Effects of electromagnetic fields on neuronal ion channels: a systematic review

Many aspects of chemistry and biology are mediated by electromagnetic field (EMF) interactions. The central nervous system (CNS) is particularly sensitive to EMF stimuli. Studies have explored the direct effect of different EMFs on the electrical properties of neurons in the last two decades, particularly focusing on the role of voltage‐gated ion channels (VGCs). This work aims to systematically review published evidence in the last two decades detailing the effects of EMFs on neuronal ion channels as per the PRISM guidelines. Following a predetermined exclusion and inclusion criteria, 22 papers were included after searches on three online databases. Changes in calcium homeostasis, attributable to the voltage‐gated calcium channels, were found to be the most commonly reported result of EMF exposure. EMF effects on the neuronal landscape appear to be diverse and greatly dependent on parameters, such as the field's frequency, exposure time, and intrinsic properties of the irradiated tissue, such as the expression of VGCs. Here, we systematically clarify how neuronal ion channels are particularly affected and differentially modulated by EMFs at multiple levels, such as gating dynamics, ion conductance, concentration in the membrane, and gene and protein expression. Ion channels represent a major transducer for EMF‐related effects on the CNS.


Introduction
As the use of electromagnetic devices continues to grow, we have had to consider ways in which such devices may impact our health. There has been a dramatic increase in exposure to different types of electromagnetic fields (EMFs), ranging from extremely low-frequency electromagnetic fields (ELF-EMFs) up to 300 Hz and radiofrequency electromagnetic fields (RF-EMFs), ranging from 10 MHz to 300 GHz, and, more recently, devices that operate at terahertz (0.3-3 THz) frequencies, but also 0 frequency fields, such as static magnetic fields (SMFs). These frequencies are mainly derived from man-made sources, such as elec-tricity power supplies, power amplifier circuits, voltage-controlled oscillators, cellular telephone antennas, and smartphones. 1,2 Exposure resulting from these sources is several orders of magnitude greater than those from natural sources, so not surprisingly, along with the increased usage of a wide range of electronic devices producing a manifold of EMF types, concern about their possible detrimental effect has been raised. Since the 1980s, various studies have investigated the link between different frequency EMFs and the risk of developing chronic diseases, such as cancer, along with cardiovascular and neurodegenerative diseases. 1,3,4 Epidemiological studies reported a significant association between exposure to ELF-EMF and the development of childhood leukemia, 5,6 and in 2011, WHO's International Agency for Research on Cancer highlighted the possible carcinogenic effect of RF-EMFs. 7 However, EMFs have also been extensively used in the treatment of many forms of neoplasia 8 and neurodegenerative diseases, 9 highlighting a possible beneficial use of these fields as therapeutic treatments for diverse chronic diseases. The effects on the central nervous system (CNS) appear to be particularly relevant, as the CNS is reliant on many voltage-dependent processes. 10 EMFs are wave-like propagating fields exerting forces on every charged object within their vicinity. They consist of two different but inextricable field components (electric and magnetic) running perpendicularly to each other and continuously inducing each other by varying in time, as described by Maxwell's equations. The electric component is the stronger, but its interaction is attenuated by endogenous charges, so its magnitude is dependent on the dielectric constant of the tissue it crosses and falls of according to an inverse square law; whereas the weaker magnetic component, despite its reduced strength, is not attenuated by charge but falls off according to an inverse cube law. Moreover, despite their differences in force, the energies associated with the electric and magnetic components within an EM wave are equal. These characteristics lead to the need for considering them as both equally relevant in the study of their effects on living systems.
Typically, we are continuously exposed to natural EMF radiation, such as every form of light radiation, cosmic microwaves, or the magnetic field generated by the planet. However, two key points must be considered when comparing natural exposure to that of artificial exposure. First, the magnitude of the two exposures is very different, with the artificial one being higher by a few to several orders of magnitude in comparison with natural fields. 1 Nevertheless, in the delicately balanced system of the CNS of higher animals, even a small increase in intensity could generate significant effects. 10 Second, many types of common man-made EMFs, in contrast to natural ones (produced by a huge number of atomic or molecular transitions that are randomly oriented between them), are polarized, meaning that they oscillate in a specific and determined plane called the "polarization plane." 11 This peculiar characteristic gives artificial EMF an additional electrostatic force deriving from the synchroniza-tion of every type of charged/polar molecule with that polarization plane (so in phase with the field).
Furthermore, the well-organized orientation of these electromagnetic waves permits the generation of constructive interference effects by polarized fields that can amplify their intensity locally, increasing the subsequent biological activity. 11 It is for these reasons that this paper focuses on the effect of artificial EMFs.
