The following thesis work aimed to explore the functioning of the central nervous system by exploiting the pathological condition of a split-brain patient, i.e. a subject undergoing surgical resection of the corpus callosum (CC), the main interhemispheric communication pathway. Previous studies aimed at understanding brain dominance have uncovered symptoms that reflect the atypicality of these patients in performing everyday actions. The pathological condition also generates quantifiable evidence in terms of structural connectivity (SC), defined as the presence of anatomical connections between brain regions, functional connectivity (FC), the statistical dependence between distant neurophysiological events and effective connectivity (EC), meaning the causal interaction between areas of the cerebral cortex. The analysis of FC can be conducted by acquiring functional magnetic resonance imaging (fMRI) data, which provides measurable electrical signals dependent on the level of oxygenation of the blood (BOLD signals). Since in the resting state the brain appears to be very active, making it possible to observe several resting-state (rs) networks (RSN), i.e. functionally connected brain regions (ROIs), it is convenient to acquire fMRI images in the resting state. In this thesis, healthy patient data from open-source databases were analysed, followed by fMRI data acquired in the resting state of a healthy subject and a split-brain patient. The data were processed to prepare them for connectivity analysis (preprocessing) and to reduce physiological and instrumental noise (denoising) as these could heavily influence the results obtained. The CONN toolbox allows the preprocessing, denoising and connectivity analysis procedures to be carried out and to obtain seed-based FC maps (SBC), in which each voxel is associated to a FC value with the selected brain regions, ROI-based FC matrices (RRC), which return the FC between each pair of selected brain regions. Furthermore, by applying the Dynamic Causal Modelling (DCM) method, once the connectivity model of the network of interest is specified, it is possible to study the EC to investigate the causality of the interactions. The results obtained are in agreement with the literature with regard to the SBC maps and RRC matrices, in which we note absent or scarce FC between the homotopic regions, i.e., specular between the two hemispheres, of all the RSNs with an exception for the sensorimotor cortex, consistent with the presence of axonal pathways alternative to the DC. The application of a fully connected EC model of the main RSN, the Default Mode Network (DMN), returned for the healthy subject parameters in agreement with previous investigations on control subjects and abnormalities in the split-brain patient. Future research could develop the study of EC in the split-brain patient for other RSNs and test different communication configurations between ROIs. The results obtained emphasise the importance of CC for interhemispheric FC, support the hypothesis of the participation of additional interhemispheric communication pathways in it, and point in the direction of grasping the details of complex brain processing.
Il seguente lavoro di tesi ha voluto esplorare il funzionamento del sistema nervoso centrale sfruttando la condizione patologica di un paziente split-brain, ovvero un soggetto sottoposto a resezione chirurgica del corpo calloso (CC), principale via di comunicazione interemisferica. Studi precedenti rivolti alla comprensione della dominanza cerebrale hanno portato alla luce sintomatologie che rispecchiano l’atipicità di questi pazienti nello svolgimento di azioni quotidiane. La condizione patologica genera, inoltre, evidenze quantificabili in termini di connettività strutturale (SC), intesa come la presenza di connessioni anatomiche tra regioni cerebrali, di connettività funzionale (FC), la dipendenza statistica tra eventi neurofisiologici distanti tra loro e di connettività effettiva (EC), definita come interazione causale tra aree della corteccia cerebrale. L’analisi della FC può essere condotta acquisendo immagini di risonanza magnetica funzionale (fMRI) che permettono di ottenere segnali elettrici misurabili dipendenti dal livello di ossigenazione del sangue (segnali BOLD). Dal momento che nello stato di riposo il cervello appare molto attivo, consentendo di osservare diverse resting-state (rs) network (RSN), ovvero regioni cerebrali (ROI) funzionalmente connesse tra loro, è conveniente acquisire immagini di fMRI allo stato di riposo. Nella presente tesi sono stati analizzati dati di pazienti sani prelevati da database open-source e successivamente dati fMRI acquisiti allo stato di riposo di un soggetto sano e di un paziente split-brain. I dati sono stati elaborati per prepararli all’analisi di connettività (preprocessing) e ridurre i rumori fisiologici e strumentali (denoising) in quanto questi potrebbero influire pesantemente nei risultati ottenuti. Il toolbox CONN permette di effettuare le procedure di