Awakening: Predicting external stimulation to force transitions between different brain states

Deco, Gustavo; Cruzat, Josephine; Cabral, Joana; Tagliazucchi, Enzo; Laufs, Helmut; Logothetis, Nikos K.; Kringelbach, Morten L.

doi:10.1073/pnas.1905534116

Cited by 215 publications

(393 citation statements)

References 61 publications

Supporting

Mentioning

314

Contrasting

Order By: Relevance

“…Traditionally, whole-brain models have been relatively successful in linking structural connectivity with functional dynamics (Breakspear, 2017;Deco and Kringelbach, 2014). This has revealed important new mechanistic principles of brain function (Deco et al, 2018;Deco et al, 2019a;Deco et al, 2019b;Deco et al, 2017e;Honey et al, 2007). Nevertheless, the present causal characterisation of whole-brain information flow offers a new avenue for generating even more useful models.…”

Section: Causal Confirmation Using Novel Generative Whole-brain Modelmentioning

confidence: 99%

Revisiting the Global Workspace: Orchestration of the functional hierarchical organisation of the human brain

Deco

Vidaurre

Kringelbach

2019

Preprint

Self Cite

View full text Add to dashboard Cite

A central, unsolved challenge in neuroscience is how the brain orchestrates function by organising the flow of information necessary for the underlying computation. It has been argued that this whole-brain orchestration is carried out by a core subset of integrative brain regions, commonly referred to as the 'global workspace', although quantifying the constitutive brain regions has proven elusive. We developed a normalised directed transfer entropy (NDTE) framework for determining the pairwise bidirectional causal flow between brain regions and applied it to multimodal wholebrain neuroimaging from over 1000 healthy participants. We established the full brain hierarchy and common regions in a 'functional rich club' (FRIC) coordinating the functional hierarchical organisation during rest and task. FRIC contains the core set of regions, which similar to a 'club' of functional hubs are characterized by a tendency to be more densely functionally connected among themselves than to the rest of brain regions from where they integrate information. The invariant global workspace is the intersection of FRICs across rest and seven tasks, and was found to consist of the precuneus, posterior and isthmus cingulate cortices, nucleus accumbens, putamen, hippocampus and amygdala that orchestrate the functional hierarchical organisation based on information from perceptual, long-term memory, evaluative and attentional systems. We confirmed the causal significance and robustness of this invariant global workspace by systematically lesioning a generative whole-brain model accurately simulating the functional hierarchy defined by NDTE. Overall, this is a major step forward in understanding the complex choreography of information flow within the functional hierarchical organisation of the human brain.

show abstract

Section: Causal Confirmation Using Novel Generative Whole-brain Modelmentioning

confidence: 99%

Revisiting the Global Workspace: Orchestration of the functional hierarchical organisation of the human brain

Deco

Vidaurre

Kringelbach

2019

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Although in all cases the highest NoEL was frequent, the most interesting patterns were observed when the frequencies for the maximum values q ∈ [80, 91], were not taken into account. When looking at the frequencies below 80, the probability grew linearly with the NoEL for the deep sleep condition, whereas it showed almost a constant behaviour for the wakefulness condition with q ∈ [1,80], as it can be seen from Fig.6C (red and blue respectively). The pattern typically observed was that the distribution for the awake state was more homogeneous along all the possible NoEL values.…”

Section: /19mentioning

confidence: 72%

“…We have seen that in deep sleep the probability of a concrete q increases with q. On the contrary, when the participants are awake the frequency remains almost constant with q ∈ [1,89].…”

Section: /19mentioning

confidence: 76%

“…Over the last couple of years, there has been an increasing interest in trying to identify the necessary and sufficient properties of different brain states 1,2 . One particularly popular theory, the Integrated Information Theory (IIT) [3][4][5][6][7] proposes a potential route into identifying the essential properties of brain states using five axioms: intrinsic existence, composition, information, integration and exclusion.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamical informational structures characterize the different human brain states of wakefulness and deep sleep

