Back

Validation of non-contact sensor quantification of heart rate and respiratory rate dynamics using real-world pretraining and label-efficient fine-tuning on polysomnograms

Gupta, K. S.; Harrington, N.; Pedros-Valls, R.; DeYoung, P.; Owens, R. L.; Orr, J. E.; King, K. R.

2026-07-04 respiratory medicine
10.64898/2026.07.01.26356016 medRxiv
Show abstract

Non-contact mechanical bed sensors can passively and longitudinally monitor the dynamics of cardiopulmonary physiology to detect changes from patient-specific baselines and facilitate care. This requires accurate longitudinal quantification of established metrics like respiratory rate (RR) and heart rate (HR) from the underlying raw waveforms, validated against ground truth labeled datasets like simultaneous polysomnography (PSG). Whereas head-to-head labeled datasets are scarce and costly to collect, unlabeled real-world datasets are often abundant. Here, we show that non-optimized heuristic algorithms can be used to soft-label large real-world data (>40M minutes across >50,000 nights) for pretraining models. This enables label-efficient fine-tuning on small numbers of head-to-head PSG-labeled datasets while maximizing generalizability and robustness to hyperparameters. The result is a highly performant validated model with mean absolute errors (MAE) of 0.6 brpm for RR and 1.1 bpm for HR across 1-minute windows. Although demonstrated in the context of bed sensor cardiopulmonary quantification, these methods are applicable to development of sensor algorithms whenever labeled ground truth data is scarce and unlabeled real-world data is abundant.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
npj Digital Medicine
118 papers in training set
Top 0.2%
23.3%
2
Scientific Reports
3612 papers in training set
Top 1%
15.7%
3
Nature Communications
5641 papers in training set
Top 23%
7.0%
4
PLOS ONE
5266 papers in training set
Top 27%
5.7%
50% of probability mass above
5
Proceedings of the National Academy of Sciences
2444 papers in training set
Top 12%
4.2%
6
Communications Medicine
113 papers in training set
Top 0.9%
3.4%
7
iScience
1154 papers in training set
Top 6%
3.4%
8
IEEE Journal of Biomedical and Health Informatics
37 papers in training set
Top 0.4%
2.9%
9
Frontiers in Physiology
106 papers in training set
Top 0.7%
2.5%
10
eBioMedicine
183 papers in training set
Top 1%
2.5%
11
PLOS Digital Health
106 papers in training set
Top 2%
2.5%
12
IEEE Access
35 papers in training set
Top 0.7%
1.8%
13
Journal of Medical Internet Research
87 papers in training set
Top 2%
1.6%
14
Nature Machine Intelligence
70 papers in training set
Top 2%
1.2%
15
Royal Society Open Science
214 papers in training set
Top 5%
1.1%
16
eLife
5828 papers in training set
Top 61%
1.0%
17
Computers in Biology and Medicine
128 papers in training set
Top 4%
1.0%
18
Frontiers in Neuroscience
256 papers in training set
Top 6%
1.0%
19
Nature Medicine
125 papers in training set
Top 3%
1.0%
20
Nature
645 papers in training set
Top 10%
0.9%
21
Biomedical Signal Processing and Control
22 papers in training set
Top 0.6%
0.9%
22
Journal of Neural Engineering
221 papers in training set
Top 2%
0.9%
23
Artificial Intelligence in Medicine
17 papers in training set
Top 0.6%
0.9%
24
Advanced Intelligent Systems
11 papers in training set
Top 0.3%
0.6%
25
Journal of Biomedical Informatics
47 papers in training set
Top 1%
0.6%
26
BMC Medical Informatics and Decision Making
43 papers in training set
Top 2%
0.6%
27
Sleep
58 papers in training set
Top 0.7%
0.6%
28
PLOS Computational Biology
1863 papers in training set
Top 23%
0.5%
29
European Respiratory Journal
59 papers in training set
Top 2%
0.5%
30
Biology Methods and Protocols
61 papers in training set
Top 3%
0.5%