Data format types Creating a study folder File naming Channel naming Importing polar heart rate
data Contents:
| Data format types |
| Creating a study folder |
| File naming |
| Channel naming |
| Importing polar heart rate
data |
Preparing your data for analysis with
ANSLAB involves the following steps:
Data
format types: ANSLAB
can read
BIOPAC Acqknowledge files (*.ACQ), Vivologic data files (*.VIV), HP VEE
data files (*.D*T), European/Biosemi Data Format files (*.EDF/*.BDF),
Vitaport data files (*.dat, *.vpd), heart rate data qcquired with POLAR
watches, ECG data acquired with MONICA fetal ecg device (*.RAW), Matlab
data files (*.MAT) and tab-separated text (TXT) files. To analyze
Vivologic or MONICA data files, no special preparation other than renaming and creating a study folder (see below)
is required. To analyze BIOPAC, VEE, EDF/BEDF and VPD data files, you
have to
inform anslab
about the way your data channels are labelled (see channel naming below). The
data formats ANS and MAT are used for saving analysis
results in anslab - they are not meant to be used as raw data format to
start analyzing from. POLAR heart rate data must be analyzed in a study
folder as described for the other data types, and renamed to match the
ANSLAB naming convention. Moreover, data files must be converted to
matlab
data format and loaded for editing as described at importing polar heart rate
data. Text files
(TXT) have to be arranged with
sample points as
rows and channels as columns. Channels can be listed at different
sampling rates, but unequal signal length have to be padded with empty
cells. Tab-delimited format is recommended.
Column-channel-associations and sampling rates have to be specified
using the anslab options dialog on the 'TXT anylsis options'-tab (shown
below). The data section of the text-file (all lines except the first
if a header row is specified) may not contain any text fields (except
for the number-equivalents 'nan' and 'inf' and exponential number
format like e.g. 5e13). [Top]
Data
can be listed at an arbitrary sampling rate (sufficient to avoid data
loss) because all anslab functions will resample the data to whatever
rate they need.
Creating a study folder:
To analyse data using anslab, you must
alway
work in a special 'study folder', that has a predefined
structure. You have to select this folder before running any kind of
analysis, as all study specific options are stored inside this folder.
To create a study folder, you have several options:
To actually select a study folder as
working directory, use the 'select'-button, on the
left column of the
anslab menu window next to the text label "working directory". After
you select the 'anslabdef.m'-file inside your
study folder, it's path is
displayed in the edit box underneath the select-button. [Top]
File
naming:
ACQ, VPD, EDF/BDF, RAW, TXT
& VEE:
ACQ, VPD, EDF/BDF, RAW and VEE raw
data files must be renamed
using the 3-character study name (e.g.
'tex') followed by a 3 digit subject number and a 2 digit session
number (for instance 'tex00101.acq'). All the programs use the specific
file name properties to deduce subject- and run-number for saving and
naming files, thus using different file names will not work!!!. You can copy the
data files to the
'raw'-folder inside the study folder, but this is not
mandatory.
| CCC | XXX | YY | .ext |
| study
identifier (for instance 'tex') |
subject#
(for instance 089) |
run#
(for instance 03) |
file
extension (for instance .ACQ) |
An example is: TAF02401.acq
VIV : Vivologic raw
data files must follow the vivologic data center naming convention,
which is depicted below. The first five characters identify the
study, then always follows a dash, then a 3 digit number for the
subject, a letter for the run, always an underscore, and last date and
time of the start of the recording. You must conform to this
convention, because all the programs use information from the filename
for saving and naming files containing the reduced data. Moreover, some
analysis types require to retrieve the file start time from the
filename.
| CXXXX | -
|
YYY | C | _ | YYYY | MM | DD | hh | mm | .viv | |
| study identifier (e.g 'p0027') |
dash |
subject# (for instance 089) |
run# as
character (a,b, c...) |
underscore |
year | month |
day |
hour |
minute |
file
extension |
|
An example is:
p0027-003b_200506231038.viv [Top]
Channel
naming:
ACQ:
In BIOPAC Acqknowledge software, channels can be labelled by the user.
To allow anslab to recognize channels relevant for each
analysis type, you must add the labels you use in your laboratory to
the set of possible labels recognized by anslab. Anslab stores these
labels in a text file called 'channelnames.m', that is located in the
'anslabutil'-folder, in the 'channelnaming'-subfolder ( \anslabfolder\ anslabutil\ channelnaming\
channelnames.m
). Initially this file is named 'channelnames.txt' to avoid
CVS-repository conflicts. Rename this file to 'channelnames.m' to allow
proper channel identification.
