Parameter estimation in a whole-brain network model of epilepsy: Comparison of parallel global optimization solvers.

The Virtual Epileptic Patient (VEP) refers to a computer-based representation of a patient with epilepsy that combines personalized anatomical data with dynamical models of abnormal brain activities. It is capable of generating spatio-temporal seizure patterns that resemble those recorded with invasive methods such as stereoelectro EEG data, allowing for the evaluation of clinical hypotheses before planning surgery. This study highlights the effectiveness of calibrating VEP models using a global optimization approach. The approach utilizes SaCeSS, a cooperative metaheuristic algorithm capable of parallel computation, to yield high-quality solutions without requiring excessive computational time. Through extensive benchmarking on synthetic data, our proposal successfully solved a set of different configurations of VEP models, demonstrating better scalability and superior performance against other parallel solvers. These results were further enhanced using a Bayesian optimization framework for hyperparameter tuning, with significant gains in terms of both accuracy and computational cost. Additionally, we added a scalable uncertainty quantification phase after model calibration, and used it to assess the variability in estimated parameters across different problems. Overall, this study has the potential to improve the estimation of pathological brain areas in drug-resistant epilepsy, thereby to inform the clinical decision-making process.

Selective Posterior Cerebral Artery Wada Better Predicts Good Memory and Naming Outcomes Following Selective Stereotactic Thermal Ablation for Medial Temporal Lobe Epilepsy Than Internal Carotid Artery Wada.

The conventional intracarotid amobarbital (Wada) test has been used to assess memory function in patients being considered for temporal lobe epilepsy (TLE) surgery. Minimally invasive approaches that target the medial temporal lobe (MTL) and spare neocortex are increasingly used, but a knowledge gap remains in how to assess memory and language risk from these procedures. We retrospectively compared results of two versions of the Wada test, the intracarotid artery (ICA-Wada) and posterior cerebral artery (PCA-Wada) approaches, with respect to predicting subsequent memory and language outcomes, particularly after stereotactic laser amygdalohippocampotomy (SLAH). We included all patients being considered for SLAH who underwent both ICA-Wada and PCA-Wada at a single institution. Memory and confrontation naming assessments were conducted using standardized neuropsychological tests to assess pre- to post-surgical changes in cognitive performance. Of 13 patients who initially failed the ICA-Wada, only one patient subsequently failed the PCA-Wada (p=0.003, two-sided binomial test with p 0 =0.5) demonstrating that these tests assess different brain regions or networks. PCA-Wada had a high negative predictive value for the safety of SLAH, compared to ICA-Wada, as none of the patients who underwent SLAH after passing the PCA-Wada experienced catastrophic memory decline (0 of 9 subjects, p <.004, two-sided binomial test with p 0 =0.5), and all experienced a good cognitive outcome. In contrast, the single patient who received a left anterior temporal lobectomy after failed ICA- and passed PCA-Wada experienced a persistent, near catastrophic memory decline. On confrontation naming, few patients exhibited disturbance during the PCA-Wada. Following surgery, SLAH patients showed no naming decline, while open resection patients, whose surgeries all included ipsilateral temporal lobe neocortex, experienced significant naming difficulties (Fisher's exact test, p <.05). These findings demonstrate that (1) failing the ICA-Wada falsely predicts memory decline following SLAH, (2) PCA-Wada better predicts good memory outcomes of SLAH for MTLE, and (3) the MTL brain structures affected by both PCA-Wada and SLAH are not directly involved in language processing.

Bifurcations and bursting in the Epileptor.

The Epileptor is a phenomenological model for seizure activity that is used in a personalized large-scale brain modeling framework, the Virtual Epileptic Patient, with the aim of improving surgery outcomes for drug-resistant epileptic patients. Transitions between interictal and ictal states are modeled as bifurcations, enabling the definition of seizure classes in terms of onset/offset bifurcations. This establishes a taxonomy of seizures grounded in their essential underlying dynamics and the Epileptor replicates the activity of the most common class, as observed in patients with focal epilepsy, which is characterized by square-wave bursting properties. The Epileptor also encodes an additional mechanism to account for interictal spikes and spike and wave discharges. Here we use insights from a more generic model for square-wave bursting, based on the Unfolding Theory approach, to guide the bifurcation analysis of the Epileptor and gain a deeper understanding of the model and the role of its parameters. We show how the Epileptor's parameters can be modified to produce activities for other seizures classes of the taxonomy, as observed in patients, so that the large-scale brain models could be further personalized. Some of these classes have already been described in the literature in the Epileptor, others, predicted by the generic model, are new. Finally, we unveil how the interaction with the additional mechanism for spike and wave discharges alters the bifurcation structure of the main burster.

Personalised virtual brain models in epilepsy.

