Thesis defense – Achille Gillig

Venue: Bâtiment A29, Amphithéâtre A, Campus Peixotto

Thesis defended in english

Achille Gillig

Neurofunctional Imaging Group,
Institut des Maladies Neurodégénératives

Supervisor: Marc Joliot

Title

Functional neuroimaging of the resting state in humans: characterization of the individual brain organization in networks

Abstract

Over the past three decades, functional neuroimaging has profoundly changed how the relationship between brain organization and cognition is studied. While classical approaches sought to localize cognitive functions in circumscribed regions, accumulating evidence has shifted the field toward a network-based understanding of brain function. Resting-state functional MRI has been central to this shift by showing that, even without an explicit task, the brain remains intrinsically organized into coherent resting-state networks. Because these networks partly resemble systems recruited during perception, action, cognition, and emotion, they have often been interpreted as a possible intrinsic architecture supporting mental processes.

However, the exact functional meaning of this intrinsic architecture remains unsettled. In particular, it is still unclear whether resting-state networks should be understood as fundamental cognitive modules involved in the offline maintenance of specific cognitive processes. This thesis addressed this question by investigating the relevance of intrinsic brain organization into resting-state networks for cognition, interindividual differences in behavior, and dementia-related cognitive decline.

The first contribution was to provide a more precise framework for interpreting the cognitive relevance of resting-state networks. To this aim, we introduced a 33-network resting-state atlas and characterized each network using meta-analytic decoding and consensus-derived cognitive labeling. This aimed to move beyond informal cognitive labels commonly assigned to resting-state networks, by providing empirically grounded associations between intrinsic networks and task-related cognitive processes. It therefore offered a framework in which the proposed cognitive relevance of resting-state networks can be tested.

The second contribution challenged the assumption that resting-state networks correspond to strictly specific cognitive modules. Using resting-state functional connectivity prediction of multiple behavioral measures, we showed that connectivity predictive of behavior did not primarily organize at the level of fine-grained cognitive processes, but around broader latent dimensions including Cognition, Positive Affect, and Negative Affect. Crucially, only connectivity predictive of Cognition was associated with a segregated intrinsic architecture, characterized by stronger within-network connectivity and reduced between-network connectivity. This pattern was particularly expressed in networks previously associated with higher-level cognitive processes. These results suggest that cognition relates to a flexible modular intrinsic architecture expressed across multiple networks.

The third contribution extended this framework to dementia-related cognitive decline. In memory-clinic patients, we showed that dementia-related steep cognitive decline was associated with global network dedifferentiation, together with a more selective loss of within-networks connectivity. This strengthens the association between network segregation and efficient cognition, while suggesting that dementia alters cognition through two complementary network-level mechanisms, with a global dedifferentiation of large-scale systems and a more selective disruption of the internal cohesiveness of specific networks.

Taken together, these studies support a nuanced view of the cognitive relevance of resting-state networks. The intrinsic architecture of the brain does not appear to map onto cognition as a set of fixed, process-specific modules. Rather, resting-state networks appear to constitute a behaviorally meaningful large-scale architecture whose segregated organization supports cognition and whose dedifferentiation accompanies cognitive decline.

Key words

Resting-state fMRI; functional connectivity; resting-state networks; cognition; predictive modeling; dementia.

Jury

• M. JOLIOT Marc Directeur de recherche, IMN, Université de Bordeaux, CNRS, CEA, Bordeaux – Directeur de thèse
• M. WASSERMANN Demian Directeur de recherche, Centre Inria de Saclay, Université Paris-Saclay, Inria, CEA, Palaiseau – Rapporteur
• M. BARTOLOMEO Paolo Directeur de recherche, Paris Brain Institute, Sorbonne Université, Inserm, CNRS, Paris – Rapporteur
• Mme LOPEZ-PERSEM Alizée Chargée de recherche, Sorbonne Université, Institut du Cerveau – Paris Brain Institute – ICM, Inserm, CNRS, AP-HP, Hôpital de la Pitié Salpêtrière, Paris – Examinatrice
• Mme CHANRAUD Sandra Maîtresse de conférences, INCIA, CNRS, Université de Bordeaux, Bordeaux – Examinatrice
• M. JOBARD Gaël Maître de conférence, IMN, Université de Bordeaux, CNRS, CEA, Bordeaux – Invité

• Place
  Bâtiment A29
  
• Dates
  On 16 October 2026 at 14:30