New head models extracted from thermal infrared (IR) images for dosimetry computations

In electromagnetic dosimetry, anatomical human models are commonly obtained by segmentation of magnetic resonance imaging or computed tomography scans. In this paper, a human head model extracted from thermal infrared images is examined in terms of its applicability to specific absorption rate (SAR) calculations. Since thermal scans are two-dimensional (2D) representation of surface temperature, this allows researchers to overcome the extensive computational demand associated with 3D simulation. The numerical calculations are performed using the finite-difference time-domain method with mesh sizes of 2 mm at 900 MHz plane wave irradiation. The power density of the incident plane wave is assumed to be 10 W/m(2). Computations were compared with a realistic anatomical head model. The results show that although there were marked differences in the local SAR distribution in the various tissues in the two models, the 1 g peak SAR values are approximately similar in the two models.     

Assessment of the exposure to WLAN frequencies of a head model with a cochlear implant

In the last few years, significant developments have taken place in the field of Wireless Local Area Networks (WLAN), and the popularity of portable devices supporting Wireless Fidelity (Wi‐Fi) is continuously growing. At the same time, the number of Active Implanted Medical Devices (AIMD) being placed in patients is widely increasing and among them, cochlear implants (CI) are becoming a common aid. The goal of this study is to investigate the effect on the electromagnetic field distribution and the specific absorption rate (SAR) due to the presence of a CI in a head model during far‐field exposure to Wi‐Fi frequencies. The head model was obtained by image segmentation, the implant was modelled as a geometric structure, and the exposure sources were modelled as a uniform plane wave (power density = 10 W/m²) at 2.4, 5.2 and 5.8 GHz. Vertical and horizontal polarizations were simulated. Conditions with and without CI were compared. The findings of that are: (1) local differences in the field distribution close to the CI, comparing the head models with or without the CI; (2) higher field strength and point SAR value in the cochlear region very close to the CI; (3) negligible differences in the field strength and point SAR value in the cochlear region far from the CI; (4) negligible variations in the average SAR values in the cochlea and head due to the presence of the CI. The results of this study conclude that insertion of a CI brings moderate localized differences in the E, H and point SAR distribution when evaluated close to the electrode array in the cochlea, while negligible differences are found in the average SAR values both in the cochlea and head, independent of frequency and wave polarization. Bioelectromagnetics 31:546–555, 2010. © 2010 Wiley‐Liss, Inc.     

SAR reduction using a single SRR superstrate for a dual-band antenna

A dual-band microstrip antenna operating at GSM 900 and GSM 1800 MHz is designed initially. Then a single split ring resonator (SRR) structure is used as a superstrate for this dual-band antenna. A circular current is induced in the SRR due to the perpendicular plane wave excitation, which in turn leads to an electric excitation coupled to the magnetic resonance. It also exhibits higher order excitations at 0.9 and 1.8 GHz which ultimately resulted in specific absorption rate (SAR) reduction of human head at both the designed frequencies of the antenna. The antenna and the SRR superstrate are printed on a 1.6 mm thick FR-4 substrate of dimension 59.6 × 49.6 mm². Analysis of the SRR using the classic waveguide theory approach is discussed. Radiation pattern of the antenna in the presence of SRR superstrate and human head is also discussed. Prototype of the antenna along with the SRR superstrate is fabricated and measured for return loss and radiation pattern. Measurement results fairly agree with the simulated results. A human head phantom is utilized in the calculation of SAR.     

Statistical multi‐path exposure method for assessing the whole‐body SAR in a heterogeneous human body model in a realistic environment

Assessing the whole‐body absorption in a human in a realistic environment requires a statistical approach covering all possible exposure situations. This article describes the development of a statistical multi‐path exposure method for heterogeneous realistic human body models. The method is applied for the 6‐year‐old Virtual Family boy (VFB) exposed to the GSM downlink at 950 MHz. It is shown that the whole‐body SAR does not differ significantly over the different environments at an operating frequency of 950 MHz. Furthermore, the whole‐body SAR in the VFB for multi‐path exposure exceeds the whole‐body SAR for worst‐case single‐incident plane wave exposure by 3.6%. Moreover, the ICNIRP reference levels are not conservative with the basic restrictions in 0.3% of the exposure samples for the VFB at the GSM downlink of 950 MHz. The homogeneous spheroid with the dielectric properties of the head suggested by the IEC underestimates the absorption compared to realistic human body models. Moreover, the variation in the whole‐body SAR for realistic human body models is larger than for homogeneous spheroid models. This is mainly due to the heterogeneity of the tissues and the irregular shape of the realistic human body model compared to homogeneous spheroid human body models. Bioelectromagnetics 34:240–251, 2013. © 2012 Wiley Periodicals, Inc.     

