Fig. 4: Classifying participants with ND-CNVs and unaffected controls from individual MEG functional networks using graph theory. | Translational Psychiatry

Fig. 4: Classifying participants with ND-CNVs and unaffected controls from individual MEG functional networks using graph theory.

From: Electrophysiological network alterations in adults with copy number variants associated with high neurodevelopmental risk

Fig. 4

a Classification performance for the three groups, using different metrics to characterize the networks (eccentricity, degree, betweenness centrality, global and local efficiency, clustering coefficient, and a pooled model combining all features). Above-chance accuracies (permutation testing) are marked with 1, 2, and 3 dots respectively for p < 0.05, p < 0.01, and p < 0.001. b How well are different ND-CNVs classified? The plot shows the mean predicted label for each of the 42 participants with ND-CNVs across 100 cross-validation iterations using pooled node features. Participants with 22q11.2 deletions are most consistently correctly identified. Two participants with schizophrenia diagnoses are marked with “S”. All error bars are ±SD.

Back to article page