- Email: [email protected]
In vitro electrical activity in human colon
GASTROENTEROLOGY
1981;81:502-8
In Vitro Electrical Activity in Human Colon M. M. CHAMBERS, K. L. BOWES, Y. J. KINGMA, C. BANNISTER, and K. R. COTE Departments of Surgery Alberta, Canada
and Electrical
Engineering,
Extracellular electrical activity was measured in vitro in 28 specimens of human colon taken at the time of operation. The mucosa was removed, and the muscle strips were mounted in an organ bath irrigated with oxygenated Krebs. Records were made from 3-7 silver/silver chloride electrodes placed directly on the circular muscle, for periods of 15-30 min. Data were stored on magnetic tape, digitized, analyzed by the fast Fourier transform method and plotted in three-dimensional form (signal power vs. frequency vs. time). Some plots showed a single frequency peak in the range 2-9 cpm, which was present throughout the study. However, multiple peaks were frequently seen and were of two types: (a) peaks of two or three closely related frequencies in the 2-9 cpm range, which divided and fused during the recording time. The power of the fused peak was greater than that of the components. And (bj peaks at frequencies which were integer multiples of the fundamental frequency; we interpret these to be the second and third harmonics of the fundamental frequency. When the fundamental frequency peak developed into several components, similar components also appeared at their individual harmonic frequencies, i.e., the second harmonic components were separated by twice the interval between the fundamental components. We conclude that only one fundamental frequency of electrical activity exists in any one site in human colon. Multiple frequencies are due either to the simultaneous recording from two or more poorly coupled electrical oscillators or to harmonics of the fundamental frequency. Received August 5,198O. Accepted April 16, 1981. Address requests for reprints to: M. M. Chambers, Department of Surgery, 11-105 Clinical Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2G3. This research was supported in part by the Medical Research Council of Canada. 0 1981 by the American Gastroenterological Association OOlS-5085/81/090502-07$02.50
University
of Alberta,
Edmonton,
Electrical activity of the stomach and small bowel has been extensively studied and is generally accepted as a major control system for contractile accolon electrical activity is tivity (1,2). However poorly understood. Electrical control activity [ECA] (also referred to as slow waves, pacesetter potential, basic electrical rhythm) has been described as intermittently present at two or more frequencies in in vivo studies of the dog and human colon (3-6). Other investigators have shown that electrical control activity is always present in human colon (73). Extracellular in vitro studies of the cat and dog colon have reported a single regular frequency that is always present (9-11). We have studied extracellular ECA in strips of human colon muscle obtained at laparotomy. Records were analyzed by fast Fourier transform and plotted in three-dimensional form (signal power vs. frequency vs. time). Evidence is offered that ECA in the human colon is always present at one frequency.
Methods Twenty-eight specimens of human colon were obat the time of operation from patients undergoing colonic resection for nonobstructing carcinoma. After resection of the diseased portion of the bowel, a 2.5cm ring of normal colon was removed from either the proximal or distal end of the future anastomotic site. Blood supply to this segment was maintained until the moment of excision. The specimen was immediately placed in oxygenated Krebs solution and transferred to the laboratory with minimum delay. Five specimens were obtained from ascending colon; 10 from transverse, 3 from descending, and 10 from sigmoid colon. In total, these specimens gave rise to 170 recording channels. tained
Each segment of colon was opened, the mucosa removed by sharp dissection and the tissue mounted in a bath perfused with oxygenated Krebs solution at 37’C. Three to seven monopolar, silver/silver chloride electrodes were placed on the submucosal surface at O.bcm intervals in a circular orientation as described previously
September 1981
ELECTRICALACTIVITYIN HUMANCOLON
($12). Recordings were made for 30-66 min on a Beckman Dynograph R411 polygraph and stored on magnetic tape. The Dynograph filters were set to give a low frequency cut-off at 0.16 Hz and a high frequency cut-off at 32 Hz. Records were examined visually by two observers independently, and an assessment was made of the pattern of the signal and its frequency. The data stored on magnetic tape were filtered with a low-pass filter at a cut-off frequency of 0.4 Hz to minimize aliasing; they were then digitized using a l-s sampling period. The data were processed, using the fast Fourier transform technique on a digital computer to obtain the frequency power spectra. A power spectrum was obtained for intervals of 256 s with each successive interval shifted by 64 s. The resulting spectra were plotted in three-dimensional form: signal power vs. frequency vs. time. This presentation allowed a good visual assessment of the frequency and duration of electrical activity in the specimens.
