Home

E.D.A.

1. 141 A Interval sec 000 46 000 47 000 48 000 49 000 50 000 51 000 52 000 53 000 54 000 55 000 56 000 57 000 58 000 59 000 60 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 73385000 73315000 67200000 68990000 69505000 73050000 69207500 65802500 71585000 72975000 70805000 70665000 69680000 70410000 72160000 HR bpm 761 839 286 969 325 136 696 182 816 220 740 908 108 215 149 bpm 0 85 1 6 1 2 0 0 0 0 1 132 0 564 050 315 912 392 224 976 083 983 You may use a time file as a reference time input If the first time of the switch r equals zero r or r0 0 1 0 RECAL searches for a time file ZMD If there is no file you are prompted for a time file name RECAL will calculate real time phasic changes for all times of the time file Together with the switch rt you may change the default settings of the pre and post intervals The next example shows a listing with the switch settings r0 0 5 and rt6 10 while DEMO G04 served as the input result file time file 1 6 5 4 3 2 1 0 UU iS DEMO 2 Interval sec 210 210 4 710 5 210 5 710 6 210 6 710 7 210 7 710 210 710 210
2. 0 0 0 0 0 0 5 Ch act label binary file results correction select 5 5 1 Y SCR left DEMO DO1 Y DEMO 601 DEMO K01 ny N 5 5 2 Y SCR right DEMO D02 Y DEMO G02 Y DEMO K02 N 5 5 3 Y EOG DEMO D03 N DEMO G03 N DEMO K03 N N 5 5 4 Y ECG DEMO D04 N DEMO G04 Y DEMO K04 Y 5 5 5 markings 05 DEMO G05 N DEMO KO5 N N 5 2 2 2 2 2 2 2 2 2 5 5 5 Any modifications entries Y N N 5 94444444444444444444444444444444444444444444444444444444444444444444444444444448 Fig 14 Upper part of EC entry screen and corresponding prompt Column 4 tells EC a name of a binary file In the case of a CODAS data type such a specification is optional because the main data file contains the original data for all channels But you may specify a binary file e g filtered data by ED to display other data than the original CODAS data for this channel In the case of a binary data type EC expects at least one binary data file to be activated i e the binary file name in column 4 must be followed by a Y in column 5 commonly the main file name If EC is unable to find a specified file the corresponding column is set to N EC uses the current DOS path Column 6 results specifies a result file which may be used to display detected events by ED together with the former specif
3. E D A EVENT DETECTION AND ANALYSIS FOR PHYSIOLOGICAL SIGNALS Manual of Version 1 04 Contents License Agreement iv 1 Basics 1 DV OTN etd 1 Main Modules and Additional Routines 1 Hardware Requirements 2 System Files 3 Dala E ute rtis ote Dr ott et are ots 4 Logie or elle Ld wed UE 4 Date Format Version Number and Support 5 2 Event Detection 6 V indicates modifications to previous versions 1992 Release 20 December 2000 E D A Manual 1 04 Interactive MOS EO CoE oh eh Ee ole LETTERE 6 Common Fields a a ke LLL Ee Ld 8 Unique Fields 11 Gap ICS te IER 12 E E a a aede sedo cone 13 Parameter Pre definitions 13 v BATCH MOGB fal E E 14 Y Memory RESOUrCes qu a RR E GN ne GU MER RE DU dos 14 Methods edunt una Sed 9945 5 09004 2408 2406 2468 24 06 2468 85064 15 Electroderm
4. time file 3 5 26 28 33 37 time markings 1 2 5 7 26 28 31 33 35 43 time window 2 27 31 33 43 total time 25 Turbo Pascal 3 40 VARCOEF 38 39 variance coefficient 38 virtual memory zero amplitude zoom window E D A Manual 1 04 14 32 32
5. 8 14 24 30 43 44 2 1992 Release 03 04 1993 correction mode 27 COSAM EXE 2 3 41 cut off frequency 10 data format 1 4 data point 4 26 32 Dataq Instruments Inc 41 42 datatype 6 8 definition file 1 13 14 23 24 26 27 30 31 43 demo files 3 24 displacement 8 11 dotted lines 27 duration 27 43 44 E D A i 1 3 5 EC 2 5 13 23 35 37 40 43 EC DEF 3 23 24 43 EC EXE 2 5 23 24 40 43 ECG 1 3 6 7 18 21 25 33 35 45 ECG1 7 13 14 18 20 43 ECG2 7 12 13 18 20 43 ED 1 6 8 11 13 14 23 25 34 35 40 43 ED DEF 1 3 13 14 43 ED EXE 1 5 14 23 40 EDA 1 3 6 8 11 13 15 16 15 17 34 35 39 42 43 45 edge 27 44 electrocardiogram 18 electrodermal activity 1 7 15 42 EMG 1 3 9 21 22 21 22 EMS 2 14 15 43 entry screen 3 5 6 8 23 28 31 event detection i 1 5 6 8 13 21 events 1 2 5 6 11 13 23 25 27 34 36 39 40 file name 4 6 8 14 23 27 29 33 35 37 38 43 45 file type 4 filter 1 8 13 15 20 21 35 43 45 filer characteristic 9 13 43 filter order 10 filter type 9 13 43 frequency 4 10 13 17 18 24 25 31 32 35 40 42 43 45 graphic 2 9 15 16 19 22 31 33 graphic colors 16 19 22 group delay 9 11 HAMM 12 Hamming 12 HANN 12 Hanning 12 hardware requirements 2 heart period 1 36 37 heart rate 1 27 36 38 heart time 34 help 3 6 8 23 25 27 29 32 34 41 HIGH 1 9 13 31 32 40 41 43 horizontal position 3
6. 710 210 710 HP sec 77500000 75640000 74955000 78000000 77715000 77370000 71000000 HR bpm 419 323 923 205 549 507 bpm 0 78 061 37 1992 Release 20 December 2000 E D A Manual 1 04 1 7 710 8 210 0 74300000 80 754 2 693 2 8 210 8 710 0 76675000 78 252 0 192 3 8 710 9 210 0 79000000 75 949 2 111 4 9 210 9 710 0 78265000 76 663 1 398 5 9 710 10 210 0 77500000 77 419 0 641 6 10 210 10 710 0 75150000 79 840 1 780 7 10 710 11 210 0 72360000 82 919 4 858 8 11 210 11 710 0 69240000 86 655 8 594 9 11 710 12 210 0 73000000 82 192 4 131 2 35 190 6 32 190 32 690 0 73060000 82 124 In case of the switch ra new result file R is generated by RECAL which may be later inspected with RECAL Compare the following listing of the file DEMO RO4 with the previous example file name binary DEMO RO04 file name ASCII number event time No off time cal IBI HR mean HR sec sec sec delta HR 1 7 210 6 4 210 0 775 77 419 0 000 2 7 210 5 4 710 0 756 79 323 0 000 3 7 210 4 5 210 0 750 80 048 0 000 4 7 210 3 5 710 0 780 76 923 0 000 5 7 210 2 6 210 0 777 77 205 0 000 6 7 210 1 6 710 0 774 77 549 0 000 7 7 210 0 7 210 0 710 84 507 78 061 8 7 210 1 7 710 0 743 80 754 2 693 9 7 210 2 8 210 0 767 78 252 0 192 10 7 210 3 8 710 0 790 75 949 2 111 11 7 210 4 9 210 0 783 76 663 1 398 12 7 210 5 9 710 0 775 77 419 0 641 13 7 210 6 10 210 0 752 79 8
7. ACCOMPANYING WRITTEN MATERIALS ARE THE SUBJECT OF COPYRIGHT IN CONSIDERATION OF PAYMENT OF THE LICENSE FEE THE HOLDERS OF THE COPYRIGHT TO THE SOFTWARE AND ACCOMPANYING WRITTEN MATERIAL GRANT YOU A NON EXCLUSIVE RIGHT TO USE THE PRODUCT THE SOFTWARE MAY BE RUN ON A SINGLE MACHINE AT A SINGLE LOCATION SO LONG AS YOU COMPLY WITH THE TERMS OF THIS LICENSE AN ARCHIVAL COPY OF THE SOFTWARE IS PERMITTED LIMITED WARRANTY AND DISCLAIMER OF WARRANTY WARRANTED IS THAT THE PRODUCT SHALL BE FREE FROM DEFECTS IN THE PHYSICAL MEDIA THE SOFTWARE AND ALL ACCOMPANYING WRITTEN MATERIALS ARE PROVIDED AS IS WITHOUT WARRANTY OF ANY KIND NO WARRANTY IS GIVEN THAT THE SOFTWARE IS FREE FROM ALL BUGS AND OMISSIONS FURTHER 15 NOT WARRANTED OR GUARANTEED AND NO REPRESENTATIONS EXPRESS OR IMPLIED ARE MADE REGARDING THE USE OR THE RESULTS OF USE OF THE SOFTWARE OR WRITTEN MATERIALS IN TERMS OF CORRECTNESS ACCURACY RELIABILITY CURRENTNESS FITNESS FOR A PARTICULAR PURPOSE OR OTHERWISE THE ENTIRE RISK AS TO THE RESULTS AND PERFORMANCE OF THE SOFTWARE IS ASSUMED BY YOU iv E D A Manual 1 04 1992 Release 03 04 1993 1 Basics Overview The software package E D A Event Detection and Analysis analyses digitized physiological recordings It was originally designed to analyze the CODAS data format for CODAS see chapter 5 p 41 but it is also possible to use a binary or ASCII format Significant events will be detected e g R waves of an
8. Esc key will force EC to read the data and to switch to the graphic display cf chapter The Graphic Interface p 28 The lt Alt X gt and lt Alt Y gt keys will save all corrections if any While the former quits the program the later returns to the upper screen and accepts a new file name Pressing lt Alt V gt will show version number release date and support address 26 E D A Manual 1 04 1992 Release 03 04 1993 Time Analysis and Time Files Very often it is necessary to look at physiological recordings in relation to stimulus presentation e g specific SCRs to slides or phasic heart rate changes Ongoing events stimuli are coded synchronically in form of rectangles in an extra channel EC checks such a time marking channel for rectangular impulses and stores the times in a time file These times serve as anchors for the Z result output file and for a time window display in the graph 5 5 5 compression 7 current time 0 000 sec 5 5 events Y 5 5 time file DEMO ZMD time window from 1 000 to 5 000 sec 5 5 channel 5 4 duration R1 0 000 sec 5 5 height 1 000 R1 R2 distance R1 R2 0 500 sec 5 5 edge 2 0 1 2 3 duration R2 1 000 sec 5 5 5 5 detect events time analysis Y N N 5 94444 444444 444 444 44 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 44 444 448 Fig 16 Lower part of entry screen for event mar
9. If you use this sample rate as tc the signal is simply rectified Larger time constants lead to a signal integration The signal amplitude is weighted in respect to the time constant No event detection is done The graphical presentation normally shows the original data together with the inte grated data For clarification the two figures below fig 12 and 13 show the signals separately Graph scaling is done in according to the range of the integrated data If you use a large tc the original data are mostly out of range 21 1992 Release 20 December 2000 E D A Manual 1 04 Fig 12 Integrated EMG for colors see Fig 13 Original EMG data for colors table 6 see table 6 Table 6 Graphic Colors for Integration method 5 0 color graphic feature 5 0 black treated data red original data 5 0 Fig 12 shows the integrated data of channel 1 EMG with an integration time constant of 50 ms No filtering and tc elimination was used The time compression factor for the graphic display was 16 Fig 13 shows the corresponding original EMG recording t demo file DEMO EMG COD holds the original EMG recording data 22 E D A Manual 1 04 1992 Release 03 04 1993 3 Event Check The EC EXE routi
10. Measures of Heart Rate Variability In Orlebeke J F Mulder G amp Doornen L J P van Eds Psychophysiology of Cardiovascular Control pp 353 374 New York Plenum Press Elbert T amp Rockstroh B 1980 Some Remarks on the Development of a Standar dized Time Constant Psychophysiology 17 504 505 Martin amp Venables P H 1980 Techniques in Psychophysiology Wiley Chiches ter Thom E 1988 Die Hamburger EDA Auswertung In Boucsein W Hrsg Elektro dermale Aktivit t Grundlagen Methoden und Anwendungen S 501 514 Berlin Springer Tietze U amp Schenk C 1978 Halbleiter Schaltungstechnik 4 Aufl Berlin Springer Trosiener H J amp Kayser J 1992 Analysis and Interactive Control of Electrodermal and Cardiac Activity A Computer Program for IBM PC s and Compatibles Poster presented at the 20th annual meeting of the DGPA June 18 20 Ham burg 42 E D A Manual 1 04 1992 Release 03 04 1993 Appendix A The Definition File ED DEF cod type of Data File COD BIN ASC 1 type of Analysis ecgl ecg2 pow int 5 0 lower border of signal range 5 0 upper border of signal range 0 time constant seconds 0 time constant bessel filter type for future versions notch filter characteristic Low High Band Notch Band rejection none 2 order of the filter 2 4 6 8 10 48 upper cutoff Frequency for LP HP filter low
