Objective: To compare traditional initial deep brain stimulation (DBS) programming with artificial intelligence (AI)-assisted automated image-guided DBS programming algorithm.

Background: DBS programming is a complex, time-intensive process requiring expert clinical adjustments to optimize therapeutic outcomes. Traditionally, initial DBS programming relies on clinician experience and trial-and-error adjustments, which can lead to variability in patient outcomes and prolonged optimization periods. Recent advancements in AI have introduced AI-assisted DBS programming, which leverage machine learning algorithms to streamline initial parameter selection.

Method: We initially assessed each patient with traditional initial monopolar programming in 4 Parkinson’s disease patients, 3 implanted in the STN and 1 in the GPi. Best stimulation parameters were then compared with those obtained with AI-assisted algorithm (Illumina 3D, Boston Scientific). This technology automatically generates patient-specific stimulation settings by selecting target/avoid regions based on neuroanatomic visual software. Resulting settings were compared in terms of active contact configuration and amplitude of stimulation. Clinical improvement was measured after both procedures using the Movement Disorders Society Unified Parkinson’s Disease Rating Scale.

Results: Traditional initial programming resulted in 8 monopolar configurations (6 ring, 2 segmented) with average amplitude of 2.08mA (0.5 SD). AI-assisted algorithm generated 8 semi-bipolar configurations (1 ring, 7 segmented) with average amplitudes of 2.46mA (0.9SD). Pulse width and frequency were kept constant across the two approaches. Clinical outcomes were equivalent after applying the two programming methods. AI-assisted algorithm, which provides settings in a matter of seconds, resulted in significantly shorter programming sessions than the traditional approach.

Conclusion: Traditional initial programming and AI-assisted algorithm generated different electrode configurations and stimulation settings. While clinical improvement was comparable between the two methodology, AI-assisted algorithm was associated with much shorter programming sessions.

« Back to 2025 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/traditional-deep-brain-stimulation-programming-versus-automated-image-guided-algorithm-in-patients-with-parkinsons-disease/