Do Electrode Properties Create a Problem in Interpreting Local Field Potential Recordings?

EditorialDo Electrode Properties Create a Problem in Interpreting Local Field Potential Recordings?Matthew J. Nelson, and Pierre PougetMatthew J. NelsonCentre de Recherche de l'Institut Cerveau-Moelle/Institut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche S975 du Centre National de la Recherche Scientifique 7225, Université Pierre et Marie Curie, Hôpital de la Salpêtrière, Paris, France; and California Institute of Technology, Pasadena, California, and Pierre PougetCentre de Recherche de l'Institut Cerveau-Moelle/Institut National de la Santé et de la Recherche Médicale, Unité Mixte de Recherche S975 du Centre National de la Recherche Scientifique 7225, Université Pierre et Marie Curie, Hôpital de la Salpêtrière, Paris, France; and Published Online:01 May 2010https://doi.org/10.1152/jn.00157.2010This is the final version - click for previous versionMoreSectionsPDF (66 KB)Download PDF ToolsExport citationAdd to favoritesGet permissionsTrack citations Local field potential (LFP) recordings within the brain have become an important tool used by neuroscientists to make inferences about the activity of a population of cells near an electrode. Each passing year analysis of LFPs in neuroscience seems to bring important new insights on the possible workings of networks in the brain to produce behavior (Buschman and Miller 2007; Canolty et al. 2006; Gregoriou et al. 2009; Liu and Newsome 2006; Lubenov and Siapas 2009; Pesaran et al. 2008; Womelsdorf et al. 2006). Indeed LFPs have become a near-ubiquitous tool in neurophysiology seemingly in use anywhere extracellular spikes are also recorded.One issue that often comes up among those who interpret LFP data is the uncertainty about how electrode impedance and other electrode parameters affect LFP recordings, presenting a potential problem in their interpretation. Indeed this is a complex question, given that current flow in the brain depends on a multitude of factors and extracellular recordings cannot uncover the precise neural events giving rise to a specific LFP voltage. Amidst this uncertainty, one commonly mentioned idea that exists today is the notion that microelectrodes of different impedances or geometries might integrate signals across space differently which could lead to differing results between experiments that use electrodes of different impedances to collect LFP data. This notion has been frequently expressed verbally by many, though direct discussions of it in the literature (Pesaran 2009) have been more rare. However, literature discussing how these electrode parameters affect spike recordings (Andersen et al. 2010; Moxon 1999; Paik et al. 2003; Ward et al. 2009) is more commonly found. Despite the relative prevalence of this question in the field, research investigating it has been lacking and no definitive answer has yet been proposed. However, we believe that the answer to this question can be found from information gleaned from a range of existing literature and published data, although most in the field are not presently aware of this. The uncertainty surrounding this issue is important to address because it creates a potential barrier for the comparison of LFP data across experiments and laboratories. Furthermore, as the interpretation of LFPs continues to move beyond its infancy further into the territory of a standard neuroscience technique, such comparisons will be increasingly common and important for building consensuses in the field.As we describe here, we believe this issue presents one of the rare cases in neuroscience in which the answer that would make the work of trying to understand the brain easier also happens to be true. That is to say, provided that the proper recording equipment is used, the impedance and geometry of microelectrode recording sites in the ranges typically used in extracellular experiments do not appreciably affect LFP recordings. Scientists in fact do not need to attend to this issue when interpreting LFP data or when comparing such results across experiments and laboratories.To defend this claim, the first point to clarify a priori is that an electrode can be considered to report the average voltage present at its uninsulated tip or recording site (Nunez and Srinivasan 2006; Robinson 1968). Indeed by using metal microelectrodes suspended in saline, we have verified that this was the case and that this model presented years ago by David Robinson does hold true (Nelson et al. 2008). Thus the only sense in which an electrode integrates a signal across space is by determining this average voltage. The shape and size of an electrode's recording site will not, for example, affect the way in which