A large body of evidence points to impairment in the decoding of emotional stimuli in Parkinson's disease (PD). These changes seem to be related to dysfunction of the mesocorticolimbic projections leading to dysfunction of the limbic cortico-subcortical loop that, throughout the basal ganglia, projects to limbic cortical regions such as the anterior cingulate cortex and the orbitofrontal cortex, which are known to be involved in emotional processing. Thus, presumably the occurrence of non-motor psychic symptoms and fluctuations in PD would also rely on the degeneration of this circuitry. Dopamine modulation of emotions has been frequently reported both in healthy subjects or PD patients, but the effect of subthalamic nucleus deep brain stimulation (STN-DBS) over mood symptoms is more controversial and there is still a paucity of data demonstrating its role in emotional processing. The objective of our study is to assess and compare the behavioral and electrophysiological effects of both dopamine and STN-DBS in PD during emotional processing. To do so, the Emotional Stroop task, a version of the classical Stroop test developed to investigate inhibition of emotional interference, will be performed by STN-DBS PD patients in four different treatment conditions while high resolution electroencephalographic cortical mapping is conducted. Using this approach, the investigators expect to better separate the mechanisms of psychic features related to either disease, dopaminergic treatment, subthalamic stimulation or the association of both, with the final aim of optimizing the clinical management of PD.

Non-motor symptoms of Parkinson's disease (PD) and particularly mood manifestations such as apathy, depression and anxiety have gained attention in the last few decades due to its high frequency and impact in patient's quality of life. In addition, these psychic non-motor features have been shown to fluctuate alike motor symptoms do, concurring to non-motor fluctuations (NMF) that can be more disabling than motor symptoms. The physiopathological basis of psychic NMF is unclear but mesolimbic dopaminergic denervation has been suggested to underlie mood manifestations during off-periods . At the same time, it has been demonstrated that PD patients suffer from abnormal emotional processing deficits in facial expression recognition and emotional prosody especially concerning negative valanced emotions, and that these impairments could also be linked to mesolimbic dopaminergic mechanisms. The mesolimbic pathway is one of the functional cortico-subcortical loops in which basal ganglia are segregated. It lies in the ventral subportion of each nucleus (ventromedial in the case of the subthalamic nucleus) and through the mediodorsal thalamus projects to the orbito-frontal cortex and the anterior cingulate cortex (ACC), limbic cortical areas that robust evidence has involved in the pathophysiology of mood disturbances and emotional processes like facial expression recognition or emotional conflict resolution. In summary, in the same way that dopamine depletion of the nigrostriatal pathway is at the origin of the parkinsonian motor triad, the lack of dopamine in the mesolimbic loop would account for non-motor psychic manifestations including apathy, anxiety, depression and disturbances in emotional decoding. A proper way to explore these phenomena is through tasks that involve exposure to affective stimuli.

The Stroop Test is a tool that allows exploring selective attention, cognitive conflict resolution (inhibition of irrelevant reading automatism) and processing speed, and is usually applied to evaluate executive functions. A modified version, the facial Emotional Stroop (ES), was developed in order to address the interference of irrelevant affective distractors. With this task, the authors demonstrated in healthy volunteers, that dorsolateral prefrontal cortex and the amygdala are the areas implicated in emotional conflict monitoring whereas rostral anterior cingulate cortex (rACC) activation associates with emotional conflict resolution through a top-down inhibition of the amygdala in order to avoid the interference of new irrelevant emotional distracters. The same group designed a variant of the task by adding non-emotional cognitive trials in which subjects had to identify the gender of the faces instead of the affect. Through this paradigm, they were able to dissociate neural nets involved in cognitive conflict resolution from those specifically involved in emotional conflict resolution. Hence, while lateral prefrontal cortex (LPFC) was shown to resolve non-emotional conflict, the pregenual part of the anterior cingulate cortex, that is, the rACC is involved in emotional conflict resolution. On the contrary, the detection of both types of conflict share activation of a common region of the dorsal anterior cingulate cortex.

By using the ES task, our group has recently demonstrated in a functional MRI study with PD patients that dopamine modulates emotional conflict resolution by "normalizing" the activity of rACC, which was hypoactivate in the off-drug state compared to the on-drug state and to healthy controls. The rACC receives massive dopamine projections from the ventral tegmental area (VTA), and thus is part of the cortico-subcortical limbic loop. Therefore, these findings would support the implication of this circuitry in emotional decoding disturbances in PD and match with the hypothesis proposed that dopaminergic mesolimbic degeneration would underlie mood manifestations and psychic NMF of the disorder. It is well demonstrated that STN-DBS improves levodopa-sensitive motor symptoms of PD. STN-DBS does not change overall cognitive function, except for some impairment in executive functions like verbal fluency but, in an experimental scenario, it has been proved to increase impulsivity during cognitive tasks in PD patients. However, its clinical effect in the affective domain of PD is highly controversial and there is only very limited data regarding emotional processing. A few available studies comparing patients before and after surgery point to a worsening in facial emotion and prosody recognition. These findings could not be replicated in a more recent work. Such studies however do not take into account the desensitization of the dopaminergic system related to marked decrease in medication after STN DBS. Because of the major changes in dopaminergic treatment and chronic desensitization all studies which do not compare stimulation conditions should be interpreted very cautiously. In our opinion, due to the large available clinical data that demonstrate behavioral changes after STN-DBS, and considering the anatomic overlapping of STN functional subdivisions, it is likely that stimulation would exert some effect over the limbic basal ganglia loop even with electrodes that are well implanted in the sensorimotor territory and thus, induce changes in both motor and affective domains, even though with some dissociation.

With this rationale, the investigators aim to reproduce in an electrophysiological perspective the functional magnetic resonance imaging (fMRI) evidences of limbic cortical regions involvement in emotional decoding processes and to differentiate them from those involved in cognitive conflict. In addition, the investigators want to confirm dopamine modulation when dealing with emotional conflict, and elucidate STN-DBS effect in behavioral and cortical electrophysiological terms.

Inclusion Criteria:

• PD patients with STN-DBS
• Healthy controls without neurologic pathology
• Age between 18 and 70 years
• Fluent in French
• Signed informed consent
• Patients covered by national health insurance

Exclusion Criteria:

• Subjects concerned by articles L1121-5 to L1121-8 from Public Health Code.
• Patients under psychotropic treatments without significant adverse event.
• MATTIS < 130
• For Healthy controls MRI anormality