Computational Models of Neurological DisorderVirtual Brain for neurological disease modeling
Introduction
Neurological disorders represent a large cost to society, $1.5 trillion/year, nearly 9% of the gross domestic product (World Health Organization (2006)). Most of brain disorders, such as migraine, epilepsies, Alzheimer’s disease, Parkinson’s disease, and major depression remain poorly treated. Their progression can be slowed down; but it is often just a temporary relief with a large inter-patient variability. For instance many patients cannot be treated with existing drugs, e.g. 30% for patients with epilepsy. These facts underscore the necessity to better understand patient-specific mechanisms underlying brain disorders. Historically personalized medicine uses heavily genetic information, but finds more and more response on the system level. Structural and functional neuroimaging play a key role and have contributed diagnostic tools, e.g. such as presurgical evaluation of epilepsy. One solution to this issue is to link the interpretation of neuroimaging signals to personalized computational brain models, which we discuss in the following.
Section snippets
Brain disease modeling
Two different mechanistic modeling approaches can be distinguished according to the desired effect, i.e. whether one looks for a preventive or a curative treatment. The preventive approach requires identifying predictive biomarkers and the causal factors responsible for the transformation of a “healthy” network into a pathological one. The curative approach includes repairing the system, or at least controlling the symptoms. The mechanisms that need to be targeted for preventive and curative
The Virtual Brain
The Virtual Brain (TVB) is a large-scale brain network model comprising a connectivity matrix between cortical and subcortical areas and network nodes representing brain areas. TVB connectivity for primates is typically derived from Diffusion-weighted Tensor Imaging (DTI), a fairly recent non-invasive technology allowing the reconstruction of the myelinated white matter fibers. Brain areas are modeled using neural population models of varying degrees of sophistication. The computational models
The Virtual Mouse Brain (TVMB)
There is a multiplicity of experimental models of neurological disorders. These models have been designed to explore mechanisms and therapeutic solutions. One can distinguish genetic from induced models. The most relevant genetic models are those for which a mutation found in human families is directly introduced in the rodent genome. Induced models mostly involve lesions (chemical or electrical), e.g. for stroke, epilepsy, Parkinson’s disease, and autism. Whether genetic or induced, these
Conclusion
Large-scale brain network modeling enables linking personalized brain models with patient-specific neuroimaging data [26], [27]. Despite the enormous neuroinformatics complexity of integrating brain data, high performance computing and mathematical modeling, modern computational neuroscience provides in silico platforms (TVB, TVMB) for the testing of hypotheses of brain function on the large-scale system level. Suitable paradigms allow exploring questions linked to the network and the
References (26)
- et al.
Depression in Alzheimer’s disease: heterogeneity and related issues
Biol.Psychiatry
(2003) - et al.
Systematic approximations of neural fields through networks of neural masses in the virtual brain
Neuroimage
(2013) Mathematical framework for large-scale brain network modelling in The Virtual Brain
Neuroimage
(2015)Functional system and areal organization of a highly sampled individual human brain
Neuron
(2015)The effect of filter size on VBM analyses of DT-MRI data
NeuroImage
(2005)The virtual epileptic patient: individualized whole-brain models of epilepsy spread
Neuroimage
(2017)Depression in Parkinson’s disease
BMJ
(2000)- et al.
Depression in epilepsy: a critical review from a clinical perspective
Nat Rev Neurol
(2011) Brain connectivity in neurodegenerative diseases–from phenotype to proteinopathy
Nat Rev Neurol
(2014)- et al.
Resting-state functional connectivity in epilepsy: growing relevance for clinical decision making
Curr Opin Neurol
(2015)
Resting state networks in temporal lobe epilepsy
Epilepsia
Functional network alterations and their structural substrate in drug-resistant epilepsy
Front Neurosci
The Virtual Brain: a simulator of primate brain network dynamics
Front Neuroinform
Cited by (6)
The Bayesian Virtual Epileptic Patient: A probabilistic framework designed to infer the spatial map of epileptogenicity in a personalized large-scale brain model of epilepsy spread
2020, NeuroImageCitation Excerpt :On the other hand, due to the potential to improve medical treatment strategies, the personalized large-scale brain network modeling has gained popularity over the recent years (Jirsa et al., 2017; Bansal et al., 2018). In the individualized whole-brain modeling approach, the patient-specific information such as anatomical connectivity obtained from non-invasive imaging techniques is combined with the mean-field models of local neuronal activity to simulate the individual’s spatio-temporal brain activity at the macroscopic scale (Bernard and Jirsa, 2016; Proix et al., 2017). The Virtual Brain (TVB; (Sanz Leon et al., 2013)) is an open-access computational framework written in Python to reproduce and evaluate the personalized configurations of the brain by using individual subject data.
Multiscale modeling for drug discovery in brain disease
2016, Drug Discovery Today: Disease ModelsCitation Excerpt :At the highest spatial scale, a pair of reviews describe modeling of interactions across brain areas. Bernard and Jirsa [1] describe The Virtual Brain (TVB; thevirtualbrain.org), a simulation tool that can simulate the entire brain as a set of interconnected neural mass models. Arle and Carlson [13] demonstrate a multi-area circuit model in their Universal Neural Circuitry simulator (UNCuS) to look at depressive disorder.
Fully Bayesian estimation of virtual brain parameters with self-tuning Hamiltonian Monte Carlo
2022, Machine Learning: Science and Technology