Le Trouble Dissociatif de l’Identité (TDI) apparaît dans la IV ^e  version du « Manuel Diagnostique et Statistique des troubles mentaux » (DSM) en 1994, et dans la 11 ^e  version de la Classification Internationale des Maladies (CIM) en 2022. Il était appelé « Trouble de personnalité multiple » dans leurs versions précédentes. Le TDI est caractérisé par la présence de deux ou plusieurs états de personnalité distincts au sein d’une même personne, chacun ayant ses propres schémas de pensée, ses comportements, et ses façons de percevoir le monde. Ces états alternent de manière récurrente, entraînant des lacunes dans la continuité de la mémoire, de la conscience, et de l’identité de la personne. Malgré cette reconnaissance internationale, plus de la moitié des psychiatres doutent encore de sa réalité (enquête réalisée lors du Congrès de l’Encéphale en 2023). Le modèle fantasmatique du TDI (patients hypnotisables, influençables, simulateurs) est encore privilégié au modèle traumatique (le TDI est une forme grave de stress post-traumatique lié à des sévices dans la petite enfance).

Afin de permettre une meilleure reconnaissance du TDI et une prise en charge des patients plus appropriée, des arguments plus ancrés dans les sciences sont nécessaires.

Cet article regroupe les conclusions de plusieurs études de neuro-imagerie structurelle, fonctionnelle et de connectivité, réalisées avec des patientes souffrant de TDI, de stress post-traumatique et des témoins sains entre 2014 et 2022, et utilisant l’IRM structurelle, l’IRM fonctionnelle ou la tomographie par émission de positrons.

Les patientes TDI peuvent être différenciées des témoins sains tant sur des anomalies structurelles que sur des anomalies fonctionnelles. Ces constatations sont renforcées par des anomalies de connectivité retrouvées chez les patientes TDI dans et entre les réseaux exécutif central, de saillance et de mode par défaut.

Les études de neuro-imagerie démontrent clairement l’existence du TDI. Le modèle traumatique du TDI doit prévaloir sur le modèle fantasmatique afin d’améliorer le diagnostic et la prise en charge de ce trouble complexe.

Dissociative Identity Disorder (DID) appeared in version IV of the Diagnostic and Statistical Manual of Mental Disorders (DSM) in 1994, and in version 11 of the International Classification of Diseases (ICD) in 2022. It was called “Multiple Personality Disorder” in their previous versions. DID is characterized by the presence of two or more distinct personality states within the same person, each with its own patterns of thought, behavior and ways of perceiving the world. These states alternate recurrently, resulting in gaps in the continuity of the person's memory, consciousness and identity. Despite this international recognition, more than half of psychiatrists still doubt its reality (survey conducted at the Congrès de l’Encéphale in 2023). The phantasmatic model of DID (hypnotizable, suggestible, malingerers) is still favored over the traumatic model (DID is a severe form of post-traumatic stress disorder linked to early severe abuse).

To enable better recognition of DID and more appropriate management of patients, more scientifically grounded arguments are needed.

This article brings together the findings of several structural, functional and connectivity neuroimaging studies conducted with female patients with DID, post-traumatic stress disorder (PTSD) and healthy controls between 2014 and 2022, using either structural MRI, functional MRI or positron emission tomography.

Patients with DID can be differentiated from healthy controls on the basis of structural and functional brain abnormalities. Structural abnormalities (Blihar et al., Chalavi et al. and Chalavi et al.): smaller cortical and subcortical volumes in the hippocampus, amygdala, parietal and frontal structures; larger white matter tracts responsible for communication between somatosensory association areas, basal ganglia and precuneus; greater striatal volume in patients with DID than in those with PTSD; much smaller volume of left inferior parietal cortex in DID patients than in PTSD patients and controls; greater reduction in hippocampal volume in DID patients than in PTSD patients and controls; abnormal shape, in both DID and PTSD patients, in different areas of the hippocampus, i.e. CA1, CA2, CA3, CA4, dentate gyrus, and subiculum; the use of probabilistic pattern recognition shows that the brain phenotype can distinguish DID patients from healthy controls with a sensitivity of 72% and a specificity of 74% (Reinders, Marquand et al.). The hypothesis that DID is a severe form of PTSD (indeed, it's virtually impossible to find a DID patient without a history of PTSD) is thus confirmed by the neuro-structural abnormalities common to both disorders, and sometimes quantitatively greater in DID than in PTSD, compared with healthy controls. Functional abnormalities : Lanius and his team have identified functional features that distinguish two forms of PTSD: simple, or excitatory, post-traumatic stress disorder (reliving the traumatic experience, intrusive symptoms, flashbacks, tachycardia, suggestive of emotional hypo-modulation): reduced activity of the medial prefrontal cortex and anterior cingulate cortex, with increased activation of the amygdala. There is thus a reduction in prefrontal inhibition of limbic regions; dissociative post-traumatic stress disorder (emotional detachment, symptoms of derealization and depersonalization, absence of tachycardia, suggestive of emotional hyper-modulation): increased activity of medial and lateral prefrontal cortex, anterior cingulate cortex, medial and inferior frontal cortex and superior and medial temporal cortex. The reduction in amygdala activity thus reflects an increase in prefrontal inhibition of limbic regions. A 2014 study (Reinders, Willemsen et al.) comparing regional cerebral blood flow in women with DID and women simulating DID, when reading neutral and traumatic scripts, showed that the identity-state-dependent patterns of neural activation in DID are the same as those in the two subtypes of excitatory and dissociative PTSD found in Lanius's neurobiological model: The traumatic script produced hypoactivation, i.e. emotional overmodulation, when the DID patients were in a neutral identity state (Apparently Normal Part); The same traumatic script produced hyper-activation, i.e. emotional under-modulation, when the DID patients were in a state of trauma-related identity (Emotional Part). DID patients therefore possess at least two different identity states, involving different subjective reactions, cardiovascular responses and brain activation patterns to a trauma-related memory scenario. These identity states are known as emotional states, alters or dissociative states. Each identity is characterized by its own pattern of perception, reaction and thought, and presents different psychobiological characteristics that the DID controls were unable to reproduce. Healthy female controls simulating DID were unable to simulate the psychobiological profiles of the neutral and trauma-related identities. Other studies have explored working memory in female DID patients compared with healthy controls attempting to simulate DID, and show that (Vissia et al.): activation in DID patients of the prefronto-parietal network, involved in working memory, is greater in the neutral identity state than in the trauma-related identity state; DID simulators are unable to simulate the neural activation patterns associated with working memory in DID individuals. Connectivity abnormalities : These findings are reinforced by the connectivity abnormalities found in DID patients within and between the central executive, salience and default mode networks. The central executive network essentially manages cognitive processes such as reasoning, attention, planning, execution and memory. The salience network enables the transition from the central executive network to the default network, or vice versa. It enables emotion to be felt on the basis of information from the internal environment. As for the default network, it is activated when we give free rein to our thoughts, without interacting with our environment. Among other things, it enables us to remember past events and simulate future ones. A 2022 study (Lebois et al.) looked at 65 women with dissociative post-traumatic stress disorder or dissociative identity disorder, compared with 26 healthy controls, neither suffering from post-traumatic stress nor with any psychiatric history. This study shows that dissociative identity disorder is characterized by hyperconnectivity within each of these three networks, hyperconnectivity between the executive and default networks, and reduced connectivity between the executive and salience networks.

Neuroimaging studies clearly demonstrate the existence of DID. The traumatic model of DID must prevail over the phantasmatic model in order to improve the diagnosis and management of this complex disorder.