Different types of EMFs have been shown to have a major impact on CNS physiology, but recent research has primarily been focused on the role of low-frequency EMFs. These forms of electromagnetic radiation are not ionizing because of the low energy contained in their quanta, in contrast to other higher frequency fields. Therefore, they do not have enough energy to remove electrons from atom and molecules, 12 but they can produce thermal effects and induce in the human body circulating currents that, if sufficiently large, could cause stimulation of nerves and muscles or affect other biological processes in many tissues. 13 ELF-EMFs were recently reported to have significant effects on synaptic plasticity of both adult 14 and newborn 15 rats, where they have been reported to modulate the development of long-term potentiation (LTP) in the hippocampus and neocortex. However, the specific impact of ELF-EMFs on LTP is controversial, as this phenomenon is both reported to be increased 14 or decreased 15,16 by this type of EMF. In addition, ELF-EMFs are known to modulate the cell cycle in several cell lines. 17 RF-EMF exposure has been linked to developmental abnormalities in specific areas of the brain 18 and to decreases in specific types of neurons, 19 and has been shown to activate the autophagic pathway. 20 Despite this, the potential beneficial impact of EMFs on CNS physiology has been extensively investigated. ELF-EMFs have been reported to increase neurogenesis both in isolated neuronal stem cells (NSCs) 21 and mouse hippocampus, 22 in addition to ameliorating the remyelination process through enhanced proliferation of NSCs. 23 On the other hand, RF-EMFs have been used in the treatment of chronic pain, 24 and studies have suggested a novel therapeutic use of low-intensity RF-EMFs in the treatment of severe brain cancer, 8 such as in recurrent glioblastoma patients, where this approach aims to replace the classical chemotherapy paradigm. 25  interfere with the physiology and functional activity of neurons. The effects exerted by EMF exposure appear to be dependent on the developmental stage of the exposed neurons. Indeed, one of the biggest EMF-related concerns that the scientific community is facing is the impact of the electromagnetic radiation on neural development, particularly as several reports have described severe effects on neural development, 18,30 although other studies have failed to observe significant effects. 31 One common explanation for this is the low thickness of a child's skull, 32 but the higher severity could be due to the higher density of stem cells present in the first stages of development. 33,34 Indeed, many studies have reported significant effects on embryonic stem cells. 30,35,36 This fact is particularly notable since a good correlation exists between the neonatal development of the CNS and late brain development in humans. 37 Nonetheless, the electromagnetic force produced by EMFs is likely to influence neuronal activities through the interaction with charged cellular components that are particularly sensitive to changes in their charge, such as voltage-gated ion channels (VGCs).
Specifically, both acute and chronic ELF-EMF exposures have been reported to increase the ion transport rate of many types of VGCs, such as voltage-gated sodium channels (VGSCs), 38 highthreshold calcium channels, and calcium-activated potassium (KCa) channels, 39 maybe directly acting on their voltage-sensing domain (VSD) and pore-forming domain, and to trigger an increase in the expression of the VGCs themselves 40 that could be caused by the same change in ion channel conduction and permeability, leading to an altered ionic equilibrium within the cell. Chronic exposure to RF-EMFs, on the contrary, has been found to cause a decrease in pan calcium channel gene and protein expression in mouse hippocampus and hypothalamus 41,42 and alter the afterhyperpolarization amplitude and spike frequency of rat Purkinje cerebellar neurons. 43 However, acute exposure to this type of field apparently fails to elicit similar effects, 39,44 suggesting a different mechanism of action with respect to ELF-EMF exposure. ELF-EMFs and RF-EMFs are not the only types of fields reported to have effects on neurons. SMFs that do not change their intensity and direction in time have been shown to deeply impact the physiological properties of several types of neurons, through their interaction with VGCs. [45][46][47] Additionally, studies have also shown some frequencies to be more biologically active than others 48 and have reported diverse biological effects for different frequencies of EMFs. 39,48,49 This difference has been linked to microthermal effects elicited by RF-EMFs (and not by ELF-EMFs), 50 although numerous studies reported nonthermal effects related to RF-EMF exposure that seem to be strictly related to modifications of the calcium signaling pathway. 41,42,51 The mechanism underlying this diversity of effects, however, has not been completely unraveled and is still a matter of debate.
The molecular mechanisms underlying the effects of EMF exposure are less well known. A role for calcium is, however, well established. Exposure to both acute and chronic ELF-EMFs promotes calcium influx, resulting in an increased intracellular calcium concentration in various types of neurons 39,40,52 that is conversely decreased following chronic exposure to RF-EMFs. 41,42,53 The involvement of calcium is remarkable because of the many related calcium signaling pathways involved in various essential neurophysiological processes, such as neural differentiation, survival, and apoptosis 54,55 that could ultimately explain the many effects of EMF exposure on learning and memory. 56 In areas of the brain involved in the modulation of these cognitive tasks, such as the hippocampus and prefrontal cortex, exposure to ELF-EMFs resulted in abnormal calcium signaling that led to a decrease in the binding between the N-methyl-d-aspartate receptor (NMDAR) and its ligand glutamate, which could directly result in important effects on synaptic plasticity. 57,58 However, the effects of EMFs reported in the literature are often conflicting. Recent studies report no increase in intracellular calcium or production of reactive oxygen species (ROS) 59 subsequent to both acute and subchronic exposure to ELF-EMFs, 31,44,59 and these disagreements also remain when considering studies focused on the same model, such as in the case of PC12 cells. 60 Also, the impact of EMFs on the CNS has been questioned, with studies reporting (in different models) no effects on the expression of synaptic receptors as NMDAR, 61 no relation between EMF exposure and brain electrical activity, 62,63 and no neurotoxic effect subsequent to EMF exposure in pre-and postnatal 84 Ann  64 However, the experimental setups are often different in many key parameters, such as the time, intensity, and type of exposure, making results difficult to compare so that there remains no consensus on the effects of EMF exposure on ion channels within the CNS. The principal aim of this review is, therefore, to provide a systematic analysis of EMF effects on neuronal ion channels.