Galadí

Pereira

Perl

et al. 2019

Preprint

View full text Add to dashboard Cite

The dynamical activity of the human brain describes an extremely complex energy landscape changing over time and its characterisation is central unsolved problem in neuroscience. We propose a novel mathematical formalism for characterizing how the landscape of attractors sustained by a dynamical system evolves in time. This mathematical formalism is used to distinguish quantitatively and rigorously between the different human brain states of wakefulness and deep sleep. In particular, by using a whole-brain dynamical ansatz integrating the underlying anatomical structure with the local node dynamics based on a Lotka-Volterra description, we compute analytically the global attractors of this cooperative system and their associated directed graphs, here called the informational structures. The informational structure of the global attractor of a dynamical system describes precisely the past and future behaviour in terms of a directed graph composed of invariant sets (nodes) and their corresponding connections (links). We characterize a brain state by the time variability of these informational structures. This theoretical framework is potentially highly relevant for developing reliable biomarkers of patients with e.g. neuropsychiatric disorders or different levels of coma. of the phase space described by a set of selected invariant global solutions of the associated dynamical system, such as stationary points (equilibria), connecting orbits among them, periodic solutions and limit cycles. The informational structure of the GA is defined as a directed graph composed of nodes associated with those invariants and links establishing their connections (see Methods and Supplementary Information for a formal rigorous definition). Informational Structure has been used to show the dependence between the topology, the value of the parameters and the state with respect to its level of integration 14 , as such pointing for the small world configuration of the brain 16 (although see recent controversies 17 ).Previous research has applied Informational Structure to population dynamics in complex networks 18,19 . Equally used for modelling mutualistic systems in Theoretical Ecology and Economy 20, 21 , Informational Structure can relate the topology of complex networks and their dynamics. In mutualistic systems when achieving robustness and life abundance -the so-called architecture of biodiversity 22, 23 -the dynamics is determined not only by the topology 24 but also by modularity 25 and the strength of the parameters 26 .These kinds of relationships are also at the heart of many important open questions in neuroscience 27,28 . Informational Structures and their continuous time-dependence on the strength of connections can allow to understand the dynamics of the system as a coherent process, whose information is structured, and potentially providing new insights into sudden bifurcations 29 .Here, we were interested in characterising human sleep, which is traditionally subdivided into different stages that alternate in the course...

show abstract

“…They can result from 260 parameter changes to the network strength or Hopf bifurcations that cause the system 261 to change its dynamics over time [10,28]. They can also be the result of adding external 262 input and stimuli into the system causing a change from the zero-input manifold and 263 altering the dynamics [3,11]. These are not mutually exclusive and could induce the 264 changes at once.…”

mentioning

confidence: 99%

Brain Network Constraints and Recurrent Neural Networks reproduce unique Trajectories and State Transitions seen over the span of minutes in resting state fMRI

Kashyap

Keilholz

2019

Preprint

View full text Add to dashboard Cite

Large scale patterns of spontaneous whole brain activity seen in resting state functional Magnetic Resonance Imaging (rsfMRI), are in part believed to arise from neural populations interacting through the structural fiber network [18]. Generative models that simulate this network activity, called Brain Network Models (BNM), are able to reproduce global averaged properties of empirical rsfMRI activity such as functional connectivity (FC) [7,27]. However, they perform poorly in reproducing unique trajectories and state transitions that are observed over the span of minutes in whole brain data [20]. At very short timescales between measurements, it is not known how much of the variance these BNM can explain because they are not currently synchronized with the measured rsfMRI. We demonstrate that by solving for the initial conditions of BNM from an observed data point using Recurrent Neural Networks (RNN) and integrating it to predict the next time step, the trained network can explain large amounts of variance for the 5 subsequent time points of unseen future trajectory. The RNN and BNM combined system essentially models the network component of rsfMRI, and where future activity is solely based on previous neural activity propagated through the structural network. Longer instantiations of this generative model simulated over the span of minutes can reproduce average FC and the 1/f power spectrum from 0.01 to 0.3 Hz seen in fMRI. Simulated data also contain interesting resting state dynamics, such as unique repeating trajectories, called QPPs [22] that are highly correlated to the empirical trajectory which spans over 20 seconds. Moreover, it exhibits complex states and transitions as seen using k-Means analysis on windowed FC matrices [1]. This suggests that by combining BNMs with RNN to accurately predict future resting state activity at short timescales, it is learning the manifold of the network dynamics, allowing it to simulate complex resting state trajectories at longer time scales. We believe that our technique will be useful in understanding the large-scale functional organization of the brain and how different BNMs recapitulate different aspects of the system dynamics. Introduction 1Over the past decade, our understanding of spontaneous whole brain activity and 2 coordination between brain regions has largely been obtained through non invasive 3 resting state functional Magnetic Resonance Imaging (rsfMRI) studies [23,31,37].Resting state, a state without an explicit task or stimulus, has surprisingly complex 5 whole brain trajectories that are well structured and highly dependent on the previous 6 brain state [1,5,29,37]. Current generative models such as Brain Network Models 7 (BNM) attempt to characterize whole brain activity as the interaction between a single 8 neural population and the activity of its network neighbors defined by its structural 9 fiber connections as measured through diffusion tensor imaging (DTI) [7,18]. Although 10 there are many variants of the model that use different sets of diffe...

show abstract

Awakening: Predicting external stimulation to force transitions between different brain states

Cited by 215 publications

References 61 publications

Revisiting the Global Workspace: Orchestration of the functional hierarchical organisation of the human brain

Revisiting the Global Workspace: Orchestration of the functional hierarchical organisation of the human brain

Dynamical informational structures characterize the different human brain states of wakefulness and deep sleep

Brain Network Constraints and Recurrent Neural Networks reproduce unique Trajectories and State Transitions seen over the span of minutes in resting state fMRI

Contact Info

Product

Resources

About