Publicación: Análisis funcional de redes cerebrales a través de nuevos paradigmas computacionales en Spiking Neural Networks
Cargando...
Fecha
2017-10-06
Autores
Editor/a
Director/a
Tutor/a
Coordinador/a
Prologuista
Revisor/a
Ilustrador/a
Derechos de acceso
Atribución-NoComercial-SinDerivadas 4.0 Internacional
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Nacional de Educación a Distancia (España). Escuela Técnica Superior de Ingeniería Informática. Departamento de Inteligencia Artificial
Resumen
El análisis conjunto de datos cerebrales multimodal (EEG, MEG, fMRI, DTI) está llamado a ser una de las grandes estrategias de búsqueda de nuevos neuro-biomarcadores en la medicina personalizada y en los futuros avances del sector. Sin embargo, hoy en día deben tratarse por separado al no existir un marco común de análisis. Por otro lado, las técnicas clásicas de análisis de redes cerebrales empiezan a presentar serios problemas para estudiar las variaciones en los patrones espaciotemporales de actividad y conectividad funcional. Este trabajo propone un nuevo entorno computacional basado en conceptos de Inteligencia Artificial, como las Redes Neuronales Pulsantes, donde los distintos tipos de señales son codificados a un dominio común para su tratamiento: los spikes. Gracias a este nuevo sistema bio-inspirado, nuevas reglas de aprendizaje sinápticos como STDP pueden ser utilizadas para capturar los patrones espacio-temporales dinámicos en la conectividad anatómico-funcional del cerebro.
The multimodal analysis of brain data (EEG, MEG, fMRI, DTI, etc.) is one of the most promising strategies to detect new neuro-biomarkers in the personalized medicine and in the future advances in the clinical sector. However, nowadays there is a lack of such multimodal platform that allows combining all this information. Moreover, is reported that the classical methods used to assess the dynamic fluctuations in the functional connectivity are not efficient and they imply some assumptions that have been proven incorrect, like stationary fluctuations. This work proposes a new computational framework based on Artificial Intelligence approaches, such as the Spiking Neural Networks, to analyze spatio-temporal patterns in the brain. The essential step in this proof-of-concept involves the coding of the activity in the brain into a new domain: the spikes. This allows the algorithm to combine different types of data and to use bio-inspired learning rules, creating a completely unsupervised multimodal platform to capture dynamic functional connectivity in the brain.
The multimodal analysis of brain data (EEG, MEG, fMRI, DTI, etc.) is one of the most promising strategies to detect new neuro-biomarkers in the personalized medicine and in the future advances in the clinical sector. However, nowadays there is a lack of such multimodal platform that allows combining all this information. Moreover, is reported that the classical methods used to assess the dynamic fluctuations in the functional connectivity are not efficient and they imply some assumptions that have been proven incorrect, like stationary fluctuations. This work proposes a new computational framework based on Artificial Intelligence approaches, such as the Spiking Neural Networks, to analyze spatio-temporal patterns in the brain. The essential step in this proof-of-concept involves the coding of the activity in the brain into a new domain: the spikes. This allows the algorithm to combine different types of data and to use bio-inspired learning rules, creating a completely unsupervised multimodal platform to capture dynamic functional connectivity in the brain.
Descripción
Categorías UNESCO
Palabras clave
Citación
Centro
E.T.S. de Ingeniería Informática
Departamento
Inteligencia Artificial