The effect of resorption cavities on bone stiffness is site dependent

Resorption cavities formed during the bone remodelling cycle change the structure and thus the mechanical properties of trabecular bone. We tested the hypotheses that bone stiffness loss due to resorption cavities depends on anatomical location, and that for identical eroded bone volumes, cavities would cause more stiffness loss than homogeneous erosion. For this purpose, we used beam–shell finite element models. This new approach was validated against voxel-based FE models. We found an excellent agreement for the elastic stiffness behaviour of individual trabeculae in axial compression (R² = 1.00) and in bending (R²>0.98), as well as for entire trabecular bone samples to which resorption cavities were digitally added (R² = 0.96, RMSE = 5.2%). After validation, this new method was used to model discrete cavities, with dimensions taken from a statistical distribution, on a dataset of 120 trabecular bone samples from three anatomical sites (4th lumbar vertebra, femoral head, iliac crest). Resorption cavities led to significant reductions in bone stiffness. The largest stiffness loss was found for samples from the 4th lumbar vertebra, the lowest for femoral head samples. For all anatomical sites, resorption cavities caused significantly more stiffness loss than homogeneous erosion did. This novel technique can be used further to evaluate the impact of resorption cavities, which are known to change in several metabolic bone diseases and due to treatment, on bone competence.     

Effects of a brain tumor in a dispersive human head on SAR and temperature rise distributions due to RF sources at 4G and 5G frequencies

In this paper, effects of a brain tumor located in a dispersive human head model on specific absorption rate (SAR) and temperature rise distributions due to different types of RF sources at 4G and 5G cellular frequencies are investigated with the use of a multiphysics model. This multiphysics model analyzes the dispersive human head with the brain tumor and provides the SAR and temperature rise distributions in the head due to the RF source operated at 4G and 5G cellular frequencies in a single finite-difference time-domain simulation. An adjacent antenna operated at 4G and 5G cellular frequencies to the human head is considered as the RF source for near-field exposure, while a plane wave field radiated by base stations operated at 4G and 5G cellular frequencies is considered as the RF source for far-field exposure. Numerical results show that the brain tumor in the head slightly affects the SAR and temperature rise distributions due to different RF sources at 4G and 5G cellular frequencies.     

Postresonance electromagnetic absorption by man and animals

A surface integral equation (SIE) method is used to calculate the specific absorption rate (SAR) in spherically capped cylindrical models irradiated by an axially incident electromagnetic plane wave (K polarization) in a frequency range for which calculations previously have not been available (80–400 MHz for man models). In the SIE method, the electromagnetic (EM) field relations are formulated in terms of electric and magnetic currents on the surface of the model. The average SAR is calculated from the far scattered EM fields by means of the forward scattering theorem. SAR data calculated by the SIE method agree with data calculated by the extended boundary condition method (EBCM) for frequencies up to 80 MHz (the upper frequency limit of the EBCM) for man models. For rat models exposed to 1–3 GHz radiation, reasonable agreement was also obtained with the limited experimental data available.     

Estimation of head tissue-specific exposure from mobile phones based on measurements in the homogeneous SAM head

The maximum spatial peak exposure of each commercial mobile phone determined in compliance with the relevant safety and product standards is publicly available. However, this information is not sufficient for epidemiological studies aiming to correlate the use of mobile phones with specific cancers or to behavioral alterations, as the dominant location of the exposure may be anywhere in the head between the chin to above the ear, depending on the phone design. The objective of this study was to develop a methodology to determine tissue-specific exposure by expanding the post-processing of the measured surface or volume scans using standardized compliance testing equipment, that is, specific absorption rate (SAR) scanners. The transformation matrix was developed using the results from generic dipoles to evaluate the relation between the SAR in many brain regions of the Virtual Family anatomical phantoms and in virtual brain regions mapped onto the homogeneous SAM head. A set of transformation factors was derived to correlate the SAR induced in the SAM head to the SAR in the anatomical heads. The evaluation included the uncertainty associated with each factor, arising from the anatomical differences between the phantoms (typically less than 6 dB (4×)). The applicability of these factors was validated by performing simulations of four head models exposed to four realistic mobile phone models. The new methodology enables the reliable determination of the maximum and averaged exposure of specific tissues and functional brain regions to mobile phones when combined with mobile phone power control data, and therefore greatly strengthens epidemiological evaluations and improves information for the consumer.     

