Term-Preterm EHG DataSet with Tocogram 1.0.0

File: <base>/#HEADER.shtml# (9,623 bytes)
<!--#set var="TITLE" value="The Term-Preterm EHG DataSet with Tocogram (TPEHGT DS)"-->
<!--#set var="USELOCALCSS" value="1"-->
<!--#include virtual="/head.shtml"-->

<!--#set var="RELEASEDATE" value="29 August 2018"-->
<!--#set var="VERSION" value="1.0"-->

<!--#set var="CITATION" value="
Jager F, Libenšek S, Geršak K (2018)
<a href=https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0202125 target=_blank>Characterization and automatic classification of preterm and term uterine records</a>. PLoS ONE 13(8): e0202125. https://doi.org/10.1371/journal.pone.0202125.
"-->
<!--#set var="CITATIONBIBTEX" value="
@article{TPEHGTDS,
  author  = {Jager F, Libenšek S, Geršak K},
  title   = {Characterization and automatic classification of preterm and term uterine records.},
  journal = {PLoS ONE},
  volume  = {13},
  number  = {8},
  year    = {2018},
  month   = {August},
  pages   = {e0202125},
  doi     = {https://doi.org/10.1371/journal.pone.0202125},
  url     = {https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0202125}
}
"-->
<!--#include virtual="/pn/cite-box.shtml"-->


<h2>Introduction</h2>

<p>
  The Term-Preterm ElectroHysteroGram DataSet with Tocogram (TPEHGT DS) contains
  26 four-signal 30-min uterine EHG records, i.e., three EHG signals accompanied by
  a simultaneously recorded external tocogram measuring mechanical uterine activity
  (TOCO signal) of pregnant women, and another five 30-min uterine records (EHG signals
  and TOCO signal) of non-pregnant women.
</p>

<h2>Data Description</h2>

<p>
  The records of the pregnant women belong to
  pregnancies that resulted in spontaneous preterm delivery (13 preterm records from
  eight pregnancies), and to pregnancies that resulted in spontaneous term delivery
  (13 term records from ten pregnancies). The main objective of the dataset is to
  provide a set of annotated contraction intervals (annotated intervals related to
  uterine contractions), and another set of annotated non-contraction intervals (dummy
  intervals, i.e., intervals out of uterine contractions). The annotated contraction
  and dummy intervals of the records allow:
</p>
<ul>
   <li> characterization of uterine records, contraction intervals, and dummy intervals
        during pregnancy; </li>
   <li> development of algorithms to classify between preterm and term uterine records,
        and/or, between labor and nonlabor uterine records.</li>
</ul>
<p>
  The dataset was developed at the Faculty of Computer and Information
  Science (Laboratory for Biomedical Computer Systems and Imaging), University of Ljubljana,
  Ljubljana, while the records were collected at the University Medical Centre Ljubljana,
  Department of Obstetrics and Gynecology. (Obtaining of the uterine records and the research
  were approved by the National Medical Ethics Committee of the Republic Slovenia:
  No. 32/01/97 and No. 108/09/09. The recording protocol (including position of
  the electrodes) was those which was also used during colecting the records of the
  <a href="https://physionet.org/physiobank/database/tpehgdb/" target="other">
  Term-Preterm EHG Database (TPEHG DB)</a>.
</p>

<p>
  The women participating in the study represented a sample of the general population.
  The TPEHGT DS contains 31 uterine records of which:
</p>
<ul>
  <li>13 preterm records are with spontaneous delivery (eight pregnancies),
           the mean delivery time was 33.7 (± 1.97) weeks, and contain 47
           annotated contraction intervals and 47 annotated dummy intervals;</li>
  <li> 13 term records are with spontaneous delivery (ten pregnancies),
           the mean delivery time was 38.1 (± 1.04) weeks, and contain 53
           annotated contraction intervals and 53 annotated dummy intervals;</li>
  <li>five records are from non-pregnant women, and contain 53 annotated
      dummy intervals.</li>
</ul>

<p>The records of pregnant women were obtained during regular check-ups in pregnancy
    around 31st week of pregnancy. The mean recording time and standard deviation of
    the records of pregnant women was 30.2 (± 2.76) weeks of pregnancy. For the manual
    annotating procedure, we used a graphic user interface and annotation editor.
    Besides visualizing original signals and annotation editing, the graphic user interface
    also allows calculating, and visualizing, spectra and spectrograms of the signals.
    Consensus about the annotated intervals was reached by two annotators.
</p>

<p>
  The EHG signals of the records were colected from the abdominal surface. The electrodes
  (AgCl2) to measure EHG signals of the records were placed in two horizontal rows,
  symmetrically above and under the navel, spaced 7 cm apart:
</p>

<ul>
  <li> the first electrode (E1) was placed 3.5 cm to the left and 3.5 cm above the navel;</li>
  <li> the second electrode (E2) was placed 3.5 cm to the right and 3.5 cm above the navel;</li>
  <li> the third electrode (E3) was placed 3.5 cm to the right and 3.5 cm below the navel;</li>
  <li> the fourth electrode (E4) was placed 3.5 cm to the left and 3.5 cm below the navel; </li>
<li> <i>the reference electrode was attached to the woman’s thigh</i>. </li>
</ul>

