Network dynamics of the brain and influence of the epileptic seizure onset zone

Abstract Graph neural networks (GNNs) and the attention mechanism are two of the most significant advances in artificial intelligence methods over the past few years. The former are neural networks able to process graph-structured data, while the latter learns to selectively focus on those parts of the input that are more relevant for the task at hand. In this paper, we propose a methodology for seizure localisation which combines the two approaches. Our method is composed of several blocks. First, we represent brain states in a compact way by computing functional networks from intracranial electroencephalography recordings, using metrics to quantify the coupling between the activity of different brain areas. Then, we train a GNN to correctly distinguish between functional networks associated with interictal and ictal phases. The GNN is equipped with an attention-based layer which automatically learns to identify those regions of the brain (associated with individual electrodes) that are most important for a correct classification. The localisation of these regions is fully unsupervised, meaning that it does not use any prior information regarding the seizure onset zone. We report results both for human patients and for simulators of brain activity. We show that the regions of interest identified by the GNN strongly correlate with the localisation of the seizure onset zone reported by electroencephalographers. We also show that our GNN exhibits uncertainty on those patients for which the clinical localisation was also unsuccessful, highlighting the robustness of the proposed approach.

Download Full-text

Influence analysis for high-dimensional time series with an application to epileptic seizure onset zone detection

Journal of Neuroscience Methods ◽

10.1016/j.jneumeth.2012.12.025 ◽

2013 ◽

Vol 214 (1) ◽

pp. 80-90 ◽

Cited By ~ 7

Author(s):

Christoph Flamm ◽

Andreas Graef ◽

Susanne Pirker ◽

Christoph Baumgartner ◽

Manfred Deistler

Keyword(s):

Time Series ◽

Epileptic Seizure ◽

High Dimensional ◽

Seizure Onset ◽

Influence Analysis ◽

Seizure Onset Zone

Download Full-text

The fence post depth electrode technique to control both brain tumors and epileptic seizures in patients with brain tumor-related epilepsy

Surgical Neurology International ◽

10.25259/sni_241_2019 ◽

2019 ◽

Vol 10 ◽

pp. 187 ◽

Cited By ~ 1

Author(s):

Yosuke Masuda ◽

Ayataka Fujimoto ◽

Mitsuyo Nishimura ◽

Keishiro Sato ◽

Hideo Enoki ◽

...

Keyword(s):

Brain Tumor ◽

Epileptic Seizure ◽

Epileptic Seizures ◽

Seizure Onset ◽

Tumor Removal ◽

Depth Electrodes ◽

Fence Post ◽

Group 2 ◽

The Brain ◽

Group 1

Background: To control brain tumor-related epilepsy (BTRE), both epileptological and neuro-oncological approaches are required. We hypothesized that using depth electrodes (DEs) as fence post catheters, we could detect the area of epileptic seizure onset and achieve both brain tumor removal and epileptic seizure control. Methods: Between August 2009 and April 2018, we performed brain tumor removal for 27 patients with BTRE. Patients who underwent lesionectomy without DEs were classified into Group 1 (13 patients) and patients who underwent the fence post DE technique were classified into Group 2 (14 patients). Results: The patients were 15 women and 12 men (mean age, 28.1 years; median age 21 years; range, 5–68 years). The brain tumor was resected to a greater extent in Group 2 than Group 1 (P < 0.001). Shallower contacts showed more epileptogenicity than deeper contacts (P < 0.001). Group 2 showed better epilepsy surgical outcomes than Group 1 (P = 0.041). Conclusion: Using DEs as fence post catheters, we detected the area of epileptic seizure onset and controlled epileptic seizures. Simultaneously, we removed the brain tumor to a greater extent with fence post DEs than without.

