SEARCH

SEARCH BY CITATION

Keywords:

  • electrophysiology;
  • filtering;
  • noise;
  • signal processing;
  • slow wave;
  • spike

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

Extracellular electrical recordings underpin an important literature of basic and clinical motility science. In the November 2011 edition of Neurogastroenterology and Motility, Sanders and colleagues reported that contraction artifacts could be recorded from in vitro murine gastric tissues using extracellular electrodes, and that true extracellular bioelectrical activity could not be detected when the contractions were suppressed. The authors interpret their findings to mean that previous extracellular studies have generally assayed contraction artifacts, rather than bioelectrical activity, and suggest that movement suppression is an obligatory control for extracellular experiments. If their interpretation is correct, these claims would be significant, requiring a reinterpretation of many studies, and posing major challenges for future in vivo and especially clinical work. However, a demonstration that motion artifacts can be recorded from murine in vitro tissue does not necessarily mean that other extracellular studies also represented artifacts. This viewpoint evaluates a recently published by Sanders and colleagues in light of the competing literature, and finds a considerable volume of evidence to support the veracity of GI extracellular electrical recordings. It is reasoned from biophysical principles, technical considerations, and experimental studies that motion artifacts cannot explain GI extracellular electrical recordings in general, and that bioelectrical fact and artifact can be readily and reliably distinguished in most contexts. Calls for obligatory motion suppression for extracellular studies are therefore not supported. However, the artifacts recorded by Sanders and colleagues nevertheless serve as a reminder that educated caution is needed when recording, filtering and interpreting extracellular data.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

Among his pioneering observations on GI extracellular recordings, Walter C. Alvarez famously stated in 1922 that: ‘The most remarkable thing about these action currents is that they are produced constantly in stomachs and intestines which, so far as the eye can detect, are absolutely motionless.’1

In the November 2011 edition of Neurogastroenterology and Motility, Sanders and colleagues prompt fresh debate about the presence and significance of motion artifacts in GI extracellular recordings.2 Using extracellular electrodes, the authors recorded ‘oscillating biopotentials’ from in vitro murine gastric tissues. They propose that these biopotentials ‘are similar to the activity recorded with extracellular electrodes from visceral organs of various species’ in previous studies. The extracellular biopotentials were eliminated when contraction suppressants (nifedipine and wortmannin) were administered, whereas intracellular slow wave activity could still be recorded with microelectrodes, indicating that the biopotentials were contraction artifacts. In addition, the artifacts were found to occur at a lower frequency than intracellular slow waves (5 vs 8 c min−1).

In an extensive discussion, the authors interpret these results to have far-reaching significance. They suggest that published extracellular recordings have, in general, assayed contractile activity rather than electrical activity, and propose that their study ‘raises questions about the suitability of extracellular recordings for analysis of slow waves.’ They recommend that movement suppression should be an obligatory control when making extracellular recordings.

These are remarkable claims, and, if correct, they would be important. Extracellular electrical recordings underpin a significant motility literature, and they continue to be developed and applied to advance normal physiology, pathophysiology, diagnosis and therapy.3–5 If extracellular recordings are generally artifacts, then such studies would need reinterpretation. A requirement for movement suppression would also pose critical challenges for new in vivo work, and especially clinical studies, where the routine administration of contraction suppressants would be difficult or impossible.

However, the finding that motion artifacts can be recorded from murine tissues does not mean that other extracellular recordings are also artifacts, or that artifact and fact are necessarily difficult to distinguish. This viewpoint examines the study by Sanders and colleagues in light of competing literature, discusses the validity of extracellular recordings, and identifies ensuing research challenges.

The Biological Basis for Extracellular Slow Wave Recordings

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

Understanding the biophysical basis of extracellular potentials enables a mechanistic interpretation and is important for differentiating signal from noise. Although slow waves propagate via a unique interstitial cell of Cajal (ICC) entrainment mechanism, the general principles of their extracellular detection are universal among bioelectric tissues, and are well established.6,7

Extracellular electrodes detect the extracellular potential fields arising from activation fronts. Axial current flow occurs from the inactive to active tissue regions, which act as a ‘source’ and ‘sink’ respectively, establishing a dipole. In typical recordings, the potentials in extracellular recordings are positive in front of the wavefront, zero at the point of activation, and negative behind the wavefront, giving rise to the classic ‘biphasic’ waveform.6–8 The point of maximal negative gradient corresponds to the arrival of the wavefront directly under the electrode, and is used to determine the local ‘time of activation.’9

