Electrical Abnormalities? Gut: First Published As 10.1136/Gut.32.2.141 on 1 February 1991

Home , Bowel resection

Gut, 1991,32, 141-146 141 Can transcutaneous recordings detect gastric electrical abnormalities? Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from

B 0 Familoni, K L Bowes, Y J Kingma, K R Cote

Abstract This study was undertaken to determine the The ability of transcutaneous recordings of ability of transcutaneous electrical recordings to gastric electrical activity to detect gastric recognise various types of gastric slow wave electrical abnormalities was determined abnormality. Implanted electrodes were used as by simultaneous measurements of gastric our 'gold standard'. The postoperative period electrical activity with surgically implanted afforded the opportunity to study both normal serosal electrodes and cutaneous electrodes in and abnormal states of gastric slow waves. six patients undergoing abdominal operations. Abnormalities in these after surgery have been Transient abnormalities in gastric electrical described using serosal electrodes."1'5 To exploit activity were seen in five of the six patients fully the potential of transcutaneous record- during the postoperative period. Recognition ings, both visual analysis and extensive filtering of normal gastric electrical activity by visual and analysis on a digital computer were per- analysis was possible 67% of the time and with formed. computer analysis 95% of the time. Ninety four per cent of abnormalities in frequency were detected by visual analysis and 93.7% by Methods computer analysis. Abnormalities involving a loss ofcoupling, however, were not recognised ELECTROGASTROGRAPHIC METHODS by transcutaneous recordings. Transcutane- Transcutaneous and serosal recordings ofgastric ous recordings of gastric electrical activity slow waves were made simultaneously on six assessed by computer analysis can usually patients undergoing laparotomy for fundoplica- recognise normal gastric electrical activity and tion, cholecystectomy, or bowel resection. The tachygastria. Current techniques, however, study had been approved by the ethics commit-

are unable to detect abnormalities in electrical tee of the University of Alberta Hospitals in http://gut.bmj.com/ coupling. October 1986 and informed written consent was obtained from each subject before the pro- cedure. Human gastric electrical activity consists of Three pairs of stainless steel wire electrodes recurrent regular depolarisations (slow waves) at 0.254 mm in diameter were implanted subseros- 2.5 to 4 cycles per minute, and intermittent high ally into the anterior surface of the antral wall,

frequency oscillations (spikes) that appear only between 15 cm and 10.5 cm from the pylorus. on September 25, 2021 by guest. Protected copyright. in association with contractions. Slow waves The Teflon coated connecting wires were begin high on the greater curvature and pass brought out through a 1 inch Penrose drain via a aborally towards the lesser curvature, accelerat- separate stab wound. Four ordinary Hewlett ing as the pylorus is approached. As spikes can Packard electrocardiography electrodes (type only occur during a portion of the slow wave 14445C) were placed on the skin in the epigastric cycle, slow waves are the ultimate determinant of region to form two pairs of bipolar recording the frequency and direction of propagation of electrodes. Each pair was aligned with the antral contractions. axis as previously described'9 or as close to it as Slow waves are best recorded in vivo by the surgical wounds permitted. surgically implanted electrodes. They have also The implanted electrodes were connected to a been recorded by percutaneous and transmuco- Beckman dynograph recorder with a low cut off sal techniques." The latter techniques are, frequency of 0-017 Hz (1 cycle/minute) and a however, noisy and recognition of gastric slow high cut off frequency of 30 Hz (1800 cycles/ waves is possible during part of the recording minute). The cutaneous electrodes were con- time only. In some patients, successful record- nected to the dynograph at a low cut off fre- ings are not obtainable. Spikes are generally not quency of 0.017 Hz (1 cycle/minute) and a high Departments of Surgery seen on transcutaneous recordings. cut offfrequency of0.08 Hz (4.8 cycles/minute). and Electrical Engineering, University In spite of the recording difficulties, abnor- The cut off frequency of 0-08 Hz was necessary of Alberta, Edmonton, malities in gastric slow wave activity have been to ensure rejection of electrocardiographic arte- Alberta, Canada described using these techniques in patients with facts from the transcutaneous recording. The B 0 Familoni K L Bowes unexplained nausea and vomiting,5 anorexia slope of this first order filter, however, was quite Y J Kingma nervosa,8 gastroparesis,`" pregnancy related gentle (6 dB/octave). The attenuation at the cut K R Cote nausea,'2 and motion sickness."'" Although offfrequency (4.8 cpm) was - 3 dB (or a factor of Correspondence to: comparing groups of patients with the same 0-71), and the attenuation of a signal at 9.6 cpm Dr Kenneth L Bowes, Department of Surgery, technique, as in these studies, lends credence to was still only -6 dB (or a factor of 0 5). MacKenzie Health Sciences the existence of an abnormality in a particular The signals at the dynograph amplifier out- Centre, Edmonton, Alberta, Canada T6G 2B7. group, the accuracy of transcutaneous electro- puts were sampled at a rate of 2 Hz and stored on Accepted for publication gastrograms in diagnosing an abnormality in a an IBM pc AT computer via a Lab Master 26 March 1990 single patient has not been determined. 200009 12 bit, 8 channel A to D converter. 142 Familoni, Bowes, Kingma, Cote