<p>
    The differences in the electrical potentials of the electrodes were recorded, producing
    three signals:
</p>
<ul>
  <li> S1 = E2 - E1 (first EHG signal);</li>
  <li> S2 = E2 - E3 (second EHG signal);</li>
  <li> S3 = E4 - E3 (third EHG signal).</li>
</ul>

<p>
    Prior to sampling, the EHG signals were filtered using an analog three-pole Butterworth
    filter with the bandwidth from 0.0 Hz to 5.0 Hz. The fourth simultaneous signal was the
    analog signal corresponding to an external tocogram (TOCO signal) measuring mechanical
    uterine pressure acquired using a cardiotocograph (model HP8030) attached at the top of
    the fundus. The analog TOCO signal was lead to one of the amplifiers of the A/D converter
    (the value of 150 μV corresponds to a pressure of 1 Pa). The sampling frequency for the
    EHG and TOCO signals was 20 samples per second per signal, with 16-bit resolution.
</p>
<p>
  The original EHG and TOCO signals of the records were further filtered using the four-pole
  band-pass digital Butterworth filter with cut-off frequencies at 0.08 Hz and 5.0 Hz
  applied bi-directionally. The records of the TPEHGT DS contain both, the original and
  filtered signals.
</p>

<h2>Files</h2>
<p>
    Record names of the TPEHGT DS are of the following format: tpehgt_TXXX, where capital
    T refers to the type of record: p - Preterm, t - Term, and n - Non-pregnant, and XXX
    is the record number of that type.
</p>
<p>
  The records of the dataset are in the <a href="/physiotools/wfdb.shtml" target="_blank">WFDB</a> format. Each records consists of three files,
  a header file (<em>.hea</em>) containing information regarding the record, the data file (<em>.dat</em>)
  containing original and filtered signal data, and the annotation file (<em>.atr</em>) containing
  annotations for the record.
</p>
<p>
  The comment section in the header files (<em>.hea</em>) includes clinical
  information, such as:
</p>
<ul>
  <li>record ID [record name];</li>
  <li>record type [Preterm/Term/Non-pregnant];</li>
  <li>pregnancy duration (gestation) [weeks];</li>
  <li>gestation duration at the time of recording [weeks];</li>
  <li>maternal age [years];</li>
  <li>number of previous deliveries (parity) [number];</li>
  <li>number of previous abortions [number];</li>
  <li>weight at the time of recording [kg];</li>
  <li>placental position [front/end];</li>
  <li>smoker [yes/no];</li>
</ul>
<p>
    If data was not available, the value is marked as None, if data was not applicable, the
    value is marked as N/A.
</p>
<p>
  The signal data in the data files (<em>.dat</em>) are in the following order:
</p>
<ul>
  <li>first EHG signal S1, unfiltered;</li>
  <li>first EHG signal S1, filtered using a 4-pole band-pass Butterworth
          filter from 0.08 Hz to 5.0 Hz, bi-directionally;</li>
  <li>second EHG signal S2, unfiltered;</li>
  <li>second EHG signal S2, filtered using a 4-pole band-pass Butterworth
          filter from 0.08 Hz to 5.0 Hz, bi-directionally;</li>
  <li>third EHG signal S3, unfiltered;</li>
  <li>third EHG signal S3, filtered using a 4-pole band-pass Butterworth
          filter from 0.08 Hz to 5.0 Hz, bi-directionally;</li>
  <li>TOCO signal, unfiltered;</li>
  <li>TOCO signal, filtered using a 4-pole band-pass Butterworth
        filter from 0.08 Hz to 5.0 Hz, bi-directionally;</li>
</ul>

<p>
  The annotations files (<em>.atr</em>) cantain manual annotations of contraction and dummy
  intervals:
<ul>
  <li>BC - the beginning of contraction interval;</li>
  <li>EC - the end of contraction interval;</li>
  <li>BD - the beginning of dummy interval;</li>
  <li>ED - the end of dummy interval;</li>
</ul>


<h2>Contact</h2>

<p>
  For further information, please contact:

  <blockquote>
    Franc Jager<br>
    Laboratory of Biomedical Computer Systems and Imaging<br>
    Faculty of Computer and Information Science<br>
    University of Ljubljana<br>
    Večna pot 113<br>
    1000 Ljubljana, Slovenia<br>
    <em>email:</em> franc.jager@fri.uni-lj.si
  </blockquote>
</p>

<h2>References</h2>

<ol>
  <li>Jager F, Libenšek S, Geršak K (2018). <a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0202125" target="_blank">Characterization and automatic classification of preterm and term uterine records</a>. PLoS ONE 13(8): e0202125. https://doi.org/10.1371/journal.pone.0202125.</li>
</ol>