Research question
Using the PICOT structure (Population, Intervention, Comparator, Outcomes, and Time), the following research question was formulated: "Are EMFs capable of influencing ion channel conductance and expression in the central nervous system?" We then followed the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines to perform this systematic review. 65 Search strategy A systematic review of the literature was carried out in April 2020 by submitting the chosen combination of keywords into three different databases: PubMed, Scopus, and Web of Science. The agreed combination with which the search was performed was "ion channel * " AND "electromagnetic field * " OR "EMF" AND "neuron." A time filter was applied to the search to isolate solely the works published in the last 15 years (2005-2020).

Search eligibility criteria
The papers resulting from the above searches were screened for the presence of duplicates and narrowed down further using the predefined inclusion and exclusion criteria based on the journal title and abstract. The chosen exclusion criteria were (1) the paper is not in English, (2) the paper is not an original research paper (i.e., it is a review article, editorial, or book chapter), or (3) the paper is not freely available (using the institutional credentials of the University of Surrey). This resulting list was then subject to an inclusion round in which we considered only original laboratory research studies, conducted on neurons, neuron-like cells, or neural tissue as a model, that met our original research questions. Thus, all the papers that did not discuss in the title or abstract the effects of EMFs on ion channel conductance or expression were excluded from the study. In summary, the inclusion criteria were (1) the paper covers original laboratory research; (2) the model of the study is neurons, neuron-like cells, or neural tissue; and (3) the paper is relevant based on its title and abstract.

Article selection and processing
The final list obtained by this iterative selection process was independently reviewed by three different investigators. The full text was obtained for each of the included articles. Where an article was not readily available, the author was personally contacted, and the manuscript of the relevant study was obtained. No papers were excluded in this step.

Quality appraisal
The Checklist of Review Criteria published by the Task Force of Academic Medicine and GEA-RIME committee 66 was used as a further inclusion instrument to highlight the quality and completeness of the papers resulting from research. This screened publications for their quality by using a set of 13 predetermined criteria. We considered papers that satisfied at least 12 out of the 13 criteria.

Data extraction
In order to facilitate the analysis and to better compare key features of the studies, a custom data extraction table was designed and used to extract relevant information. The features evaluated were: material used (cell type and eventual modifications, brain regions, and animal provenance); EMF type and exposure time; EMF generation system (type of exposure system and specifics); delivery mode; techniques used to study ion channels; type of ion channel studied; key results and discussion; and proposed mechanism of EMF effect on the ion channel.