Individuals with drug-resistant focal epilepsy are candidates for surgical treatment as a curative option. Before surgery can take place, the patient must have a presurgical evaluation to establish whether and how surgical treatment might stop their seizures without causing neurological deficits. Virtual brains are a new digital modelling technology that map the brain network of a person with epilepsy, using data derived from MRI. This technique produces a computer simulation of seizures and brain imaging signals, such as those that would be recorded with intracranial EEG. When combined with machine learning, virtual brains can be used to estimate the extent and organisation of the epileptogenic zone (ie, the brain regions related to seizure generation and the spatiotemporal dynamics during seizure onset). Virtual brains could, in the future, be used for clinical decision making, to improve precision in localisation of seizure activity, and for surgical planning, but at the moment these models have some limitations, such as low spatial resolution. As evidence accumulates in support of the predictive power of personalised virtual brain models, and as methods are tested in clinical trials, virtual brains might inform clinical practice in the near future.

Delineating epileptogenic networks using brain imaging data and personalized modeling in drug-resistant epilepsy.

Precise estimates of epileptogenic zone networks (EZNs) are crucial for planning intervention strategies to treat drug-resistant focal epilepsy. Here, we present the virtual epileptic patient (VEP), a workflow that uses personalized brain models and machine learning methods to estimate EZNs and to aid surgical strategies. The structural scaffold of the patient-specific whole-brain network model is constructed from anatomical T1 and diffusion-weighted magnetic resonance imaging. Each network node is equipped with a mathematical dynamical model to simulate seizure activity. Bayesian inference methods sample and optimize key parameters of the personalized model using functional stereoelectroencephalography recordings of patients' seizures. These key parameters together with their personalized model determine a given patient's EZN. Personalized models were further used to predict the outcome of surgical intervention using virtual surgeries. We evaluated the VEP workflow retrospectively using 53 patients with drug-resistant focal epilepsy. VEPs reproduced the clinically defined EZNs with a precision of 0.6, where the physical distance between epileptogenic regions identified by VEP and the clinically defined EZNs was small. Compared with the resected brain regions of 25 patients who underwent surgery, VEP showed lower false discovery rates in seizure-free patients (mean, 0.028) than in non-seizure-free patients (mean, 0.407). VEP is now being evaluated in an ongoing clinical trial (EPINOV) with an expected 356 prospective patients with epilepsy.

Focal non-invasive deep-brain stimulation with temporal interference for the suppression of epileptic biomarkers.

Introduction: Neurostimulation applied from deep brain stimulation (DBS) electrodes is an effective therapeutic intervention in patients suffering from intractable drug-resistant epilepsy when resective surgery is contraindicated or failed. Inhibitory DBS to suppress seizures and associated epileptogenic biomarkers could be performed with high-frequency stimulation (HFS), typically between 100 and 165 Hz, to various deep-seated targets, such as the Mesio-temporal lobe (MTL), which leads to changes in brain rhythms, specifically in the hippocampus. The most prominent alterations concern high-frequency oscillations (HFOs), namely an increase in ripples, a reduction in pathological Fast Ripples (FRs), and a decrease in pathological interictal epileptiform discharges (IEDs). Materials and methods: In the current study, we use Temporal Interference (TI) stimulation to provide a non-invasive DBS (130 Hz) of the MTL, specifically the hippocampus, in both mouse models of epilepsy, and scale the method using human cadavers to demonstrate the potential efficacy in human patients. Simulations for both mice and human heads were performed to calculate the best coordinates to reach the hippocampus. Results: This non-invasive DBS increases physiological ripples, and decreases the number of FRs and IEDs in a mouse model of epilepsy. Similarly, we show the inability of 130 Hz transcranial current stimulation (TCS) to achieve similar results. We therefore further demonstrate the translatability to human subjects via measurements of the TI stimulation vs. TCS in human cadavers. Results show a better penetration of TI fields into the human hippocampus as compared with TCS. Significance: These results constitute the first proof of the feasibility and efficiency of TI to stimulate at depth an area without impacting the surrounding tissue. The data tend to show the sufficiently focal character of the induced effects and suggest promising therapeutic applications in epilepsy.

Identifying spatio-temporal seizure propagation patterns in epilepsy using Bayesian inference.

Focal drug resistant epilepsy is a neurological disorder characterized by seizures caused by abnormal activity originating in one or more regions together called as epileptogenic zone. Treatment for such patients involves surgical resection of affected regions. Epileptogenic zone is typically identified using stereotactic EEG recordings from the electrodes implanted into the patient's brain. Identifying the epileptogenic zone is a challenging problem due to the spatial sparsity of electrode implantation. We propose a probabilistic hierarchical model of seizure propagation patterns, based on a phenomenological model of seizure dynamics called Epileptor. Using Bayesian inference, the Epileptor model is optimized to build patient specific virtual models that best fit to the log power of intracranial recordings. First, accuracy of the model predictions and identifiability of the model are investigated using synthetic data. Then, model predictions are evaluated against a retrospective patient cohort of 25 patients with varying surgical outcomes. In the patients who are seizure free after surgery, model predictions showed good match with the clinical hypothesis. In patients where surgery failed to achieve seizure freedom model predictions showed a strong mismatch. Our results demonstrate that proposed probabilistic model could be a valuable tool to aid the clinicians in identifying the seizure focus.

