Archives of Physiology and Biochemistry 138 1-3445/95/10301-0000$6.00 1995, VO~.103, NO. 4, pp. 436-441 0 Swets & Zeitlinger NON~NVASIVEELECTROGASTROGRAPHY PART2 HUMANELECTROGASTROGRAM E. Atanassova’, I. Daskalov2, I. Dotsinsky2, I. Christov2 and A. Atanassova’ ‘Institute of Physiology, Bulgarian Academy of Sciences; 2Laboratory of Biomedical Engineering, Bulgarian Academy of Sciences, Sofia, Bulgaria ABSTRACT The electrical activity of human stomach muscle wall - electrogastrogram (EGG),was led offby surface (cutane- ous) electrodes placed on the abdominal wall and recorded on an electrogastrograph. A method for complete elimination of the cardiac artefact was elaborated and successfully implemented. It consists of a preliminary elimination of the QRS complex, based on its higher amplitude and slope. The eliminated intervals were replaced by linear segments. A subsequent low-pass filtering allowed to obtain a high quality EGG signal. The electrical activity of the stomach of healthy volunteers was characterized by waves with a frequency of 3.35 _+ 0.09 cpm during the quiescent periods and 2.99 2 0.14 cpm during the activity periods of the migrating myoelectricat complex (MMC). Bearing in mind the correlation between the bursts of spike potentials with the slow waves in the For personal use only. dog EGMG and the high-amplitude waves, characterizing the periods of activity in the dog EGG, we can better differentiate the periods of quiescence and activity by the amplitude of the waves. The wave amplitude during periods of quiescence was 81.13 f 20.61 pV, significantly different from the wave amplitude during periods of activity being 164.74 _+ 43.34 pV (n=7).Thus with this method in visual inspection it is possible to identify MMC of the human stomach by the changes in the amplitude of the waves in the EGG. KEY WORDS: stomach, electrical activity, cutaneous electrodes, EGG, visual interpretation, human gastric MMC. INTRODUCTION respiration, small intestinal activity, movements, skin- electrode noise, etc.). This requires computer process- Electrogastrography (EGG) is a non-invasive meth- ing of the data in order to separate the gastric signals Archives of Physiology and Biochemistry Downloaded from informahealthcare.com by JHU John Hopkins University on 10/03/11 od for recording the electrical activity of the gastric from the artifacts (Abell & Malagelada, 1988). How- muscle wall. This method could be very useful for ever, “computer analysis ... is somewhat risky for its the diagnosis of gastric motility disorders but it is potential to *create* nonexistent data” (Abell & still at experimental level (Chen & McCallum, 1991 a) Malagelada, 1988). The combination of clear record- mainly because of the relatively low quality of the ings amenable to visual inspection and computer- EGG records impaired by artifacts (cardiac activity, assisted analysis would offer the best of all possibil- ities (Abell & Malagelada, 1988). Correspondence to: Prof. Elena Atanassova, Institute of The purpose of this work was to obtain very high Physiology, Acad. G. Bonchev Str. B1. 23, Sofia 1113, quality EGG recordings intended for assessment of Bulgaria. the changes in stomach electrical activity in order to HUMANELECTROGASTROGRAM 437 characterize the periods of quiescence and activity applied with a selectable cut off frequency of 0.2 or of the migrating myoelectrical complex (MMC) and 0.4 Hz. Thus the electrocardiogram signals were re- to show the end of the evacuation of the gastric con- moved entirely. tents. A large number of electrocardiograms were in- vestigated to define the appropriate criteria for iden- tification of the high and/or steep waves. The results METHODS pointed out that the beginning of these waves must be set when the slope of the signal exceeded the Investigations were made on 26 healthy volunteers level of 2.5 pV/ms. The end of the same waves was - 14 females and 12 males, mean aged 32.4 (17 - 58) detected by a more sophisticated algorithm because years, without any gastrointestinal complaints. The of the risk of false detection in larger and longer procedure of the investigation was explained and their elevated or depressed quasi horizontal ST segments. signed consent was obtained. The persons were ly- An interval with a slope continuously lower than 2.5 ing in supine position, kindly requested not to move, pV/ms during 250 ms was searched for. Once this without speaking. requirement was met, the end of the high andlor steep The electrogastrogram (EGG) was recorded by cu- wave was marked off 250 ms back on the time scale. taneous electrodes on the abdominal wall (Fig. 1). The Investigations were made on 8 - 10 hours fasting position of the active electrodes was as follows: elec- persons: i) after 2 - 3 hours of recording