Thesis Estimación de función respuesta hemodinámica a una muestra amplia de sujetos sanos
Loading...
Date
2025-10
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Electrónica
Departament
Campus
Campus Casa Central Valparaíso
Abstract
La función respuesta hemodinámica (HRF) constituye un elemento central en el análisis de señales de resonancia magnética funcional (fMRI), dado que describe la relación entre la actividad neuronal y las variaciones en el flujo sanguíneo cerebral. Sin embargo, determinar una HRF representativa en sujetos sanos sigue siendo un desafío, debido a la variabilidad interindividual y a las limitaciones de los modelos canónicos. El objetivo general de este trabajo es estimar un conjunto representativo de funciones de respuesta hemodinámica (HRFs) en una muestra amplia de sujetos sanos. Para ello, se plantearon tres objetivos específicos: (i) seleccionar un conjunto de bases de datos públicas de fMRI, (ii) desarrollar un pipeline de procesamiento para obtener HRFs a partir de dichas bases, y (iii) caracterizar las HRFs en función de los atributos de los sujetos. La metodología se basa en el preprocesamiento de imágenes fMRI mediante SPM12 en MATLAB, junto con la implementación de un modelo Fuzzy GLM y técnicas de mínimos cuadrados para la estimación de las HRFs. Asimismo, se utiliza la función doble gamma como base analítica para describir la dinámica temporal de la respuesta. Los resultados lograron evidenciar la variabilidad que hay intersujeto e intrasujeto, tanto en comparaciones como teslaje como estímulo, como entre datasets.
The hemodynamic response function (HRF) is a central element in the analysis of functional magnetic resonance imaging (fMRI) signals, as it describes the relationship between neuronal activity and variations in cerebral blood flow. However, determining a representative HRF in healthy subjects remains a challenge due to interindividual variability and the limitations of canonical models. The general objective of this work is to estimate a representative set of hemodynamic response functions (HRFs) from a large sample of healthy subjects. To this end, three specific objectives were defined: (i) to select a set of publicly available fMRI datasets, (ii) to develop a processing pipeline to extract HRFs from these datasets, and (iii) to characterize the HRFs as a functionof subject attributes. The methodology is based on fMRI preprocessing using SPM12 in MATLAB, combined with the implementation of a Fuzzy GLM model and least-squares techniques for HRF estimation. In addition, the double-gamma function is employed as an analytical basis to describe the temporal dynamics of the response. The results showed the variability present both between subjects and within subjects, in comparisons such as tessellation and stimulus, as well as across datasets.
The hemodynamic response function (HRF) is a central element in the analysis of functional magnetic resonance imaging (fMRI) signals, as it describes the relationship between neuronal activity and variations in cerebral blood flow. However, determining a representative HRF in healthy subjects remains a challenge due to interindividual variability and the limitations of canonical models. The general objective of this work is to estimate a representative set of hemodynamic response functions (HRFs) from a large sample of healthy subjects. To this end, three specific objectives were defined: (i) to select a set of publicly available fMRI datasets, (ii) to develop a processing pipeline to extract HRFs from these datasets, and (iii) to characterize the HRFs as a functionof subject attributes. The methodology is based on fMRI preprocessing using SPM12 in MATLAB, combined with the implementation of a Fuzzy GLM model and least-squares techniques for HRF estimation. In addition, the double-gamma function is employed as an analytical basis to describe the temporal dynamics of the response. The results showed the variability present both between subjects and within subjects, in comparisons such as tessellation and stimulus, as well as across datasets.
Description
Keywords
Resonancia magnética funcional (fMRI), Función respuesta hemodinámica (HRF), Preprocesamiento, Mínimos cuadrados, Actividad neuronal, Función doble gamma, Modelo Fuzzy GLM