Results We recorded a total of 170 channels of electrical activity from the original 28 specimens of human colon. The records were very noisy, and only 8.2% of the channels could be satisfactorily analyzed visually. This proportion was significantly improved by using computer analysis, which allowed 44.1% of the channels to be assessed (Table 1). As a group, specimens taken from the ascending colon gave the least noisy records and hence the most reliable analysis. In the present study, a recognizable regular frequency could be intermittently distinguished visually in one or two channels in each study. This frequency was in the range 2-9/min with the rare exception which was seen in 2 of the 170 channels in which we saw fast activity of approximately 22 cpm, which lasted at most 90 s at a time. We could only visually recognize one frequency in the same electrode at any one time. The results are in sharp contrast to our experience with canine colon, using identical equipment and technique (9) in which a single, regular, universally present frequency of ECA could easily be recognized visually. Table 1. The Distribution of Colon Specimens and their Analyses
Specimen Ascendingcolon Transverse Descending Sigmoid TOTALS Percent of total
Total channels analyzed visually
Total channels analyzed by computer 25
so
4 6 3 1
170
14
75
Total channels 33 59 18
8.2
18 8 24
44.1
503
Using computer analysis of the original 28 specimens, we were able to analyze five specimens of ascending, five of transverse, two of descending, and eight of sigmoid colon: 44.1% of the total channels. A single frequency peak indicating a single frequency of electrical activity was seen in 25% of the analyses with the power of the peak diminishing so as to almost disappear at times (Figure l), while the remaining 75% showed multiple frequencies which were of two distinct types. In approximately 50% of these analyses, two or more almost identical frequencies occurred which fluctuated with time, often fusing and redividing, with increases in power during fusion and diminution of power when divided (Figure 2). The frequencies of this group ranged from 2 to 9/ min. In the other half of analyses with multiple frequencies, peaks occurred at higher frequencies which were integer multiples of the fundamental frequency of electrical activity (Figure 3). When the fundamental frequency peak developed into several components, similar components also appeared at their individual harmonic frequencies, and therefore the second harmonic components were separated by twice the interval between the fundamental components (Figure 4). Some specimens showed a harmonic frequency peak with greater power than that of the fundamental frequency (Figure 5). On no occasion did the computer analysis show two significantly different frequencies which could not be interpreted as harmonics. Simple spectra from fast Fourier transformational analysis showed that recordings from adjacent recording channels produced peaks which were at similar but not identical frequencies (Figure 6). The variation in frequency was well within the range to be expected from any biologic system. The spectrum shown in Figure 7 illustrates the second and third harmonic peaks recorded from a specimen of human transverse colon. No significant gradient of activity was seen over the length of the entire colon. Frequencies in the descending colon were significantly higher than elsewhere, but we were only able to obtain two such specimens and therefore only analyzed 8 channels (Table 2).
Discussion Previous studies of the electrical activity in human colon have reported two distinct frequencies or frequency ranges, each of which was present for only a portion of the recording time (4,5,7). Similar findings were initially reported in studies of canine colon (6). More recent in vitro studies of canine colon, however, have demonstrated a single regular electrical frequency that was always present (9). It was suggested that the noisy records and intermit-
GASTROENTEROLOGY
504
CHAMBERS ET AL.
Figure
1. Three-dimensional plot (signal power vs. frequency vs. time) after fast Fourier transformational analysis of data obtained from an extracellular electrode placed on the mucosal side of human colon tissue in vitro, after removal of the mucosa. A single frequency peak is seen throughout the study of approximately 4 cpm. Between minutes 17-20 of the recording, the power of the peak diminishes, which corresponds to the waxing and waning phenomenon which can be recognized visually in the original recording.
plot Fig;ure 2. A three-dimensional showing a single frequency of 3.5 cpm dividing after 20 min into two or three closely-related frequencies. The power of the single frequency peak is greater than that of the divided peaks.
PO
Vol. 81, No. 3
September
ELECTRICAL ACTIVITY IN HUMAN COLON
1961
505
Figure 3. Data plotted in three-dimensional form which shows signal peaks occurring at higher frequencies which are integer multiples of the fundamental frequency. We interpret these higher frequencies to be the second and third harmonics of the fundamental frequency.
P(
0
j
i
i
FREQUENCY
13
I?