Search results
The initial search yielded 157 titles from PubMed, 38 titles from Scopus, and 19 titles from Web of Science, for a total of 214 ( Fig. 1). Removal of duplicates reduced this number to 175, and the subsequent exclusion and inclusion rounds led to the final number of 24 articles.
Following quality appraisal (Table 1), there were 21 publications deemed of suitable quality for inclusion in this review.

Types of EMFs studied
Central to any discussion related to EMFs is the definition of particular parameters, such as the frequency and intensity of these fields, as they are directly correlated with the different effects that such fields have on cells. Accurately defining the type of field delivered and its exposure time is, therefore, necessary to allow any reasonable comparison. To mimic the effects of the principal human-generated sources of EMFs, relative EMF exposures have been used. The most commonly studied fields (57% of total papers) were ELF-EMFs at 50 Hz, delivered with various magnetic intensities, among which the most used did not exceed 1 mT (33% of total cases). ELF-EMFs were preferentially delivered following an acute or subchronic (≤48 h) delivery (75% of cases). This trend was inverted in the case of RF-EMF exposure (33% of total), where 57% of studies used chronic exposure to simulate the daily exposure to fields derived from mobile phone antennas.

Experimental platforms
Many different systems have been used to generate the EMFs, ranging from conventional EMF generators to custom-made devices arranged to fit particular experimental queries. Helmholtz coils, consisting of two circular electromagnets placed along a common axis, were the preferred generator for ELF-EMFs (67% of papers), while the Wave Exposer V20 custom device (extensively described in Ref. 53) was the most common generator for RF-EMFs (43%), maybe by virtue of the multiple studies published by the same team. Other highly represented RF-EMF generators were the Coplanar Waveguide (CPW) (28.5%), consisting of a single conducting track printed onto a dielectric substrate and arranged to be in the middle of a pair of return conductors, and mobile phones (28.5%). All EMF types and exposure systems used are summarized in Figure 2 and more extensively reported in Table 2.

Choice of biological system
The majority of the studies (95%) chose to focus on rodent-derived cell lines or tissues, among which Note: A total of 13 criteria were assessed, based on the quality checklist formulated by the Checklist of Review Criteria established by the Task Force of Academic Medicine and the GEA-RIME Committee. 66 Twenty-two out of the 24 papers met at least 12 out of the 13 criteria and were included in the study.
the rat (R. norvegicus) represented the favored source. However, the present study revealed great variability in the specific type of rodent cell lines. Sixty-seven percent of the studies used cell culture as a model, 71% of which had been isolated from neural tissue (among which 30% were from the cortex, 20% were from the cerebellum and dorsal ganglion, and 10% were from the hippocampus, trigeminal nerve, and entorhinal cortex), while 29% used cancer cell lines. The lines used in this latter case were PC12 cells (derived from rat pheochromocytoma), NG108-15 (a hybrid between mouse neuroblastoma and rat glioma), SH-SY5Y (derived from human neuroblastoma), and F11 (a hybrid between mouse neuroblastoma and rat dorsal root ganglion neurons). A single study used the AtT20 D16V cell line, derived from the pituitary gland. Notably, brain-derived slices have been used only in 19% of the papers. Among the different regions of the CNS studied, the hippocampus was used the most (24% of the articles), followed by the cortex and cerebellum (14% each). A small proportion of studies derived their cells and tissues from the hypothalamus, trapezoid body, entorhinal cortex, pituitary gland, trigeminal nerve, and spinal cord. These results are summarized in Table 2.

Ion channels studied
Seventy-six percent of articles focused on different types of calcium channels (P/Q, N, R, but mainly L subtypes), with 24% reporting effects on VGSCs and only 19% studying the role of potassium channels (A-type K + , delayed rectifier K + , M-type K + , KCa, fast-inactivating transient (IK, A), and dominant-sustained (IK,V) channels). Some studies also investigated the role of GABAA, HCN, TRPA1, and TRPV1 channels in the modulation of EMF effects on the cell. (Note that the studies reviewed here do not always explore only one ion channel, thus the percentages reported cannot always be summed to 100%.)