11. Trosiener Berlin 9444444444444444444441444444444444444444444444444444414444444444444444444444 1992 Release 20 December 2000 E D A Manual 1 04 2 Event Detection One of your first steps in analysis is the detection of interesting events in your data e g the occurrence of R waves in ECG or maxima in EDA One of the best ways to become familiar with our programs is to test them with our demo data files Interactive Mode Start the program by typing ED lt you will begin in the interactive mode of the program as you will see later there is a batch mode as well After the display of the identification screen normally it takes 5 seconds but you can skip this by pressing the lt Return gt key the entry screen appears and ED prompts for information The Enter key terminates the entry process in the current field Throughout the prog ram the F1 key will give you an on line help If you press Ctrl Break the program quits as soon as possible In the entry screen you are prompted for the file name of your data file data type desired method 64444444444444444444 F1 Help Ctrl Break Exit 444444444444 14 41 25 47 5 5 5 File DEMO COD 5 5 Datatype COD with f 200 00 Hz T 0 1 h m 40 01 s 5 5 Method EDA 5 Fig 1 Partial ED entry mask File Name and Data Type Remember that you can choose different types of raw data The data files can be one of three different types ASCII Binary o
12. and Calculations The Result File Converting Utility The RECAL EXE routine is a multi purpose program to operate on result files By typing RECAL or with as any call parameter you will get the following help screen EC Result File Converting Utility Version 1 04 01 07 1993 Syntax RECAL lt switches gt lt input file gt lt output file gt switches file converting ASCII gt binary default binary gt ASCII index show sample rate index of events in listing default show event number n lt t1 gt lt t2 gt number or events optional from time stat time related statistics only with n lt amp 1 gt to time lt t2 gt p lt t1 gt lt t2 gt IBI heart time in seconds pk 11 ul IBI plausibility control 11 lower HR limit interrupts listing ul upper HR limit interrupts listing r t0 tu IBI phasic changes real time t0 reference time default times from time file ZMD tu calculation time default 1 sec gt creates a new result file R rt tl t2 specifications for IBI phasic changes real time tl calculations before t0 default 5 sec t2 calculations after t0 default 10 sec z convert time file this help RECAL expects at least a result file name input file gt to convert the binary information into readable ASCII code All ASCII output to the screen may be directed to an ASCII text file if you specify
13. channel one file type refers to the type of data or to the kind of result the file exists of Table 1 lists the possibilities of file type E D A Manual 1 04 1992 Release 03 04 1993 Table 1 E D A file names binary data converted ASCII data ED result control data ED D binary data filtered or time constant eliminated data ED TET ASCII overflow listing ED result detected events ED K result file corrected events EC result file real time related phasic heart rates RECAL 72 result file corrected events including time markings ZMD time file EC Remarks In brackets the routine which establishes the respective files file name channel number Date Format Version Number and Support All dates in this manual and in the software are entered mm dd yyyy month day year In case of inquiries communicate with us see contact addresses on page iii and refer to the respective routine its version number and release date For example the following logo parts of the EC EXE shown with A t V in entry screen cf p 26 and ED EXE routines show the respective version numbers 1 04 and 0 26 and release dates Jan 7 1993 and Oct 12 1992 6444444EC Version 1 04 Release 01 07 1993 E4444447 5 Event Check 5 5 Copyright C J rgen Kayser 1992 5 5 00 00 00 step Event Detection Rel 10 12 92 5 00 00 00 total c H J
14. for identification purposes results binary output Range from to Sometimes a signal may be cut because amplifiers were overranged All parts of data outside a specified range are inspected for signal cuts Cut is defined here as 3 or more successive points with equal amplitude All cuts are reported in an error file The number of cuts is shown in the field Overflows This is an example part of an error file listing raw data from DEMO EMG COD Begin Width Value Begin Width Point Points uS mV s s E D A Manual 1 04 1992 Release 03 04 1993 4330 3 8 21 2 16 0 0015 4963 3 8 21 2 48 0 0015 10378 4 8 17 5 19 0 0020 12233 7 8 21 6 12 0 0035 12244 3 8 17 6 12 0 0015 Time Constant The specification of this field is necessary The RC time constant fc must be gt 0 0 It is defined as the time a signal needs to decrease to 37 of its stationary value A tc of 0 0 indicates that your signal is not affected by a tc A tc 0 0 leads to an elimination of the effect of the specified This is done by computing the signal as if it was affected by a tc of 100 seconds This is nearly equivalent to DC amplification Some manufacturers of psychophysiological equipment use the term time constant in a very confusing and unusual manner e g AC time constant In doubt determine the RC constant of your equipment by empirical tests Filter Filters are predefined as the Bessel type which are recommen
15. help you may switch to different help messages by pressing PgUp or lt PgDn gt the Esc key quits the help You will get a different set of help messages depending on whether you are working at the entry screen or at the graphic display 23 1992 Release 20 December 2000 E D A Manual 1 04 The emergency interrupt key lt Ctrl Break gt will quit the program as soon as possible Open files will be closed and occupied memory will be emptied Use this key only if you in trouble because in contrast to the normal termination of EC with the lt gt key no data are saved The Definition File A definition file holds the particular parameters you need for EC EXE e g significant labels for your data the display or operation mode or the kind of printer you use Normally this information is stored in an ASCII text file named EC DEF but you may also specify another file name when calling the program cf chapter Getting Started p 23 If you have no definition file EC will take default settings You may alter the default settings on line An example file EC DEF is given in Appendix B p 43 to show how such a definition file might look like All possible presettings are commented and are further explained in this chapter If you have copied the demo files DEMO to your hard disk you can try to run EC with DEMO COD DEMO DEF lt or simply DEMO lt The Entry Screen The EC entry mask ser
16. lt cursor left gt or lt cursor right gt keys the cross moves and the actual value and the corresponding time are displayed in the measuring panel cf fig 21 Together with the lt Ctrl gt key the cursor movement is speeded up In fact when determining an amplitude Measuring hannel 2 you need to specify two data points for the base value onset and the position of curser other for the maximum of the ampli tude Pressing Space fixes one data point and moves to the other data point which is represented by a second cross Again you can move the second cross with the help of the cursor keys As before Space will fix this data point and jump back to the first data point and so on Values and times of Vii m both data point together with their diffe base value rences are shown on the panel amplitude o k zero value If you have specified any compression larger than 1 EC will zoom the area around the actual data point when pressing Space This gives you the opportunity to select each data value by moving a new cross within the zoom ax window Pressing lt Space gt or lt Esc gt Fig 21 Measuring an event Zoom window closes the zoom window and shifts to with cursor right hand and corresponding the other data point status panel left hand After choosing two data points the key Enter enters your decision as a choice for a new result If you confirm your decision after prompting
17. overridden by specifying a smaller segment length 644444444 Segment Length 444444447 Data file 20001 data pts EMS allows 11487 data pts We recommend 20001 data pts Disk will be used Decrease i of data points to use EMS Your choice 20001 data pts lt Oi Oi Ci Oi lt Oi Ci Oi Oi Oi 9444444444444444444444444444444448 Fig 3 Segment length mask Methods Electrodermal Activity EDA The meaning of all values in the entry mask is described under in the former section Common Fields p 8 As result of the detection process you receive the time of the maxima in relation to the beginning of your measurement the amplitudes of the maxima the maxima types and the half value times if computing is possible cf Results 13 and Appendix p 45 It is strongly recommended to use a disk cache as SMARTDRV EXE or a hardware cache be cause this speeds up processing extremely when using virtual storage on disk The EDA section was originally based on the algorithms described by Thom 1988 but sub stantially revised and extended to obtain a maximum of accuracy and reliability 15 1992 Release 20 December 2000 E D A Manual 1 04 Hit one key to proceed exit graphics with Fig 4 Graphical presentation of the EDA method results cf Appendix C p 45 Graphic colors are given in table 3 Fig 4 shows a normal EDA recording treated with a six pole 2 Hz bessel fil
18. represent the absolute maxima of the R wave complex Under some circumstances e g small sample rates the absolute maximum does not represent the best estimation of the time of the R wave ECG2 calculates the time of the R waves as the best estimation i e it repre sents a virtual maximum Use this method only if ECG1 results are insufficient The procedure is very time consuming and it might be very confusing for presentation purposes if the calculated maximum is not the absolute maximum POW calculates the power spectrum of the selected data channel The data are divided in successive segments of 1024 data points The spectra are linear averaged and the result is only presented visually INT integrates the selected data channel The integration time constant can be selected over a wide range The detection process is skipped After you have specified the main file name the data type and the desired method you are prompted with the question OK Y N If you type N you can modify all of your entries you must use the lt gt key to skip through the fields 1992 Release 20 December 2000 E D A Manual 1 04 64444444444444444444 F1 Help Ctrl Break Exit 444444444444 14 41 25 47 5 5 5 File DEMO COD 5 5 Datatype with f 200 00 Hz T 0 1 h m 40 01 s 5 5 Method EDA 5 5 Channel number 1 of 5 5 5 Range from 5 00 to 5 00 1 5 5 Time constant 1 00 s BESSEL LOW filt 6 order even 5 5 Fi