Recording was started in the recovery room and between 2.0 and 4 5 cycles/minute.'341920 Both continued for up to 120 hours. coupled (phaselocked) and uncoupled abnor-

malities were observed. The coupled abnormali- Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from ties included tachygastria, irregular slow waves, ANALYSIS OF RECORDINGS and electrical 'shutdown'. Tachygastria was All channels were subjected to both visual and defined as stable slow wave with a frequency computer analysis. The recorded signals were higher than 4.5 cpm.2s61012 Irregular slow waves visually inspected for regularity in the slow wave were characterised by variations of 10-60% in period, frequency, and presence of abnormali- the period ofsuccessive cycles ofthe slow waves. ties. The transcutaneous recording was An electrical 'shutdown' was assumed to occur inspected blind - that is, without reference to when no slow waves were seen in any channel for the serosal recordings. The percentage of time 40 seconds or more (corresponding to a maxi- during which the recorded signal was normal and mum slow wave frequency ofless than 1-5 cpm). the number and type of abnormalities recorded Two types of uncoupled (not phaselocked) were determined. These were then compared gastric slow waves were seen. In the first type the with a similar individual analysis of the serosal different sites had different frequencies and the recordings to determine the correlation and signals were irregular. In the second type the accuracy ofthe transcutaneous method. same frequency was evident at different sites, We defined normal gastric slow waves in but the signals were not coupled. accordance with several previous reports of Computer analysis consisted of sequential gastric slow waves in healthy people - that is, power spectrum analysis over successive 128 regular activity with a fundamental frequency second 50% overlapping intervals and cross r~~~ 1 minE1 20iuV 1 min 'v * ? 1.\ ~\A I20,u 1V F-4---.4. 2 , E2 I 200,uV

I 200,uV --- --p r 7-77rrv

E3 http://gut.bmj.com/

- 0 - wr- v - 0S S - w- 0 # f- - w - Eo >H1 \WXQAu\4,mn.4 °I20,uV I200uV

E31 El on September 25, 2021 by guest. Protected copyright. I200DV F.91 min 0.1 I_.*1 minwIm E2 E2 I200sV