Download Full-text

Epileptic seizure propagation from the second somatic sensory area to the fronto-medial region, by insular redistribution. A case report and a connectome description

Journal of Epileptology ◽

10.1515/joepi-2015-0023 ◽

2015 ◽

Vol 23 (1) ◽

pp. 61-67

Author(s):

Attila Balogh ◽

Péter Halász ◽

Dániel Fabó ◽

Lóránd Erőss

Keyword(s):

Case Report ◽

Epilepsy Surgery ◽

Epileptic Seizure ◽

Motor Area ◽

Seizure Onset ◽

3T Mri ◽

Sensory Area ◽

Seizure Onset Zone ◽

Video Eeg

SUMMARY Introduction. The seizure propagation phenomenon by inducing remote symptoms brings several difficulties in finding the seizure onset and delineating the epileptic network which should be taken into consideration in epilepsy surgery. By demonstrating a difficult (MRI negative) epilepsy surgery case explored with invasive presurgical evaluation we highlight the importance to recognise the secondary sensory area and to explore the the parieto-opercular-insular-medial frontal network in certain cases. A further conclusion is the consideration of the redistributory role of the insula as a special structure in the cerebral connectome, having a role in epileptic network organisation. Aims. To support the role of the insula in the organisation of an opercular – medial frontal epileptic network and to confirm Penfield’s the “second somatic sensory leg area” by way of a case report. We try to give an up to date exploration of our patient’s remote epileptic seizures by way of a connectome. Methods. The epileptic disorder was studied with intensive video EEG monitoring and two times 3T MRI. Interictal FDG (fluorodeoxyglucose) PET was also undertaken. Beside the scalp EEG and computerized frequency analysis, the evaluation was performed by invasive EEG with 2 grids and 2 strips and an insular deep electrode in addition. Electrical cortical stimulation and cortical mapping were also undertaken. Results. The video-EEG study revealed the complex seizure semiology. The left sided global somatosesensory aura in the leg, followed supplementary motor area manifestations represented a remote seizure. The seizure onset zone and the symptomatogenic zone were localised by the invasive electrophysiology. With the insular deep electrode we succeeded to explore the propagation of ictal activity to the insula and later to frontal medial surface. The PET, the negative 3T MRI results and the postprocessing morphometry confirmed the lesional origin and localised the epileptogenic area to the second somato-sensory field where a dysgenesis was located. Conclusions. By preoperative invasive video-EEG evaluation, the second somato-sensory leg area was delineated as the seizure onset zone. The resection of this area by IIb type cortical dysgenesis, resulted in a complete relief of the seizures. The invasive video-EEG revealed the peculiar role of the insula in the propagation of the epileptic seizure from the second sensory leg area to the ipsilateral fronto-medial supplemetary motor area. Our results, confirm, that the insula has a relay or node function on the parietal opercular-fronto-medial epileptic network. The connectome of the insula is a further additive of the scale-free features of the remote epileptic networks.

Download Full-text

The rate of contradictory lateralization of the epileptic seizure onset zone between ictal and interictal brain perfusion SPECT

10.1055/s-0041-1726755 ◽

2021 ◽

Author(s):

I Apostolova ◽

M Jaber ◽

J Taherpour ◽

S Stodieck ◽

S Klutmann ◽

...

Keyword(s):

Epileptic Seizure ◽

Brain Perfusion ◽

Seizure Onset ◽

Brain Perfusion Spect ◽

Seizure Onset Zone ◽

Perfusion Spect

Download Full-text

SozRank: A new approach for localizing the epileptic seizure onset zone

PLoS Computational Biology ◽

10.1371/journal.pcbi.1005953 ◽

2018 ◽

Vol 14 (1) ◽

pp. e1005953 ◽

Cited By ~ 10

Author(s):

Yonathan Murin ◽

Jeremy Kim ◽

Josef Parvizi ◽

Andrea Goldsmith

Keyword(s):

Epileptic Seizure ◽

Seizure Onset ◽

New Approach ◽

Seizure Onset Zone

Download Full-text

The Insula: A Stimulating Island of the Brain

Brain Sciences ◽

10.3390/brainsci11111533 ◽

2021 ◽

Vol 11 (11) ◽

pp. 1533

Author(s):

Inès Rachidi ◽

Lorella Minotti ◽

Guillaume Martin ◽

Dominique Hoffmann ◽

Julien Bastin ◽

...