19. the result is automatically selected for the output result file The Esc key will cancel the measuring Sometimes you do not need two data points e g when specifying the R wave peak or another zero amplitude In this case you simply press the key 0 otherwise you have to set both data points exactly to the same position Note that when correcting ECG results you should not measure the R wave amplitude Calculations of RECAL EXE refer to the difference of time onset time of a particular amplitude minus rise time cf chapter Calculations and more p 36 Normally you take the R wave peak as the reference time and therefore the rise time should be zero In contrast to the key A t which will always insert a new result the lt nsert gt key will modify an existing result if activated In this case you may change the base value and 32 E D A Manual 1 04 1992 Release 03 04 1993 the maximum of the activated result The cursor will start at the base value of the parti cular result and swap to its maximum Time Markings If you are working with a time file cf chapter Time Analysis and Time Files p 27 the current time marker number and time is displayed at the top left corner between file name and channel label cf fig 17 All manual result selection will store this time as the event time i e the results of the Z files are related to this marker time cf example 2 in chapter Listing a Result File p 35 When
20. the rise time A half value time of 1 is coded when an estimation was impossible normally type 0 of reaction A half value time of 2 is coded if there are not enough points for an estimation Electrocardiogram ECG Most fields of the ECG screen are known cf Interactive Mode p 6 One field is specific for method ECG2 R wave enhancement Remember that you can choose between two different methods for the detection of R waves 1 and ECG2 ECG1 determines the absolute maxima in the neighbourhood of a likely R wave A convolu tion of the data with a template of specified frequency is done in ECG2 A frequency value between 8 and 10 Hz is normal You can determine the frequency when you measure the width of the R wave where frequency equals the reciprocal of double width in seconds By convolution we get an enhancement of the relevant parts of the R wave Under some circumstances this leads to a better recognition of the maxima The computed time values of ECG2 do not always point to the real maxima i e values found in your data but to hypothetical time values Sometimes this estimation is better than the time of the absolute maxima primarily if you have used a slow sample rate so that the maximum lies between two data points Method ECG1 is normally 18 E D A Manual 1 04 1992 Release 03 04 1993 sufficient Hit one key to proceed exit graphics with Fig 8 ECG demo data channel 4 of DEMO COD Graphic
21. working with time markings the Insert key will activate the cursor cross in the middle of the time window when no result is formerly activated You may jump to different time markers by either using the F9 key choose any marker number within the range of detected markers or the keys Tab and Shift Tab to skip to the next and previous time marker respectively It is also possible to insert or delete a time marker manually The key F8 insert a new time marker at the specified time The key F7 deletes the time marker Both keys adjust the time file The keys F7 to F9 as well as the Tab keys are only functional when using a time file Changing Graphic Color Settings The default color settings of EC may be altered by the user Pressing the lt Shift F1 gt key toggles between eight different background colors using EGA or VGA The Shift F2 key activates a left panel to modify the colors of the currently activated channel The keys Space Home and End toggle between fifteen foreground colors EGA VGA for the graph line the selected and the unselected results respec tively The Esc key quits the channel color panel You will save the current color settings with lt Shift F3 gt to a file named EC PAL When starting the program EC searches in the current DOS directory for a file EC PAL with the color settings 33 1992 Release 20 December 2000 E D A Manual 1 04 4 Results
22. 0 34 36 40 IHR 38 39 INT 7 13 43 integration 1 3 7 21 22 47 1992 Release 20 December 2000 48 interactive mode 6 10 12 14 15 18 23 label 24 25 29 33 44 45 LOW 1 8 10 13 17 31 40 43 manual scaling 12 maximum 7 8 11 13 15 20 25 26 29 32 35 43 45 mean sum of squares of successive differences 38 memory 2 14 24 method 6 8 11 13 14 16 19 21 22 MSSD 38 39 NOTCH 1 9 10 13 21 43 on line help 3 6 23 26 29 onset 1 32 35 45 open files 24 26 parameter pre definitions 13 Parzen 12 phasic changes 34 37 POW 7 13 48 power spectrum 1 3 7 11 20 21 precision 28 print file 30 31 printer 24 30 31 45 PSCREEN EXE 2 3 31 R wave 7 12 13 18 19 32 33 35 43 R wave enhancement 12 18 RAM 2 3 14 43 range 4 7 8 12 13 21 26 29 30 33 36 43 44 range overflows 18 RECAL EXE 2 3 28 33 34 40 recovery time 16 17 20 34 35 RECT 12 rectangular window 12 21 result file 1 3 5 13 25 30 38 40 44 45 RESULT PAS 3 40 rise time 17 33 36 sample rate 2 4 8 18 21 24 25 28 34 36 41 45 scaling 12 21 29 30 44 scaling factor 29 30 44 SCR 1 17 18 25 34 44 scrolling 30 segment 12 14 16 19 21 43 selection 11 25 31 33 44 45 speed 2 8 11 13 35 43 45 STDEV 38 39 System files 3 tc 9 21 22 threshold 8 11 13 34 35 43 45 time analysis 26 28 33 43 time axis 29 time constant 1 3 5 7 10 13 21 23 35 42 43 45
23. 23 91 480 1 1 860 2 485 1 690 1 220 24 94 125 1 1 381 2 307 1 060 0 865 25 97 720 1 2 274 3 096 1 355 1 020 46 E D A Manual 1 04 Index lt gt 30 lt gt 30 lt 0 gt 32 lt Alt A gt 30 lt Alt l gt 32 33 lt Alt V gt 5 26 lt Alt X gt 24 26 lt Alt Y gt 26 lt Backspace gt 30 lt Ctrl End gt 30 lt Ctrl Home gt 30 lt Ctrl PgDn gt 30 lt Ctrl PgUp gt 30 lt Ctrl PrtScr gt 30 31 lt Cursor down gt 26 30 lt Cursor left gt 31 32 lt Cursor right gt 31 32 lt Cursor up gt 26 30 lt Delete gt 30 31 lt End gt 30 33 lt Esc gt 23 26 28 30 33 F1 6 23 25 29 31 lt F10 gt 28 30 lt F2 gt 29 lt F3 gt 30 lt F4 gt 30 lt F5 gt 30 lt F6 gt 30 F7 33 F8 33 F9 33 Home 30 33 Insert 32 33 lt Shift F1 gt 33 lt Shift F2 gt 33 lt Shift F3 gt 33 lt Shift Tab gt 33 lt Space gt 32 33 lt Tab gt 33 acceleration 8 11 13 43 actual time 29 adjustment factor 30 44 amplitude criterion 1 15 17 arbitrary scaling factor 29 ASCII 1 2 4 7 10 13 23 24 28 34 36 38 45 BAND 1 9 10 13 21 43 Band pass 9 10 base value 25 32 34 batch mode 6 10 13 15 Bessel filter 9 11 15 binary data 5 10 23 25 44 45 binary output 1 8 10 11 13 43 BORLAND 40 bpm 39 calculations 2 13 28 33 34 36 38 calibration file 3 4 41 channel number CO processor CODAS 1 4 6 8 23 26 41 42 compression 8 12 15 19 22 26 27 32 43 5
24. 40 1 780 14 7 210 7 10 710 0 724 82 919 4 858 15 7 210 8 11 210 0 692 86 655 8 594 16 7 210 9 11 710 0 730 82 192 4 131 17 35 190 6 32 190 0 731 82 124 0 000 18 35 190 5 32 690 0 735 81 633 0 000 19 35 190 4 33 190 0 742 80 917 0 000 Time Related Statistics The switch stat calculates several time specific statistics for 5 i e mean MEAN standard deviation STDEV variance coefficient VARCOEF mean sum of squares of successive differences MSSD instantaneous heart rate IHR For a detailed explanation of these statistical parameters and their relevance we recommend the article of van Dellen Aasman Mulder amp Mulder 1985 as a reference In contrast to the switch n all calculations refer to the time interval between the first and the last heart beat For clarification look at the following three examples corresponding commands are entered after the DOS prompt 38 E D A Manual 1 04 C EDA DEMO gt recal demo g04 n15 75 stat 86 events 15 000 75 000 sec gt 86 000 n min DEMO G04 n 85 IBI 15 435 74 660 sec gt 86 112 n min MEAN 696 7647058824 msec STDEV 43 9338374472 msec VARCOEF 6 3054051212 MSSD 1245 8333334122 IHR 1 4352047277 Hz ok C EDA DEMO gt recal demo g04 n15 75 n 86 events 15 000 75 000 sec gt 86 000 n min in DEMO G04 ok C EDA DEMO gt recal demo g04 n stat n 143 events in DEMO G04 142 IBI 0 375 99 925 s
25. D EXE and EC EXE consisting of real type values range 2 9 x 10 to 1 7 x 1098 accuracy 11 12 digits i e 1 sign bit 39 bit fraction and 8 bit exponent The first data point holds the sample rate the following data points represent the continuous data flow In addition ASCII files may be used as a data source only for ED EXE which converts these files to the binary format Each ASCII value must be separated by a carriage return and a line feed ASCII decimal codes 13 and 10 i e each line contains one value The first line value consists of the total amount of data points the second value contains the sampling rate in Hertz The following lines illustrate an ASCII file Note that there is 1 no need for an exponential notation and 2 the comments in the right column are not allowed to occur in the ASCII file 20001 line 1 20001 data points 2 0000000000E 02 line 2 frequency 200 Hz 6 3461538462 02 line 3 first data point at real time 0 sec 6 3781324248 02 line 4 second data point at 0 005 sec 6 6665196610 02 1 2444866899 01 1 2791017588 01 line 20003 last data point at 100 000 sec Logic of File Names All output generated by the modules is named as follows file name file type channel file name is a string of 1 to 8 characters file type a one character string and channel a two character string channel codes the number of the channel to which the file refers e g 01 for
26. ECG or SCRs of an EDA recor ding and stored in a result file These events can be displayed with the original signal Alterations or corrections of these events are possible The detected events or results are stored in relation to the beginning of recording i e in real time Besides a listing of the results it is therefore possible to look at particular parts of the recording in order to calculate several tonic or phasic parameters of interest from the events e g SCR frequencies heart rate heart period phasic heart rate second by second If one of the recording channels contains time markings e g for stimulus onset the program is able to detect these and it is possible to consider the events in relation to these time markings In short E D A enables the user to generate parameters from the raw signal the sort of data needed for statistical purposes Main Modules and Additional Routines The package incorporates several independent programs First there is the routine ED EXE Event Detection Till now it is possible to investigate recordings of electro cardiography ECG and electrodermal activity EDA to integrate the data e g for electromyography EMG and to run a power spectrum for descriptive purposes The user may specify several presets prior to the investigation e g type of analysis ECG EDA integration power spectrum time constant amplitude criterion kind of filter high low band notch none ED EXE can s
