Each record includes a header (.hea) file, a short text file that contains information about the types of signals, calibration constants, the length of the recording, and (in the last line of the file) the age, gender, and weight (in kg) of the subject. In this database, all 16 subjects were male, aged 32 to 56 (mean age 43), with weights ranging from 89 to 152 kg (mean weight 119 kg). Records slp01a and slp01b are segments of one subject's polysomnogram, separated by a gap of about one hour; records slp02a and slp02b are segments of another subject's polysomnogram, separated by a ten-minute gap. The remaining 14 records are all from different subjects.
All recordings include an ECG signal, an invasive blood pressure signal (measured using a catheter in the radial artery), an EEG signal, and a respiration signal (in most cases, from a nasal thermistor). The six- and seven-channel recordings also include a respiratory effort signal derived by inductance plethysmography; some include an EOG signal and an EMG signal (from the chin), and the remainder include a cardiac stroke volume signal (see reference below) and an earlobe oximeter signal. The table below summarizes the contents of each record:
| Record | length | AHI | ECG | BP | EEG | Resp (nasal) |
Resp (pleth [1]) |
EOG | EMG | SV | SO2 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| slp01a | 2:00 | 17.0 | 0 | 1 | 2 (C4-A1) | - | 3 (S) | - | - | - | - |
| slp01b | 3:00 | 22.3 | 0 | 1 | 2 (C4-A1) | - | 3 (S) | - | - | - | - |
| slp02a | 3:00 | 34.0 | 0 | 1 | 2 (O2-A1) | 3 | - | - | - | - | - |
| slp02b | 2:15 | 22.2 | 0 | 1 | 2 (O2-A1) | 3 | - | - | - | - | - |
| slp03 | 6:00 | 43.0 | 0 | 1 | 2 (C3-O1) | 3 | - | - | - | - | - |
| slp04 | 6:00 | 59.8 | 0 | 1 | 2 (C3-O1) | 3 | - | - | - | - | - |
| slp14 | 6:00 | 30.7 | 0 | 1 | 2 (C3-O1) | 3 | - | - | - | - | - |
| slp16 | 6:00 | 53.1 | 0 | 1 | 2 (C3-O1) | 3 | - | - | - | - | - |
| slp32 | 5:20 | 22.1 | 0 | 1 | 2 (C4-A1) | 3 | 4 (C) | 5 | 6 | - | - |
| slp37 | 5:50 | 100.8 | 0 | 1 | 2 (C4-A1) | 3 | 4 (A) | 5 | 6 | - | - |
| slp41 | 6:30 | 60 [2] | 0 | 1 | 2 (C4-A1) | 3 | 4 (A) | 5 | 6 | - | - |
| slp45 | 6:20 | 5 [2] | 0 | 1 | 2 (C3-O1) | 3 | 4 (A) | 5 | 6 | - | - |
| slp48 | 6:20 | 46.8 | 0 | 1 | 2 (C3-O1) | 3 | 4 (C) | 5 | 6 | - | - |
| slp59 | 4:00 | 55.3 | 0 | 1 | 2 (C3-O1) | 3 | 4 (A) | - | - | 5 | 6 |
| slp60 | 5:55 | 59.2 | 0 | 1 | 2 (C3-O1) | 4 | 3 (A) | - | - | 5 | 6 |
| slp61 | 6:10 | 41.2 | 0 | 1 | 2 (C3-O1) | - | 3 (A) | - | - | 4 | 5 |
| slp66 | 3:40 | 65.5 | 0 | 1 | 2 (C3-O1) | 3 | 4 (A) | - | - | 5 | 6 |
| slp67x | 1:17 | 0.7 | 0 | 1 | 2 (C3-O1) | 3 | 4 (C) | - | - | 5 | 6 |
[1] Respiratory effort signal from inductance plethysmography: sum (S),
abdominal (A), or chest (C) signal
[2] Estimated from visual review; apnea annotations unavailable
Entries in the table indicate, for each record, the signal numbers of the signals named at the top of the table. Note that the usual ordering of the respiratory signals is reversed for record `slp60'. These signal numbers may be used to select specific signals for examination using rdsamp (1). Durations are given in hours and minutes.
Each record includes two annotation files. The `.ecg' files contain beat annotations, and the `.st' files contain sleep stage and apnea annotations. Information about types of annotations in the `.ecg' files is included in the WFDB Programmer's Guide. Annotations in the `.st' files are all NOTE annotations, which contain the sleep staging and apnea information in their aux fields. Each annotation in the `.st' files applies to the thirty seconds of the record that follow the annotation. The coding scheme is:
| aux | meaning |
|---|---|
| W | subject is awake |
| 1 | sleep stage 1 |
| 2 | sleep stage 2 |
| 3 | sleep stage 3 |
| 4 | sleep stage 4 |
| R | REM sleep |
| H | Hypopnea |
| HA | Hypopnea with arousal |
| OA | Obstructive apnea |
| X | Obstructive apnea with arousal |
| CA | Central apnea |
| CAA | Central apnea with arousal |
| L | Leg movements |
| LA | Leg movements with arousal |
| A | Unspecified arousal |
| MT | Movement time |
Ichimaru Y, Moody GB. Development of the polysomnographic database on CD-ROM. Psychiatry and Clinical Neurosciences 1999 Apr; 53(2):175-177.
Lee JM, Kim DJ, Kim IY, Park KS, Kim SI. Detrended fluctuation analysis of EEG in sleep apnea using MIT/BIH polysomnography data. Comput Biol Med 2002 Jan; 32(1):37-47. [Abstract]
Rigney DR, Goldberger AL, Ocasio WC et al. Multichannel physiological data description and analysis. In: Weigend AS, Gerschenfeld NA (eds.), Time Series Prediction, pp 105-129. Addison-Wesley, 1994.
Garpestad E, Katayama H, Parker JA, Ringler J, Lilly J, Yasuda T, Moore RH, Strauss HW, Weiss JW. Stroke volume and cardiac output decrease at termination of obstructive apneas. J Appl Physiol 73(5):1743-1748, 1992 Nov. [Abstract]
Ichimaru Y, Clark KP, Ringler J, Weiss WJ. Effect of sleep stage on the relationship between respiration and heart rate variability. Computers in Cardiology 1990 1990; 17:657-660.
Tamaki N, Yasuda T, Moore RH, Gill JB, Boucher CA, Hutter AM, Gold HK, Strauss HW. Continuous monitoring of left ventricular function by an ambulatory radionuclide detector in patients with coronary artery disease. JACC 1988; 12(3):669-679. [This reference describes the method used to obtain the left ventricular stroke volume signal included in some of the recordings in the database; it does not describe or refer to the database itself.] [Abstract]