Post-stroke cognitive impairment (PSCI) is a common source of invisible disability. The actual techniques of cognitive remediation have been reported to have only a small effect, justifying the need to foster research in this field. Real-time functional MRI neurofeedack (rt-fMRI NF) is a procedure of brain-machine interface supporting self-regulation of brain activity, with promising results in the treatment of mental disorders. The main objective of the study is to evaluate in a pilot study, the ability of stroke patients to learn to self-regulate their brain activity in a priviledged direction, using rt-fMRI NF

Stroke is a major source of cognitive impairment including slowdown of processing speed, executive functions disturbances, attention and memory loss, that can lead to invisible disability, even before the ultimate stage of dementia. The traditional techniques of cognitive remediation have been reported to have a small effect, and european scientific societies stated on the need to foster research in the field of PSCI, with the goal of preventing cognitive decline. PSCI mechanisms arising from functional MRI studies include a disconnexion of brain networks, with a loss of segregation between networks and integration within networks, centered on the frontoparietal network. Rt-fMRI NF is a psychophysiological procedure of brain-machine interface enabling to a subject to learn to self-regulate his brain activity in a desired direction, based on the brain haemodynamic response. Promising results in the treatment of mental disorders have been reported using rt-fMRI NF, but data are scarce in the field of cognitive rehabilitation after stroke. Based on the mechanisms of PSCI reported in fMRI studies, the goal of the present study is to evaluate the ability of stroke patients to learn to self-regulate a brain target using rt-fMRI NF. The brain target is defined by the difference of activation between the central executive network (CEN, or frontoparietal network) and the default mode network (DMN), known to be highly anticorrelated with the CEN in the healthy brain. The experimental procedure will include four training sessions of rt-fMRI NF (1/week), preceded by a localizer session for the identification of the target networks, and followed by a tranfer session where the subjects will be asked to try to self-regulate their brain activity but without receiving feedback. A control group will perform the four fMRI sessions but without receiving feedback. The two groups will receive cognitive remediation therapy as treatment as usual. Resting state scans and diffusion tensor imaging seqences will be performed before and after the intervention, to measure the changes in functional and structural connectivity. A neuropsychological assessment will be also performed before and after the intervention, to measure the changes in cognitive performances.

Inclusion Criteria:

• Men or women
• Right handed
• 18 ≤ years old < 75
• first-ever ischemic stroke on brain MRI in the past 3 to 6 months ;
• Persisting cognitive complaint 3 months after the ischemic stroke, proved by a MoCA score ≤ 26 (if educational level ≥ 12 years, or ≤ 25 if educational level < 12 years), and involving at least one of the following cognitive domains : executive functions, attention, and processing speed ;
• Ischemic stroke located in the territory of the middle cerebral artery and /or basal ganglia/thalamus ;
• NIHSS ≤ 5, without aphasia or severe visual disability empeding the understanding of the instructions, the neurofeedback task achievement, and/or the neuropsychological assessment ;
• No prestroke disability defined by a modified Rankin Scale ≤ 2 ;
• For women of childbearing potential, use of effective contraception and negative pregnancy test.
• Written informed consent by the patient.
• Coverage by the French Social Insurance

Exclusion Criteria:

• Previous cognitive impairment ;
• History of psychiatric disorder ;
• Current use of psychotropic medication ;
• MRI contraindication ;
• Pregnancy or breastfeeding ;
• Chronic disease empeding the follow-up of the subject during the lenght of the study ;
• Non fluently-French speakers ;
• Patient under legal protection.
• Patient unable of giving personal consent
• Emergency situation
• 1st MRI incomplete, not performed or not analyzable