Fig. 5: Behavioral prediction models using individualized tOC. | Communications Medicine

Fig. 5: Behavioral prediction models using individualized tOC.

From: Individualized cerebellar damage predicts the presence of behavioral disorders in children with brainstem tumors

Fig. 5

A, B Prediction of behaviors in glioma patients using ridge regression models based on tOC. Scatter plots show the relationship between tOC and social problems (A), as well as withdrawal behaviors (B). Each dot represents an individual sample, with observed behavioral scores plotted against tOC values. The solid line depicting the ridge regression fit and a shaded area indicating the 95% confidence interval. The Pearson correlation coefficient (r) quantifies the linear relationship between tOC and behavioral scores. Prediction of social problems (C) and withdrawal behaviors (D) in previously unseen individuals using tOC (total outlier count) with five-fold cross-validation. Points represent test samples (observed values on the x-axis; predicted values on the y-axis), with colors distinguishing folds. Lines depict fold-specific regression fits, and shaded areas indicate confidence intervals. Prediction accuracy is measured by the mean r across folds, with significant p-values for both behaviors. tOC the total outlier count.

Back to article page