EEG Topologies with Neo4j: Interpolation-Driven Graph Learning for Chronic Pain Detection
summary
Implements a Neo4j mult-tier EEG graph that maps a canonical sensor topology to study sessions, uses Cypher inverse-distance interpolation to create edges between mismatched electrode layouts, and runs both cloud and edge inference pipelines. Tested first on one balanced pain/healthy dataset then on larger open cohorts, the system preserves regional connectivity and absolute band-power signatures and retains discriminative signal with fewer electrodes, enabling portable GNN training and low-resource pain detection.
AI-written summary · about
chapters (8)
- 0:00Introduction and pain detection challenge
- 1:14EEG data harmonization with Neo4j
- 3:28Variability and topology representation
- 5:14Multi-tier graph model in Neo4j
- 7:01Interpolation for device alignment
- 8:25Testing and validation results
- 10:20Responsible data handling and ethics
- 12:13Collaboration and future steps
related talks
13mEnhanced EEG Data Classification with 3D SNN and On-Line Learning ESN
32mIrina Rish, Research Staff, IBM T.J. Watson Research Center
42mGraph Theory Takes on the Opioid Epidemic
29mMaking Neuroimage Analysis Easier to Understand
11mConnected brains
29mEnhancing Business Process Anomaly Detection With Neo4j and Graph Neural Networks
09mLightning Talk: Bridging the Gap: Engineering Compliant...
28mRole of Knowledge Graphs in Mental Health Diagnosis and Cure