27. Release 03 04 1993 When getting into the graphic mode for the first time or after changing the actual time EC will read the data In the gra phic the lt F1 gt key always shows an on line help the lt F2 gt redraws the screen The main part of the screen displays the signals of the recordings If there are results within the time range of the display the results are integra ted in each graph with ortho gonal lines together with the last two digits of the result number The top left corner of the screen exhibits the main file name cf fig 18 Below the file name you see the number of the currently activated channel together with the corresponding label Result parameters of this channel are listed below number time of maximum amplitude in units Fig 18 Top left of graphic display file name and currently activated channel At the bottom of the screen a time axis informs you about the time scaling start and end time of the current display and the per centage of data read so far cf fig 19 The right panel displays the scaling of each channel in real units The number at the bottom of each scale represents an arbi trary scaling factor which may range from 1 to 9999 see next section Fig 19 Bottom right of graphic display time axis channel scaling percentage of data read Controlling the Appearance of a Single Channel If you want to select another channel instead of the one currently activated yo
28. ULT PAS Regard the comments in the source code RESULT PAS C 1992 H J Trosiener amp J Kayser The following declarations refer to the file structure of result files created by ED EXE EC EXE and RECAL EXE The first record of a result file consists of lt ResultType gt 2 the second record of lt ResultType gt 3 and all following records are of lt ResultType gt 1 Depending on the kind of lt CodeType gt the variables of lt ResultType gt 1 differ from the meaning of the variable name which hold only for lt CodeType gt NormalType TYPE CodeType NormalType SlopeType LatencyType IBI_RealtimeType ResultType RECORD CASE INTEGER OF 1 Index LongInt TimePoint REAL f AmplitudeType INTEGER Amplitude REAL SlopeMaximum REAL RiseTime REAL RecoveryTime REAL 2 Min Ampl Crit REAL 4 Min Inter REAL Min Slope Crit REAL Percentage REAL Code CodeType 3 TimeConstant REAL SampleRate REAL FilterDelay REAL Frequency low REAL Frequency high REAL END 40 E D A Manual 1 04 1992 Release 03 04 1993 VAR ResultFile FILE of ResultType 5 Modifying CODAS Data Files The COSAM EXE routine allows you to reduce the original sample rate of a CODAS data file in order to reduce the amount of data to extract parts from the original file and to join different parts to anew CODAS data file By typing COSAM you wil
29. al Activity EDA 15 Electrocardiogram E GG 18 Power SDOCIF fTi aues acreedor orl 20 rie 21 3 Event Check 23 Getting Started aad a s aas se dons der 23 On line Help and Interrupt Key 23 v The D finition File ER S EP EUREN SEE EN 24 The Ent Ue aca par PIE 24 The UPPErSCheen EA AIO UE Teo 24 v The Lower Screen 26 Time Analysis and Time Files 27 The Graphic Interface c c 28 VIEWING the sort os 28 Y Controlling the Appearance of a Single Channel 29 Scrolling the Current Display 30 Printing BS CROC ord te erat Seu eM MEM 30 Selection er Results 22220 2 Joe ee Gy ace AN 31 Measuring Inserting and Revising Results 32 Time Markings eo COD E 33 Changing Graphic Color 05 33 4 Results and Calculations 34 The Result File Converting Util
30. an output file name Listing a Result File Example 1 shows an EDA result file created by ED The column type refers to the kind of SCR i e 1 codes a normal SCR 2 codes an SCR with an interpolated recovery time 2 because of a new SCR and 0 codes an SCR with no recovery at all recovery time 2 is set to 1 000 The column time holds the real time of the maximum of the SCR amplitude The onset of the SCR base value is this time minus the rise time file name binary DEMO GO1 file name ASCII DEMO EDA threshold maximum 0 0250 threshold overlayed maximum 0 0250 threshold speed slope 0 0750 34 E D A Manual 1 04 percentage of maximum onset 5 00 time constant in seconds 0 98 delay in sample points 43 00 number time sec type amplitude sample rate 1992 Release 03 04 1993 filter frequency slope maximum ok 200 00 Hz 2 00 Hz rise recovery time 2 1 210 1 080 1 350 1 387 1 265 1 110 1 125 3 393 0 710 0 325 1 070 0 790 1 695 1 235 1 065 0 865 1 355 1 025 Example 2 shows an EDA output result file created by EC while using time markings Only a few results from the input result file have been selected The column recovery time 2 is replaced by event time i e the actual reference time of the time file when selecting a particular result The column slope maximum is replaced by latency i e time minus rise time minus event time As there has been an interactive correction cf chapter M
31. ay of one second The discrete anchor is a edge within this pattern coded by a number between zero and three cf schema of fig 16 If all pattern times match zero an even edge number 0 or 2 forces EC to search all rising edges odd number 1 or 3 to look for all falling edges All of this spe cifications may be stored in advance in a definition file cf chapter The Definition File p 24 27 1992 Release 20 December 2000 E D A Manual 1 04 When searching for time markings EC checks times with a precision of 2 x sample rate but at least 25 ms An ongoing time analysis can be interrupted by the lt Esc gt key If you then confirm the interruption the time file will contain all time points dis covered prior to interruption otherwise the analysis will be continued You may switch between reference time points in the graphic display cf chapter Time Markings p 33 In the entry screen you are no longer allowed to choose an current time but you can do so from the graphic display using lt 10 gt A time file may be converted to ASCII code with the RECAL EXE routine cf chapter Calculations and Other Options p 34 The Graphic Interface Viewing the Display COD 7 210 sec SCR left 5 Fig 17 The whole EC graphic display showing 5 active channels DEMO COD The lt Esc gt key toggles between the entry screen and the graphic display cf fig 17 28 E D A Manual 1 04 1992
32. colors are given in table 5 Table 5 Graphic Colors for ECG method 5 0 color graphic feature 5 0 black treated data yellow points that are shifted to the next segment red plus Maximum R wave light blue angle sides of the R wave 5 0 Fig 8 shows the results of ECG1 No filtering was done and all maxima above 0 4 are selected as R waves No time compression was requested As a result you get the time values of the R waves and their amplitudes number time type amplitude slope rise recovery sec maximum time 2 1992 Release 20 December 2000 E D A Manual 1 04 3 1 855 9 1 291 0 000 0 000 0 000 4 2 575 9 1 216 0 000 0 000 0 000 5 3 330 9 1 223 0 000 0 000 0 000 6 4 130 9 1 217 0 000 0 000 0 000 7 4 900 9 1 261 0 000 0 000 0 000 Fig 9 shows an ECG not part of the demo data where some other parts of the ECG have the same amplitude as the R waves The program detects the maxima without further intervention Hit one key to proceed exit graphics with Fig 9 R waves together with large T waves for colors see table 5 In both methods ECG1 and ECG2 the type of maximum is set to 9 The slope maxi mum rising and half recovery time not computed and set to 0 0 Power Spectrum The power spectrum is used to control the effec
33. ded by constant group delay and low overswing Conceptually the filter is designed as recursive IIR filter The precise determination of signals phase shift by bessel filters makes this type most interesting in event related reaction designs Furthermore a Bessel filter has low signal distortion passband ripple but less selectivity no sharp cut offs Filter Characteristic You can choose high pass low pass band pass and band rejection notch filters LOW Low pass filters are used to eliminate high frequencies like EMG in some recordings HIGH High pass filters are used to eliminate low frequencies e g slow potential moving artifacts BAND Band pass filters reject low and high frequencies and allow only frequen cies in the passband from flo to fhi to pass like EEG filters NOTCH Band rejection filters cut out a frequency band from flo to fhi like line noise 50 Hz 60 Hz Linear transformations of time constant affected data are discussed e g by Elbert amp Rockstroh 1980 and Thom 1988 5 See e g Cooklll amp Miller 1992 or Tietze amp Schenk 1978 for digital filtering 1992 Release 20 December 2000 E D A Manual 1 04 Filter Order The sharpness of the cut offs and the distortion increase with the order of the filter If you need a very selective filter then choose a high order The order of the filters are selectable between 2 and 10 for high and low pass filters in steps of 2 For both band filt
34. dered to be of interest The values are normalized to speed and acceleration per second In some special cases the distinction between reaction and recovery shifts can be forced by variation of acceleration threshold Then only reactions which fitto a combined criterion of speed and acceleration are reported Normally the parameters are not used Window power spectrum Windowing means a weighting data to reduce spectral leakage error in the Fast Fourier Transform FFT This becomes necessary since the waveforms are sampled for a finite length of time Discontinuities at the end of the segments produce spectral errors We have implemented some common windows Rectangular window No weighting Leakage is very severe HANN Hanning window This window is most useful where good frequency resolu tion is needed and amplitude accuracy is not important HAMM Hamming window This window is useful for resolving closely spaced frequencies and amplitude accuracy is not important PARZ Parzen window The data are weighted with a triangle 11 1992 Release 20 December 2000 E D A Manual 1 04 R wave Enhancement ECG2 ECG2 convolves the data with a template of a specified frequency and enhances the R waves By convolution we get an enhancement of the relevant parts of the R wave Under some circumstances this leads to a better recognition of the maxima but the reported time values will not correspond necessarily to the absolute maximum i
35. e G01 n y show results n do not show results file name for corrected results 01 y y correct results n do not correct results Selection of results y all n none CHANNELO4 start entries of channel 4 y y show data of channel n ignore data of channel ECG label comment max 12 characters binary file name D01 y y use binary data n do not use binary data result file name G01 n show results do not show results file name for corrected results K01 y y correct results n do not correct results y selection of results y all n none CHANNELO5 start entries of channel 5 y y show data of channel n ignore data of channel markings label comment max 12 characters binary file name D01 n y use binary data n do not use binary data result file name G01 n y show results n do not show results file name for corrected results 01 y correct results n do not correct results 9 selection of results all n none the entry PRINTER n initializes the printer n printer number with three digits e g 001 the default setting is a standard printer up to now the following printer are supported A 001 Epson MX 100 002 Epson 80 Epson 80F T 003 Epson LQ 850 1050 004 IBM 4207 002 and compatible IBM printers P 005 Hewle