Data-driven method to infer the seizure propagation patterns in an epileptic brain from intracranial electroencephalography.

Surgical interventions in epileptic patients aimed at the removal of the epileptogenic zone have success rates at only 60-70%. This failure can be partly attributed to the insufficient spatial sampling by the implanted intracranial electrodes during the clinical evaluation, leading to an incomplete picture of spatio-temporal seizure organization in the regions that are not directly observed. Utilizing the partial observations of the seizure spreading through the brain network, complemented by the assumption that the epileptic seizures spread along the structural connections, we infer if and when are the unobserved regions recruited in the seizure. To this end we introduce a data-driven model of seizure recruitment and propagation across a weighted network, which we invert using the Bayesian inference framework. Using a leave-one-out cross-validation scheme on a cohort of 45 patients we demonstrate that the method can improve the predictions of the states of the unobserved regions compared to an empirical estimate that does not use the structural information, yet it is on the same level as the estimate that takes the structure into account. Furthermore, a comparison with the performed surgical resection and the surgery outcome indicates a link between the inferred excitable regions and the actual epileptogenic zone. The results emphasize the importance of the structural connectome in the large-scale spatio-temporal organization of epileptic seizures and introduce a novel way to integrate the patient-specific connectome and intracranial seizure recordings in a whole-brain computational model of seizure spread.

Non-thermal Electroporation Ablation of Epileptogenic Zones Stops Seizures in Mice While Providing Reduced Vascular Damage and Accelerated Tissue Recovery.

In epilepsy, the most frequent surgical procedure is the resection of brain tissue in the temporal lobe, with seizure-free outcomes in approximately two-thirds of cases. However, consequences of surgery can vary strongly depending on the brain region targeted for removal, as surgical morbidity and collateral damage can lead to significant complications, particularly when bleeding and swelling are located near delicate functional cortical regions. Although focal thermal ablations are well-explored in epilepsy as a minimally invasive approach, hemorrhage and edema can be a consequence as the blood-brain barrier is still disrupted. Non-thermal irreversible electroporation (NTIRE), common in many other medical tissue ablations outside the brain, is a relatively unexplored method for the ablation of neural tissue, and has never been reported as a means for ablation of brain tissue in the context of epilepsy. Here, we present a detailed visualization of non-thermal ablation of neural tissue in mice and report that NTIRE successfully ablates epileptic foci in mice, resulting in seizure-freedom, while causing significantly less hemorrhage and edema compared to conventional thermal ablation. The NTIRE approach to ablation preserves the blood-brain barrier while pathological circuits in the same region are destroyed. Additionally, we see the reinnervation of fibers into ablated brain regions from neighboring areas as early as day 3 after ablation. Our evidence demonstrates that NTIRE could be utilized as a precise tool for the ablation of surgically challenging epileptogenic zones in patients where the risk of complications and hemorrhage is high, allowing not only reduced tissue damage but potentially accelerated recovery as vessels and extracellular matrix remain intact at the point of ablation.

Optimization of surgical intervention outside the epileptogenic zone in the Virtual Epileptic Patient (VEP).

Studies to improve the efficacy of epilepsy surgery have focused on better refining the localization of the epileptogenic zone (EZ) with the aim of effectively resecting it. However, in a considerable number of patients, EZs are distributed across multiple brain regions and may involve eloquent areas that cannot be removed due to the risk of neurological complications. There is a clear need for developing alternative approaches to induce seizure relief, but minimal impact on normal brain functions. Here, we develop a personalized in-silico network approach, that suggests effective and safe surgical interventions for each patient. Based on the clinically identified EZ, we employ modularity analysis to identify target brain regions and fiber tracts involved in seizure propagation. We then construct and simulate a patient-specific brain network model comprising phenomenological neural mass models at the nodes, and patient-specific structural brain connectivity using the neuroinformatics platform The Virtual Brain (TVB), in order to evaluate effectiveness and safety of the target zones (TZs). In particular, we assess safety via electrical stimulation for pre- and post-surgical condition to quantify the impact on the signal transmission properties of the network. We demonstrate the existence of a large repertoire of efficient surgical interventions resulting in reduction of degree of seizure spread, but only a small subset of them proves safe. The identification of novel surgical interventions through modularity analysis and brain network simulations may provide exciting solutions to the treatment of inoperable epilepsies.