Notably, the ‘biopotentials’ recorded by Sanders and colleagues can be readily differentiated from extracellular slow wave activity, being small amplitude ‘spike-like oscillations’ instead of periodic biphasic events.2 The biopotential artifact time course (approximately 5 s) also matched the duration of intracellular slow waves and video-mapped contractions, which is a much longer duration than that of extracellular slow waves.3,10,11 Therefore, the recordings by Sanders and colleagues should not be considered representative of extracellular recordings in general from various species.

‘Expected Kinetics’ of Extracellular Recordings

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

In their discussion, Sanders and colleagues present a theoretic study on the ‘expected kinetics’ of extracellular slow wave activity.2 On the basis of Ca2+-sensitive dye studies of ICC excitation, they derive an anticipated extracellular slow wave time course of ‘approximately 100 ms.’ This derivation is used to support several claims, such as that published extracellular data ‘seem inconsistent with the kinetics of slow waves’, and that ‘transients displayed in many studies … may be the result of muscle movement.’

However, the assumptions underlying this theoretic study warrant scrutiny, as Ca2+-sensitive dye studies do not accurately reflect what is recorded by extracellular electrodes. For example, the authors assume that a 400 μm square field of tissue adequately reflects the area ‘sensed’ by an extracellular electrode of diameter 300 μm. In reality, extracellular electrodes detect electrical fields at hundreds of μm in all directions beyond the electrode.7,12,13 Although this relationship reduces at an exponential rate, sensitivity is greatly aided using modern high-fidelity digital amplifier systems. The large size of the tissue field detected using extracellular electrodes was effectively demonstrated in the recent recording of ‘double potentials’ at the center of gastric slow wave re-entry.12 Double potentials reflect the simultaneous detection of wavefronts from across both sides of a functional conduction block, at distance from the electrode.12,14

Another assumption in the study by Sanders and colleagues was that murine ileum ICC networks are representative of large animal gastric ICC networks, to which the ‘expected kinetics’ were generalized.2 However, normal human corpus has a more complex layered structure with a less well-defined ICC-MY network,15 and slow wave propagation may also occur through ICC-IM,16 in close proximity to surface electrodes.

The anticipated extracellular slow wave time course of ‘approximately 100 ms’ is therefore likely to be a substantial underestimate of the expected slow wave duration. Future studies should improve this work, and it would be valuable to employ quantitative methods that appropriately incorporate the relevant extracellular field biophysics.

Variability in Extracellular Slow Wave Morphology

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

The suggestion by Sanders and colleagues that extracellular waveform variability is explained by electrode ‘movement trajectories,’2 rather than by true bioelectrical phenomena, echoes classic debates in cardiac and neural electrophysiology.17,18 Although movement artifacts may contaminate extracellular recordings, waveform variability can again be better explained by biophysical and technical considerations.

Conduction velocity, for example, has a direct relationship with extracellular slow wave amplitude, such that signal-to-noise ratio (SNR) increases when waves travel faster. This relationship can be explained by proportionality between velocity and transmembrane current passing extracellularly,19 but not by motion artifacts. Species and regional differences also contribute to variability; for example, waveforms in the gastric pacemaker region and distal antrum (where conduction is fast) show a higher amplitude and consistent high SNR, whereas normal waves recorded from the corpus have reduced SNR, and can be ‘fractionated’ (having several components).3,10,11 Fractionation could result from heterogeneity in activation front propagation through layered tissue structures, anisotropy, wavefront curvature, and tissue discontinuities.7

Waveform variability is more pronounced during gastric dysrhythmia.4,5,12 This variability is reproducible, and can be explained by bioelectrical mechanisms such as the emergence of rapid anisotropic conduction, interference at wavefront collisions5,12 wavefront fragmentation into multiple wavelets,5 and double potentials.12

Technical factors also contribute to waveform variability. Different morphologies are achieved by unipolar and bipolar approaches.20 Digitization, amplification and filtering can introduce distortions,13 and variable electrode contact, metal choice, shielding, and contact junction design are important considerations. Flexible printed circuit board (PCB) arrays, for example, are suited to human use, but record at a reduced and more variable SNR than custom-built platforms.21