36. e enough points for an es timation Even if there are enough points particular circumstances can lead to enor mous errors e g a critical sample frequency a critical AD converter resolution or a very low amplitude criterion A visual inspection or a programmatical control of plausib ility is strongly recommended i Fig 5 SCR of type 2 Critical estima Fig 6 SCR of type 2 Estimated half tion of half value time maximum 4 for recovery time for colors see table 3 colors see table 3 Classification of Maxima After Boucsein 1988 1992 we distinguish three different types of maxima in the EDA method cf table 4 Table 4 SCR Classification 5 0 1 Normal maximum half value time could be computed Type 2 Overlayed maximum half value time had to be estimated Type 0 Overlayed maximum half value time could not be estimated because the SCR has no distinct maximum see 1st reaction in fig 7 Boucsein 1988 1992 labels this type different as type 5 0 1992 Release 20 December 2000 E D A Manual 1 04 Hit one key to proceed exit graphics with Fig 7 Reaction of type 0 1 SCR The half value times are always computed if possible but no plausibility check is made The graphical presentation of the half value time is suppressed if the value is greater than twice
37. easuring p 32 there is a 1 in the column type file name binary file name ASCII number Example 3 shows an result file created by ED Note time sec DEMO Z01 type amplitude 2 2 263 1 0 331 1 1 913 1 0 485 latency peak of the R wave and that there is always a rise time of zero file name binary DEMO G04 file name ASCII threshold maximum 0 5000 threshold overlayed maximum 0 5000 threshold speed slope 0 0750 percentage of maximum onset 5 00 time constant in seconds 0 00 delay in sample points 0 00 number time type amplitude sec 1 0 375 9 1 216 2 1 130 9 1 230 10 7 200 9 1 201 11 7 910 9 1 209 sample rate filter frequency slope maximum 200 00 Hz 0 00 Hz rise recovery time 2 0 000 0 000 0 000 0 000 0 000 0 000 0 000 0 000 that time holds the 35 1992 Release 20 December 2000 E D A Manual 1 04 Options You may use the switch a to reconvert from an ASCII file generated by RECAL to a binary file but this will only work if the ASCII listing contains the head entries i e sample rate etc The switch index will show the sample rate index of each event instead of the event number Example 4 shows a time file DEMO ZMD listing generated with the switch z 7 21000 sec 35 19000 sec 59 24000 sec 82 21500 sec Calculations The switches n stat p r and pk are used to calculate on a result file Number of Events The switch n co
38. ec gt 85 585 n min MEAN 701 0563380281 msec STDEV 45 4794435129 msec VARCOEF 6 4872737105 MSSD 1128 5460992791 IHR 1 4264188850 Hz ok Plausibility Check 1992 Release 03 04 1993 The switch pk is used to check the IBls for extra beats systoles or artifacts For example the switch 0k55 85 will list all IBls and their respective HR but will pause the listing whenever an IBI of an HR below 55 bpm or over 85 bpm appears C EDA D EMO gt recal demo g04 pk55 85 H w H HR bpm 1 130 0 75500 79 470 1 855 0 72500 82 759 2 575 0 72000 83 333 3 325 0 75000 80 000 4 125 0 80000 75 000 4 900 0 77500 77 419 5 645 0 74500 80 537 6 425 0 78000 76 923 7 200 0 77500 77 419 7 910 0 71000 84 507 8 675 0 76500 78 431 9 465 0 79000 75 949 10 240 0 77500 77 419 10 990 0 75000 80 000 11 680 0 69000 86 957 39 1992 Release 20 December 2000 E D A Manual 1 04 lt Return gt 16 12 410 0 73000 82 192 17 13 180 0 77000 77 922 18 13 950 0 77000 77 922 19 14 695 0 74500 80 537 20 15 435 0 74000 81 081 21 16 125 0 69000 86 957 lt Return gt Result File Structure Experienced programmers can use the result files created by ED EXE EC EXE and RECAL EXE to calculate their own parameters For this reason the storage format of the events in the result file is given as a part of a BORLAND Turbo Pascal source code in the file RES
39. ected to the printer or to a print file The printing process is displayed on the screen cf fig 20 and may be cancelled by pressing Esc You may setup your printer in the definition file cf chapter The definition file p 24 or on line after Ctrl 30 E D A Manual 1 04 1992 Release 03 04 1993 PrtScr gt type Esc and then the printer 22 code use lt F1 gt for help 0006 If you confirm the prompt question which also shows the printer model you are asked whether to send the screen image to a file or to the printer Files are named SRCEENOO PRN SCREEN01 PRN etc in regard of existing print files You may print the files later with the screen file printing utility PSCREEN EXE Typing LL PSCREEN lt will show the message Fig 20 Screen printing in progress below Remember to delete your screen print files after printing when you are no longer using them EC Screen File Printing Utility Version 1 04 01 07 1993 Syntax PSCREEN lt PrintFile gt PRN lt 1 gt default is Example PSCREEN 01 2 print file SCREENO1 PRN at LPT2 Selection of Results EC will indicate with different colors whether a particular result is selected for the output result file or not Per default the selected results are highlighted i e they are in the same color as the graph line you may change the color settings cf chapter Changing Graphic Color Se
40. er for Bandfilters notch 52 upper cutoff frequency for bandfilters and bandrejection none binary Output tcor filo filc orig none 4 Threshold maximum smallest amplitude of a maximum 0 Threshold overlayed maximum 0 Threshold speed 0 Threshold acceleration 0 frequency of r wave for enhancement or time constant of integrator Only for methods ECG2 and INT 0 segment length DataPoints EMS RAM disk 0 automatic use B The Definition File EC DEF 1 EC DEF parameter settings definitions of Kayser 01 07 1993 EC EXE searches for a file EC DEF another file name may be specified as the second call parameter when starting the program This file is an example of a parameter setting file oo empty lines are ignored comments start with a semicolon the order of all parameters for the head settings is fixed head settings occur only once and refer to all channels the default is coded by a minus sign sequenced settings may be left i e it is possible to have no head settings at all default settings head settings rate of compression default is 1 not smaller than 1 whole number event related analysis using time markings file name of time markings ZMD channel number for time analysis start of time window sec in relation to a time marker end of time window sec
41. ers this is equivalent to 1 to 5 order Cut off Frequency f flo fhi High and low pass filters need the cut off frequency as a characteristic parameter Cut off frequency is the value where the signal gain is 3 Db down From this value the asymptotic gradient of the frequency characteristic has a value of n x 20 dB decade with n order of the filter Band filters also need the upper cut off frequency Binary Output save data You have different options to save your data in binary format ORIG The original data are saved This is normally done if you have ASCII data TCOR The time constant is eliminated from your original data FILO You save the filtered signal and if requested the adjusted time constant too The group delay of the filter is not adjusted FILC You save the filtered signal and also if requested the adjusted time con stant The data are adjusted in respect to the group delay of the filter This is done by cutting the first points in the length of the group delay shift and adding values at the end of the data The last value is repeated until the original sampling time is reached NONE data are saved In the interactive mode existing binary data files with the same name are not destroy ed A message appears on the screen that prompts you to refuse or accept overwriting of binary data In batch mode data files are overwritten unconditionally Displacement group delay shift The group delay
42. for EC EXE EC DBF ED DBT on line help of the entry screen EC HLP on line graphic command key reference UTIL EXE PSCREEN EXE to print stored screens RECAL EXE to list and to calculate on result files COSAM EXE to modify CODAS data files RESULT PAS result file structure Turbo Pascal code Example Files DEMO EXE DEMO COD CODAS data file DEMO CAL the corresponding calibration file DEMO DEF the corresponding user default setup cf EC DEF DEMO D01 time constant eliminated data channel 1 DEMO D02 time constant eliminated data channel 2 DEMO G01 EDA results of channel 1 DEMO G02 EDA results of channel 2 DEMO G04 ECG results of channel 4 DEMO ZMD the corresponding time file to channel 5 DEMO EMG COD CODAS data file with an EMG recording DEMO EMG CAL the corresponding calibration file Copy the system files after extraction from the archives to a subdirectory e g C EDA of your hard disk If you would like to try the examples mentioned in this manual then copy the demo files to your hard disk e g CAEDA or C EDA DEMO 1992 Release 20 December 2000 E D A Manual 1 04 Data Format ED EXE and EC EXE work on CODAS data files which are stored in an 12 bit integer format The data is optionally calibrated to significant units e g umho or Mv and the calibration constants are stored in a CODAS calibration file same file name with the extension CAL Alternatively binary files may serve as an input to E
43. hannels will be initialized Due to the architecture of DOS a maximum of 15 open files at a time is possible If you exceed this limit i e try to open more than 15 files at a time the initialization is terminated with the message too many open files Note that you will need files in addition to the ones you have specified one for each correction the time file the CODAS data file and the on line help file The Lower Screen Once you have established your channel entries you will look at a modified screen cf fig 15 5 5 5 compression 7 current time 0 000 sec 5 5 events N 5 5 5 5 5 5 5 5 5 5 5 5 lt Enter gt change lt Esc gt graph lt Alt X gt end lt Alt Y gt new file 5 94444444444444444444444444444444444444444444444444444444444444444444444444444443 Fig 15 Lower part of EC entry screen and main menu By pressing lt Enter gt you can adjust the settings of the lower screen The compression rate fixes the time resolution of the graphic display e g to display only every 7 data point as a single pixel cf fig 15 A compression of 1 is the normal and largest time resolution Current time represents the time of the first data point at the left border of the graphic display You may specify any time within the range of the data file Events sets the mode of correction i e to use time markings as references for controlling results of a recording channel cf chapter Time Analysis and Time Files p 27 The
