Overview The MIMIC PERform datasets contain physiological signals recorded from critically-ill patients during routine clinical care. Specifically, the datasets contain the following signals: electrocardiogram (ECG) photoplethysmogram (PPG) impedance pneumography (imp), also known as respiratory (resp) (in some cases) arterial blood pressure (abp) They also contain some fixed parameters, stored in the 'fix' field. The datasets were extracted from the MIMIC III Waveform Database. Further details of the datasets are provided in the documentation accompanying the ppg-beats project, which is available at: https://ppg-beats.readthedocs.io/en/latest/ . Datasets The following datasets are available: MIMIC PERform AF Dataset: Recordings from 35 critically-ill adults during routine clinical care, categorised as either AF (atrial fibrillation, 19 subjects) or non-AF (16 subjects). Matlab format (AF subjects, non-AF subjects) WFDB format (AF subjects, non-AF subjects) CSV format (AF subjects, non-AF subjects) MIMIC PERform Training Dataset: Recordings from 200 patients during routine clinical care, who are categorised as either adults (100 subjects) or neonates (100 subjects). Matlab format (all data, adults, neonates) WFDB format (all data, adults, neonates) CSV format (all data, adults, neonates) MIMIC PERform Testing Dataset: Recordings from 200 patients during routine clinical care, who are categorised as either adults (100 subjects) or neonates (100 subjects). Matlab format (all data, adults, neonates) WFDB format (all data, adults, neonates) CSV format (all data, adults, neonates) MIMIC PERform Large Training Dataset: 5,248 Recordings from 681 patients during routine clinical care, all of whom are adults. Recordings are 32 seconds long, with reference respiratory rates (denoted rr) derived from the imp signals using this algorithm.(added in v.2.0) Matlab format MIMIC PERform Large Testing Dataset: 1,257 Recordings from 167 patients during routine clinical care, all of whom are adults. Recordings are 32 seconds long, with reference respiratory rates (denoted rr) derived from the imp signals using this algorithm. (added in v.2.0) Matlab format Dataset Samples The following dataset samples are available: Truncated MIMIC PERform Dataset: MIMIC PERform Training Dataset - Data from the first 10 patients in the (5 adults and 5 neonates) in Matlab format. MIMIC PERform Testing Dataset - Data from the first 10 patients in the (5 adults and 5 neonates) in Matlab format. MIMIC PERform AF Dataset - Data from the first 10 AF and non-AF patients in Matlab format (AF patients, non-AF patients). 1-minute sample of adult data: Extracted from a non-AF subject in the MIMIC PERform AF Dataset, available in Matlab format. 1-minute sample of neonate data: Extracted from a neonate subject in the MIMIC PERform Training Dataset, available in Matlab format. 1-minute sample of AF data: Extracted from an AF subject in the MIMIC PERform AF Dataset, available in Matlab format. 1-minute sample of noisy data: Extracted from a non-AF subject in the MIMIC PERform AF Dataset, available in Matlab format. Citation When using these datasets, please cite the following publication: Charlton PH et al. Detecting beats in the photoplethysmogram: benchmarking open-source algorithms. Physiological Measurement 2022. DOI: 10.1088/1361-6579/ac826d Acknowledgments Each dataset is accompanied by a licence which acknowledges the source(s) of the data - please see the individual licenses for these acknowledgements.

This work was supported by British Heart Foundation (BHF) grants [FS/20/20/34626] and [PG/15/104/31913], and an EPSRC Impact Acceleration Award.