Table 5 Comparison of angle algorithm performance for the BBVS dataset by the limb on which the device was worn. All participants wore devices on their dominant (dw) and non-dominant (ndw) wrist as well as on their thigh. The best performance metrics were obtained for the non-dominant wrist device, but thigh wearables gave the least time differences overall in terms of total sleep time (TST), sleep onset and offset. BBVS TST for diaries mean ± 95% CI = 7.739 ± 0.073 hours (464.34 ± 4.38 minutes).
From: Detecting sleep outside the clinic using wearable heart rate devices
Sleep parameter | Metric | Angle change algo. (ndw) | Angle change algo.(dw) | p-value | Angle change algo. (Thigh) | p-value |
|---|---|---|---|---|---|---|
Value (mean ± 95% CI) | Value (mean ± 95% CI) | ndw-dw | Value (mean ± 95% CI) | ndw-thigh | ||
Total sleep time | Time difference (min.) | 222.64 ± 7.78 | 218.96 ± 7.92 | 0.271 | 214.90 ±8.08 | 0.048 |
MSE | 0.16 ± 0.00 | 0.16 ± 0.00 | 0.052 | 0.16 ± 0.00 | 0.291 | |
Cohen’s kappa | 0.58 ± 0.01 | 0.59 ± 0.01 | 0.027 | 0.58 ± 0.01 | 0.650 | |
Sleep onset | Time difference (min.) | −100.99 ± 6.63 | −96.00 ± 7.01 | 0.167 | −96.07 ± 7.85 | 0.250 |
Sleep offset (Wake Up) | Time difference (min.) | 121.65 ± 7.05 | 122.96 ± 7.43 | 0.727 | 118.82 ± 8.11 | 0.501 |