Controlling seizure propagation in large-scale brain networks.

Information transmission in the human brain is a fundamentally dynamic network process. In partial epilepsy, this process is perturbed and highly synchronous seizures originate in a local network, the so-called epileptogenic zone (EZ), before recruiting other close or distant brain regions. We studied patient-specific brain network models of 15 drug-resistant epilepsy patients with implanted stereotactic electroencephalography (SEEG) electrodes. Each personalized brain model was derived from structural data of magnetic resonance imaging (MRI) and diffusion tensor weighted imaging (DTI), comprising 88 nodes equipped with region specific neural mass models capable of demonstrating a range of epileptiform discharges. Each patient's virtual brain was further personalized through the integration of the clinically hypothesized EZ. Subsequent simulations and connectivity modulations were performed and uncovered a finite repertoire of seizure propagation patterns. Across patients, we found that (i) patient-specific network connectivity is predictive for the subsequent seizure propagation pattern; (ii) seizure propagation is characterized by a systematic sequence of brain states; (iii) propagation can be controlled by an optimal intervention on the connectivity matrix; (iv) the degree of invasiveness can be significantly reduced via the proposed seizure control as compared to traditional resective surgery. To stop seizures, neurosurgeons typically resect the EZ completely. We showed that stability analysis of the network dynamics, employing structural and dynamical information, estimates reliably the spatiotemporal properties of seizure propagation. This suggests novel less invasive paradigms of surgical interventions to treat and manage partial epilepsy.

Defining epileptogenic networks: Contribution of SEEG and signal analysis.

Epileptogenic networks are defined by the brain regions involved in the production and propagation of epileptic activities. In this review we describe the historical, methodologic, and conceptual bases of this model in the analysis of electrophysiologic intracerebral recordings. In the context of epilepsy surgery, the determination of cerebral regions producing seizures (i.e., the "epileptogenic zone") is a crucial objective. In contrast with a traditional focal vision of focal drug-resistant epilepsies, the concept of epileptogenic networks has been progressively introduced as a model better able to describe the complexity of seizure dynamics and realistically describe the distribution of epileptogenic anomalies in the brain. The concept of epileptogenic networks is historically linked to the development of the stereoelectroencephalography (SEEG) method and subsequent introduction of means of quantifying the recorded signals. Seizures, and preictal and interictal discharges produce clear patterns on SEEG. These patterns can be analyzed utilizing signal analysis methods that quantify high-frequency oscillations or changes in functional connectivity. Dramatic changes in SEEG brain connectivity can be described during seizure genesis and propagation within cortical and subcortical regions, associated with the production of different patterns of seizure semiology. The interictal state is characterized by networks generating abnormal activities (interictal spikes) and also by modified functional properties. The introduction of novel approaches to large-scale modeling of these networks offers new methods in the goal of better predicting the effects of epilepsy surgery. The epileptogenic network concept is a key factor in identifying the anatomic distribution of the epileptogenic process, which is particularly important in the context of epilepsy surgery.

Individual brain structure and modelling predict seizure propagation.

See Lytton (doi:10.1093/awx018) for a scientific commentary on this article.Neural network oscillations are a fundamental mechanism for cognition, perception and consciousness. Consequently, perturbations of network activity play an important role in the pathophysiology of brain disorders. When structural information from non-invasive brain imaging is merged with mathematical modelling, then generative brain network models constitute personalized in silico platforms for the exploration of causal mechanisms of brain function and clinical hypothesis testing. We here demonstrate with the example of drug-resistant epilepsy that patient-specific virtual brain models derived from diffusion magnetic resonance imaging have sufficient predictive power to improve diagnosis and surgery outcome. In partial epilepsy, seizures originate in a local network, the so-called epileptogenic zone, before recruiting other close or distant brain regions. We create personalized large-scale brain networks for 15 patients and simulate the individual seizure propagation patterns. Model validation is performed against the presurgical stereotactic electroencephalography data and the standard-of-care clinical evaluation. We demonstrate that the individual brain models account for the patient seizure propagation patterns, explain the variability in postsurgical success, but do not reliably augment with the use of patient-specific connectivity. Our results show that connectome-based brain network models have the capacity to explain changes in the organization of brain activity as observed in some brain disorders, thus opening up avenues towards discovery of novel clinical interventions.

Synchronization in networks with random interactions: theory and applications.

Synchronization is an emergent property in networks of interacting dynamical elements. Here we review some recent results on synchronization in randomly coupled networks. Asymptotical behavior of random matrices is summarized and its impact on the synchronization of network dynamics is presented. Robert May's results on the stability of equilibrium points in linear dynamics are first extended to systems with time delayed coupling and then nonlinear systems where the synchronized dynamics can be periodic or chaotic. Finally, applications of our results to neuroscience, in particular, networks of Hodgkin-Huxley neurons, are included.