--
(CPM1
Figure
P(
0
3
6
FREQUENCY
9
ICPHI
12
15
4. A three-dimensional plot showing a fundamental frequency of approximately 6 cpm, which divides and fuses into several components during the course of the study. The second harmonic frequency at approximately 12 cpm also has components which fluctuate in parallel with those of the fundamental frequency, for most of the study. The components of the second harmonic signal occur at frequencies which are harmonic to the corresponding fundamental frequency components, and therefore the components of the harmonic are separated by twice the frequency difference which separates the fundamental frequency components.
506
CHAMBERS
GASTROENTEROLOGY
ET AL.
Vol. 81, No. 3
-----
a. egg
_I _A
-
---
-- -E---.
_._.
-~
---_
-_-_
_
_-_
-~-
--.
-A
.-
-__
_-_
~-
-__.
.---_._:
.----
-
_
~__~~
Figure 5. Data presented in three-dimensional form which shows, between the 9 and 19 min of recording, a harmonic frequency peak which has considerably more power than does the fundamental peak.
___. _ ___ _
-
17. r7g
1-J -
---
FREOUENCY
tent appearance of a recognizable regular frequency in in vivo studies were due to the simultaneous recording from several poorly coupled oscillators. In the present study in which we measured extracellular electrical activity in human colon tissue in vitro, we obtained very noisy records. This was in striking contrast to our studies using identical equipment and techniques in canine colon where a clear record was always obtained (9). Tissue ischemia was unlikely to be responsible as the blood supply to the tissue studies was not interrupted until it was excised. The tissue was transported in oxygenated Krebs, and recordings were made within minutes of removing the specimen. The difference may partially be due to the obvious anatomic differences between canine and human colon. It is theoretically possible that muscle contractions could be more common in human colon specimens and introduce a motion artifact, although these were not observed. We believe the difference, is due at least in part, to an inherent difference in the character of human colon muscle electrical oscillators. Three-dimensional frequency power spectra (signal power vs. frequency vs. time) were plotted and showed the common occurrence of two or more closely related frequencies. When present, these fre-
-.--_
---L -.-__
(CPM1
-
I i
-
quencies often melded and divided during the period of recording. This observation supports the theory that the recordings emanate from small poorly coupled electrical osicllators, as was suggested in canine colon (12). When adjacent oscillators have an identical frequency (tight coupling), only a single peak is seen. When coupling is poor, the simultaneous recording of several oscillating regions would result in either several closely related frequencies or in a generally noisy record. The low percentage of records, which could be analyzed even with the computer, suggests that the coupling in human colon is generally poor. In the dog, the use of small contact electrodes in the in vitro recordings instead of the relatively large ones used in vivo, resulted in a marked improvement in the recordings. The same small electrodes used in the human colon did not improve the records to the same degree, which suggests that the oscillating regions are very small. Higher frequencies of electrical activity were often observed, but with two exceptions (see below) analysis showed that these were always at frequencies which were integer multiples of the fundamental frequency. We interpret these higher frequencies to be second and third harmonics of the fundamental frequency. When the fundamental frequency
September
1983
ELECTRICAL ACTIVITY IN HUMAN COLON
POWER
Table
--
2.
The Mean Frequencies of the ECA and Second Harmonic in Different Parts of Human Colon
Spectra from 3 Recordmg Channels
I
Mean frequency Mean frequency Specimen Ascending colon Transverse Descending Sigmoid
i.e., if the second
22
7.80
----FREQUENCY (CPM)
Figure 6. Spectra of data from three adjacent recording channels after fast Fourier transformational analysis, showing similar, but not identical, frequencies in the three channels. The specimen was human transverse colon.
peak developed into several components, the harmonic frequency also developed several similar components, each of which was at the individual harmonic frequency of the fundamental component:
PC
ER
11 50
-FREQUENCY (CPM)
Figure 7. A spectrum of data from human transverse colon, after fast Fourier transformational analysis. Second and third harmonic frequency peaks ace clearly seen.
