Joana Cabral

Role of local network oscillations in resting-state functional connectivity

Spatio-temporally organized low-frequency fluctuations (<0.1 Hz), observed in BOLD fMRI signal during rest, suggest the existence of underlying network dynamics that emerge spontaneously from intrinsic brain processes. Furthermore, significant correlations between distinct anatomical regions-or functional connectivity (FC)-have led to the identification of several widely distributed resting-state networks (RSNs). This slow dynamics seems to be highly structured by anatomical connectivity but the mechanism behind it and its relationship with neural activity, particularly in the gamma frequency range, remains largely unknown. Indeed, direct measurements of neuronal activity have revealed similar large-scale correlations, particularly in slow power fluctuations of local field potential gamma frequency range oscillations. To address these questions, we investigated neural dynamics in a large-scale model of the human brains neural activity. A key ingredient of the model was a structural brain network defined by empirically derived long-range brain connectivity together with the corresponding conduction delays. A neural population, assumed to spontaneously oscillate in the gamma frequency range, was placed at each network node. When these oscillatory units are integrated in the network, they behave as weakly coupled oscillators. The time-delayed interaction between nodes is described by the Kuramoto model of phase oscillators, a biologically-based model of coupled oscillatory systems. For a realistic setting of axonal conduction speed, we show that time-delayed network interaction leads to the emergence of slow neural activity fluctuations, whose patterns correlate significantly with the empirically measured FC. The best agreement of the simulated FC with the empirically measured FC is found for a set of parameters where subsets of nodes tend to synchronize although the network is not globally synchronized. Inside such clusters, the simulated BOLD signal between nodes is found to be correlated, instantiating the empirically observed RSNs. Between clusters, patterns of positive and negative correlations are observed, as described in experimental studies. These results are found to be robust with respect to a biologically plausible range of model parameters. In conclusion, our model suggests how resting-state neural activity can originate from the interplay between the local neural dynamics and the large-scale structure of the brain.

Exploring the network dynamics underlying brain activity during rest

Since the mid 1990s, the intriguing dynamics of the brain at rest has been attracting a growing body of research in neuroscience. Neuroimaging studies have revealed distinct functional networks that slowly activate and deactivate, pointing to the existence of an underlying network dynamics emerging spontaneously during rest, with specific spatial, temporal and spectral characteristics. Several theoretical scenarios have been proposed and tested with the use of large-scale computational models of coupled brain areas. However, a mechanistic explanation that encompasses all the phenomena observed in the brain during rest is still to come. In this review, we provide an overview of the key findings of resting-state activity covering a range of neuroimaging modalities including fMRI, EEG and MEG. We describe how to best define and analyze anatomical and functional brain networks and how unbalancing these networks may lead to problems with mental health. Finally, we review existing large-scale models of resting-state dynamics in health and disease. An important common feature of resting-state models is that the emergence of resting-state functional networks is obtained when the model parameters are such that the system operates at the edge of a bifurcation. At this critical working point, the global network dynamics reveals correlation patterns that are spatially shaped by the underlying anatomical structure, leading to an optimal fit with the empirical BOLD functional connectivity. However, new insights coming from recent studies, including faster oscillatory dynamics and non-stationary functional connectivity, must be taken into account in future models to fully understand the network mechanisms leading to the resting-state activity.

Modeling the outcome of structural disconnection on resting-state functional connectivity.

A growing body of experimental evidence suggests that functional connectivity at rest is shaped by the underlying anatomical structure. Furthermore, the organizational properties of resting-state functional networks are thought to serve as the basis for an optimal cognitive integration. A disconnection at the structural level, as occurring in some brain diseases, would then lead to functional and presumably cognitive impairments. In this work, we propose a computational model to investigate the role of a structural disconnection (encompassing putative local/global and axonal/synaptic mechanisms) on the organizational properties of emergent functional networks. The brains spontaneous neural activity and the corresponding hemodynamic response were simulated using a large-scale network model, consisting of local neural populations coupled through white matter fibers. For a certain coupling strength, simulations reproduced healthy resting-state functional connectivity with graph properties in the range of the ones reported experimentally. When the structural connectivity is decreased, either globally or locally, the resultant simulated functional connectivity exhibited a network reorganization characterized by an increase in hierarchy, efficiency and robustness, a decrease in small-worldness and clustering and a narrower degree distribution, in the same way as recently reported for schizophrenia patients. Theoretical results indicate that most disconnection-related neuropathologies should induce the same qualitative changes in resting-state brain activity.

