Skip to main content

Table 2 Statistical descriptions and p values for all 7 selected features

From: Detection of mild cognitive impairment in Parkinson’s disease using gradient boosting decision tree models based on multilevel DTI indices

ID

Feature type

Brain region

PD-MCI (n = 68)

PD-NC (n = 52)

P value

1

LDHk

PCT

0.807 ± 0.050

0.819 ± 0.048

0.186

2

MD

CH. R

(0.699 ± 0.066) ×10 − 3

(0.673 ± 0.064) ×10 − 3

0.035*

3

LDHs

Fornix

0.783 ± 0.118

0.801 ± 0.122

0.430

4

LDHs

ML.L

0.827 ± 0.082

0.842 ± 0.060

0.250

5

LDHs

SLF.R

0.814 ± 0.053

0.820 ± 0.058

0.603

6

LDHk

ST. L

0.717 ± 0.068

0.739 ± 0.069

0.088

7

MD

RIC.R

(0.798 ± 0.075) ×10 − 3

(0.778 ± 0.086) ×10 − 3

0.191

  1. The ID numbering order is consistent with the SHAP value. “.R and.L” in the text indicate the right and left sides, respectively. All data are represented as the mean ± SD, and we applied two-sample t tests for comparisons between groups. *P < 0.05. PD-MCI Parkinson’s disease with mild cognitive impairment, PD-NC Parkinson’s disease with normal cognition, LDHk LDH using Kendall’s coefficient concordance, LDH LDH using Spearman’s rank correlation coefficient, MD mean diffusivity, RIC retrolenticular part of the internal capsule, PCT pontine crossing tract, ML medial lemniscus, ST fornix (cres)/stria terminalis, SLF Right superior longitudinal fasciculus, CH cingulum (hippocampus)