Table 3.
Performance Across Demographics. (B) Performance of the Sleep Staging Model on the MGH Wireless Dataset for Different Demographics
| Performance across demographics, on MGH dataset, using wireless signals as inputs | |||
|---|---|---|---|
| Demographic | No. of samples | Accuracy (%) | Cohen’s Kappa |
| All | 849 | 80.4 ± 8.2 | 0.681 ± 0.140 |
| Gender | |||
| Male | 480 | 80.4 ± 7.9 | 0.679 ± 0.138 |
| Female | 363 | 80.2 ± 8.6 | 0.681 ± 0.144 |
| Age | |||
| Young (<40) | 202 | 81.5 ± 8.7 | 0.709 ± 0.138 |
| Middle (40,60) | 287 | 81.0 ± 8.2 | 0.693 ± 0.137 |
| Senior (≥60) | 360 | 79.2 ± 7.7 | 0.655 ± 0.141 |
| Race | |||
| Asian | 35 | 78.6 ± 7.9 | 0.648 ± 0.135 |
| Black | 64 | 79.0 ± 8.3 | 0.659 ± 0.145 |
| White | 639 | 80.7 ± 8.1 | 0.684 ± 0.141 |
| Others | 111 | 80.1 ± 8.3 | 0.683 ± 0.135 |
NA means “not applicable.” All MROS participants are males with age ≥60.
Sex groups: male, female
Age groups: young (age < 40), middle (40 ≤age < 60), senior (age ≥ 60)
Race groups: Asian, black, white, and others