Figure 6

The R-parameter quantitatively characterizes ASD severity. Plots (A–D) of R-parameter vs. psychiatric test scores. (A) R-parameter is possitively correlated with IQ score (16 subjects). Data points are labeled by the subject index according to their R value rank. The black line denotes a linear regression fit. The significance p value for the Pearson Correlation between the R-parameter and IQ is 0.0680 (0.0104 if excluding the outlier subject #2 in the plot.) The significance p value for Kendal Rank correlation between R-parameter and IQ is 0.0714 (0.0208 if excluding subject #2). (B) The R-parameter is possitively correlated with Vineland Adaptive Behavior Composite score (17 subjects). The significance p value for the correlation between the R-parameter and Vineland Composite Score is 0.0064 for Pearson Correlation and 0.0044 for Kendal Rank Correlation. (C) R-parameter is negatively correlated with ADI-R sum scores (17 subjects). The significance p value for the correlation between R-parameter and ADI-R sum score is 0.0663 for Pearson Correlation (0.0346 if excluding subject #14) and 0.0569 for Kendal Rank Correlation (0. 0189 if excluding subject #14). (D) The R parameter is negatively correlated with ADOS combined score in module 2–4 (11 subjects). The ADOS combined score is calculated as a linear combination of ADOS Sect. B-D scores and ADOS module number. The weights were obtained from the Canonical Correlation analysis: 2SectB + 11SectC + 9SectD-5 ModuleNumber. The significance p value for the Kendal Rank Correlation between R-parameter and the ADOS combined number is 0.0264. (E) Further validation comes from including ASD subjects with labels given by the schools as low functioning or high functioning: a clear separation is found between high functioning (6 subjects) from ASD functioning (8 subjects).