Multimodal MRI Study of Human Brain Connectivity: Cognitive Networks

Abstract

Aquesta tesi inclou 6 estudis en els quals s’han utilitzat diversos mètodes d’imatge per ressonància magnètica (IRM) per caracteritzar la connectivitat cerebral i la seva relació amb l’estat cognitiu en joves sans, persones d’edat avançada i pacients amb malaltia d’Alzheimer. Les tècniques d’IRM permeten estudiar la connectivitat estructural a partir del Tensor de Difusió (DTI), i la connectivitat funcional mitjançant la IRM funcional (IRMf), mesurada durant l’execució d’una tasca o durant el repòs. S’ha descrit un conjunt de xarxes neuronals, que inclou la xarxa neuronal per defecte (o Default Mode Network, DMN).

Utilitzant IRMf en un grup de joves sans, es va trobar que la connectivitat de la DMN, durant el repòs previ a una tasca correlacionava amb el rendiment. A més, l’estat de les xarxes implicades en la tasca també correlacionà amb l’execució.

En un altre grup de joves sans, es van estudiar les xarxes implicades en la memòria de treball amb estímuls facials o espaials. Es van descriure regions i connexions específiques per la memòria de treball amb cares, que implicaren el fusiforme i la regió frontal inferior. Posteriorment, es van estudiar les xarxes de la memòria de treball en l’envelliment. Els subjectes d’edat avançada amb correcta execució mostraren increments d’activitat en regions frontals bilaterals, indicant mecanismes compensatoris.

També es varen estudiar patrons de connectivitat global, utilitzant IRM en repòs i graph-theory en un grup de 98 subjectes d’edat avançada. L’edat correlacionà amb un decrement de la connectivitat funcional en les connexions llargues i un increment en les curtes. Els subjectes més grans presentaren més clustering funcional, sobretot en regions frontals i parietals, indicant un increment de la segregació. Les mesures obtingudes es van relacionar amb la memòria visual i verbal.

A continuació es vàren estudiar els patrons d’interacció de les xarxes neuronals també en el context de l’envelliment. La connectivitat entre els components de la DMN correlacionà negativament amb l’edat, mentre que la connectivitat entre diferents sistemes augmentà amb l’edat. Aquestes interaccions es relacionaren amb la memòria i les funcions executives.

Finalment, es va utilitzar la IRMf en subjectes amb una mutació en el gen Presenilin-1, causant de la malaltia d’Alzheimer. Es van trobar alteracions funcionals en la DMN i en una tasca de memòria episòdica, fins i tot en aquells subjectes que estaven lluny de l’edat d’inici de la malaltia.

En resum, hem pogut demostrar la utilitat de la IRMf per a estudiar xarxes de connectivitat cerebral i hem pogut concloure que aquestes correlacionen amb l’estat cognitiu i amb l’execució de les tasques. A més, s’han descrit alteracions en la connectivitat durant l’envelliment sà i en la malaltia d’Alzheimer, fins i tot abans de l’aparició de símptomes clínics. La DMN apareix com un sistema altament implicat en cognició en subjectes joves sans, però també com el sistema més vulnerable tant en l’envelliment com en la malaltia d’Alzheimer.

INTRODUCTION: This thesis has been elaborated as a compendium of 6 research studies, in which we have used a variety of methods related with Magnetic Resonance Imaging (MRI) with the objective to characterize brain connectivity and its relationship with cognition in young and aged subjects and in preclinical Alzheimers Disease (AD).

Brain Connectivity refers to any pattern of links connecting different areas of the brain. It can be stud­ied at its functional level, by using functional MRI (fMRI), which measures the statistical dependence between brain activity at different regions, or at its structural level, with Diffusion Tensor Imaging (DTI), which estimates the directionality of white matter .ber tracts.

FMRI can be studied during the execution of a task (task-fMRI) or during resting state (rs-fMRI). With task-fMRI, it is possible to obtain patterns of brain co-activations, measured by the Blood-Oxigen Level Dependent (BOLD) signal. On the other hand, rs-fMRI allows measuring patterns of brain con­nectivity, known as Resting State Networks (RSNs), as correlations between spontaneous oscillations. Functional connectivity has been used to describe a set of networks in the brain, including the Default Mode Network (DMN), which is deactivated during task and shows high activity levels at rest.

RESULTS: In the .rst study, we included 16 healthy young subjects, under rs-fMRI and during a Working Memory (WM) task-fMRI. The connectivity of the DMN at rest correlated with the perfor­mance in the subsequent task. During task-fMRI, we described the DMN and the WM network. The connectivity between these networks was highly negative in the most demanding blocks and during .xation, which also correlated with performance. In addition, within-network connectivity increased with cognitive demands.

In the second study, we analyzed the networks involved in two different WM tasks, with visuoper­ceptive (faces) and visuospatial (squares) stimuli. We used task-fMRI and DTI. We found that the fusiform and the inferior frontal gyrus were selectively activated for faces, and that the DTI measures of the tracts connecting these regions correlated with task performance for the facial-WM.

In the third study, we analyzed a sample of young and old subjects during a WM task. We found patterns of increased activity in the WM networks in aging that were related to successful performance, indicating compensatory mechanisms. The results of this study supported and extended previous research on fMRI and aging.

In study 4, we included a sample of 98 healthy elders, during rs-fMRI. We used graph-theory to an­alyze whole-brain patterns of connectivity. Age correlated with a loss of functional connectivity in long-range connections and an increase in functional connectivity in short-range connections. These changes resulted in increases in clustering coef.cient and larger average minimum path-length of the global network, indicating higher functional segregation and less integration. Increases in clustering were located in frontal and parietal regions and they correlated negatively with visual and verbal memory functions.

In study 5, we used a sample of 76 healthy elders during rs-fMRI and we analyzed functional interac­tions between the components of the main RSNs. The anterior and posterior components of the DMN were less connected in older subjects, and the connectivity between different networks increased with age, indicating alterations in the global architecture of functional networks. RSN-interactions corre­lated with the results in memory and executive functions. Finally, in the last study, we included a sample of asymptomatic and symptomatic carriers of the PSEN1 mutation, which causes Familial AD. We found altered functional connectivity and brain ac­tivity at rest and during memory encoding that appear even before the onset of AD.

CONCLUSIONS: Overall, we proved the usefulness of fMRI to study brain connectivity networks and we concluded that connectivity signals correlate with cognitive status and with task performance. In addition, we described alterations in connectivity during the healthy aging process and in AD, even before the disease onset. Finally, we identi.ed the DMN as the system showing the highest correlations in healthy subjects but also as a core target for aging and AD.