Skip to main content
. 2024 Jan 31;14:2570. doi: 10.1038/s41598-024-52910-x

Figure 2.

Figure 2

Optimized diffusional kurtosis metrics performance by receiver operating characteristic (ROC) curves. ROC curves were used to differentiating targeted diseased group. Ranked by AUC values, performance of optimized kurtosis metrics was displayed in (A) differentiating SIVD from AD by WMA [AUC = .964 (95% CI 0.932–0.997)], WMA + THA [AUC = .958 (95% CI 0.918–0.998)], and THA [AUC = .909 (95% CI 0.847–0.971)], (B) differentiating SIVD from NC by WMA [AUC = .987 (95% CI 0.969–1.000)], WMA + THA [AUC = .971 (95% CI 0.942–1.000)], and THA [AUC = .965 (95% CI 0.930–1.000)], and (C) differentiating AD from NC by WMA + THA [AUC = .953 (95% CI 0.908–0.997)], WMA [AUC = .937 (95% CI 0.886–0.988)], and THA [AUC = .885 (95% CI 0.881–0.958)]. WMA = White matter atlas. THA = Segregated thalamus analysis. SIVD = Subcortical ischemic vascular disease. AD = Alzheimer's disease. NC = normal cognition. WMA = White Matter Atlas. T = Segregated thalamus. ROC = Receiver operating characteristic curves. AUC = Area under curve. CI = Confidence interval.