Determination of electric current distributions in animals and humans exposed to a uniform 60-Hz high-intensity electric field

The thermographic method for determining specific absorption rate (SAR) in animals and models of tissues or bodies exposed to electromagnetic fields was applied to the problem of quantifying the current distribution in homogeneous bodies of arbitrary shape exposed to 60-Hz electric fields. The 60-Hz field exposures were simulated by exposing scale models of high electrical conductivity to 57.3-MHz VHF fields of high strength in a large 3.66 × 3.66 × 2.44-m TE₁₀₁ mode resonant cavity. After exposure periods of 2–30 s, the models were quickly disassembled so that the temperature distribution (maximum value up to 7 °C) along internal cross-sectional planes of the model could be recorded thermographically. The SAR, W′, calculated from the temperature changes at any point in the scale model was used to determine the SAR, W, for a full-scale model exposed to a 60-Hz electric field of the same strength by the relation W = (60/ f² · (σ′/σ) · W′ where f′ is the model exposure frequency, σ′ is the conductivity of the scale model at the VHF exposure frequency, and σ is the conductivity of the full-scale subject at 60 Hz. The SAR was used to calculate either the electric field strength or the current density for the full-scale subject. The models were used to simulate the exposure of the full-scale subject located either in free space or in contact with a conducting ground plane. Measurements made on a number of spheroidal models with axial ratios from 1 to 10 and conductivity from 1 to 10 s/m agreed well with theoretical predictions. Maximum current densities of 200 nA/cm² predicted in the ankles of man models and 50 nA/cm² predicted in the legs of pig models exposed to 60-Hz fields at 1kV/m agreed well with independent measurements on full-scale models.     

Salt impacts on organic carbon and nitrogen leaching from senesced vegetation

Senesced vegetation is exposed to a wide range of salt concentrations in surface waters resulting from human activities which include deicing salts and irrigation water chemistry. Both dissolved organic carbon (DOC) and salt concentrations are rising in northern hemisphere watersheds, yet there has been little investigation of sodium as a potential mechanism for DOC increases. The objective of this study was to investigate the impact of solution sodicity and salinity on DOC and dissolved organic nitrogen (DON) leaching from five types of senesced and cut vegetation. Vegetation was soaked for 24 h in a series of sodium chloride (NaCl)–calcium chloride (CaCl₂) solutions with sodium adsorption ratios (SAR) of 2, 10, or 30 and electrical conductivities of 0.1 dS m^?1 through 3.0 dS m^?1. Vegetation was also soaked in a sodium bicarbonate (NaHCO₃) solution at SAR = 30 and stream water from local watersheds with a range of sodicity and salinity. The mass of both DOC and DON released increased as SAR increased in the NaCl solutions, but the total salinity had inconsistent effects on DOC and DON release. NaHCO₃ leached similar amounts of DOC and DON as NaCl. The SAR of the stream water solutions was able to explain 88 % of the variability in DOC leached from vegetation (p < 0.05). The results indicated that sodicity, quantified by SAR, had a significant impact on DOC and DON leaching from senesced vegetation and could be a potential mechanism to explain the observed increases in surface water DOC.     

SAR versus Sinc: What is the appropriate RF exposure metric in the range 1–10 GHz? Part I: Using planar body models

This is the first of two articles addressing the most appropriate crossover frequency at which incident power flux density (S_inc) replaces the spatial peak value of the specific energy absorption rate (SAR) averaged over 1 or 10 g (i.e., peak 1 or 10 g SAR) as the basic restriction for protecting against radiofrequency (RF) heating effects in the 1–10 GHz range. Our general approach has been to compare the degree of correlation between these basic restrictions and the peak induced tissue temperature rise (ΔT) for a representative range of population/exposure scenarios. In this article we particularly address the effect of human population diversity in the thickness of body tissue layers at eight different sites of the body. We used a Monte Carlo approach to specify 32000 models (400 models for each of 8 body sites for 10 frequencies) which were representative of tissue thicknesses for age (18–74 years) and sex at the eight body sites. Histogram distributions of S_inc and peak 1 and 10 g SAR corresponding to a peak 1 °C temperature rise were obtained from RF and thermal analyses of 1D multiplanar models exposed to a normally incident plane wave ranging from 1 to 10 GHz in thermo‐neutral environmental conditions. Examination of the distribution spread of the histograms indicated that peak SAR was a better predictor of peak tissue temperature rise across the entire 1–10 GHz frequency range than S_inc, as indicated by the smaller spread in its histogram distributions, and that peak 10 g SAR was a slightly better predictor than peak 1 g SAR. However, this result must be weighed against partly conflicting indications from complex body modeling in the second article of this series, which incorporates near‐field effects and the influence of complex body geometries. Bioelectromagnetics 31:454–466, 2010. © 2010 Wiley‐Liss, Inc.     