E 3-t

31------,

Eo men o1 u'[-c mi E o El 1 1 I200mV I200iuV E2 2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ E3 --'n < t V0t- E3tttt Figure 1: Gastric electrical activity recorded simultaneously by cutaneous (E0) and serosal (E1-E3) electrodes. (A) Normal gastric slow waves (GSW) (2-8 cpm) recorded by both cutaneous and serosal electrodes. (B) Tachygastria (6 cpm) evident in both serosal and cutaneous records. E1 has become loosely implanted and displays low signal to noise ratio. (C) Irregular, but coupled, GSWevident in both serosal and cutaneous records. (D) GSW 'shutdown' marked by absence ofdiscernible signal in serosal and cutaneous electrodesfollowed by return ofslow waves in cutaneous and one serosal electrode. The third serosal electrode in this patient hadpreviously broken off. (E) Uncoupled regular GSW with differentfrequencies in different parts ofthe stomach (E1, E2 2-9 cpm, E3 4-2 cpm). The cutaneous electrode ignores the frequency ofE3 but accurately reflects E1 and E2. (F) Uncoupled and irregular GSW. The boken lines in the serosal recording link successive slow waves in adjacent channels. The variable pattern ofthese lines indicates uncoupling. The cutaneous electrode's signal does not accurately reflect the actual gastric events. Can transcutaneous recordings detectgastric electricalabnonnalities? 143

correlation. Fast Fourier transform with a room (0 to 1 hour postoperatively), the slow Hamming window were calculated for the 256 waves in all six patients remained regular, with a

samples in each time frame to determine the mean (SD) frequency 2191 (0.48) cpm (0049 Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from signal frequencies. Power spectrum analysis (0.008) Hz) (Fig IA). breaks down a signal, permitting determination Slow wave abnormalities were seen in five offrequencies ofany rhythmical activities even if patients shortly after the recovery room period; the signal is corrupted by random noise;2' a the abnormalities lasted up to 10.75 (4-61) hours difference in the fundamental frequencies ofany postoperatively in four, and persisted until the two signals indicates that they are not phase implanted electrodes were removed (120 hours) locked. Cross correlation determines which fre- in the fifth. We had suspected from intraluminal quencies, if any, are common to two given and transcutaneous measurements before the signals. This method was employed to determine operation that this fifth patient had gastric whether the slow wave was coupled or not.20 electrical abnormalities. Periods of nursing care or patient movement One hundred and fifty six separate episodes of as marked on the recording paper by the nursing slow wave abnormalities were recorded from the staff were excluded from the analysis. The serosa (Table I). Coupled (tachygastria, irregular implanted electrodes were removed with the slow wave, electrical 'shutdown') and uncoupled drain when the latter was no longer needed 72 to abnormalities were observed. Tachygastria was 120 hours after its insertion. the commonest abnormality detected in both the transcutaneous and gastric recordings (Fig iB). Eleven episodes of irregular slow waves lasting Results an average of seven minutes were recorded (Fig IC). Electrical 'shutdown' was recorded five SEROSAL RECORDINGS times by the serosal electrodes with an average Serosal electrodes gave excellent recordings of duration of 3.5 (2.6) minutes (Fig ID). gastric slow waves. Out of a total recording Nine episodes of uncoupled activity were duration of491 .9 hours, 7.3 hours were unusable recorded. In four there were different frequen- because of motion artefacts from nursing care. cies at separate sites (Figs 1E and F). Data analysis was restricted to the remaining 484.6 hours ofrecording. Normal slow wave was recorded for most of CUTANEOUS RECORDINGS the recording time (76 84%). In the recovery Cutaneous recordings reflected actual gastric electrical events poorly (Table II). Normal

TABLE I Gastric slozv wave (GSW) abnormalities recorded slow waves were recognised by cutaneous http://gut.bmj.com/ at (A) serosal electrodes, (B) cutaneous electrodes, and (C) electrodes 67% of the time. those recorded simultaneously at both sites only Surprisingly, tachygastria was recognised more often (91% of No of Duration (min) Total the time). Irregular electrical activity and episodes (mean (SD)) (hrs) electrical 'shutdown' pauses were often not (A) Serosal electrodes recognised; they were recognised at the Tachygastria 131 45.9 (74.3) 100-22 cutaneous IrregularGSW 11 7(2.6) 1-28 recording only 63% and 67% of the