Experimental techniques
Forty-eight percent of total papers used the patch clamp technique to assess the effects of EMF exposure on ion channels. In particular, the whole-cell recording configuration appeared to be the most frequently used (90%), with only one study using the single channel recording configuration (10%), maybe due to the technical limitation of this configuration despite its great relevance for the study of single ion channel conductance. In 60% of cases, researchers added various channel inhibitors and agonists to better discriminate between different      inhibited the ELF-EMF-induced I Na increase. The Na V channel steady-state activation curve was significantly shifted toward hyperpolarization by ELF-EMF stimulation but remained unchanged by MT in cerebellar GCs that were either exposed or not exposed to ELF-EMF. The inhibitory effects of MT on ELF-EMF-induced Na V activity was greatly reduced by the calmodulin inhibitor KN93. Ca 2+ imaging showed that MT did not increase the basal intracellular Ca 2+ level, but it significantly elevated the intracellular Ca 2+ level evoked by the high Na + stimulation in cerebellar GCs that were either exposed or not exposed to ELF-EMF.   125-mT SMF showed no effect on the peak current density and I-V relationship of IK, A, and IK, V activation, but changes were found in the inactivation kinetics of both types of VGPCs between the SMF exposure and control groups A biological membrane would be deformed in an SMF and the ion channels on the membrane would be affected. The alterations of ion channel activity caused by SMF exposure are indirect. The primary effect of magnetic fields is to induce rotation and reorientation of the membrane lipid molecules and such reorientation could affect conformational changes of ion channels  channel currents, among which the most commonly used were the sodium channel blocker tetrodotoxin (TTX) and the potassium channel blockers 4aminopyridine (4-AP) and tetraethylammonium (TEA). The first is a potent neurotoxin derived from pufferfish that binds the pore domain (PD) of fast VGSCs, blocking sodium conductance. 67 The second is a relatively selective blocker of members of the K v 1 (Shaker, KCNA) family of voltagegated potassium channels (VGPCs), 68 which have interestingly been reported to directly potentiate ion conduction through voltage-gated calcium channels (VGCCs). 69 The third is thought to physically enter the pore of VGPCs and KCa channels, blocking potassium conductance. 70 Fifty-two percent of papers also measured changes in intracellular calcium levels using various methodologies, including single-cell Fura-2 AM (54%), Fluo-3 AM (18%), and Fluo-4 AM (18%), and fluorescence cell labeling with fluorescent probes (Indo/SNARF) (9%). In addition to this, 14% and 9% of studies evaluated the impact of EMFs on ion channel expression using western blots and quantitative RT-PCR, respectively. Finally, an isolated study used short interfering RNA (siRNA) to target ion channels, while the other one used Fourier transform infrared spectroscopy to assess the effect of EMF exposure on the α-helices of ion channels (Fig. 3).

Experimental results
The most commonly recorded effect evoked by ELF-EMFs in neurons was an increase in the basal calcium concentration (reported in 42% of papers focusing on these types of fields). Nonetheless, the results of acute ELF-EMF exposure appear to be diverse. In addition to a rise in intracellular calcium concentration, altered gating dynamics of highthreshold calcium channels and calcium-activated potassium channels as well as increased activity and insertion in the membrane of Na + channels were described. On the other hand, two papers reported a lack of effects on both the calcium homeostasis and the electrophysiological properties of the cell under this condition. However, under chronic exposure, all the papers reported an increase in the intracellular calcium levels at intensities greater than 1000 μT, along with increases in the gene and protein expression of transmembrane calcium channels.  Acute exposure to RF-EMFs does not appear to exert a significant effect according to the two papers focusing on this type of exposure. However, under chronic exposure, 60% of studies reported a decrease in calcium channel expression and one study reported effects on the electrophysiological properties of neurons, such as altered afterhyperpolarization amplitude, spike frequency, half width, and first spike frequency.
A single study on direct current fields (DCFs) showed membrane polarization in hippocampal CA1 pyramidal neurons subsequent to DCF exposure that was, however, independent of any alteration in the dynamics of VGCs. This paper, as the ones investigating the role of SMFs, has been included since the electric and magnetic components are never totally separable (although in these cases the magnetic component of the field is likely to be negligible due to the constant magnetic flux generated by the DC). Acute exposure to SMFs is reported to have effects on both the inactivation kinetics of VGPCs and the intracellular basal calcium level. Finally, one other study 71 established a direct proportionality between the frequency of EMFs used (ELF-EMFs and RF-EMFs) and the displacement of protein α-helices of different ion channels.