the volun- trode 2 was placed just under the left lower rib, on the teers were given standardized meal; 20 - 30 min lat- line between the nipple and the navel; electrode 1 - 3 er the investigation continued for 40 - 60 min. ii) cm above it, on the chest; 3 - on the line between 2 after 1 -hour of recording the same type of food was and navel; 4 - under the right lower rib, opposite 2. given to the volunteers, the recording began 20 - 30 The reference electrode (R) was placed to the right of min later and lasted 2 - 3 hours. the navel. Conventional type disposable pre-jelled The standardized meal consisted of two sandwiches stick-on electrocardiographic electrodes were used. with ham or sausages, a cup of fruit juice and coffee The records were made on an 2-channel electro- (proteins - 47 g, 192,7 Kcal; fats - 3,4 g, 31,62 Kcal; gastrograph, designed by us. A method for complete carbohydrates - 115 g, 4713 Kcal; total - 695,82 For personal use only. elimination of the cardiac artefact was elaborated Kcal). and successfully implemented. It consists of a pre- The amplitude and the frequency of the gastric liminary elimination of the QRS complex, based on waves in the EGG was measured in selected inter- its higher amplitude and slope. The eliminated inter- vals: i) low-amplitude waves ii) high-amplitude vals were replaced by linear segments. A subsequent waves; iii) waves 20 - 30 min after feeding. The low-pass filtering allowed to obtain high quality EGG amplitude of the waves was measured in mm and signals. The input signals were analog-to-digital transformed in pV. The data were statistically proc- converted with a relatively high sampling rate of essed and the means If: SEM were calculated. The 200 Hz allowing adequate digital acquisition of both statistical significance of the differences between the electrocardiogram (ECG) and EGG signals. The re- slow wave frequency and amplitude in the groups sidual mains interference was additionally removed was shown using the Student’s t-test using an artifact subtraction procedure in real time (Christov & Dotsinsky, 1988; Levkov ef al., 1984) The total elimination of the heart electrical activ- RESULTS Archives of Physiology and Biochemistry Downloaded from informahealthcare.com by JHU John Hopkins University on 10/03/11 ity artifacts began with the identification of high and steep waves. These were usually the QRS complex- Waves with a frequency of about 3 cpm were ob- es but might be large and steep T waves too. Their tained in all recordings from fasting volunteers. In 8 participation in direct averaging of the signal could of the cases highest amplitude signals were led off be disturbing, because the high amplitudes will not with a position of the two active electrodes along the be suppressed radically, despite the relatively good antral axis [combinations 1 - 2 or 2 - 3 (Fig. l)]. In frequency separation from the EGG signals. There- 16 cases however the higher wave amplitudes were fore, these waves were cancelled and substituted by obtained by 1 - 4 or 2 - 4 and in 2 studies - with 3 - 4 linear segments. Then a digital averaging filter was leads. 438 E. ATANASSOVAE7 AL. Periods with relatively low-amplitude waves, last- ing about 20 - 30 min were observed in the continu- ous records (Fig. 2A). We measured the amplitude of 100 well expressed waves from the records of 7 persons where the mean value and SEM were 8 1.13 f 20.61 pV. (Table 1). After these periods waves with higher amplitude compared to that of the former ones began to appear in the EGG and they characterized other periods - of gastric activity (Fig. 2B). The mean amplitude and SEM of these waves were 164.74 f 43.34 pV (n=7) (Table 1). The statistical processing showed a significant difference between the amplitudes of the waves in these two cases (Student’s t-test p < 0.05). In some studies there was an evident transition from high- amplitude to low-amplitude waves (Fig. 2C). In oth- er cases, e.g. the one of Figure 3A,B, the amplitude of the slow waves was lower compared with the one in Figure 2A,B. The transition from low-amplitude to high-amplitude slow waves was also evident (Fig. 3C). It seems that the low amplitude waves charac- terized the period of quiescence, while the high-am- Fig. 1. SCHEME OF THE POSITION OF THE ELECTRODES ON THE AB- plitude waves expressed the period of activity of the DOMINAL WALL. MMC. Usually the transition from one period to an- other was more gradual compared to that in Figures For personal use only. Table 1. Frequency and Amplitude of the Slow waves in the EGG of healthy volunteers. Fasting state After feeding Low -amp]. High-ampl. Waves 20-40 Low-ampl. High-ampl. Name waves waves Name min after waves waves feeding u A PV u A pv u A pv 1.K.S. 3.283 58.44 2,955 154.38 M.K.