Exploring mechanisms of spontaneous functional connectivity in MEG: How delayed network interactions lead to structured amplitude envelopes of band-pass filtered oscillations

Spontaneous (or resting-state) brain activity has attracted a growing body of neuroimaging research over the last decades. Whole-brain network models have proved helpful to investigate the source of slow (<0.1 Hz) correlated hemodynamic fluctuations revealed in fMRI during rest. However, the mechanisms mediating resting-state long-distance correlations and the relationship with the faster neural activity remain unclear. Novel insights coming from MEG studies have shown that the amplitude envelopes of alpha- and beta-frequency oscillations (~8-30 Hz) display similar correlation patterns as the fMRI signals. In this work, we combine experimental and theoretical work to investigate the mechanisms of spontaneous MEG functional connectivity. Using a simple model of coupled oscillators adapted to incorporate realistic whole-brain connectivity and conduction delays, we explore how slow and structured amplitude envelopes of band-pass filtered signals - fairly reproducing MEG data collected from 10 healthy subjects at rest - are generated spontaneously in the space-time structure of the brain network. Our simulation results show that the large-scale neuroanatomical connectivity provides an optimal network structure to support a regime with metastable synchronization. In this regime, different subsystems may temporarily synchronize at reduced collective frequencies (falling in the 8-30 Hz range due to the delays) while the global system never fully synchronizes. This mechanism modulates the frequency of the oscillators on a slow time-scale (<0.1 Hz) leading to structured amplitude fluctuations of band-pass filtered signals. Taken overall, our results reveal that the structured amplitude envelope fluctuations observed in resting-state MEG data may originate from spontaneous synchronization mechanisms naturally occurring in the space-time structure of the brain.

Neural plasticity in human brain connectivity: the effects of long term deep brain stimulation of the subthalamic nucleus in Parkinson's disease

Background Positive clinical outcomes are now well established for deep brain stimulation, but little is known about the effects of long-term deep brain stimulation on brain structural and functional connectivity. Here, we used the rare opportunity to acquire pre- and postoperative diffusion tensor imaging in a patient undergoing deep brain stimulation in bilateral subthalamic nuclei for Parkinson’s Disease. This allowed us to analyse the differences in structural connectivity before and after deep brain stimulation. Further, a computational model of spontaneous brain activity was used to estimate the changes in functional connectivity arising from the specific changes in structural connectivity. Results We found significant localised structural changes as a result of long-term deep brain stimulation. These changes were found in sensory-motor, prefrontal/limbic, and olfactory brain regions which are known to be affected in Parkinson’s Disease. The nature of these changes was an increase of nodal efficiency in most areas and a decrease of nodal efficiency in the precentral sensory-motor area. Importantly, the computational model clearly shows the impact of deep brain stimulation-induced structural alterations on functional brain changes, which is to shift the neural dynamics back towards a healthy regime. The results demonstrate that deep brain stimulation in Parkinson’s Disease leads to a topological reorganisation towards healthy bifurcation of the functional networks measured in controls, which suggests a potential neural mechanism for the alleviation of symptoms. Conclusions The findings suggest that long-term deep brain stimulation has not only restorative effects on the structural connectivity, but also affects the functional connectivity at a global level. Overall, our results support causal changes in human neural plasticity after long-term deep brain stimulation and may help to identify the underlying mechanisms of deep brain stimulation.

Effects of lesions on synchrony and metastability in cortical networks

At the macroscopic scale, the human brain can be described as a complex network of white matter tracts integrating grey matter assemblies - the human connectome. The structure of the connectome, which is often described using graph theoretic approaches, can be used to model macroscopic brain function at low computational cost. Here, we use the Kuramoto model of coupled oscillators with time-delays, calibrated with respect to empirical functional MRI data, to study the relation between the structure of the connectome and two aspects of functional brain dynamics - synchrony, a measure of general coherence, and metastability, a measure of dynamical flexibility. Specifically, we investigate the relationship between the local structure of the connectome, quantified using graph theory, and the synchrony and metastability of the models dynamics. By removing individual nodes and all of their connections from the model, we study the effect of lesions on both global and local dynamics. Of the nine nodal graph-theoretical properties tested, two were able to predict effects of node lesion on the global dynamics. The removal of nodes with high eigenvector centrality leads to decreases in global synchrony and increases in global metastability, as does the removal of hub nodes joining topologically segregated network modules. At the level of local dynamics in the neighbourhood of the lesioned node, structural properties of the lesioned nodes hold more predictive power, as five nodal graph theoretical measures are related to changes in local dynamics following node lesions. We discuss these results in the context of empirical studies of stroke and functional brain dynamics.

Functional connectivity dynamically evolves on multiple time-scales over a static structural connectome: Models and mechanisms

Over the last decade, we have observed a revolution in brain structural and functional Connectomics. On one hand, we have an ever-more detailed characterization of the brains white matter structural connectome. On the other, we have a repertoire of consistent functional networks that form and dissipate over time during rest. Despite the evident spatial similarities between structural and functional connectivity, understanding how different time-evolving functional networks spontaneously emerge from a single structural network requires analyzing the problem from the perspective of complex network dynamics and dynamical systems theory. In that direction, bottom-up computational models are useful tools to test theoretical scenarios and depict the mechanisms at the genesis of resting-state activity. Here, we provide an overview of the different mechanistic scenarios proposed over the last decade via computational models. Importantly, we highlight the need of incorporating additional model constraints considering the properties observed at finer temporal scales with MEG and the dynamical properties of FC in order to refresh the list of candidate scenarios.

