Ujjwal Pasupulety’s Post

View profile for Ujjwal Pasupulety, graphic

Data Scientist @ Frank OCD Lab | AI for Mental Health | MS-CS @ USC

One Man Journal Club : Day 17 Adaptive Deep Brain Stimulation(DBS) This study tested a dual algorithm for adaptive DBS to account for sleep effects on biomarkers in Parkinson's patients. 4 Parkinson's patients had electrodes implanted in subcortical and cortical areas. One detector used cortical alpha and theta power to track sleep state with a single threshold. Another detector tracked Parkinsonian motor signs using subcortical biomarkers relative to two thresholds. Detector parameters were optimized using patient training data across wake/sleep and medication cycles. Performance was tested using algorithm state logs downloaded daily from patients at home over 47 days. The sleep detector achieved high concordance (mean 88%) with patient reported sleep state. It successfully switched to sleep mode during sleep across 47 days of testing. The motor sign detector adjusted stimulation based on medication fluctuations when patients were awake. Discriminating sleep is crucial for chronic embedded adaptive DBS, as biomarkers are influenced by sleep. This study demonstrates a crucial development in the field of adaptive deep brain stimulation for neurological and psychiatric indications. The dual detector design elegantly overcomes this issue by allowing independent tracking of sleep and disease states on different timescales appropriate for each. Enabling adaptive DBS systems to respond intelligently to sleep could improve outcomes and quality of life for patients. #neurosurgery #deepbrainstimulation #adaptivelearning #parkinsons #mentalhealth

Ujjwal Pasupulety

Data Scientist @ Frank OCD Lab | AI for Mental Health | MS-CS @ USC

9mo

The full paper

  • No alternative text description for this image
Like
Reply

To view or add a comment, sign in

Explore topics