Personalization of hybrid brain models from neuroimaging and electrophysiology data

Sanchez-Todo, Roser; Salvador, Ricardo; Santarnecchi, Emiliano; Wendling, Fabrice; Deco, Gustavo; Ruffini, Giulio

doi:10.1101/461350

Cited by 16 publications

(20 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In such networks, nodes correspond to cortical or sub-cortical brain regions and edges correspond to either structural (i.e., direct connections) or functional (i.e., through synaptic or ephaptic interactions) couplings between these regions. Several computational studies have developed whole-brain network models to explore the relationship between brain function and its underlying connectivity [1,2,3,4,5,6,7,8].…”

mentioning

confidence: 99%

The 2D Ising model, criticality and AIT

Ruffini¹,

Deco

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this short note we study the 2D Ising model, a universal computational model which reflects phase transitions and critical phenomena, as a framework for establishing links between systems that exhibit criticality with the notions of complexity. This is motivated in the context of neuroscience applications stemming from algorithmic information theory (AIT). Starting with the original 2D Ising model, we show that — together with correlation length of the spin lattice, susceptibility to a uniform external field — the correlation time of the magnetization time series, the compression ratio of the spin lattice, the complexity of the magnetization time series — as derived from Lempel-Ziv-Welch compression—, and the rate of information transmission in the lattice, all reflect the effects of the phase transition, which results in spacetime pockets of uniform magnetization at all scales. We also show that in the Ising model the insertion of sparse long-range couplings has a direct effect on the critical temperature and other parameters. The addition of positive links extends the ordered regime to higher critical temperatures, while negative links have a stronger, disordering influence at the global scale. We discuss some implications for the study of long-range (e.g., ephaptic) interactions in the human brain and the effects of weak perturbations in neural dynamics.

show abstract

mentioning

confidence: 99%

The 2D Ising model, criticality and AIT

Ruffini¹,

Deco

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Neural mass models. Neural mass (semi-empirical) models (NMM), first developed in the early seventies by W. Freeman (17) and F. Lopez de Silva (2), provide a physiologically grounded description of the average synaptic activity and firing rate of a neural population (1)(2)(3)(4)(5)(6). NMMs are increasingly used for local and whole-brain modeling in neurology (e.g., epilepsy (18,19) or Alzheimer's disease (20)) and for understanding and optimizing the effects of transcranial electrical brain stimulation (tES) (6,7,21,22).…”

Section: Significancementioning

confidence: 99%

“…Neural mass (semi-empirical) models (NMM), first developed in the early seventies by W. Freeman (17) and F. Lopez de Silva (2), provide a physiologically grounded description of the average synaptic activity and firing rate of a neural population (1)(2)(3)(4)(5)(6). NMMs are increasingly used for local and whole-brain modeling in neurology (e.g., epilepsy (18,19) or Alzheimer's disease (20)) and for understanding and optimizing the effects of transcranial electrical brain stimulation (tES) (6,7,21,22). The central conceptual elements in this framework are the synapse, which is seen to transduce incoming activity (quantified by firing rate) into a mean membrane potential perturbation in the receiving neuron population, and the sigmoid function transforming population membrane potential to output (mean) firing rate with due account for threshold and saturation effects (see (16) for a nice introduction to the Jansen-Rit model).…”

Section: Significancementioning

confidence: 99%

Analysis and extension of exact mean-field theory with dynamic synaptic currents

Ruffini¹

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Neural mass models such as the Jansen-Rit system provide a practical framework for representing and interpreting electrophysiological activity (1-6) in both local and global brain models (7). However, they are only partly derived from first principles. While the post-synaptic potential dynamics in NMM are inferred from data and can be grounded on diffusion physics (8-10), Freeman's "wave to pulse" sigmoid function (11-13) is used to transduce mean population membrane potential into firing rate rests on a weaker theoretical standing. On the other hand, Montbrio et al (14, 15) derive an exact mean-field theory from a quadratic integrate and fire neuron model under some simplifying assumptions (MPR), connecting microscale neural mechanisms and meso/macroscopic phenomena. The MPR model can be seen to replace Freeman's sigmoid function with a pair of differential equations for the mean membrane potential and firing rate variables, providing a mechanistic interpretation of NMM semi-empirical sigmoid parameters. In doing so, it sheds light on the mechanisms behind enhanced network response to weak but uniform perturbations: in the exact mean-field theory, intrinsic population connectivity modulates the steady-state firing rate transfer function in a monotonic manner, with increasing self-connectivity leading to higher firing rates. This provides a plausible mechanism for the enhanced response of densely connected networks to weak, uniform inputs such as the electric fields produced by non-invasive brain stimulation. Finally, we complete the MPR model by adding the equations for delayed post-synaptic currents, bringing together the MPR and NMM formalisms into a unified exact mean-field theory (NMM2) displaying rich dynamical features. As an example, we analyze the dynamics of a single population model, and a model of two coupled populations with a simple excitation-inhibition (E-I) architecture, showing it displays rich dynamics with limit cycles, period doubling, bursting behavior, and enhanced sensitivity to external inputs.

show abstract

“…In addition, and equally importantly, we used here an interaction model that does not consider the complexity or spatial distribution of pyramidal neurons, or the effects on other types of neurons, or the complexities associated to the dynamical nature of neural physiologymuch as it is done in brain stimulation research, with some justification [14,19,34] for the analysis of tES effects. Only recently the effects of tES have been studied in computational models of the brain [19,86,87] using the lambda-E model discussed above, but yet ignoring the intricacies of micro-cortical network circuitry. Our modeling work and EMOD inherits all these limitations: this is a first approach that will be improved in the future.…”

Section: Limitations Of the Study And Future Directionsmentioning

confidence: 99%

Realistic modeling of mesoscopic ephaptic coupling in the human brain

Ruffini

Salvador²,

Tadayon

et al. 2020

PLoS Comput Biol

Self Cite

View full text Add to dashboard Cite

Several decades of research suggest that weak electric fields may influence neural processing, including those induced by neuronal activity and proposed as a substrate for a potential new cellular communication system, i.e., ephaptic transmission. Here we aim to model mesoscopic ephaptic activity in the human brain and explore its trajectory during aging by characterizing the electric field generated by cortical dipoles using realistic finite element modeling. Extrapolating from electrophysiological measurements, we first observe that modeled endogenous field magnitudes are comparable to those in measurements of weak but functionally relevant self-generated fields and to those produced by noninvasive transcranial brain stimulation, and therefore possibly able to modulate neuronal activity. Then, to evaluate the role of these fields in the human cortex in large MRI databases, we adapt an interaction approximation that considers the relative orientation of neuron and field to estimate the membrane potential perturbation in pyramidal cells. We use this approximation to define a simplified metric (EMOD1) that weights dipole coupling as a function of distance and relative orientation between emitter and receiver and evaluate it in a sample of 401 realistic human brain models from healthy subjects aged 16-83. Results reveal that ephaptic coupling, in the simplified mesoscopic modeling approach used here, significantly decreases with age, with higher involvement of sensorimotor regions and medial brain structures. This study suggests that by providing the means for fast and direct interaction between neurons, ephaptic modulation may contribute to the complexity of human function for cognition and behavior, and its modification across the lifespan and in response to pathology.

show abstract

Personalization of hybrid brain models from neuroimaging and electrophysiology data

Cited by 16 publications

References 74 publications

The 2D Ising model, criticality and AIT

The 2D Ising model, criticality and AIT

Analysis and extension of exact mean-field theory with dynamic synaptic currents

Realistic modeling of mesoscopic ephaptic coupling in the human brain

Contact Info

Product

Resources

About