*********** Introductory information

## File name: 2017-04 Metadata_Variation of Korotkoff stethoscope sounds
  -- the file (KorS identified percentage) summarises results of all the identified Korotkoff beats as percentages from 10 folds for all 140 subjects. For each subject, there were 3 values from 3 repeat measurements.
  -- column A: Fold number
  -- column B-L: <-3, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6    Identified Korotkoff beats as percentage at SBP, and above and below SBP
  -- column N-X: -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, >3  Identified Korotkoff beats as percentage at DBP, and above and below DBP
  -- row 13: the mean of Korotkoff beat results as percentage from 10 folds of 140 subjects
  -- row 14: the Stand Deviation (SD) of Korotkoff beat results as percentage from 10 folds of 140 subjects
  -- row 15: the Standard Error of Mean (SEM) of Korotkoff beat results as percentage from 10 folds of 140 subjects

SBP Systolic Blood Pressure
DBP Diastolic Blood pressure
All 140 subjects were devided into 10 groups randomly. 

## Key words: Blood pressure; Convolutional neural network; Korotkoff sound. 

 

*********** Methodological information

## A brief method description - what the data is, how and why it was collected, and how it was processed
  --  140 healthy subjects were studied. For each subject, 3 BP repeat measurements were performed in total. 
  --  During linear cuff deflation, the Stethoscope sound was recorded by a microphone, with the stethoscope terminal connected to the microphone and placed on the antecubital fossa of the forearm. The Cuff Pressure signal was recorded by a pressure sensor connected to the cuff. The Oscillometric Pressure was derived from the cuff pressure signal. All signals were digitally recorded at 2 kHz and 16 bits, and stored to a computer for offline processing.
  --  a specific CNN structure was designed to identify the Korotkoff sound beats from stethoscope sounds. 

## There was no missing data.

 

*********** Funding information

## The data was collected at Newcastle University, UK. Anonymised data were analysed at Sichuan University, Chengdu, Sichuan, China 

## This work was supported by the Engineering and Physical Sciences Research Council(EPSRC) Healthcare Partnership Award (reference number EP/I027270/1), and EPSRC standard grant (reference number EP/F012764/1).