Filtering of Extracellular Recordings

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

Noise is ubiquitous during bioelectrical recordings, and judicious filtering can de-emphasize artifacts. However, the generalization by Sanders and colleagues that ‘there are larger amplitude, larger dV/dt signals occurring in cells surrounding pacemaker cells in visceral smooth muscles that must be deemphasized by filtering’2 should be questioned. In vivo cardiac extracellular recordings are readily achieved from a forcefully beating heart, and typical slow waves can also be easily recorded without filtering during contractions, as demonstrated by early investigators such as Alvarez and Bozler, who lacked access to modern filters.1,22

Slow wave signal content in extracellular recordings includes slow transients (<1-Hz spectrum), and faster transients of higher frequency.9,23 Noise can be broadly classified as low-frequency drift and high-frequency interference, and one common filtering strategy has been to use a high pass cut-off of approximately 0.05 Hz to reduce baseline drift, and a low pass cut-off of approximately 2 Hz to reduce high-frequency noise.3,24,25 This author’s group prefers a moving median filter for baseline drift, and a Savitzky–Golay filter for high-frequency noise, to minimize loss of information and waveform distortion.23

Sanders and colleagues employed a 3 or 5–100 Hz band-pass filter in the majority of their murine extracellular recordings.2 This filter could capture some of the fastest slow wave transients (i.e., the rapid downstroke); however, it is an imperfect filter that potentially introduces bias, as important slow wave components below 3 or 5 Hz range are excluded, whereas competing high-frequency artifacts are emphasized and accentuated. More appropriate filtering strategies could help to minimize artifact contamination in extracellular recordings, and optimal approaches could be investigated.

Artifacts and Extracellular Recordings

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

It is well known that artifacts can potentially contaminate GI extracellular recordings and cutaneous EGG.26 However, in practice, most artifacts can be readily distinguished from signals of interest if educated care is taken. For example, extrinsic and non-GI biologic noise sources, such as cable movements, electronic interference, and cardiac and respiratory noise, register as simultaneous deflections in electrograms, whereas slow waves show a time-lag between adjacent channels.3,10

Contraction artifacts are potentially more problematic than other artifacts, because in stomach they propagate at the same velocity as slow waves.2 However, many previous observations also support the reliable differentiation of extracellular contraction artifacts from slow waves. Lammers et al. recorded slow waves and peristaltic contractions propagating simultaneously in different directions in in vitro intestine.27 The slow waves (of classic biphasic appearance) could be readily distinguished from peristaltic waves, which displayed multiple long-duration oscillations. Alvarez, Puestow and Bozler all observed the striking fact that their extracellular potentials were very similar during fasting and fed states, after the onset of vigorous contractions, indicating that contractions contributed very little to their recordings.1,28,29 Berkson et al. demonstrated that electrical recordings precede contractions,20 meaning they cannot be motion transduction, a finding verified by others.30–32 Xing et al. performed simultaneous electrical and strain gauge recordings, and showed that contractions may be absent during tachyarrhythmia, whereas the extracellular potentials continued unabated.33

In Vitro Recording Methodologies

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

An array of valuable small animal models is now available for GI electrophysiology research, and methods for reliably achieving in vitro extracellular recordings is a subject of importance to the field. In particular, multielectrode mapping in such models has potential for investigating the consequences of cellular and molecular defects for tissue and whole-organ function.

Sanders and colleagues could not resolve extracellular slow wave activity in their murine in vitro tissues, even after suppression of contraction artifacts, leading them to conclude that extracellular slow wave recording is generally difficult or impossible.2 However, if slow waves were present intracellularly, then an extracellular field must exist, and will be detectable if the tissue is sufficiently viable to support coordinated activations, and the experimental methods are adequately sensitive. It is possible that the recording methods applied in their study may have been better suited to in vivo large animal work, where the amplitude of the extracellular potentials are larger and the SNR is greater.10,11

In recent times, Lammers et al. have been the predominant exponents of in vitro extracellular recordings, including working in mouse, rat, rabbit and cat. Interestingly, Lammers et al. have recommended specific adaptations for in vitro recordings, to preserve tissue quality and viability, including very high-tissue bath perfusion rates (100–400 mL min−1), using electrode wires that extend 3–5 mm beyond their platform to allow free tissue perfusion, and approximating the electrodes only very lightly against the tissue.27,34 Lammers et al.,35 and Bortoff 8 also reported achieving extracellular slow wave recordings via electrodes placed ‘just above’ the serosal surface, which would seem to a preclude a major motion artifact signal component.