it responds to distant as opposed to nearby voltage sources.Second, if the proper recording equipment is used, the voltage that is ultimately amplified and recorded will not be appreciably electrically affected by the electrode's impedance. Indeed in previous work we demonstrated this to be the case (Nelson et al. 2008). Recorded voltages in saline using electrode impedances spanning the range typically used in extracellular experiments were independent of the electrode's impedance when using a headstage with a high (>1 GΩ) input impedance. This does require some attention from neurophysiologists though. For example another commercially available headstage we tested had a lower input impedance that led to electrode impedance–dependent signal distortions. Fortunately, when they occur these distortions can be corrected post hoc (Nelson et al. 2008), as reported in several recent publications (Gregoriou et al. 2009; Siegel et al. 2009).A separate question from whether the impedance of an electrode electrically influences the recorded voltage is the question of whether the size and geometry of the recording site have an impact on the recorded LFPs. Within the framework that we have described here, one can conceive how during a recording these parameters could affect the average voltage present over the whole uninsulated site and thus affect the recorded values.An electrode's impedance is of course highly dependent on the size of its uninsulated surface area. Indeed impedance, which is more easily measured, is essentially used only as a proxy to describe an electrode's uninsulated surface area, which is of more interest to neuroscientists. For spike recordings, for example, the general viewpoints in the field tend to be that larger recording sites spanning more of the extracellular space will record more neurons at a given time (Andersen et al. 2010; Moxon 1999; Paik et al. 2003; Ward et al. 2009), whereas smaller recording sites will have more specificity and record individual neurons that are better isolated (Moxon 1999; Paik et al. 2003). These views are not established facts, however, and there is at least some contrary evidence in the literature (Suner et al. 2005).The impact on spike recording set aside, it may be tempting for one to make the claim that LFPs are affected by recording site size and geometry. However, a closer look reveals that current estimates of the spread of LFP signals suggest that these potentials in the brain vary on a spatial scale larger than the size of uninsulated electrode recording sites used in extracellular recordings. For instance, Katzner and colleagues (2009) recently showed that LFPs are more local than previously thought, but even their newly lowered estimate still suggests that LFPs primarily originate from sources within 100 μm of a given point, with noticeable contributions coming from up to within 250 μm. Meanwhile, even a 300 kΩ at 1 kHz tungsten electrode, which is a very low impedance by extracellular recording standards and thus has a very large recording site, corresponds to a surface area of only 1,850 μm2 (Yaeli et al. 2009). For a conical electrode at even an extreme angle of 10°, this would result in an uninsulated height near 80 μm, as illustrated in Fig. 1. Similar data published for platinum/iridium electrodes (Lemon 1984; Tielen et al. 1971) ultimately yield the same conclusion, further suggesting that this will hold true for other metals as well. Changing the size and shape of a recording site within this range would not likely appreciably affect the average potential found over the site based on current estimates of the spread of LFPs.Fig. 1.Electrode recording site sizes relative to the spread of local field potentials (LFPs). Even for a low impedance (300 kΩ) electrode with an extremely fine taper angle (10°), the uninsulated recording site length is smaller than that of current estimates of the SD of the spread of LFP signals. Larger taper angles or higher impedances correspond to even smaller recording site lengths. The average potential across the electrode's tip should thus not be appreciably affected by variations of recording site size within the range of electrodes used for extracellular recordings. Data for the spread of LFPs come from Katzner et al. (2009), whereas data for electrode recording site sizes can be found in Lemon (1984), Tielen et al. (1971), and Yaeli et al. (2009).Download figureDownload PowerPointThus we feel that evidence available in the literature suggests that the impedance, size, and shape of the recording sites of microelectrodes will not affect the recording of LFPs. This holds for microelectrodes spanning the range of impedances typically used in extracellular recording experiments, provided that