Keyword(s):

Anterior Insula ◽

Homologous Region ◽

Seizure Onset ◽

Functional Brain Mapping ◽

Axonal Conduction ◽

Seizure Onset Zone ◽

Clinical Responses ◽

History Of ◽

Posterior Insula ◽

The Brain

Direct cortical stimulation (DCS) in epilepsy surgery patients has a long history of functional brain mapping and seizure triggering. Here, we review its findings when applied to the insula in order to map the insular functions, evaluate its local and distant connections, and trigger seizures. Clinical responses to insular DCS are frequent and diverse, showing a partial segregation with spatial overlap, including a posterior somatosensory, auditory, and vestibular part, a central olfactory-gustatory region, and an anterior visceral and cognitive-emotional portion. The study of cortico-cortical evoked potentials (CCEPs) has shown that the anterior (resp. posterior) insula has a higher connectivity rate with itself than with the posterior (resp. anterior) insula, and that both the anterior and posterior insula are closely connected, notably between the homologous insular subdivisions. All insular gyri show extensive and complex ipsilateral and contralateral extra-insular connections, more anteriorly for the anterior insula and more posteriorly for the posterior insula. As a rule, CCEPs propagate first and with a higher probability around the insular DCS site, then to the homologous region, and later to more distal regions with fast cortico-cortical axonal conduction delays. Seizures elicited by insular DCS have rarely been specifically studied, but their rate does not seem to differ from those of other DCS studies. They are mainly provoked from the insular seizure onset zone but can also be triggered by stimulating intra- and extra-insular early propagation zones. Overall, in line with the neuroimaging studies, insular DCS studies converge on the view that the insula is a multimodal functional hub with a fast propagation of information, whose organization helps understand where insular seizures start and how they propagate.

Download Full-text

The critical role of spiny stellate cells in seizure onset based on dynamic analysis of a neural mass model

10.1101/2021.07.19.452876 ◽

2021 ◽

Author(s):

Saba Tabatabaee ◽

Fariba Bahrami ◽

Mahyar Janahmadi

Keyword(s):

Bifurcation Analysis ◽

Epileptic Seizure ◽

Critical Role ◽

Stellate Cells ◽

Neural Mass Model ◽

Seizure Onset ◽

Mass Model ◽

Neural Mass ◽

The Brain

Increasing evidence has shown that excitatory neurons in the brain play a significant role in seizure generation. However, spiny stellate cells are cortical excitatory non-pyramidal neurons in the brain which their basic role in seizure occurrence is not well understood. In the present research, we study the critical role of spiny stellate cells or the excitatory interneurons (EI), for the first time, in epileptic seizure generation using an extended neural mass model introduced originally by Taylor and colleagues in 2014. Applying bifurcation analysis on this modified model, we investigated the rich dynamics corresponding to the epileptic seizure onset and transition between interictal and ictal states due to the EI. Our results indicate that the transition is described by a supercritical Hopf bifurcation which shapes the preictal activity in the model and suggests why before seizure onset, the amplitude and frequency of neural activities increase gradually. Moreover, we showed that 1) the altered function of GABAergic and glutamatergic receptors of EI can cause seizure, and 2) the pathway between the thalamic relay nucleus and EI facilitates the transition from interictal to the ictal activity by decreasing the preictal period. Thereafter, we considered both sensory and cortical periodic inputs to drive the model responses to various harmonic stimulations. Our results from the bifurcation analysis of the model suggest that the initial stage of the brain might be the main cause for the transition between interictal and ictal states as the stimulus frequency changes. The extended thalamocortical model shows also that the amplitude jump phenomenon and nonlinear resonance behavior result from the preictal stage of the brain. These results can be considered as a step forward to a deeper understanding of the mechanisms underlying the transition from brain normal activities to epileptic activities.

Download Full-text