Structural connectivity in schizophrenia and its impact on the dynamics of spontaneous functional networks

The neuropathology of schizophrenia remains unclear. Some insight has come from modern neuroimaging techniques, which offer an unparalleled opportunity to explore in vivo the structure and function of the brain. Using functional magnetic resonance imaging, it has been found that the large-scale resting-state functional connectivity (rsFC) in schizophrenia--measured as the temporal correlations of the blood-oxygen-level-dependent (BOLD) signal--exhibit altered network topology, with lower small-world index. The origin of these rsFC alterations and link with the underlying structural connectivity remain unclear. In this work, we used a computational model of spontaneous large-scale brain activity to explore the role of the structural connectivity in the large-scale dynamics of the brain in health and schizophrenia. The structural connectomes from 15 adolescent patients with early-onset schizophrenia and 15 age- and gender-matched controls were built from diffusion tensor imaging data to detect the white matter tracts between 90 brain areas. Brain areas, simulated using a reduced dynamic mean-field model, receive excitatory input from other areas in proportion to the number of fibre tracts between them. The simulated mean field activity was transformed into BOLD signal, and the properties of the simulated functional networks were analyzed. Our results suggest that the functional alterations observed in schizophrenia are not directly linked to alterations in the structural topology. Instead, subtly randomized and less small-world functional networks appear when the brain operates with lower global coupling, which shifts the dynamics from the optimal healthy regime.

Single or multiple frequency generators in on-going brain activity: A mechanistic whole-brain model of empirical MEG data.

ABSTRACT During rest, envelopes of band‐limited on‐going MEG signals co‐vary across the brain in consistent patterns, which have been related to resting‐state networks measured with fMRI. To investigate the genesis of such envelope correlations, we consider a whole‐brain network model assuming two distinct fundamental scenarios: one where each brain area generates oscillations in a single frequency, and a novel one where each brain area can generate oscillations in multiple frequency bands. The models share, as a common generator of damped oscillations, the normal form of a supercritical Hopf bifurcation operating at the critical border between the steady state and the oscillatory regime. The envelopes of the simulated signals are compared with empirical MEG data using new methods to analyse the envelope dynamics in terms of their phase coherence and stability across the spectrum of carrier frequencies. Considering the whole‐brain model with a single frequency generator in each brain area, we obtain the best fit with the empirical MEG data when the fundamental frequency is tuned at 12 Hz. However, when multiple frequency generators are placed at each local brain area, we obtain an improved fit of the spatio‐temporal structure of on‐going MEG data across all frequency bands. Our results indicate that the brain is likely to operate on multiple frequency channels during rest, introducing a novel dimension for future models of large‐scale brain activity. HighlightsResting‐state MEG reveals correlated amplitude envelopes between brain areas.Envelope functional connectivity spans a range of carrier frequencies.Local emergence of carrier oscillations modeled with a Hopf bifurcation model.Each brain area may resonate at one or multiple fundamental frequencies.Multiple resonant frequencies outperform the single frequency scenario. Graphical abstract Figure. No Caption available.

Novel fingerprinting method characterises the necessary and sufficient structural connectivity from deep brain stimulation electrodes for a successful outcome

Deep brain stimulation (DBS) is a remarkably effective clinical tool, used primarily for movement disorders. DBS relies on precise targeting of specific brain regions to rebalance the oscillatory behaviour of whole-brain neural networks. Traditionally, DBS targeting has been based upon animal models (such as MPTP for Parkinson’s disease) but has also been the result of serendipity during human lesional neurosurgery. There are, however, no good animal models of psychiatric disorders such as depression and schizophrenia, and progress in this area has been slow. In this paper, we use advanced tractography combined with whole-brain anatomical parcellation to provide a rational foundation for identifying the connectivity ‘fingerprint’ of existing, successful DBS targets. This knowledge can then be used pre-surgically and even potentially for the discovery of novel targets. First, using data from our recent case series of cingulate DBS for patients with treatment-resistant chronic pain, we demonstrate how to identify the structural ‘fingerprints’ of existing successful and unsuccessful DBS targets in terms of their connectivity to other brain regions, as defined by the whole-brain anatomical parcellation. Second, we use a number of different strategies to identify the successful fingerprints of structural connectivity across four patients with successful outcomes compared with two patients with unsuccessful outcomes. This fingerprinting method can potentially be used pre-surgically to account for a patient’s individual connectivity and identify the best DBS target. Ultimately, our novel fingerprinting method could be combined with advanced whole-brain computational modelling of the spontaneous dynamics arising from the structural changes in disease, to provide new insights and potentially new targets for hitherto impenetrable neuropsychiatric disorders.