bridge centrality
play

BRIDGE CENTRALITY: IDENTIFYING BRIDGE SYMPTOMS IN PSYCHOPATHOLOGY - PowerPoint PPT Presentation

BRIDGE CENTRALITY: IDENTIFYING BRIDGE SYMPTOMS IN PSYCHOPATHOLOGY NETWORKS Payton Jones Harvard University Masters Thesis Presentation Mental Disorder Comorbidity 1 6 3 4 7 8 2 9 5 10 Depression Social Anxiety Bridge Nodes 1


  1. BRIDGE CENTRALITY: IDENTIFYING BRIDGE SYMPTOMS IN PSYCHOPATHOLOGY NETWORKS Payton Jones Harvard University Master’s Thesis Presentation

  2. Mental Disorder Comorbidity 1 6 3 4 7 8 2 9 5 10 Depression Social Anxiety

  3. Bridge Nodes 1 6 3 4 7 8 2 9 5 10

  4. Social Anxiety Depression

  5. MOTIVATING QUESTION: CAN WE DETECT BRIDGE NODES IN COMPLEX NETWORKS?

  6. BRIDGE CENTRALITY bridge strength bridge betweenness bridge closeness bridge expected influence

  7. Bridge Strength 1 6 3 4 7 8 2 9 5 10

  8. Bridge Betweenness 1 6 3 4 7 8 2 9 5 10

  9. Bridge Betweenness 1 6 3 4 7 8 2 9 5 10

  10. Bridge Closeness 1 6 3 4 7 8 2 9 5 10

  11. Bridge Expected Influence 1 6 3 4 7 8 2 9 5 10

  12. IS BRIDGE CENTRALITY ACCURATE?

  13. Study 1: Detection in simulations Trials: 1. Adding noise 2. Adding causal direction 3. Adding uncertainty about clusters 2 x 2 x 2 x 500 = 4,000 simulations

  14. Sensitivity Specificity 8% misses 15% false alarms 85% correct 92% hits rejections

  15. IS BRIDGE CENTRALITY USEFUL?

  16. Study 2: Simulation of contagion 1. Start with one community activated 2. Allow the activation to spread over many iterations

  17. Study 2: Simulation of contagion 1. Start with one community activated 2. Allow the activation to spread over many iterations 3. Treat a limited number of nodes 4. Bridge centrality vs. previous metrics

  18. Control 2 Bridge Centrality Control 1

  19. DOES BRIDGE CENTRALITY WORK IN REAL LIFE SITUATIONS?

  20. RE-ANALYSIS: • 18 published networks • Bridge symptoms reported by researchers • Bridge symptoms indicated by bridge centrality

  21. HOW CAN I USE BRIDGE CENTRALITY?

  22. BRIDGE CENTRALITY: BEYOND MENTAL DISORDERS? • Social Anxiety – Depression • Grief – Growth • Borderline Personality – Narcissistic Personality • State-Trait Anxiety – Neuroticism • Cognitive – Affective (Attitudes)

  23. GENERATE A NETWORK • network <- qgraph(cor(data), • graph=“ glasso ”, • sampleSize=1000)

  24. DEFINE COMMUNITIES • communities <- c(rep(“1”, 11), rep(“2”, 9)) 1 2 2 1 1 1 2 1 2 1 2 1 2 1 1 2 2 1 1 2

  25. BRIDGE CENTRALITY • b <- bridge(network, communities) • plot(b, include=“Bridge Strength”)

  26. THANK YOU! Read the preprint: https://osf.io/c5dkj/ Use the R package: networktools

Recommend


More recommend