Figure 3.

Plots of standardized loadings of CCA analysis relating sleep disturbances and multidomain predictors in the ABCD study two-year follow-up data. We performed a similar CCA analysis on ABCD data from the two-year follow-up visit to assess the consistency of findings observed in baseline data. The y-axis lists the predictor variables (middle) and their respective categories (left), and the x-axis shows standardized beta weight values for these predictors. Error bars represent the 95% CI based on bootstrap analysis. Statistical significance is indicated by darkly shaded bars. On the first CV (left column), we once again saw multidomain psychopathology as the strongest predictor of difficulty initiating and maintaining sleep and excessive daytime somnolence. For the second CV (right column), we once again observed that BMI was strongly predictive of sleep breathing disorders. Abbreviations: CCA, canonical correlation analysis; RAVLT, Rey Auditory Verbal Learning Test; ER, emergency room.