Skip to main content

Table 5 The correlation coefficients between estimated sleep parameters and ground truth before energy score thresholding for each dataset

From: InsightSleepNet: the interpretable and uncertainty-aware deep learning network for sleep staging using continuous Photoplethysmography

 

MESA

CFS

CAP

TST

r = 0.969, r2 = 0.938

r = 0.913, r2 = 0.834

r = 0.810, r2 = 0.654

Light sleep fraction

r = 0.671, r2 = 0.450

r = 0.796, r2 = 0.768

r = 0.827, r2 = 0.684

Deep sleep fraction

r = 0.270, r2 = 0.073

r = 0.889, r2 = 0.790

r = 0.936, r2 = 0.877

REM sleep fraction

r = 0.809, r2 = 0.573

r = 0.781, r2 = 0.611

r = 0.790, r2 = 0.624