You can modify the list in channelnames.m manually using a text editor
(e.g. using
the matlab editor by typing 'edit channelnames' on the command prompt),
or you can edit this file indirectly using the 'channel naming'-tab
from the options dialog in the 'tools'-menu (*tools*options*channel
naming). If you edit this file directly, be sure to preserver
correct matlab syntax. As the label lists are stored as multiline
character arrays, you must padd the right side with empty spaces, so
that all entry lines have exactly the same length! You can see this
directly if you are using a font with uniform character sizes (e.g.
courier, as shown in the example below, if your browser support this
font):
The channel naming tab of the options dialog allows you to edit the
names without bothering about the length of the labels. Anslab
automatically converts the entries to equal sized lines and saves them
in the 'channelnames.m' -file. Separate new entries with commata, and
hit 'apply' or 'apply & close' to save your changes. Note that you
will loose your settings by changing to a different option tab (e.g.
the ecg options) without applying first.
The channel naming tab of the options dialog allows you to edit the
names without bothering about the length of the labels. Anslab
automatically converts the entries to equal sized lines and saves them
in the 'channelnames.m' -file.
VIV:
Vivologic channel names cannot be changed, settings therefore do not
have to modified by you.
RAW: Monica data is read from previously exported files, which must be named as they are named automatically by the MONICA software (xxxx.ch1, xxxx.ch2, xxxx.ch3 and xxxx.ch4). Moreover, fetal beat-by-beat data must be stored in an associated text file called "xxxx.ibi". All these files belonging to a *.RAW data file must live in the same directory as the RAW-file for anslab to find them.
VEE:
HP VEE channel names, signal types, sample rates and calibration
information (mult, div and offset) are defined in the
'veechannelnames.m'-file in the folder
\anslabutil\channelnaming\. This file initially is named
'veechannelnames.txt' to avoid CVS repository conflicts. Rename this
file to 'veechannelnames.m' to allow proper channel identificaiton.
Adjust the content of this file corresponding to
your data files. Note that anslab identifies the channels to load
exclusively by the Ch#, that is 'hardcoded' in the fileheader of
each datafile, and retrieves scaling information based on this number
from the 'veechannelnames.m'-file.
VPD:
Vitaport channel names are defined in the 'vpdchannelnames.m'-file.
This file initially is named 'vpdchannelnames.txt' to avoid CVS
repository conflicts. Rename this file to 'veechannelnames.m' to allow
proper channel identificaiton.Like the Acqknowledge channelnames-file,
it contains only
signal type information, calibration data is taken from the file
header.
EDF/BDF:
EDF channel names are defined in the 'edfchannelnames.m'-file, BDF
channel names in the 'bdfchannelnames.m'-file.
These files initially are named 'edfchannelnames.txt' and
'bdfchannelnames.txt' to avoid CVS
repository conflicts. Rename this file to
'edfchannelnames.m'/'bdfchannelnames.m' to allow
proper channel identification. Like the Acqknowledge channelnames-file,
it contains only
signal type information, calibration data is taken from the file
header. [Top]
Rereferencing data in a BDF data file
As non-EEG data (e.g. ECG) in a Biosemi data file is usually
saved referenced to the common mode sense (CMS) electrode but needs to
be rereferenced to a specific second sensor before analysis, the
channelname file for BDF data is checked for reference information. You
can specificy a channel to be rereferenced to some other channel
by putting both channelnames in a row in the channelnames file
separated by three dashes ("---"). So the ECG section of your
"bdfchannelnames.m" file might read:
EcgLabels = [...
'Ekg1---Ekg2';...
'EKG
';...
'EC
';...
'ECG Lead I ';...
'ECG
'];
indicating
that the channel "Ekg1" is to be rereferenced to "Ekg2" before analysis.
To convert polar heart rate text files, first create a subfolder in
your study folder. Second, create and load a batchfile containing
the paths of the text files to import as described at create batchfile and load
filematrix . Select 'Import polar data' from the 'tools'-menu. You
are prompted to select a folder, where the converted data files will be
stored. Choose the just created 'import' folder and wait while anslab
converts the files. [Top]