headstages of the proper input impedance are used (Nelson et al. 2008). Of course it would be possible to produce an extremely low impedance microelectrode that may indeed report different voltages than typical microelectrodes, but it is not clear that anything would be gained by the use of such an electrode. Indeed larger electrode recording sites essentially only serve as low-pass spatial filters (Nunez and Srinivasan 2006) and it would seem that the more precise spatial measurement provided by microelectrodes currently in use would be preferred in any reasonable scientific application.Those who might disagree with the claim we have presented here should be clear in their reasons for doing so. The framework that we have presented describing the circuit properties of microelectrode recordings is generally accepted, although many neuroscientists may be unaware of it. Assuming that one then does not disagree with this framework, disagreeing with our claim amounts to the counterclaim that LFP voltage profiles in the brain change appreciably on a spatial scale smaller than a few tens of microns, which is the scale of microelectrode recording sites in use today. Thus for this counterclaim to be true, it is important to note that current beliefs about the spread of LFPs must then be wrong. Even if hypothetically LFPs did vary on such a small spatial scale, the very precise placement of an electrode within the brain would then likely be of more importance than the size and geometry of its recording sites. In such a case it is unclear that it would even be possible or desirable to attempt to capitalize on such a fine spatial structure of the local field potential.Although the electrode parameters of impedance and recording site size likely have no effect on the recorded LFP, this says nothing about the effects of the mechanical disruption introduced by the electrode. Such effects would likely be dependent on electrode parameters such as size and taper angle, although an electrode's impedance or the size of its uninsulated region should not directly influence this. Additional caveats to consider are the effects of increased thermal noise and loss of signal due to shunt capacitance, both of which affect recorded voltages more as electrode impedance increases (Cogan 2008). However, we believe that again the evidence suggests these will not have sufficiently large effects on LFP recordings to be of notable concern for recordings made within typical ranges of parameters. With respect to thermal noise, its magnitude is dependent on signal bandwidth as well as electrode impedance, and LFP recordings typically take place over a limited bandwidth. Using electrode impedance spectra data from our previous work (Nelson et al. 2008) and the well-established formula for thermal noise (Johnson 1928; Nyquist 1928) we can derive estimates of the amount of thermal noise theoretically expected for different microelectrodes in the frequency band from 0 to 150 Hz given their average impedance in this frequency band, which is a generous bandwidth for LFPs. The expected root mean square value of this thermal noise was 7.7 μV for an electrode with a 3.3 MΩ impedance at 1 kHz and 4.2 μV for an electrode with a 500 kΩ impedance at 1 kHz. The different amount of noise between these electrodes, which approximately span the range of electrode impedances in use today, is all but negligible compared with the size of most LFP results of interest. With respect to signal loss through shunt pathways, the impact of shunt capacitance in recordings serves to effectively compromise the input impedance of the headstage at high frequencies (Cogan 2008; Nelson et al. 2008; Robinson 1968), which results in electrode impedance–dependent signal loss and distortions at those frequencies. However, this impact declines as frequency decreases. Indeed during test recordings with metallic resistors and electrodes suspended in saline, we found that as frequency decreased, shunt capacitance ceased to influence recordings starting in the 300 to 400 Hz range (Nelson et al. 2008), well above the LFP frequency ranges of interest. We should be clear, however, that shunt capacitance does depend on how the recording equipment is arranged and the impacts could be different in some recording configurations. However, this has yet to be tested in the literature and the available evidence we mentioned suggests this will not be a problem for LFP frequencies.There is still a great deal of interesting future research that is needed to better understand the details of what can be inferred about neural activity from a given LFP trace. Indeed there are many curiosities and unknowns related to this. Nonetheless, we hope that neuroscientists interpreting LFP data can more