Are Intracellular and Extracellular Frequencies ‘Mismatched’?

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

Sanders and colleagues observed an interesting ‘mismatch’ in the frequencies of murine intracellular slow wave activity and contraction artifacts.2 The authors also point to previous human gastric in vitro intracellular studies showing a slow wave frequency of 6 c min−1,36 and suggest that the normal human gastric slow wave frequency of approximately 3 c min−1 has been ‘underestimated by extracellular recordings,’ due to artifact contamination.2

However, several other sources of evidence provide convincing support for the 3 c min−1 human gastric slow wave frequency that is recorded extracellularly in vivo. MRI recordings of human gastric contractions closely match the pattern, velocity and frequency (approximately 3 c min−1) of extracellular electrical mapping data,3,37 consistent with the understanding that gastric slow waves are associated with tight excitation–contraction coupling.38 Magnetogastrography, an independent technique that measures magnetic fields associated with bioelectrical fields, shows congruity with extracellular frequencies.39 In gastric pacing, stimuli are delivered at frequencies slightly higher than the extracellularly measured rate, typically 3.3 c min−1 in humans,40 and pacing would not work if extracellular recordings were artifacts misrepresenting the true intrinsic rate. Moreover, paced waves interact with native waves in a manner exactly consistent with their extracellularly recorded frequencies.41

With these in vivo studies supporting the veracity of extracellular frequencies, conflicting intracellular studies on isolated tissues in vitro must be interpreted cautiously. More attention is needed as to whether in vitro experimental conditions may affect frequency.42 Negative chronotropic influences may have been eliminated in the in vitro preparations, or positive chronotropic effects may have been active, as a result of variables such as tissue condition, temperature, stretch, electrode impalement, prostaglandins, and neural influences.42–45

‘Spike’ Activity and Extracellular Recordings

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

The movement artifacts noted by Sanders and colleagues do not resemble extracellular slow waves, but more closely resemble spike activity. However, competing evidence should again be considered before accepting the authors’ generalization that ‘spikes appear to be caused by contraction artifacts.’2

In early studies of extracellular spike activity, Bozler employed flexible cotton-wick electrodes that closely followed the intestinal movements.22 Spikes were readily recorded using these electrodes; however, in certain tissues such as corpus and ureter, strong contractions consistently occurred in the total absence of spikes.29 If spikes were contraction artifacts, then they should have been present regardless of the type of tissue being studied. More recent canine gastric extracellular studies support these observations, again showing spikes predominantly in the antrum rather than the corpus.10,30,31 As noted by Sanders and colleagues, intracellular canine gastric studies have similarly shown action potentials to mainly feature in distal antral tissues.2 The concordance between these results suggests that many extracellular recordings have faithfully represented intracellular electrical events at the tissue level, rather than representing contraction artifacts.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References

The finding by Sanders and colleagues that motion artifacts can be recorded from murine in vitro gastric tissues offers a reminder that extracellular recordings should be evaluated carefully, particularly when atypical. However, a large volume of competing literature argues against the authors’ claims that extracellular GI recordings are in general artifact. On the basis of theoretical, technical and experimental considerations, extracellular GI electrical recordings cannot be explained as motion artifacts, and claims that extracellular slow wave detection is biophysically implausible are not supported. Although noise is ubiquitous in bioelectrical recordings, fact and artifact can be readily and reliably distinguished in extracellular studies if filtering is appropriate and educated care is taken. Therefore, motion suppression is not a necessary requirement for extracellular recordings, and extracellular studies should continue to underpin important motility advances in coming years.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. The Biological Basis for Extracellular Slow Wave Recordings
  5. ‘Expected Kinetics’ of Extracellular Recordings
  6. Variability in Extracellular Slow Wave Morphology
  7. Filtering of Extracellular Recordings
  8. Artifacts and Extracellular Recordings
  9. In Vitro Recording Methodologies
  10. Are Intracellular and Extracellular Frequencies ‘Mismatched’?
  11. ‘Spike’ Activity and Extracellular Recordings
  12. Conclusions
  13. Acknowledgments
  14. Conflicts of Interest
  15. Author Contributions
  16. References