Discussion
Biological effects of EMFs are widely reported in the literature, and extreme low frequencies in particular have been shown to have many effects, including changes in VGC conductance and neurotransmitter expression.
The effects of EMFs on VGCs are important due to the key role of these transmembrane proteins in physiological processes in the cell and particularly in the CNS, where they are at the center of the regulation of a myriad of neurophysiological processes, ranging from the generation of action potentials (APs) to synaptic transmission. 72 These transmembrane proteins are highly conserved throughout diverse biological kingdoms 73 and all share a similar structure, consisting of a variable number of subunits arranged to form a pore through which ions can travel in the direction of different electrochemical gradients. 74 The functionality of these VGCs is mediated by the VSD, a specialized region of the protein that is rich in charged residues and can trigger a conformational change in the channel, modulating its entire configuration. 75,76 Properties central to the functionality of VGCs are ion selectivity and ion conductance. 77 Ion selectivity is key to the specificity of each VGC family and is regulated by a portion of the protein known as the selectivity filter, embedded in the upper extracellular side of the PD. This selectivity filter discriminates between different ions based on their diameter (of a few angstroms) and charge. 78 Ion conductance reflects, on the other hand, the inverse of channel resistance to ion flow, or in other words the number of ions that the VGCs translocate in a specific period of time. Different VGCs belonging to the same family can have the same selectivity but greatly differ in their conductance. 79 A great diversity of VGCs exist but three main families can be identified on the basis of their distinct selectivity for a particular permeant ion. First, VGSCs are responsible for the initiation of APs in excitable cells, such as neurons, 70 and represent a common target for the therapy of epilepsy and many neurodegenerative diseases. 80,81 Second, VGPCs are extremely important in shaping APs, where they are responsible for membrane's repolarization, and in the general modulation of neuronal excitability. 77 Their essential role is testified to by the existence of numerous channelopathies 79 in which a mutation in the genes encoding for structural subunits of these channel complexes is enough to cause severe conditions, ranging from epilepsy and related disorders to numerous forms of ataxia and dyskinesia. 82 Finally, VGCCs are involved in the modulation of an extremely broad variety of neurological processes, including (but not limited to) modulation of neurotransmitter release and intersynaptic short-and long-term communication, neuronal plasticity, neurite outgrowth, and gene expression. 83 Due to the major role of calcium as a second messenger, VGCC dysfunctions have been implicated in a wide variety of CNS pathologies, including epilepsy, neurodegenerative diseases, neuropathic pain, and neuropsychiatric disorders. 84 The particular electrical sensitivity of VGCs themselves makes them a perfect target for EMF effects. First, the charges located on the S4 helix voltage sensor in the VSD 85 are particularly exposed to the electrostatic forces resulting from an applied EMF, and this type of stimulation could easily trigger a displacement in these charges, similar to the one generated by the depolarizing wave of an AP, irregularly gating the channel. 11 Having the same direction at a particular time, the magnitude of this shift becomes greatly amplified. Furthermore, the location of the VSD, embedded in surrounding lipid environment, needs to be considered. Indeed, the forces exerted by an EMF tend to align the permanent phospholipid dipoles, which are normally randomly oriented because of thermal excitation, producing new fields. 86 Finally, transmembrane proteins allow a higher permeation of the EMF with respect to the surrounding membrane since the resistance of the membrane is extremely high and displays a dielectric constant up to 120 times lower than the aqueous phase of the cytoplasm where most of the charges are located. 87 96 Ann The effect of EMFs on VGCs is exerted at multiple levels VGCs represent a perfect candidate for the transduction of the EMF effects on neural tissue. Here, we summarize the evidence for EMF-related effects on these ion channels in the studies analyzed in this review. Both ELF-EMFs and RF-EMFs have been found to modulate VGCs in many ways, including their expression, [40][41][42]58 gating dynamics, 38,39,43,88,89 and insertion into the membrane. 38 An increase 40 (or a decrease, as reported for RF-EMF exposure) 41,42 in VGC density could explain the altered ion flux and account for the many secondary effects reported in many papers, including, but not limited to, activation of the autophagic pathway, 41 altered spike frequencies and AP firing, 43 and facilitated vesicle endocytosis and synaptic plasticity. 40 However, this mechanism cannot account for the rapid effects elicited by acute exposure to ELF-EMFs, such as increased level of ROS, 52,60 altered firing rate, 39 and spontaneous intracellular calcium variations, 60 due to the longer time (ranging from tens of seconds to days) required by the cell to modify its gene expression patterns. Thus, an effect on the voltage-sensing and gating dynamics of VGCs is likely to be involved. Indeed, a shift in the steady state of VGSCs (Na v ) was observed in two different studies, 38,89 and single-ion channel studies revealed altered gating dynamics for both high-threshold VGCCs and KCa channels subsequent to acute ELF-EMF exposure. 39 Similar alterations were further observed in the inactivation kinetics of different types of VGPCs after acute exposure to SMFs. 46 Interestingly, one of the studies analyzed in this review reported a direct correlation between the frequency of EMFs (ELF-EMFs and RF-EMFs) and the displacement of the α-helices in ion channels, 71 which could be related to the changes in gating dynamics reported elsewhere. 38,39,43,88,89 In summary, it appears that the effect of EMFs on VGCs is exerted at multiple levels, one being a rapid modulation in the transport dynamics of VGC proteins, and the other being changes in both their gene and protein expression and density in the membrane. The latter seem to require a prolonged exposure to EMFs and could result from the same intracellular ionic concentration shift extensively reported as one of the major effects of EMF exposure. On the other hand, three studies reported no (or at least no significant) changes to VGC transport dynamics related to EMF exposure. However, it must be noted that these studies used very low field intensities (≤1000 μT) 31,59 or a particular combination of radio frequencies and acute exposure 44 that other studies have shown to be less effective, 39 thus explaining the lack of identifiable effects reported. Finally, the different ion channels in different brain regions must be taken into account. The only paper proposing a VGC-independent mechanism (based on the release of intracellular Ca 2+ stores) for the effect triggered by EMF exposure on calcium homeostasis also points out how the model used is characterized by a poor expression of VGCCs. 90 The role of calcium The important role of calcium in the induction of EMF exposure-related effects was already established 91 and is further confirmed in this review by the many papers reporting an increase in the basal level of calcium following EMF exposure. 39,40,52,88 This is not surprising, since calcium is universally known as the most important and widespread regulator of neurological processes, such as neural differentiation, survival, apoptosis, 54,55 neurotransmitter release, excitability, and synaptic plasticity. 92 The complex and different calcium-related signaling pathways have variable importance in different areas of the brain. This very reason could explain the complex and contrary effects reported in the literature. Such an increase in intracellular calcium influx could be due to different mechanisms, and the papers reviewed here propose many mechanisms: a direct effect on the mean opening time of VGCCs, 39,88 both an increase 40 or a decrease 41,42 in their expression (depending on the frequency of the field and exposure time), and the augmented release of glutamate that through the activation of NMDAR would stimulate such an increase in the hippocampus. 58 This last hypothesis is particularly intriguing because it could easily explain the many reported effects that the EMFs exert on learning, memory, and synaptic plasticity. 57,93 The particularly relevant role of VGCCs in the transduction of EMF effects on neurons seems to be attributable to the great number of effects produced by an increase in intracellular calcium levels, as previously suggested. 94 Indeed, the few studies focusing on other VGCs, such as VGPCs and VGSCs, reviewed here 38,39,43,46,90,95 reported significant effects specific to these ion