preprocessing, denoising e analisi di connettività e di ottenere mappe di FC basate sui semi (SBC), in cui ad ogni voxel è associato un valore di FC con le regioni cerebrali selezionate, matrici di FC basate sulle ROI (RRC), che 2 restituiscono la FC tra ciascuna coppia di regioni cerebrali selezionate. Inoltre, applicando il metodo del Dynamic Causal Modelling (DCM), una volta specificato il modello di connettività della network di interesse, è possibile studiare la EC per investigare la causalità delle interazioni. I risultati ottenuti risultano concordi con la letteratura per quanto concerne le mappe SBC e matrici RRC, nelle quali notiamo assente o scarsa FC tra le regioni omotopiche, cioè speculari tra i due emisferi, di tutte le RSN con un’eccezione per la corteccia sensorimotoria, compatibilmente alla presenza di percorsi assonali alternativi al CC. L’applicazione di un modello di EC completamente connesso della principale RSN, la Default Mode Network (DMN), ha restituito per il soggetto sano parametri concordi con investigazioni precedenti su soggetti controllo e anomalie nel paziente split-brain. Future ricerche potrebbero sviluppare lo studio della EC nel paziente split-brain per le altre RSN e testare diverse configurazioni di comunicazione tra ROI. I risultati ottenuti rimarcano l’importanza del CC per la FC interemisferica, corroborano l’ipotesi della partecipazione alla stessa di ulteriori vie di comunicazione interemisferiche e vanno nella direzione di cogliere i particolari della complessa elaborazione cerebrale.
Analisi della connettività interemisferica mediante il toolbox SPM
MARCANTONI, TOMMASO
2022/2023
Abstract
The following thesis work aimed to explore the functioning of the central nervous system by exploiting the pathological condition of a split-brain patient, i.e. a subject undergoing surgical resection of the corpus callosum (CC), the main interhemispheric communication pathway. Previous studies aimed at understanding brain dominance have uncovered symptoms that reflect the atypicality of these patients in performing everyday actions. The pathological condition also generates quantifiable evidence in terms of structural connectivity (SC), defined as the presence of anatomical connections between brain regions, functional connectivity (FC), the statistical dependence between distant neurophysiological events and effective connectivity (EC), meaning the causal interaction between areas of the cerebral cortex. The analysis of FC can be conducted by acquiring functional magnetic resonance imaging (fMRI) data, which provides measurable electrical signals dependent on the level of oxygenation of the blood (BOLD signals). Since in the resting state the brain appears to be very active, making it possible to observe several resting-state (rs) networks (RSN), i.e. functionally connected brain regions (ROIs), it is convenient to acquire fMRI images in the resting state. In this thesis, healthy patient data from open-source databases were analysed, followed by fMRI data acquired in the resting state of a healthy subject and a split-brain patient. The data were processed to prepare them for connectivity analysis (preprocessing) and to reduce physiological and instrumental noise (denoising) as these could heavily influence the results obtained. The CONN toolbox allows the preprocessing, denoising and connectivity analysis procedures to be carried out and to obtain seed-based FC maps (SBC), in which each voxel is associated to a FC value with the selected brain regions, ROI-based FC matrices (RRC), which return the FC between each pair of selected brain regions. Furthermore, by applying the Dynamic Causal Modelling (DCM) method, once the connectivity model of the network of interest is specified, it is possible to study the EC to investigate the causality of the interactions. The results obtained are in agreement with the literature with regard to the SBC maps and RRC matrices, in which we note absent or scarce FC between the homotopic regions, i.e., specular between the two hemispheres, of all the RSNs with an exception for the sensorimotor cortex, consistent with the presence of axonal pathways alternative to the DC. The application of a fully connected EC model of the main RSN, the Default Mode Network (DMN), returned for the healthy subject parameters in agreement with previous investigations on control subjects and abnormalities in the split-brain patient. Future research could develop the study of EC in the split-brain patient for other RSNs and test different communication configurations between ROIs. The results obtained emphasise the importance of CC for interhemispheric FC, support the hypothesis of the participation of additional interhemispheric communication pathways in it, and point in the direction of grasping the details of complex brain processing.File | Dimensione | Formato | |
---|---|---|---|
Tesi_MARCANTONI_FINALE.pdf
embargo fino al 25/10/2026
Dimensione
2.49 MB
Formato
Adobe PDF
|
2.49 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/15302