Electrical shutdown 5 3.5 (2.6) 0-29 time respectively. Uncoupled gastric electrical on September 25, 2021 by guest. Protected copyright. Uncoupled I 4 217-3 (92A4) 14-49 events were not in any cutaneous Uncoupled II 5 52 (2.7) 0.43 recognised Total 156 116-71 recording. During uncoupling cutaneous (B) Cutaneous electrodes electrodes did not reflect slow Tachygastria 127 47 (74.9) 99.48 accurately gastric IrregularGSW 12 7-3(2.3) 1-46 wave (Fig 1E and F). Electrical shutdown 3 3-6(2.7) 0-18 Computer the Uncoupled I 0 0 (0) 0 00 analysis dramatically improved Uncoupled II 0 0 (0) 0 00 ability of transcutaneous recordings to detect Total 142 101-12 events For (C) Abnornalities recorded simultaneously by both methods coupled gastric (Table II). instance, Tachygastria 122 46.3 (74.6) 94-14 normal slow waves were recognised 95% of the Irregular GSW 9 7-1(2.2) 1-07 time. slow waves and Electrical shutdown 3 3-6(2.7) 0-18 recording Irregular Uncoupled I 0 0 (0) 0.00 electrical 'shutdown' were not diagnosed with Uncoupled II 0 0 (0) 0.00 greater was Total 134 95.39 precision, however, and uncoupling again not recognised. Detection ofirregular slow

TABLE II Accuracy oftranscutaneous recordings in diagnoses ofnormal and abnormalgastric slow wave (GSW) (visual analyses) Cutaneous electrical recording(%) Gastric slow wave activity Irregular Elctrical (Serosal electrodes) (hrs) Normal Undecipherable Tachygastria GSW shutdown Uncoupled (A) Visual analyses Normal 369-05 67-01 31.02 1-96 0.01 0 0 Tachygastria 100.4 0 9.11 90-89 0 0 0 Iregular GSW 1-4 0-71 35.71 1-43 62.86 0 0 Electrical shutdown 0.3 0 26-67 13.33 0 66-67 0 UncoupledI 15-1 35-1 43-71 20.53 0-66 0 0 Uncoupled II 0.4 0 97.5 2-5 0 0 0 (B) Computer analyses Normal 367.88 95.44 3 85 0.71 0 0 0 Tachygastria 100-22 0 5 82 93.93 0-25 0 0 IrregularGSW 1-29 0.78 15-5 0.73 82.95 0 0 Electrical'shutdown 0-29 0 27.59 10-34 0 62-07 0 Uncoupled I 14-25 22-46 57.54 18-95 1-05 0 0 Uncoupled II 0 43 0 97-67 2.33 0 0 0 144 Familoni, Bowes, Kingma, Cote 0 Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from

EO O I20/uV 1 mi In \ CL A C E 0~ B D

6 9 12 15 24.12 166-1'

L- W a) 0: 0 CL 0

15 0 3 6 9 12 15 127.31 , 2.35

5L a, 0~ 0 http://gut.bmj.com/

A 111 : 11 1 -"- - ' - 1 1 0 3 6 9 12 15 0 3 6 9 12 15 Frequency (cpm) Frequency (cpm) Figure 2: (A) Cutaneous (Eo and serosal (E3) recordings ofgastric slow wave (GSW) during a period ofshiftingfrequencies. Sequentialfrequency power spectra of128 second overlapping intervals (A, C, D, E) ofEo and were obtained and are B, E3 on September 25, 2021 by guest. Protected copyright. illustrated in 2A-F. The amplitude scales are arbitrary. The majorpeaks are labelled with theirfrequencies. (B) Eo and E3 have thefundamentalfrequency of4.23 cpm in common. The term fundamental is the repetitionfrequency ofa signal. When this periodic signal is highly non-sinusoidal in shape the harmonic content increases - that is, integrar multiples or overtones ofthefundamentalfrequency are seen.25 Note that the harmonics ofE3 are prominent and thefundamental is small. (C) Eo and E3 deviatefrom the common frequency in B. The wide peak ofEo at 3177 cpm indicates shiftingfrequencies. (D) Eo and E3 are both settling to a lower common frequency. (E and F) Eo and E3 have a sharpfundamental peak in common at 2135 cpm.