freely compare results and continue the important work of using field potentials to understand the brain without worrying about the possible effect of electrode impedance and recording site geometry as one more unknown for concern.DISCLOSURESNo conflicts of interest are declared by the authors.REFERENCES Andersen RA , Hwang EJ , Mulliken GH . Cognitive neural prosthetics. Annu Rev Psychol 61: 169–190, 2010. Crossref | PubMed | ISI | Google Scholar Buschman TJ , Miller EK . Top-down versus bottom-up control of attention in the prefrontal and posterior parietal cortices. Science 315: 1860–1862, 2007. Crossref | PubMed | ISI | Google Scholar Canolty RT , Edwards E , Dalal SS , Soltani M , Nagarajan SS , Kirsch HE , Berger MS , Barbaro NM , Knight RT . High gamma power is phase-locked to theta oscillations in human neocortex. Science 313: 1626–1628, 2006. Crossref | PubMed | ISI | Google Scholar Cogan SF . Neural stimulation and recording electrodes. Annu Rev Biomed Eng 10: 275–309, 2008. Crossref | PubMed | ISI | Google Scholar Gregoriou GG , Gotts SJ , Zhou HH , Desimone R . High-frequency, long-range coupling between prefrontal and visual cortex during attention. Science 324: 1207–1210, 2009. Crossref | PubMed | ISI | Google Scholar Johnson J . Thermal agitation of electricity in conductors. Phys Rev 32: 97–109, 1928. Crossref | Google Scholar Katzner S , Nauhaus I , Benucci A , Bonin V , Ringach DL , Carandini M . Local origin of field potentials in visual cortex. Neuron 61: 35–41, 2009. Crossref | PubMed | ISI | Google Scholar Lemon R . A method for the continuous recording of acid gastric secretion in the rat. In: Methods for Neuronal Recording in Conscious Animals. New York: Wiley–Interscience, 1984, p. 74–75. Google Scholar Liu J , Newsome WT . Local field potential in cortical area MT: stimulus tuning and behavioral correlations. J Neurosci 26: 7779–7790, 2006. Crossref | PubMed | ISI | Google Scholar Lubenov EV , Siapas AG . Hippocampal theta oscillations are travelling waves. Nature 459: 534–539, 2009. Crossref | PubMed | ISI | Google Scholar Moxon KA . Multichannel electrode design: considerations for different applications. In: Methods for Neural Ensemble Recordings, edited by , Nicolelis MA. Boca Raton, FL: CRC Press, 1999, p. 375–405. Google Scholar Nelson MJ , Pouget P , Nilsen EA , Patten CD , Schall JD . Review of signal distortion through metal microelectrode recording circuits and filters. J Neurosci Methods 169: 141–157, 2008. Crossref | PubMed | ISI | Google Scholar Nunez PL , Srinivasan R . Electric fields of the brain. In: The Neurophysics of EEG (2nd ed.). New York: Oxford Univ. Press, 2006, p. 298–305. Crossref | Google Scholar Nyquist H . Thermal agitation of electric charge in conductors. Phys Rev 32: 110–113, 1928. Crossref | Google Scholar Paik SJ , Park Y , Cho D . Roughened polysilicon for low impedance microelectrodes in neural probes. J Micromech Microeng 13: 373–379, 2003. Crossref | ISI | Google Scholar Pesaran B . Uncovering the mysterious origins of local field potentials. Neuron 61: 1–2, 2009. Crossref | PubMed | ISI | Google Scholar Pesaran B , Nelson MJ , Andersen RA . Free choice activates a decision circuit between frontal and parietal cortex. Nature 453: 406–409, 2008. Crossref | PubMed | ISI | Google Scholar Robinson DA . The electrical properties of metal microelectrodes. Proc IEEE 56: 1065–1071, 1968. Crossref | ISI | Google Scholar Siegel M , Warden MR , Miller EK . Phase-dependent neuronal coding of objects in short-term memory. Proc Natl Acad Sci USA 106: 21341–21346, 2009. Crossref | PubMed | ISI | Google Scholar Suner S , Fellows MR , Vargas-Irwin C , Nakata GK , Donoghue JP . Reliability of signals from a chronically implanted, silicon-based electrode array in non-human primate primary motor cortex. IEEE Trans Neural Syst Rehabil Eng 13: 525–541, 2005. Crossref | ISI | Google Scholar Tielen AM , Giesen C , Mollevanger W . Some mechanical and electrical characteristics of metal microelectrodes. Inst Med Phys TNO Rep 2.3.65/1 MEI, 1971. Google Scholar Ward MP , Rajdev P , Ellison C , Irazoqui PP . Toward a comparison of microelectrodes for acute and chronic recordings. Brain Res 1282: 183–200, 2009. Crossref | PubMed | ISI | Google Scholar Womelsdorf T , Fries P , Mitra PP , Desimone R . Gamma-band synchronization in visual cortex predicts speed of change detection. Nature 439: 733–736, 2006. Crossref | PubMed | ISI | Google Scholar Yaeli S , Binyamin E , Shoham S . Form-function relations in cone-tipped stimulating microelectrodes. Front Neuroeng 2: 13: 1–8, 2009. Google ScholarAUTHOR NOTESAddress for reprint requests and other correspondence: M. Nelson, CRICM, UMR S975, INSERM/Université Pierre et Marie Curie, Neurologie et Thérapeutique Expérimentale, Hôpital de la Salpêtrière, 47 boulevard de l'Hôpital, 75651 Paris Cedex 13, France (E-mail: matthew.