n 5 5 2 8
of ECA/min 4.42 f 4.31 f 6.10 f 4.40 f
1.50 1.06 0.43 1.68
of second hacmonic/min 8.48 f 3.00 7.60 f 0.44 11.70 f
1.39
7.06 f 2.56
harmonic is seen, each of its components occurs at the second harmonic frequency of the corresponding fundamental component and will be separated by twice the frequency difference of the fundamental frequency components (Figure 4). In two of the 170 recording channels we saw brief periods of 22 cpm activity. Van Merwyk and Duthie (13) found slow wave activity at 22 f 4 cpm in human longitudinal muscle. Although our electrodes were placed on the circular muscle, perhaps the rare high frequency we see reflects an occasional measurement from the longitudinal muscle due to a damaged, or unusually thin, circular muscle layer. In some specimens, the harmonic frequency had significantly greater power than the fundamental frequency. In Figure 5, for example, the first 9 min of recording were very noisy; this was followed by a lo-min period in which the harmonic frequency peak at 7 cpm had a greater power than the fundamental peak at 3.5 cpm. This arises due to the a.c. coupling of the recording equipment and to the use of electrodes which have a capacitive effect (14). In this situation the harmonic components in the resulting wave form will be enhanced relative to the fundamental component. If the original signal already contains a strong second harmonic component, the ratio between signal power at the harmonic frequency and the fundamental frequency will become even greater. The computer-generated power spectrum will then show little or no power at the fundamental frequency and significant power at the second harmonic frequency (15). Occasionally the power of the fundamental frequency diminishes so as to almost disappear (Figure 1). This corresponds to a waxing and waning phenomenon which can be recognized visually in the original recordings and occurs when two signals at very similar frequencies are recorded simultaneously. If the waxing and waning periods are of the same order of magnitude as the periods for which the power spectra are obtained, then the signal power at the fundamental frequency can vary greatly from one spectrum to the next. This study suggests that essentially a single universally present ECA frequency exists in human co-
LL 0
507
508
CHAMBERS ET AL.
lon. Higher frequencies recorded intermittently in in vivo studies are probably second or third harmonics of the fundamental frequency. Coupling between electrical oscillators is frequently poor and varies considerably with time.
References 1. Bortoff A. Myogenic control of intestinal
2.
3.
4.
5. 6.
motility. Physiol Rev 1976;56:416-34. Daniel EE, Irwin J. Electrical activity of gastric musculature. In: Code CF, ed. Handbook of physiology, sect. 6, alimentary canal. Washington, DC: Am Physiol Sot, 1968;4:1969-84. Couturier D, Roze C, Couturier-Turpin MH, et al. Electromyography of the colon in situ. An experimental study in man and in the rabbit. Gastroenterology 1969;56:317-22. Snape WJ, Carlson GM, Cohen S. Colonic myoelectric activity in the irritable bowel syndrome. Gastroenterology 1976: 70~326-30. Taylor I, Duthie HL, Smallwood R, et al. Large bowel myoelectric activity in man. Gut 1975;16:808-14. Vanasin B, Ustach TJ, Schuster MM. Electrical and motor activity of human and dog colon in vitro. Johns Hopkins Med J 1974;134:201-10.
GASTROENTEROLOGY
Vol. 81, No. 3
7. Sarna SK, Bardakjian BL, Waterfall WE, et al. Human colonic electrical control activity (ECA). In: Christensen J, ed. Proc of the 7th Int Symposium on Gastrointestinal Motility, Iowa City. New York: Raven Press, 1979403-10. 6. Sarna SK, Bardakjian BL, Waterfall WE, et al. Human colonic electrical control activity (ECA). Gastroenterology 1960; 78:1526-36. 9. Shearin NL, Bowes KL, Kingma YJ. In vitro electrical activity in canine colon, Gut 1979;20:780-6. 10. Weinbeck M, Christensen J, Weisbrodt NW. Electromyography of the colon in the unanaesthetized cat. Am J Dig Dis 1972;17:356-62. 11. Christensen J, Caprilli R, Lund GF. Electrical slow waves in circular muscle of cat colon. Am J Physiol 1969;217:771-6. 12. Bowes KL, Shearin NL, Kingma. YJ, et al. Frequency analysis of electrical activity in dog colon, In: Duthie HL, ed. Gastrointestinal motility in health and disease. Proc of the 6th Int Symposium on Gastrointestinal Motility. MTP Press Ltd, 1978:251-68 13. Van Merwyk AJ, Duthie HL. Characteristics of human colonic smooth muscle in vitro. In: Christensen J, ed. Proc of the 7th Int Symposium on Gastrointestinal Motility, Iowa City. New York: Raven Press, 1979473-8. 14. Geddes LA, Baker LE. Principles of Applied Biomedical Instrumentation. New York: John Wiley and Sons Inc, 1666. 15. Malmstadt HV, Enke CG, Crouch SR. Electronic measurements for scientists. Menlo Park, California: WA Benjamin Inc, 1974.
