Figure 5.

Schematic of network architecture at baseline and at twelve months, responders vs. nonresponders. ANOVA of network loadings (principal components of network measures) demonstrated a statistically significant interaction between treatment status and time, indicating an alteration in network measures following treatment in those that responded clinically, and suggesting a link between electrophysiology and clinical effect. There was no statistically significant effect of time alone. Interestingly, there was evidence of an effect of responder independent of time, suggesting baseline differences in groups that were not apparent on clinical assessment. This may allow EEG biomarkers to be developed which are able to identify patients who are likely to respond to treatment. Detailed visualisations of network measures are shown in Figures S11–S14.