44. how how your choices influence the detection of events e g using filters with different upper or lower borders or using dif ferent amplitude criteria A validated setup the presets can be stored in a definition file ED DEF to run the routine in a batch job for several data files e g overnight ED EXE also transforms the original data i e you can choose between a binary output file containing the filtered signal the original signal converted from ASCII or CODAS or the signal with the time constant eliminated All features and the usage of the ED The first version of E D A was presented at the 20th annual meeting of the German Psychophy siological Society DGPA in Hamburg Trosiener amp Kayser 1992 Martin amp Venables 1980 is recommended as a standard handbook for description and evaluation of methods in Psychophysiology 1992 Release 20 December 2000 E D A Manual 1 04 routine are documented in chapter 2 Second the EC EXE routine Event Check displays the data original filtered time constant eliminated together with the detected events Up to eight data channels can be displayed at a time therefore making it possible to inspect the recording to compare the activity in several channels by overlaying the data curves to change the resolution separately for each channel to jump to different time points in the recording to revise the detected results change add or delete an event and to select or to rejec
45. ied signal Again if EC is unable to find the file the corresponding column is set to N Column 8 corrections gives a name to a new output file i e an interactively checked and or corrected result file In the graphic display you may select or de select a particular event from the input result file for the output correction file Further you are able to change an event i e to change the base value or the maximum of an event or to insert a new event cf chapter Measuring p 32 If you want to have the possibility to create a new output result file you must specify Y in column 9 Before overwriting an existing file EC will ask for information Results files generated by EC may serve as input result files in another EC session If you have decided to generate a correction file column 10 select permits you to determine to accept a priori all results from the input file for the output file cf chapter Selection of Results p 31 After you have specified the main file name and the data type you are prompted with the question any modifications on entries Y N lf you type Y you can modify any of these entries use the keys cursor up and cursor down to change the channel Enter to skip to the next entry and Esc to return to the prompt You may restore 25 1992 Release 20 December 2000 E D A Manual 1 04 the parameter settings from a definition file cf The Definition File p 24 After typing N all c
46. in relation to a time marker height of markings a base of zero is supposed 43 1992 Release 20 December 2000 E D A Manual 1 04 44 0 duration sec of the 1st rectangular impulse 0 any time 0 5 duration sec between both impulses 0 any time 1 0 duration sec of the 2nd rectangular impulse 0 any time 2 time marking in pattern 0 rise edge first rectangular impulse 1 decay edge first rectangular impulse 2 rise edge second rectangular impulse 3 decay edge second rectangular impulse if each duration equals zero an even number 0 or 2 will lead to a detection of all rise edges an odd number 1 or 3 will detect all decay edges adjustment factor settings AF02 0 01 adjustment factor channel 02 data divided by 100 AF03 0 1 adjustment factor channel 03 scaling factor settings default auto scaling are coded by the initial characters SF followed the channel number two digits and a not significant character e g double point i e the scaling factor value must be on column 6 15 scaling factor settings range 1 to 9999 SF01 100 scaling factor channel 01 SF02 100 scaling factor channel 02 SF03 200 scaling factor channel 03 SF04 50 scaling factor channel 04 parameter settings of a particular channel channel entries are introduced by CHANNEL n n cha
47. ity 34 Listihd a ResulbPlle o ou e downs 0 60074 Dus 34 Options of cater i seat PE uie pu prs 36 Calc latioS eri wd le Sad pen hn de ni a de a pes Wn 36 Number of Evers euge acera ck ucc 36 Heart Period Heart Rate and 36 Phasic Heart Bale 37 Time Related IBI 38 v Plavisibility aus Medi ars 39 Fiesult File SIKU COG EACUS 40 5 Modifying CODAS Data Files 41 E D A Manual 1 04 1992 Release 03 04 1993 References 42 Appendix 43 V A The Definition File ED DEF 43 The Definition File EC DEF 43 C EDA Result File Listing Example 45 Index 47 Jurgen Kayser Hans J rgen Trosiener Poschingerstr 22 Motzstr 9 D 12157 Berlin D 10777 Berlin Tel 49 0 30 855 95 07 49 0 30 216 95 71 449 0 30 314 25273 Mail kays0O230 mailszrz zrz tu berlin de trosienr zrzsp3 gp tu berlin de 1992 Release 20 December 2000 E D A Manual 1 04 LICENSE AGREEMENT SOFTWARE LICENSE THE SOFTWARE AND THE
48. kings related correction mode After specifying Y for events EC prompts you to choose a name for a time file The default is the main file name with an extension If there is an existing file EC will load the file and will display the total number of marking times and the first time refe rence cf fig 16 The borders of the time window in the graphic display serve only as a visual help two dotted lines They are related to the current time reference A time window border of zero indicates visually the current reference time point If there is no time file you may create a new file For this purpose you must specify an active channel which keeps the rectangular impulses EC assumes a base line of zero for this channel and searches for changes from this base line which are at least b of the height of the rectangles Rectangles of a fixed size only represent a binary information on or off but the duration of a rectangle can give additional information EC may analyze a pattern of two following rectangles and consider the duration of both rectangles as well as the duration between the rectangular impulses As you see in fig 16 you may specify this pattern in the entry fields duration distance R1 R2 A zero time codes any time i e ignore this part of the pattern For instance a pattern of 0 000 0 000 1 000 searches for 1 sec impulses a pattern of 2 000 1 000 0 500 searches for a 2 sec impulses followed by z sec impulse with a del
49. l get a command help screen CODAS File Operating Utility Version 1 04 01 07 1993 Syntax COSAM arg lt inp gt lt out gt lt gt arg arguments r interactive data reduction to a smaller sample rate e extract parts j join with second CODAS input file i information about CODAS input file this help inp CODAS input file out CODAS output file joi second CODAS input file for joining If you are creating a new CODAS data file i e using one of the arguments r e or j then the calibration file CAL of the CODAS input file is copied 8 CODAS package Computer based Oscillograph and Data Acquisition System is a software and hardware utility that turns a personal computer into a high performance data acquisition sys tem Dataq Instruments Inc 1988 41 1992 Release 20 December 2000 E D A Manual 1 04 References Boucsein W 1988 Elektrodermale Aktivitat Grundlagen Methoden und Anwendun gen Berlin Springer English translation 1992 Electrodermal Activity New York Plenum Press Cooklll E W amp Miller G A 1992 Digital Filtering Background and Tutorial for Psychophysiologists Psychophysiology 29 350 367 Dataq Instruments Inc 1988 CODAS User s Manual Software Release Level 4 Manual Revision H 825 Sweitzer Ave Akron Ohio 44311 USA van Dellen H J Aasman J Mulder L J M amp Mulder G 1985 Time Domain versus Frequency Domain
50. l range 1 0 time constant in seconds 0 no time constant bessel filter type for future versions low filter characteristic Low High Band Notch Band rejection none 6 order of the filter 2 4 6 8 10 2 upper cutoff Frequency for LP HP filter lower for band filters notch 0 upper cutoff frequency for band filters and band rejection none binary Output tcor filo filc orig none 025 Threshold maximum smallest amplitude of a maximum Threshold overlayed maximum Threshold speed Threshold acceleration 4444 13 1992 Release 20 December 2000 E D A Manual 1 04 200 frequency of r wave for enhancement or time constant of integrator 0 segment length DataPoints in EMS RAM disk 0 automatic use In the real definition files only one line is used for each parameter Appendix A p 43 shows as another example a typical definition file for the ECG1 method Any error suspends execution and an error message appears on the screen Batch Mode You can use ED in batch mode as well Thus ED has to be used with several call parameters Parameters which define batch mode processing C channel number i input file name If both parameters are present and the data file can be opened successfully then processing continues in batch mode Parameters which work in batch mode and in interactive mode d data type overrides all other specifications m method overrides all othe
51. lter f flow 2 00 Hz fhi Hz save data NONE 5 5 Displacement 43 00 pts t 0 215 s 5 5 Threshold maximum 0 025 uS Spec ParamsN Y N 5 5 Threshold overlayed maximum 0 025 uS 5 5 Threshold speed 0 000 uS s 1 5 5 Threshold acceleration 0 000 uS s 5 5 5 5 Auto scaleN ylo 6 000 yhi 6 000 Overflows 5 5 Maximum number at time s Point number 5 5 Time compression factor 1 9 1 5 0 0 5 00 00 00 step Event Detection Rel 1 04 01 07 93 5 5 00 00 00 total c H J Trosiener Berlin 5 944444444444444444444N4444444444444444444444444444444844444444444444444444448 Fig 2 Complete ED entry screen with demo data Depending on your method the lower part of the screen is filled with different fields Some fields are unique to the method most are common to all methods cf fig 2 t our demo example choose the EDA option as method and after confirmation the screen will change to the EDA screen If the data files have been opened successfully the lower part of the screen is filled with the appropriate entry fields The sample rate and the over all sample time of your recording is displayed cf top of fig 2 Common Fields Channel Number Channel number must always be specified If you have CODAS files the number of data channels is displayed and the requested data channel is selected For all other files the channel number is
52. n the R wave complex Graphics Results may be displayed graphically A segment of 256 or 256 x time compression points 1 99 is used Time compression means that more data points are used for a clearer presentation You can request graphics only in interactive mode by typing when graphics y n appears on the screen The binary result files produced in graphical mode should not be used further because the recursive filter algorithm leads to some problems if the segment lengths are small We recommend choosing convenient parameters and checking this graphically You should then run the same job without graphics to produce best results Scaling The graph can be scaled automatically Auto scale Your data segment is displayed so that it fits exactly into the graphical surface This is very useful if you have some unpredictably high signal amplitudes or if you want to inspect the visual results exactly Manual scaling requires the specification of upper yhi and lower ylo border of the scaling Data out of that range are inspected but not displayed 12 E D A Manual 1 04 1992 Release 03 04 1993 Results Range Overflows Cutting of your data by overranged amplifiers might occur This is detected automati cally and reported in ASCII error file The report gives information about the beginning and the length of the cut signal in units of data points and seconds The amplitude is reported and the number of cuts is repo