[email protected]fr). Download PDF Back to Top Next FiguresReferencesRelatedInformation Cited ByNew thin-film surface electrode array enables brain mapping with high spatial acuity in rodents28 February 2018 | Scientific Reports, Vol. 8, No. 1Locally induced neuronal synchrony precisely propagates to specific cortical areas without rhythm distortion16 May 2018 | Scientific Reports, Vol. 8, No. 1The Summating Potential Is a Reliable Marker of Electrode Position in Electrocochleography: Cochlear Implant as a Theragnostic ProbeEar & Hearing, Vol. 39, No. 4Neural Interfaces for Intracortical Recording: Requirements, Fabrication Methods, and Characteristics7 December 2017 | Frontiers in Neuroscience, Vol. 11Computational analysis of network activity and spatial reach of sharp wave-ripples15 September 2017 | PLOS ONE, Vol. 12, No. 9Magnitude and behavior of cross-talk effects in multichannel electrophysiology experimentsMatthew J. Nelson, Silvana Valtcheva, and Laurent Venance14 July 2017 | Journal of Neurophysiology, Vol. 118, No. 1Brain-Machine Interfaces: From Basic Science to Neuroprostheses and NeurorehabilitationMikhail A. Lebedev and Miguel A. L. Nicolelis8 March 2017 | Physiological Reviews, Vol. 97, No. 2Filter based phase distortions in extracellular spikes30 March 2017 | PLOS ONE, Vol. 12, No. 3Extrapolating meaning from local field potential recordingsJournal of Integrative Neuroscience, Vol. 16, No. 1Local Field Potentials: Myths and Misunderstandings15 December 2016 | Frontiers in Neural Circuits, Vol. 10Hybrid Scheme for Modeling Local Field Potentials from Point-Neuron Networks20 October 2016 | Cerebral Cortex, Vol. 26, No. 12Electrophysiological effects of non-invasive Radio Electric Asymmetric Conveyor (REAC) on thalamocortical neural activities and perturbed experimental conditions11 December 2015 | Scientific Reports, Vol. 5, No. 1Subthalamic, not striatal, activity correlates with basal ganglia downstream activity in normal and parkinsonian monkeys23 August 2016 | eLife, Vol. 5Energy and Potassium Ion Homeostasis during Gamma Oscillations16 June 2016 | Frontiers in Molecular Neuroscience, Vol. 9Local Field Potential Analysis for Closed-Loop NeuromodulationStimulation of the subthalamic nucleus engages the cerebellum for motor function in parkinsonian rats15 August 2014 | Brain Structure and Function, Vol. 220, No. 6Modelling and Analysis of Electrical Potentials Recorded in Microelectrode Arrays (MEAs)31 March 2015 | Neuroinformatics, Vol. 13, No. 4Stimulus-dependent spiking relationships with the EEGAdam C. Snyder and Matthew A. Smith1 September 2015 | Journal of Neurophysiology, Vol. 114, No. 3ViSAPy: A Python tool for biophysics-based generation of virtual spiking activity for evaluation of spike-sorting algorithmsJournal of Neuroscience Methods, Vol. 245Revealing neuronal function through microelectrode array recordings6 January 2015 | Frontiers in Neuroscience, Vol. 8High-frequency oscillations in Parkinson's disease: Spatial distribution and clinical relevance10 July 2014 | Movement Disorders, Vol. 29, No. 10A Detailed and Fast Model of Extracellular RecordingsNeural Computation, Vol. 25, No. 5In Vitro and In Vivo Recording of Local Field Potential Oscillations in Mouse Hippocampus1 September 2012The use of macroelectrodes in recording cellular spiking activityJournal of Neuroscience Methods, Vol. 206, No. 1Dielectrophoresis-Based Integrated Lab-on-Chip for Nano and Micro-Particles Manipulation and Capacitive DetectionIEEE Transactions on Biomedical Circuits and Systems, Vol. 6, No. 2Physical model of coherent potentials measured with different electrode recording site sizesMatthew J. Nelson, and Pierre Pouget1 March 2012 | Journal of Neurophysiology, Vol. 107, No. 5How Local Is the Local Field Potential?Neuron, Vol. 72, No. 5Neural control of cursor trajectory and click by a human with tetraplegia 1000 days after implant of an intracortical microelectrode array24 March 2011 | Journal of Neural Engineering, Vol. 8, No. 2On the Origin of the Extracellular Field Potential in the Nucleus Laminaris of the Barn Owl (Tyto alba)Paula T. Kuokkanen, Hermann Wagner, Go Ashida, Catherine E. Carr, and Richard Kempter1 October 2010 | Journal of Neurophysiology, Vol. 104, No. 4How Global Are Olfactory Bulb Oscillations?Leslie M. Kay, and Philip Lazzara1 September 2010 | Journal of Neurophysiology, Vol. 104, No. 3 More from this issue > Volume 103Issue 5May 2010Pages 2315-2317 Copyright & PermissionsCopyright © 2010 the American Physiological Societyhttps://doi.org/10.1152/jn.00157.2010PubMed20220081History Published online 1 May 2010 Published in print 1 May 2010 Metrics