Parkinson's disease is the second most common neurodegenerative disease, after Alzheimer's disease. In France, more than 270,000 people live with Parkinson's disease, and this number could double by 2050. For some patients, deep brain stimulation (DBS) can significantly reduce tremors and dyskinesias that hinder daily life. This procedure involves implanting electrodes in deep areas of the brain, such as the subthalamic nucleus or the globus pallidus, depending on the case. However, determining the most relevant target to achieve the desired therapeutic effect while minimizing the stimulation of neighboring tissues remains a major challenge. Discrepancies of a few millimeters can strongly influence the outcome, as the human brain exhibits individual anatomical variations. Professor Jean Régis, head of the functional neurosurgery department at Timone in Marseille and president of the WSSFN 2026 congress, emphasized the need to precisely locate a "sweet spot" offering the best compromise between safety and effectiveness. A French startup, Rebrain, is working on this challenge by training an AI on imaging data and clinical data from patients who have previously undergone successful surgery. From anatomical landmarks identified on MRI scans, supervised mathematical models analyze variability between patients and propose an area of interest that could help surgeons better determine the target. Rebrain, created in 2021, is the culmination of a decade of work conducted at the University Hospital of Bordeaux and with Inria on this targeting strategy. The strategy is now being evaluated in the PARKEO 2 clinical trial, a multicenter and randomized trial involving 128 patients with Parkinson's disease. It compares a targeting strategy using this approach to the strategy usually used in the participating centers for deep brain stimulation. The study evaluates, after one year, the effectiveness of the targeting strategy, its accuracy, its safety, and its impacts for patients, surgical teams, and the healthcare system. Results will be revealed on October 1st during a symposium at the WSSFN 2026 congress in Marseille, before a scientific publication. These results could allow the deployment of the technology and its extension to other indications of functional neurosurgery, notably essential tremor. Rebrain also indicates developing new models for targeting the internal globus pallidus, notably in dystonia.