1981;81:502-8
In Vitro Electrical Activity in Human Colon M. M. CHAMBERS, K. L. BOWES, Y. J. KINGMA, C. BANNISTER, and K. R. COTE Departments of Surgery Alberta, Canada
and Electrical
Engineering,
Extracellular electrical activity was measured in vitro in 28 specimens of human colon taken at the time of operation. The mucosa was removed, and the muscle strips were mounted in an organ bath irrigated with oxygenated Krebs. Records were made from 3-7 silver/silver chloride electrodes placed directly on the circular muscle, for periods of 15-30 min. Data were stored on magnetic tape, digitized, analyzed by the fast Fourier transform method and plotted in three-dimensional form (signal power vs. frequency vs. time). Some plots showed a single frequency peak in the range 2-9 cpm, which was present throughout the study. However, multiple peaks were frequently seen and were of two types: (a) peaks of two or three closely related frequencies in the 2-9 cpm range, which divided and fused during the recording time. The power of the fused peak was greater than that of the components. And (bj peaks at frequencies which were integer multiples of the fundamental frequency; we interpret these to be the second and third harmonics of the fundamental frequency. When the fundamental frequency peak developed into several components, similar components also appeared at their individual harmonic frequencies, i.e., the second harmonic components were separated by twice the interval between the fundamental components. We conclude that only one fundamental frequency of electrical activity exists in any one site in human colon. Multiple frequencies are due either to the simultaneous recording from two or more poorly coupled electrical oscillators or to harmonics of the fundamental frequency. Received August 5,198O. Accepted April 16, 1981. Address requests for reprints to: M. M. Chambers, Department of Surgery, 11-105 Clinical Sciences Building, University of Alberta, Edmonton, Alberta, Canada T6G 2G3. This research was supported in part by the Medical Research Council of Canada. 0 1981 by the American Gastroenterological Association OOlS-5085/81/090502-07$02.50
University
of Alberta,
Edmonton,
Electrical activity of the stomach and small bowel has been extensively studied and is generally accepted as a major control system for contractile accolon electrical activity is tivity (1,2). However poorly understood. Electrical control activity [ECA] (also referred to as slow waves, pacesetter potential, basic electrical rhythm) has been described as intermittently present at two or more frequencies in in vivo studies of the dog and human colon (3-6). Other investigators have shown that electrical control activity is always present in human colon (73). Extracellular in vitro studies of the cat and dog colon have reported a single regular frequency that is always present (9-11). We have studied extracellular ECA in strips of human colon muscle obtained at laparotomy. Records were analyzed by fast Fourier transform and plotted in three-dimensional form (signal power vs. frequency vs. time). Evidence is offered that ECA in the human colon is always present at one frequency.
Methods Twenty-eight specimens of human colon were obat the time of operation from patients undergoing colonic resection for nonobstructing carcinoma. After resection of the diseased portion of the bowel, a 2.5cm ring of normal colon was removed from either the proximal or distal end of the future anastomotic site. Blood supply to this segment was maintained until the moment of excision. The specimen was immediately placed in oxygenated Krebs solution and transferred to the laboratory with minimum delay. Five specimens were obtained from ascending colon; 10 from transverse, 3 from descending, and 10 from sigmoid colon. In total, these specimens gave rise to 170 recording channels. tained
Each segment of colon was opened, the mucosa removed by sharp dissection and the tissue mounted in a bath perfused with oxygenated Krebs solution at 37’C. Three to seven monopolar, silver/silver chloride electrodes were placed on the submucosal surface at O.bcm intervals in a circular orientation as described previously
September 1981
ELECTRICALACTIVITYIN HUMANCOLON
($12). Recordings were made for 30-66 min on a Beckman Dynograph R411 polygraph and stored on magnetic tape. The Dynograph filters were set to give a low frequency cut-off at 0.16 Hz and a high frequency cut-off at 32 Hz. Records were examined visually by two observers independently, and an assessment was made of the pattern of the signal and its frequency. The data stored on magnetic tape were filtered with a low-pass filter at a cut-off frequency of 0.4 Hz to minimize aliasing; they were then digitized using a l-s sampling period. The data were processed, using the fast Fourier transform technique on a digital computer to obtain the frequency power spectra. A power spectrum was obtained for intervals of 256 s with each successive interval shifted by 64 s. The resulting spectra were plotted in three-dimensional form: signal power vs. frequency vs. time. This presentation allowed a good visual assessment of the frequency and duration of electrical activity in the specimens.
