Spersonalizowane symulacje wirtualnego mózgu i wyniki stymulacji nerwu błędnego u dzieci z epilepsją: badanie wieloośrodkowe z projektu CONNECTiVOS
Personalized Virtual Brain Simulations and Vagus Nerve Stimulation Outcomes in Pediatric Epilepsy: A Multicenter Study From the CONNECTiVOS Collaboration
W skrócie
Badacze zbadali, czy komputerowe modele indywidualnego mózgu dziecka mogą przewidzieć, czy implant stymulujący nerw błędny pomoże w leczeniu trudnej do opanowania epilepsji. Odkryli, że dzieci, u których leczenie działało, miały inne cechy elektryczne w mózgu niż te, u których nie działało, szczególnie w obszarach odpowiedzialnych za hamowanie aktywności mózgowej. Wyniki sugerują, że spersonalizowane komputerowe modele mózgu mogą w przyszłości pomóc lekarzom wybrać właściwe leczenie dla każdego dziecka.
Oryginalny abstract (angielski)
OBJECTIVE: Although vagus nerve stimulation (VNS) is the most common implantable therapy for drug-resistant epilepsy (DRE), there is a need to develop novel methods to predict treatment response before surgery. An integrated approach combining structural and functional neuroimaging data to infer biophysical brain features at the mesoscopic scale could provide insights into brain-network differences related to response. The Virtual Brain (TVB) offers a computational framework to simulate individual subject brain dynamics using biophysical neural mass models to estimate synaptic transmission properties in the brain. This study aims to (i) assess the ability of TVB-based modeling to capture individual brain dynamics in pediatric DRE, (ii) evaluate the relationship between predictability of brain dynamics using a biophysical approach and VNS outcome, and (iii) identify regional inhibitory features as potential biomarkers of VNS response. MATERIALS AND METHODS: Preimplantation functional and diffusion magnetic resonance imaging data were acquired and used to construct personalized virtual brains using the reduced Wong-Wang model. The model was optimized for individual participants, and both local and global parameters of the models were compared between responders and nonresponders to VNS. RESULTS: A total of 38 children with DRE undergoing VNS implantation were included in this multicenter study, including 16 responders and 22 nonresponders. Predictability of the functional connectivity using the biophysical model was significantly lower in nonresponders, implicating atypical brain dynamics in VNS outcomes. TVB simulations showed stronger inhibitory synaptic weights in critical regions of the vagal afferent network in responders compared with nonresponders, including the thalamus, cingulate, and frontal cortices. CONCLUSIONS: This study constitutes the first work exploring TVB in pediatric DRE and VNS. Personalized brain dynamics simulations revealed distinct inhibitory patterns between responders and nonresponders, offering new insights into the interindividual variability in VNS outcomes and mechanisms of action of VNS therapy.