waves by power spectrum analysis was con- pattern in the power spectra analysis as shown in tingent on the observer detecting particular Figure 2B to F. At the height of the transition, trends in sequential frames of the power spectra two distinct slow wave frequencies that were not and was therefore quite subjective. For example, harmonically related (2.85 and 4.69 cpm) were the variations in frequency of the irregular slow seen (Fig 2D). waves (Fig 2) resulted in an unstable frequency When the transcutaneous signal was normal,

TABLE III Gastric slow waves (GSW) during various patterns oftranscutaneous recordings Gastric electrcal activity (%) Irregular Electrcal Uncoupled Uncoupled Cutaneous recording (hrs) Normal Tachygastra GSW shutdown I II (A) Visual analyses Normal 251-04 94-8 3 0-6 0 2-1 0 Undecipherable 127-81 88-2 7-4 0-2 01 4-1 0.1 Tachygastria 103-96 13 81-5 0.3 0-02 4-9 0-27 IrregularGSW 0.97 4-12 2-06 71-13 0 22-68 0 Electrical shutdown 0-17 0 23.53 0 70.59 0 11-76 Uncoupled 0 0 0 0 0 0 0 (B) Computer analyses Normal 341-06 97.34 1.55 0-02 0 1-08 0 Undecipherable 39-2 76.52 6-89 1-73 0.05 14-29 0.51 Tachygastria 103-62 5.4 88-88 0-21 0 5-31 0.19 IrregularGSW 0-63 26-98 1.59 47-62 0 17-46 6.35 Electrical shutdown 0-26 3.85 11-54 0 65 38 0 19-23 Uncoupled 0 0 0 0 0 0 0 Can transcutaneous recordings detectgastric electrical abnormalities? 145

8.3 Cutaneous recording can recognise 67% of EO normal gastric slow waves, 90% of tachygastria, 62% ofirregularities in gastric electrical activity, Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from and 67% of electrical 'shutdown'. If computer analysis is used these figures alter to 95%, 93%, 82%, and 62% respectively. j j A 11-32 When cutaneous electrodes indicate a normal a) record they are accurate 95% of the time. A 0 cutaneous recording oftachygastria is confirmed 82% ofthe time. Irregularities in slow waves and j'l~~~~~~~11 electrical 'shutdown' are accurate reflections of the true situation in 71 and 70% of the time. A4-23 I . These figures are altered by computer analysis to 97%, 89%, 47%, and 65%. Cutaneous records, however, cannot pick up uncoupling (Fig 3). Homogeneous linear con- _ ducting models of the human torso2' 23 suggest 12 15 that this is probably because electrogastrographs Frequency (cpm) represent the summation of all activity and Figure 3: Frequency power spectra ofuncoupled regular electrical~~~~~~~activity (Fig JA). The reflect the major mass of tissue. If most of the cutaneous recording Eo duplicated E 1, but ignored E3. Eo (cutaneous) ,:and E 1 (serosal) exhirbit the same strongfundamentalfrequency at 2.85 cpm. E1 also contains several harmonics oft) uis stomach is normally coupled, a regular record at frequency. Serosal electrode E3 however does not contain a component at 2 85 cpm, but instead a normal or fast frequency is obtained (Fig IE) it has peaks at 4-23 cpm. and a very significant underlying gastric disorder could be missed. the actual slow waves were usually normal al[so Computer analysis is obviously of consider- (94-8%: Table III). Unfortunately, a lar.ge able value in interpreting gastric electrical portion of transcutaneous recordings we~re activity recordings. Without it 26% of records undecipherable. The transcutaneous recording are undecipherable, whereas with it only 8% of tachygastria less accurately reflected the tr*ue cannot be deciphered. With computer analysis, gastric event (80%). Transcutaneous recordinlgs transcutaneous gastric electrical activity is in ofirregular slow waves and electrical 'shutdowin' certain instances of definite value. If a normal were even less precise. No noticeable impro%ve- recording is obtained it is highly likely that the ment occurred in the detection of these abn(or- patient's gastric slow wave activity is normal.