Results We recorded a total of 170 channels of electrical activity from the original 28 specimens of human colon. The records were very noisy, and only 8.2% of the channels could be satisfactorily analyzed visually. This proportion was significantly improved by using computer analysis, which allowed 44.1% of the channels to be assessed (Table 1). As a group, specimens taken from the ascending colon gave the least noisy records and hence the most reliable analysis. In the present study, a recognizable regular frequency could be intermittently distinguished visually in one or two channels in each study. This frequency was in the range 2-9/min with the rare exception which was seen in 2 of the 170 channels in which we saw fast activity of approximately 22 cpm, which lasted at most 90 s at a time. We could only visually recognize one frequency in the same electrode at any one time. The results are in sharp contrast to our experience with canine colon, using identical equipment and technique (9) in which a single, regular, universally present frequency of ECA could easily be recognized visually. Table 1. The Distribution of Colon Specimens and their Analyses
Specimen Ascendingcolon Transverse Descending Sigmoid TOTALS Percent of total
Total channels analyzed visually
Total channels analyzed by computer 25
so
4 6 3 1
170
14
75
Total channels 33 59 18
8.2
18 8 24
44.1
503
Using computer analysis of the original 28 specimens, we were able to analyze five specimens of ascending, five of transverse, two of descending, and eight of sigmoid colon: 44.1% of the total channels. A single frequency peak indicating a single frequency of electrical activity was seen in 25% of the analyses with the power of the peak diminishing so as to almost disappear at times (Figure l), while the remaining 75% showed multiple frequencies which were of two distinct types. In approximately 50% of these analyses, two or more almost identical frequencies occurred which fluctuated with time, often fusing and redividing, with increases in power during fusion and diminution of power when divided (Figure 2). The frequencies of this group ranged from 2 to 9/ min. In the other half of analyses with multiple frequencies, peaks occurred at higher frequencies which were integer multiples of the fundamental frequency of electrical activity (Figure 3). When the fundamental frequency peak developed into several components, similar components also appeared at their individual harmonic frequencies, and therefore the second harmonic components were separated by twice the interval between the fundamental components (Figure 4). Some specimens showed a harmonic frequency peak with greater power than that of the fundamental frequency (Figure 5). On no occasion did the computer analysis show two significantly different frequencies which could not be interpreted as harmonics. Simple spectra from fast Fourier transformational analysis showed that recordings from adjacent recording channels produced peaks which were at similar but not identical frequencies (Figure 6). The variation in frequency was well within the range to be expected from any biologic system. The spectrum shown in Figure 7 illustrates the second and third harmonic peaks recorded from a specimen of human transverse colon. No significant gradient of activity was seen over the length of the entire colon. Frequencies in the descending colon were significantly higher than elsewhere, but we were only able to obtain two such specimens and therefore only analyzed 8 channels (Table 2).
Discussion Previous studies of the electrical activity in human colon have reported two distinct frequencies or frequency ranges, each of which was present for only a portion of the recording time (4,5,7). Similar findings were initially reported in studies of canine colon (6). More recent in vitro studies of canine colon, however, have demonstrated a single regular electrical frequency that was always present (9). It was suggested that the noisy records and intermit-
GASTROENTEROLOGY
504
CHAMBERS ET AL.
Figure
1. Three-dimensional plot (signal power vs. frequency vs. time) after fast Fourier transformational analysis of data obtained from an extracellular electrode placed on the mucosal side of human colon tissue in vitro, after removal of the mucosa. A single frequency peak is seen throughout the study of approximately 4 cpm. Between minutes 17-20 of the recording, the power of the peak diminishes, which corresponds to the waxing and waning phenomenon which can be recognized visually in the original recording.
plot Fig;ure 2. A three-dimensional showing a single frequency of 3.5 cpm dividing after 20 min into two or three closely-related frequencies. The power of the single frequency peak is greater than that of the divided peaks.
PO
Vol. 81, No. 3
September
ELECTRICAL ACTIVITY IN HUMAN COLON
1961
505
Figure 3. Data plotted in three-dimensional form which shows signal peaks occurring at higher frequencies which are integer multiples of the fundamental frequency. We interpret these higher frequencies to be the second and third harmonics of the fundamental frequency.
P(
0
j
i
i
FREQUENCY
13
I?
