may gain new insights into various aspects of these diseases: how they are caused, how they progress, how they can be treated, etc..

With respect to future applications, the relevance of the models we develop also becomes apparent from the latter observation. If these models can eventually be reliably fitted to the data from a single individual, they could also be used in future clinical methods that focus on the patient as an individual. In the end, the investigation of these models may thus pave the way to individualized diagnoses, prognoses and treatment plans.


Institute of Neuroscience and Medicine (INM-7)
52425 Jülich