malities after computer analysis. Similarly, tachygastria will be an accurate reflec- http://gut.bmj.com/ tion ofthe situation in the stomach in nine out of 10 subjects. The commonest abnormality Discussion reported in patients with gastric electrical abnor- The concept of cutaneous electrodes diagnosiing malities is, in fact, tachygastria.5 8 10 It might be abnormalities in gastric electrical activity 1has asked whether tachygastria is ofmuch functional great appeal. Electrocardiograms are obtainled importance. Although a frequency of 6 per

using similar techniques and there is litttle minute indicates a change in the state of the on September 25, 2021 by guest. Protected copyright. question of their great value in the diagnosis of gastric oscillators, it is not established that cardiac disease. The electrogastrogra[m, contractions occurring at a frequency of 6 per however, differs significantly from the elect]ro- minute are less efficient than those occurring at 3 cardiogram. The amplitude of the signal is per minute. extremely low (average 0.05 mV on the electro- The important abnormalities in gastric slow gastrogram v 1. 6 mV for the R wave on the wave activity are probably those involving an electrocardiogram). This necessitates a consid[er- orad spread or a loss of coupling of the electrical ably increased gain and therefore an apprecialble slow waves. As gastric slow wave determines the increase in artefact. Part ofthis artefact is caussed frequency and direction of propagation of by the motion of respiration or movement of 1the contractions, these abnormalities could have a gut wall and part is probably from superimpc)si- profound effect on gastric contractions and tion of signals arising in multiple sites in the emptying. If the gastric oscillators were coupled small bowel and colon. Unlike the heart, only a in an orad direction, contractions could only minority of the signals recordable (slow waves) move in that direction resulting in delayed are associated with contractions. emptying and possibly vomiting. A loss ofcoupl- Definite abnormalities in gastric electriical ing in the antrum would produce incoordinated activity have been described in only a few contractions and retard gastric emptying con- conditions and usually with non-implantted siderably. Unfortunately, this study indicates electrodes. In all such recordings, a good recoird- that neither of these abnormalities can be ing may not be obtained for much of the tinne. accurately assessed with transcutaneous record- The abnormalities described are usually prestent ings. The practical value of the test is, therefore, for a portion of the recording time only and 1the severely limited. Future endeavours should incidence of noisy unanalysable records is hig,h. focus on techniques for improving non-invasive It is possible that some of the abnormalities recordings ofgastric electrical activity before too described are recording artefacts. Some recoird- much reliance is placed on recordings made ing artefacts may be secondary to the abnornnal using current techniques. state and not the cause of it - for example, We offer a final note ofcaution. Often a record heaving before vomiting or gastric dilatation or is noisy. It is tempting to interpret apparent displacement from any cause. irregularities in transcutaneous recording to be a 146 Familoni, Bowes, Kingma, Cote

true reflection of real gastric electrical activity. 11 Telander RL, Morgan KG, Kreulen DL, Schmalz PF, Kelly KA, Szurszewski JK. Human gastric atony with tachy- Such interpretations could be dangerous, par- gastria and gastric retention. Gastroenterology 1978; 75: 497-