--
(CPM1
Figure
P(
0
3
6
FREQUENCY
9
ICPHI
12
15
4. A three-dimensional plot showing a fundamental frequency of approximately 6 cpm, which divides and fuses into several components during the course of the study. The second harmonic frequency at approximately 12 cpm also has components which fluctuate in parallel with those of the fundamental frequency, for most of the study. The components of the second harmonic signal occur at frequencies which are harmonic to the corresponding fundamental frequency components, and therefore the components of the harmonic are separated by twice the frequency difference which separates the fundamental frequency components.
506
CHAMBERS
GASTROENTEROLOGY
ET AL.
Vol. 81, No. 3
-----
a. egg
_I _A
-
---
-- -E---.
_._.
-~
---_
-_-_
_
_-_
-~-
--.
-A
.-
-__
_-_
~-
-__.
.---_._:
.----
-
_
~__~~
Figure 5. Data presented in three-dimensional form which shows, between the 9 and 19 min of recording, a harmonic frequency peak which has considerably more power than does the fundamental peak.
___. _ ___ _
-
17. r7g
1-J -
---
FREOUENCY
tent appearance of a recognizable regular frequency in in vivo studies were due to the simultaneous recording from several poorly coupled oscillators. In the present study in which we measured extracellular electrical activity in human colon tissue in vitro, we obtained very noisy records. This was in striking contrast to our studies using identical equipment and techniques in canine colon where a clear record was always obtained (9). Tissue ischemia was unlikely to be responsible as the blood supply to the tissue studies was not interrupted until it was excised. The tissue was transported in oxygenated Krebs, and recordings were made within minutes of removing the specimen. The difference may partially be due to the obvious anatomic differences between canine and human colon. It is theoretically possible that muscle contractions could be more common in human colon specimens and introduce a motion artifact, although these were not observed. We believe the difference, is due at least in part, to an inherent difference in the character of human colon muscle electrical oscillators. Three-dimensional frequency power spectra (signal power vs. frequency vs. time) were plotted and showed the common occurrence of two or more closely related frequencies. When present, these fre-
-.--_
---L -.-__
(CPM1
-
I i
-
quencies often melded and divided during the period of recording. This observation supports the theory that the recordings emanate from small poorly coupled electrical osicllators, as was suggested in canine colon (12). When adjacent oscillators have an identical frequency (tight coupling), only a single peak is seen. When coupling is poor, the simultaneous recording of several oscillating regions would result in either several closely related frequencies or in a generally noisy record. The low percentage of records, which could be analyzed even with the computer, suggests that the coupling in human colon is generally poor. In the dog, the use of small contact electrodes in the in vitro recordings instead of the relatively large ones used in vivo, resulted in a marked improvement in the recordings. The same small electrodes used in the human colon did not improve the records to the same degree, which suggests that the oscillating regions are very small. Higher frequencies of electrical activity were often observed, but with two exceptions (see below) analysis showed that these were always at frequencies which were integer multiples of the fundamental frequency. We interpret these higher frequencies to be second and third harmonics of the fundamental frequency. When the fundamental frequency
September
1983
ELECTRICAL ACTIVITY IN HUMAN COLON
POWER
Table
--
2.
The Mean Frequencies of the ECA and Second Harmonic in Different Parts of Human Colon
Spectra from 3 Recordmg Channels
I
Mean frequency Mean frequency Specimen Ascending colon Transverse Descending Sigmoid
i.e., if the second
22
7.80
----FREQUENCY (CPM)
Figure 6. Spectra of data from three adjacent recording channels after fast Fourier transformational analysis, showing similar, but not identical, frequencies in the three channels. The specimen was human transverse colon.
peak developed into several components, the harmonic frequency also developed several similar components, each of which was at the individual harmonic frequency of the fundamental component:
PC
ER
11 50
-FREQUENCY (CPM)
Figure 7. A spectrum of data from human transverse colon, after fast Fourier transformational analysis. Second and third harmonic frequency peaks ace clearly seen.