53. ne displays the data original filtered time constant eliminated together with events normally detected by ED EXE and allows checking of the events with optional modification or correction Thus this routine may create new event result files Getting Started If you type EC lt you will get the following message EC Version 1 04 01 07 1993 Syntax EC lt fil gt lt def gt DEF lt opt gt lt fil gt data file name def file name for definitions default extension DEF opt options this help lb binary data codas data You may specify your input data file as the first call parameter and the data type either by the option e g b for binary data or by the extension of the data file e g B N for binary data see also p 6 You may set up your parameters for EC with the help of an ASCII text file You tell EC the name of this definition file with the second call parameter cf chapter The Definition File p 24 If there is no second call parameter EC will search per default for a file named EC DEF When calling the program by typing EC lt you will start in the interactive mode EC prompts you to name an input data file and to specify the data type CODAS data or binary data The Enter and Esc keys quit the entry field On line Help and Interrupt Key You are allowed to press the F1 key throughout the program This key will give you an on line help While using the on line
54. nnel number with two digits e g 01 max 8 channels are possible entries for a particular channel are optional if a channel is mentioned all entries for that channel have to be explicit the default is coded a minus sign CHANNELO1 start entries of channel 1 y Y show data of channel n ignore data of channel SCR left label comment max 12 characters binary file name D01 y y use binary data n do not use binary data result file name G01 y y show results n do not show results file name for corrected results K01 y correct results n do not correct results n Selection of results y all n none CHANNELO2 start entries of channel 2 y Y show data of channel n ignore data of channel SCR right label comment max 12 characters binary file name D01 y y use binary data n do not use binary data result file name G01 y y show results n do not show results file name for corrected results K01 y correct results n do not correct results n Selection of results y all n none CHANNELO3 start entries of channel 3 E D A Manual 1 04 1992 Release 03 04 1993 y y show data of channel n ignore data of channel EOG label comment max 12 characters binary file name D01 n y use binary data n do not use binary data result file nam
55. of a filter is the phase shift of the incoming signal which is proportional to the signal frequency Bessel filter group delay is practically frequency independent and determined by the order of the filter A high order produces a large group delay or shift of the incoming signal We estimate the group delay and report it in data points and seconds Our estimate is the time the outcoming signal of the filters needs to raise from 0 to 50 of the stationary value We must adjust our detected events with this E D A Manual 1 04 1992 Release 03 04 1993 value This is already done in all our ED result files If you save data for further use with the FILO option the binary output data are not adjusted Unique Fields Some fields are specific to the method s some very sophisticated The special parameters can improve the detection of events Threshold Maximum A maximum with an amplitude below this value is rejected e g a value of 0 5 means that only the amplitude of R waves equal or above this value are reported Threshold Overlayed Maximum special The amplitude of an overlayed EDA reaction must have at least this value Default value is the maximum threshold For an explanation of overlayed maximum refer to fig 5 and 6 p 17 Threshold Speed Acceleration special Speed 1 derivate and acceleration 274 derivate control the selection of reactions Only reaction above the predefined speed and acceleration are consi
56. r Codas The file name extension can define the type of data if you use ASC for ASCII files BIN for binary files and for Codas files If one of these extensions is specified then the field Datatype is automa tically set to that type If your data file has another extension or if the extension does not correspond to your type of data the type must be specified individually If you have multiple channels recorded and your data type is not COD then you must choose one channel for analysis and prepare it as a binary or ASCII file E D A Manual 1 04 1992 Release 03 04 1993 Examples for valid filenames CHANNEL1 BIN VALUES2 ASC EXPER3 COD CHANNEL4 VP024 DATA0701 EXP t examples please use DEMO COD as filename This file contains of 5 channels cf table 2 Table 2 Recording Channels of DEMO COD 5 0 Recording Unit 5 0 1 electrodermal activity EDA on left hand uS 2 electrodermal activity EDA on right hand uS 3 horizontal eye movements EOG 4 heart activity ECG mV 5 time markings of stimulus exposure Mark 5 0 Method EDA calculates the time the amplitudes and the half value time of EDA reactions The type of EDA reaction is determined ECG1 calculates the time of the R wave The calculated values
57. r specifications p optional another alternate definition file than ED DEF very useful in case of frequent alterations of the predefinitions If a definition file ED DEF or an alternate file or the channel number or the filename are missing the program automatically switches to the interactive mode A wrong channel number stops execution with an error message Example ED EXE iDEMO COD 4 pECG DEF MECG1 lt Memory Resources The program decides whether it can use RAM EMS or virtual memory on disk for storage of the data segments The decision depends on available resources If there is not enough RAM or EMS and the anticipated consumption of disk space exceeds available disk space then the program is terminated with an error message You can override the automatic decision process in two ways 14 E D A Manual 1 04 1992 Release 03 04 1993 f you are in the interactive mode and not using graphic display mode where some kind of fixed length segments are used you are prompted for change of the segment length if EMS is not sufficient for storage cf fig 3 In batch mode a segment length parameter can be set If it is lt gt 0 and gt 1000 then this value is used as the segment length A value greater than the maximum number of data points is truncated Besides its usefulness to force EMS usage the consequence of an early program termination produced by a small amount of available disk space can be
58. refined units you can multiply your original data by any factor by using the keys F6 or Alt A value smaller than one will divide the original data which will fit for a coarse differentiation Values larger than one increase the original data which should be done with very refined input data The adjustment process will not affect your data and it operates only as an adaptation for the graphic display Again you can preset the adjustment factor for each channel via a definition file cf the example in the chapter The Definition File p 24 where the adjustment factor settings are coded by the initial characters AF followed by the channel number two digits and a non significant character e g a double point and the adjustment factor value There is no limit for the adjustment factor a value equalling one has no effect With the help of the key Home Ctrl Home End and Ctrl End you move the graph in a vertical direction up and down respectively Scrolling the Current Display To look through the whole recording the keys lt gt and lt gt move the current display half a screen back and forth respectively i e the former centre of the screen will be left or right in the next screen By pressing the F10 key you may jump to any time position start time of the display within the recording range Printing a Screen Typing Ctrl PrtScr forces EC to draw a pixel image of the screen The drawing may be dir
59. rted in the field Overflows of the entry mask Event Detection All detected events SCRs R waves and related values are written to a result file Two result files are generated One main holds time amplitude type etc of events the other holds corresponding speeds and accelerations to give ideas for the sophisti cated special parameters The result files represent the input for program EC Our EDA example produces the main binary result file DEMO GO01 This is listed by the program RECAL with the command HECAL DEMO GO01 lt cf chapter 4 Results and Calculations p 34 Appendix C p 45 holds this listing Parameter Pre definitions All parameters can be preset if you use the ASCII file ED DEF Although all parameters have some default values which are set automatically by the program it is not very handy to change some or all parameters manually for each new kind of analysis Each time the program starts it searches for ED DEF and uses if found all valid pre definitions as default in the entry mask and for the batch mode If ED DEF is not present the program uses only internal presets and the batch mode is rejected The following example of ED DEF shows you how such a definition file might look like for EDA method It comments all possible presettings cod type of Data File COD BIN ASC eda type of Analysis eda ecgl ecg2 pow int 5 0 lower border of signal range 5 0 upper border of signa
60. ry of 640 Kb RAM on an ordinary IBM PC or any compatible PC For the EC EXE routine a color monitor with a VGA or EGA interface is recommended and is absolutely necessary if you are displaying more than one channel For the ED EXE routine such a color screen is recommended The other routines will not need a special monitor type Depending on the amount of data a large fast hard disk is recommended ED EXE will use available extended memory EMS which will speed up the analysis All routines will operate with an installed co processor which will also increase speed So best is a machine with a 486 CPU built in co processor 16 MB RAM 200 MB hard disk and a VGA adapter but the package will work with a 286 AT no co processor E D A Manual 1 04 1992 Release 03 04 1993 640 Kb RAM 20 MB hard disk and a EGA adapter only much slower System Files Before copying the files you are advised to make a back up copy of the E D A distribution diskette and to store the original data in a safe place as a reference The E D A distribution diskette contains four self extracting archive files which store the following files archive names in the left column System Files ED EXE ED EXE to analyze your raw data EDA ECG integration power spectrum ED DEF user defined default settings for ED EXE ED DBF ED DBT on line help of the entry menu EC EXE EC EXE to check the recorded and analyzed data EC DEF user defined default settings