ticularly ifthey led to therapeutic interventions. 501. Gut: first published as 10.1136/gut.32.2.141 on 1 February 1991. Downloaded from 12 Koch KL, Creasey GW, Dwey A, Vasey M, Stern RM. Gastric dysrhythmia and nausea of pregnancy [Abstract]. Dig Dis This work was supported in part by grant MA-5946 from the Sci 1987; 32: 909. Medical Research Council ofCanada. 13 Stern RM, Koch KL, Liebowitz HW, Lindblad IM, Shupert CL, Stewart WR. Tachygastria and motion sickness. Aviat Space Environ Med 1985; 56: 1074-7. 1 Hamilton JW, Bellahsene BE, Reichelder M, Webster JG, 14 Stern RM, Koch KL, Stewart WR, Lindblad IM. Spectral Bass P. Human electrogastrograms: comparison of surface analysis of tachygastria recorded during motion sickness. and mucosal recordings. DigDisSci 1986; 31: 33-9. Gastroenterology 1987; 92: 92-7. 2 Abell TL, Malagelada J-R. Glucagon evoked gastric 15 Rague BW, Oman CM. Use of a microcomputer system for dysrhythmias in humans shown by an improved electro- running spectral analysis of EGG's to predict the onset of gastrographic technique. Gastroenterology 1985; 88: 1932- motion sickness. Proc IEEE 9th Annual ConfofEng in Med 40. and Biol Society, 1987: 87-90. 3 Bellahsene BE, Hamilton JW, Webster JG, Bass P, Reich- 16 Nelsen TS, Kohatsu S. Clinical electrogastrography and its felder M. An improved method for recording and analyzing relationship to gastric surgery. AmJ'Surg 1968; 116: 215-22. the electrical activity of the human stomach. IEEE Trans 17 Sarna SK, Bowes KL, Daniel EE. Post-operative gastric BiomedEng 1985; 32: 911-5. electrical control activity in man. In: Daniel EE, ed. 4 Familoni BO, Kingma YJ, Bowes KL. A study of trans- Gastrointestinal motility. Vancouver: Mitchell Press, 1974: cutaneous and intraluminal measurements of gastric 73-83. electrical activityin humans. MedBiolEngComput 1987; 25: 18 McIntyre JA, Deitel M, Baida M, Jail S. The human 397-402. electrogastrogram at operation: a preliminary report. Can J 5 Geldof H, Van Der Schee EJ, Van Blankenstein M, Grashuis Surg 1%9; 12: 275-84. JL. Electrogastrographic study of gastric myoelectrical 19 Mirrizi N, Scafoglieri U. Optimal direction of the electro- activity in patients with unexplained nausea and vomiting. gastrographic signal in man. MedBiolEngComput 1983; 21: Gut 1986; 27: 799-808. 385-9. 6 You CH, Chey WY. Study ofelectromechanical activity ofthe 20 Familoni BO, Kingma YJ, Bowes KL. Noninvasive assess- stomach in humans and in dogs with particular attention to ment ofhuman gastric motor function. IEEE Trans Biomed tachygastria. Gastroenterology 1980; 86: 1460-8. Eng 1987; 34: 30-6. 7 You CH, Lee KY, Chey WY, Mengur R. Electrogastrographic 21 Ramirez W. The FFTfundamentals and concepts. New York: study of patients with unexplained nausea, bloating and Prentice-Hall, 1985. vomiting. Gastroenterology 1980; 79: 311-4. 22 Code CF, Szurszewski JH, Kelly KA, Smith IB. A concept 8 Abell TL, Malagelada J, Lucas AR, et al. Gastric electro- of control of gastrointestinal motility. Handbook of mechanical and neurohormonal function in anorexia physiology: alimentary canal, Section 6. Vol. 5. Washington nervosa. Gastroenterology 1987; 93: 958-65. DC: American Physiological Society. 9 Abell TL, Camilleri M, Malagelada J-R. High prevalence of 23 Plonsey R. Bioelectric phenomena. New York: McGraw-Hill, gastric electrical dysrhythmias in diabetic gastroparesis. 1979:202-9. Gastroenterology 1985; 88: 1299. 24 Kingma YJ, Durdle NG, Bowes KL, Kocylovski M, Szmidt J. 10 Pfister CJ, Hamilton JW, Nagel N, Bass P, Webster JG, Size of electrical oscillating regions in the canine colon. In: Tompkins WJ. Use of spectral analysis in the detection of Christensen J, ed. Gastrointestinal motility. New York: frequency differences in the electrogastrograms of normal Raven Press, 1980: 425-31. and diabetic subjects. IEEE Trans Biomed Eng 1988; 35: 25 Kingma YJ. Spectral analysis ofgastrointestinal electrical signals. 935-41. Automedica, 1987: 237-48. http://gut.bmj.com/ on September 25, 2021 by guest. Protected copyright.