Different effects for different fields
An extensive literature exists reporting a myriad of EMF exposure-related effects on many biological processes, ranging from cell differentiation, survival, and changes in gene expression 96,97 to effects on cell membranes and signal transduction pathways. 30 However, many other studies indicated the absence of significant effects elicited by these fields. 59,61 A possible explanation for the different effects reported could be related to the fact that the way in which EMFs interact with the body depends on what combination of frequencies are used and the related wavelengths.
It is well known that the effects of exposure to EMFs differ significantly based on the exposure intensities and the exposure time, 1,98 and because of this any reasonable comparison must be made between groups having the same experimental conditions. In this study, we found different types of EMFs employed, although the two most commonly represented categories were ELF-EMFs and RF-EMFs, in line with the well-documented biological relevance of these fields. 94,99,100 It is important to point out that the effects exerted by these two types of exposures are not equal due to the intrinsic electrical properties of the neuronal membrane. For instance, electrical phenomena involving a redistribution of charges in the membrane subsequent to EMF exposure, such as counterion polarization and phospholipid reorientation, are not likely to occur in RF-EMF exposure, due to the high inertia of charged particles at this high frequency. 101 Moreover, pulsed EMFs are often reported to be more active relative to static EMFs, which are characterized by a continuous electromagnetic wave to which the cell could be more adapted, 102 and they could affect the gating properties of VGCs since these proteins are intrinsically sensitive to minimal electrical variations. 100 Likewise, the effect of SMFs could similarly influence VGCs through a deformation of the membrane involving a reorientation of the phospholipid bilayer, as suggested by Rosen's study. 103 Indeed, both of the studies reviewed here that focused on SMF effects reported effects on VGCs, specifically on the gating dynamics of VGCCs 47 and the inactivation dynamics of VGPCs. 46 Lastly, it is worth mentioning that the frequencyrelated impact of the various type of EMFs has not been totally clarified, and theories exist suggesting that only specific frequencies would relevantly interact with the cell. [104][105][106] However, although many different types of fields have been used, the frequencies used were similar (specifically 50 Hz for ELF-EMFs and 835 and 900 Hz for RF-EMFs).