61. t the detected events Further you can investigate a particular channel for time markings i e rectangular impulses which are simplest the ascents or descents or more complicatedly a pattern of two rectangular impulses determined by the time of and between the impulses The detected discrete time points perform as anchors in order to display a related time frame or time window on the screen e g to collect inter actively only those events which fit in your time window The EC routine as well as the PSCREEN EXE routine which prints stored graphic screens while working with EC are documented in detail in chapter 3 The RECAL EXE routine results and calculations converts the result files to ASCII code or calculates several parameters from the detected and or corrected events e g frequencies for a particular part of the recording Of course it is possible to direct the ASCII listing to an output file and to calculate own parameters from the results or even better from the original result files for such a calculation Refer to chapter 4 concerning the RECAL routine The COSAM EXE routine CODAS sampling works on CODAS data files i e to extract parts of a data file to reduce a sampling rate to combine two data files or simply to inform you about the number of channels sample rate length of recording etc The COSAM routine is documented in chapter 5 Hardware Requirements All routines work in a DOS environment with a base memo
62. ter before the detection process All maxima and overlayed maxima exceeding a particular amplitude criterion here gt 0 025 uS are treated as meaningful In the graphical presentation the y axis scale is fixed with a time compression value of 3 Table 3 Graphic Colors for EDA method 5 0 color graphic feature 5 0 black treated data yellow points that are shifted to the next segment green angle sides of computed half recovery time red plus maximum reaction white plus half value time white cross x computed half value time light blue angle sides of the maximum reaction Sub p PT A ded dede pep PEA Five maxima were detected in this example The 1 3 and 5 maximum are of type E D A Manual 1 04 1992 Release 03 04 1993 1 and the half value times could be computed directly The 27 and 4 maximum are of type 2 and the half value times had to be estimated In fig 4 you can see that for maxi mum 2 the algorithm leads to a valid estimation The presentation of the estimation of the 4 maximum was suppressed because the computed value is not very realistic cf fig 5 An estimation is only realistic if we can extract a gradient from the data which leads to a practical determination of an approximation which is near to the assumed original value cf fig 6 This is only possible if there ar
63. ts of the requested filtering As in the other methods you choose a data channel and specify the filter A window can be requested Windowing means a weighting of your data to reduce spectral leakage error in the Fast Fourier Transform FFT Normally this is not needed 20 E D A Manual 1 04 1992 Release 03 04 1993 With the fields and fhigh you can change the x scaling of the graph If you specify e g 40 and 60 Hz you reduce the presentation to the range of 50 Hz noise without harmonics The y scaling is done automatically You can choose logarithmic or linear scaling The method itself divides your data in segments of 1024 points and computes the power spectrum of each segment The segments are averaged without any weighting The averaged power spectrum remains on your screen Fig 10 Power spectrum of ECG chan Fig 11 Power spectrum of ECG chan nel nel with notch filter 48 52 Hz Fig 10 shows the power spectrum of the ECG channel 4 without any treatment The scaling is logarithmic and a rectangular window was used You can see that the signal is superposed by a large amount of 50 Hz noise If you choose a 2 order band rejection filter notch between 48 and 52 Hz the reduction of 50 Hz noise is obvious cf fig 11 Integration The integration method integrates the chosen data channel The minimum of the used integration time constant fc has to be the sample rate
64. tt Packard LaserJet II Hewlett Packard DeskJet PLUS 006 Panasonic KX P1092 printer 004 initializes printer An EDA Result File Listing Example file name binary 01 file name ASCII DEMO EDA threshold maximum 0 0250 threshold overlayed maximum 0 0250 threshold speed slope 0 0000 percentage of maximum onset 5 00 time constant in seconds 1 00 sample rate 200 00 Hz 45 1992 Release 20 December 2000 E D A Manual 1 04 delay in sample points 43 00 filter frequency 2 00 Hz number time type amplitude slope rise recovery sec maximum time 2 1 1 295 1 0 441 0 668 1 235 1 080 2 4 360 2 1 331 1 762 1 350 1 373 3 7 070 1 3 675 4 910 1 255 1 105 4 9 835 2 2 254 3 630 1 120 3 249 5 10 990 1 0 324 0 698 0 710 0 325 6 16 790 2 6 364 7 326 1 610 20 091 7 25 845 2 4 584 5 843 1 490 2 174 8 27 465 1 1 311 2 776 0 775 0 500 9 36 645 1 0 054 0 080 1 050 0 580 10 39 675 1 2 043 2 753 2 100 1 265 11 42 365 1 1 576 2 812 0 965 0 730 12 44 475 1 1 670 2 939 0 975 0 700 13 49 160 1 1 533 2 817 0 935 0 690 14 51 275 1 0 416 0 922 0 740 0 430 15 53 640 1 1 821 2 725 1 160 0 785 16 62 705 1 0 482 0 662 1 170 1 485 17 67 980 0 1 441 2 129 1 195 1 000 18 68 970 1 1 635 2 649 0 985 0 675 19 75 510 1 2 498 2 822 1 520 1 005 20 77 680 1 0 981 2 015 0 825 0 650 21 81 895 1 0 279 0 519 0 945 0 670 22 86 350 1 0 154 0 234 1 135 0 800
65. ttings p 33 Depending on column 10 of the entry screen the results of a channel are either selected of unselected cf chapter The Upper Screen p 24 You will activate a result in the display by pressing cursor left left arrow key or cursor right right arrow key which shifts between adjacent results An activated result is highlighted in white in the list on the left as well as in the graph Pressing Return will select or de select the activated result for the result output file depending on the former status quo An acoustic signal of either a high now selec ted or low now unselected frequency indicates your operation An activated result can be deleted by pressing Delete You have to confirm your decision because the activated result will be removed from this EC session and it cannot be reinserted but the original input result file will not be affected Note that when working with time markings results are not implicitly selected when fitting the time window You have to do this explicitly by activating the respective results and pressing lt Return gt which is followed by the high frequency tone Measuring Inserting and Revising Results 31 1992 Release 20 December 2000 E D A Manual 1 04 The keys lt nsert gt or lt Alt gt are used to activate the measuring panel on the left of the screen as well as a small cross in the middle of the screen The cross behaves like a cursor if you press the
66. u simply have to type the number of this channel 1 8 Only the channel displayed at the left panel may be modified 29 1992 Release 20 December 2000 E D A Manual 1 04 The keys Cursor up Cursor down PgUp PgDn Ctrl PgUp lt Ctrl PgDn adjust the scaling factor of this channel Alternatively you use the F4 or lt Alt S keys and specify the scaling factor directly press Backspace or Delete to change the former scaling factor and enter the new The key F3 recalculates the scaling factor and the best fitting horizontal position of the graph line this is the initial default display The F5 key allows you to set the scaling factor of all measur ing channels at once i e of channels creating a new correction result file Per default all channel are auto scaled which is exactly what the key F3 does But itis also possible to preset the scaling factor for each channel via a definition file cf the example in the chapter The Definition File p 24 In such a file scaling factor settings are coded by the initial characters SF followed by the channel number two digits and a non significant character e g double point and the scaling factor value The scaling factor has to be in the range 1 to 9999 Under some circumstances you need to adjust your input data in order to get a good resolution and scaling on the graphic display If your data represents very coarse or very
67. unts the number of events n 24 results DEMO GO1 Together with a time range specification only the events fitting the time range are regarded The critical time for fitting is column time minus column rise time For example n50 0 100 0 calculates on a 50 seconds interval n 73 results 50 000 100 000 sec gt 87 600 n min in 04 Heart Period Heart Rate and IBI The same logic holds for the switch p which calculates heart periods HP and heart rate HR on IBls i e the time between adjacent results IBI time sec HP sec HR bpm 1 20 810 0 72000 83 333 2 21 560 0 75000 80 000 3 22 285 0 72500 82 759 4 23 015 0 73000 82 192 5 23 765 0 75000 80 000 6 24 480 0 71500 83 916 36 E D A Manual 1 04 14 IBI n ok Phasic Heart Rate Changes 25 145 25 765 26 365 26 970 27 620 28 360 29 065 29 760 15 20 000 30 00 0 66500 0 62000 0 60000 0 60500 0 65000 0 74000 0 70500 0 69500 0 226 774 000 174 308 081 106 331 Sec in DEMO GO4 1992 Release 03 04 1993 The switch r calculates on IBIs phasic changes for real time For example the switch r50 0 calculates for the reference time 50 0 seconds for steps of 1 seconds default heart period and heart rate Additionally EC calculates a mean heart rate for the interval before the reference time and the difference heart rate to the mean heart rate tet tot NW gt oa aS gt 10Y 0
68. ves as a status report of your work It is logically divided into an upper and a lower part In the upper screen the user establishes fixed settings for the work with a particular data file In the lower screen it is possible to change several para meters throughout the run of EC The Upper Screen Besides the file name and the data type the top of the screen informs you about the sample frequency sample rate and the total file length cf fig 14 If the frequency per channel exceeds 1000 Hz time will be displayed in units of milliseconds ms Below the file information you will get a channel listing of several columns Depending on your data type there will be a display of eight binary data or of the total number of channels CODAS data The first column codes the channel number Ch ranges from 1 to 8 The second column codes whether a channel is active i e to be included in the graphic act may be Y or The third column holds a user defined label for the channel which will be shown in the graphic max twelve characters Columns 4 6 8 hold file names the adjacent columns 5 7 9 labelled with a question 24 E D A Manual 1 04 1992 Release 03 04 1993 mark initiate the files vs 644444444444444444444444444444444 F1 help 4444444444444444444444444444444447 5 file name DEMO COD total time 100 000 sec 5 5 data type COD frequency 200 00 Hz sample rate 0 005 sec 5 0 0 0

E.D.A.

Contents

Download Pdf Manuals

Related Search

Related Contents