n 5 5 2 8
of ECA/min 4.42 f 4.31 f 6.10 f 4.40 f
1.50 1.06 0.43 1.68
of second hacmonic/min 8.48 f 3.00 7.60 f 0.44 11.70 f
1.39
7.06 f 2.56
harmonic is seen, each of its components occurs at the second harmonic frequency of the corresponding fundamental component and will be separated by twice the frequency difference of the fundamental frequency components (Figure 4). In two of the 170 recording channels we saw brief periods of 22 cpm activity. Van Merwyk and Duthie (13) found slow wave activity at 22 f 4 cpm in human longitudinal muscle. Although our electrodes were placed on the circular muscle, perhaps the rare high frequency we see reflects an occasional measurement from the longitudinal muscle due to a damaged, or unusually thin, circular muscle layer. In some specimens, the harmonic frequency had significantly greater power than the fundamental frequency. In Figure 5, for example, the first 9 min of recording were very noisy; this was followed by a lo-min period in which the harmonic frequency peak at 7 cpm had a greater power than the fundamental peak at 3.5 cpm. This arises due to the a.c. coupling of the recording equipment and to the use of electrodes which have a capacitive effect (14). In this situation the harmonic components in the resulting wave form will be enhanced relative to the fundamental component. If the original signal already contains a strong second harmonic component, the ratio between signal power at the harmonic frequency and the fundamental frequency will become even greater. The computer-generated power spectrum will then show little or no power at the fundamental frequency and significant power at the second harmonic frequency (15). Occasionally the power of the fundamental frequency diminishes so as to almost disappear (Figure 1). This corresponds to a waxing and waning phenomenon which can be recognized visually in the original recordings and occurs when two signals at very similar frequencies are recorded simultaneously. If the waxing and waning periods are of the same order of magnitude as the periods for which the power spectra are obtained, then the signal power at the fundamental frequency can vary greatly from one spectrum to the next. This study suggests that essentially a single universally present ECA frequency exists in human co-
LL 0
507
508
CHAMBERS ET AL.
lon. Higher frequencies recorded intermittently in in vivo studies are probably second or third harmonics of the fundamental frequency. Coupling between electrical oscillators is frequently poor and varies considerably with time.
References 1. Bortoff A. Myogenic control of intestinal
2.
3.
4.
5. 6.
motility. Physiol Rev 1976;56:416-34. Daniel EE, Irwin J. Electrical activity of gastric musculature. In: Code CF, ed. Handbook of physiology, sect. 6, alimentary canal. Washington, DC: Am Physiol Sot, 1968;4:1969-84. Couturier D, Roze C, Couturier-Turpin MH, et al. Electromyography of the colon in situ. An experimental study in man and in the rabbit. Gastroenterology 1969;56:317-22. Snape WJ, Carlson GM, Cohen S. Colonic myoelectric activity in the irritable bowel syndrome. Gastroenterology 1976: 70~326-30. Taylor I, Duthie HL, Smallwood R, et al. Large bowel myoelectric activity in man. Gut 1975;16:808-14. Vanasin B, Ustach TJ, Schuster MM. Electrical and motor activity of human and dog colon in vitro. Johns Hopkins Med J 1974;134:201-10.
GASTROENTEROLOGY
Vol. 81, No. 3
7. Sarna SK, Bardakjian BL, Waterfall WE, et al. Human colonic electrical control activity (ECA). In: Christensen J, ed. Proc of the 7th Int Symposium on Gastrointestinal Motility, Iowa City. New York: Raven Press, 1979403-10. 6. Sarna SK, Bardakjian BL, Waterfall WE, et al. Human colonic electrical control activity (ECA). Gastroenterology 1960; 78:1526-36. 9. Shearin NL, Bowes KL, Kingma YJ. In vitro electrical activity in canine colon, Gut 1979;20:780-6. 10. Weinbeck M, Christensen J, Weisbrodt NW. Electromyography of the colon in the unanaesthetized cat. Am J Dig Dis 1972;17:356-62. 11. Christensen J, Caprilli R, Lund GF. Electrical slow waves in circular muscle of cat colon. Am J Physiol 1969;217:771-6. 12. Bowes KL, Shearin NL, Kingma. YJ, et al. Frequency analysis of electrical activity in dog colon, In: Duthie HL, ed. Gastrointestinal motility in health and disease. Proc of the 6th Int Symposium on Gastrointestinal Motility. MTP Press Ltd, 1978:251-68 13. Van Merwyk AJ, Duthie HL. Characteristics of human colonic smooth muscle in vitro. In: Christensen J, ed. Proc of the 7th Int Symposium on Gastrointestinal Motility, Iowa City. New York: Raven Press, 1979473-8. 14. Geddes LA, Baker LE. Principles of Applied Biomedical Instrumentation. New York: John Wiley and Sons Inc, 1666. 15. Malmstadt HV, Enke CG, Crouch SR. Electronic measurements for scientists. Menlo Park, California: WA Benjamin Inc, 1974.