Relevance of EMF exposure time
It is worth noting that, in the studies analyzed here, many different exposure times have been used. This variability might account for the many different and sometimes opposite effects reported. The time dependency of EMF exposure-related effects is well known. A 1973 study by Tolgskaya and Gordon reported how, in the first months of exposure to radio waves, the morphological and physiological effects on animals brain are poor and modest, becoming evident and irreversible after longer exposure. 107 Most papers reviewed here investigated acute (up to 2 h) or subchronic exposure (from 2 to 48 h) and could, therefore, have overlooked the effects elicited by longer exposures. Interestingly, all the papers except one 31 reported significant effects after chronic (>48 h) exposure, whatever the type of field used. On the other hand, the rapid increase in intracellular calcium reported in many papers after ELF-EMF exposure seems to go against this line of thinking, pointing toward a direct effect on VGCs. These reasons, in line with the different and various effects of EMF exposure reported, seem to suggest that EMFs could act through more than one mechanism, to differentially influence particular brain areas or neuronal populations according to the exposure time.

Effects of EMFs on neural development
As stated above, there is great interest regarding EMF-related effects on neural development since less differentiated cells have been proposed to be particularly susceptible to these fields. Supporting this idea, one of the studies reviewed here reported significant effects on the electrical properties of Purkinje cerebellar neurons subsequent to exposure during development (6 h per day for the gestation period) to RF-EMFs. 43 On the other hand, de Groot and colleagues failed to observe important anomalies in an ELF-EMF development exposure model, although the short exposure time (7 days) that did not cover the entire length of mouse pregnancy  60 Finally, VGCC expression dynamics are far from stable throughout development, ranging from a prevalent expression of T-type calcium channels in the initial stages to the higher presence of Nand L-type channels in the mature neuron. 108 This differential expression and the preferential effect of EMF on specific types of VGCCs should, therefore, be considered in the interpretation of every study involving EMF exposure.

Limitations of this study
This study investigates a complex field, with sometimes conflicting results. The many variables that influence the impact of EMF exposure on neural tissue, such as the physiological state of the cell, its developmental stage, and the various physical characteristics of the many fields involved, complicate the reproducibility and often impede a consistent comparison between different studies. In spite of having highlighted some recurring patterns in the reported results, this review is, therefore, limited by the intrinsic differences of the studies reviewed.

Conclusion
The studies reviewed here show VGCs as an important transducer of the effect of EMFs in neurons, and the central role played by these proteins in the regulation of important biological processes, central in the regulation of brain physiology, sheds a light on the influence that modern exposure to EMFs could have on human health. While a diverse range of biological systems were used, cell lines were the preferred option, and VGCCs were the most studied ion channels, in line with their central role in the regulation of many physiological processes in neurons. However, many other VGCs have been shown to be affected by EMFs and the results are often conflicting. In spite of the controversy, this systematic review reports significant correlation between EMFs and multiple changes in the electrophysiological properties of diverse neuronal tissues, and these results, if interpreted well, could pave the way to a new understanding of the relationship between electromagnetic stimulation and brain functions. In conclusion, we systematically demonstrate how the complex effects of EMFs in neuronal ion channels are exerted at multiple levels and how their significance in the alteration of neuronal functions is strictly dependent on different parameters relative to the type of field used and the studied cell or tissue. Improved experimental reproducibility will be key to any advances in this field, and the development of new experimental procedures capable of measuring the small but profound way in which certain types of EMF exposure seem to affect our brain might help us to establish whether it is harmful and its therapeutic potential. We hope this work will help in improving our knowledge about the molecular dynamics of neuronal VGCs, which will be key both for any progress in the treatment of neurodegenerative diseases and for an advancement in the general understanding of the relationship between technological progress and cellular dynamics.