Human head dynamic response to side impact by finite element modeling

The dynamic response of the human head to side impact was studied by 2-dimensional finite element modeling. Three models were formulated in this study. Model I is an axisymmetric model. It simulated closed shell impact of the human head, and consisted of a single-layered spherical shell filled wiht an inviscid fluid. The other two models (Model II and III) are plane strain models of a coronal section of the human head. Model II approximated a 50th percentile male head by an outer layer to simulate cranial bone and an inviscid interior core to simulate the intracranial contents. The configuration of Model III is the same as Model II but more detailed anatomical features of the head interior were added, such as, cerebral spinal fluid (CSF); falx cerebri, dura, and tentorium. Linear elastic material properties were assigned to all three models. All three models were loaded by a triangular pulse with a peak pressure of 40 kPa, effectively producing a peak force of 1954 N (440 lb). The purpose of this study was to determine the effects of the membranes and that of the mechanical properties of the skull, brain, and membrane on the dynamic response of the brain during side impact, and to compare the pressure distributions from the plane strain model with the axisymmetric model. A parametric study was conducted on Model II to characterize fully its response to impact under various conditions.(ABSTRACT TRUNCATED AT 250 WORDS)     

Enhanced absorption of millimeter wave energy in murine subcutaneous blood vessels

The aim of the present study was to determine millimeter wave (MMW) absorption by blood vessels traversing the subcutaneous fat layer of murine skin. Most calculations were performed using the finite-difference time-domain (FDTD) technique. We used two types of models: (1) a rectangular block of multilayer tissue with blood vessels traversing the fat layer and (2) cylindrical models with circular and elliptical cross-sections simulating the real geometry of murine limbs. We found that the specific absorption rate (SAR) in blood vessels normally traversing the fat layer achieved its maximal value at the parallel orientation of the E-field to the vessel axis. At 42 GHz exposure, the maximal SAR in small blood vessels could be more than 30 times greater than that in the skin. The SAR increased with decreasing the blood vessel diameter and increasing the fat thickness. The SAR decreased with increasing the exposure frequency. When the cylindrical or elliptical models of murine limbs were exposed to plane MMW, the greatest absorption of MMW energy occurred in blood vessels located on the lateral areas of the limb model. At these areas the maximal SAR values were comparable with or were greater than the maximal SAR on the front surface of the skin. Enhanced absorption of MMW energy by blood vessels traversing the fat layer may play a primary role in initiating MMW effects on blood cells and vasodilatation of cutaneous blood vessels.     

A numerical and experimental comparison of human head phantoms for compliance testing of mobile telephone equipment

A new human head phantom has been proposed by CENELEC/IEEE, based on a large scale anthropometric survey. This phantom is compared to a homogeneous Generic Head Phantom and three high resolution anatomical head models with respect to specific absorption rate (SAR) assessment. The head phantoms are exposed to the radiation of a generic mobile phone (GMP) with different antenna types and a commercial mobile phone. The phones are placed in the standardized testing positions and operate at 900 and 1800 MHz. The average peak SAR is evaluated using both experimental (DASY3 near field scanner) and numerical (FDTD simulations) techniques. The numerical and experimental results compare well and confirm that the applied SAR assessment methods constitute a conservative approach.     

Cortical Source Analysis of High-Density EEG Recordings in Children

EEG is traditionally described as a neuroimaging technique with high temporal and low spatial resolution. Recent advances in biophysical modelling and signal processing make it possible to exploit information from other imaging modalities like structural MRI that provide high spatial resolution to overcome this constraint¹. This is especially useful for investigations that require high resolution in the temporal as well as spatial domain. In addition, due to the easy application and low cost of EEG recordings, EEG is often the method of choice when working with populations, such as young children, that do not tolerate functional MRI scans well. However, in order to investigate which neural substrates are involved, anatomical information from structural MRI is still needed. Most EEG analysis packages work with standard head models that are based on adult anatomy. The accuracy of these models when used for children is limited², because the composition and spatial configuration of head tissues changes dramatically over development³. In the present paper, we provide an overview of our recent work in utilizing head models based on individual structural MRI scans or age specific head models to reconstruct the cortical generators of high density EEG. This article describes how EEG recordings are acquired, processed, and analyzed with pediatric populations at the London Baby Lab, including laboratory setup, task design, EEG preprocessing, MRI processing, and EEG channel level and source analysis.      

Setup and dosimetry for exposing anaesthetised pigs in vivo to 900 MHz GSM mobile phone fields

The aim of this study was a dosimetrical analysis of the setup used in the exposure of the heads of domestic pigs to GSM-modulated radio frequency electromagnetic fields (RF-EMF) at 900 MHz. The heads of pigs were irradiated with a half wave dipole using three different exposure routines; short bursts of 1-3 s at two different exposure levels and a continuous 10-min exposure. The electroencephalogram (EEG) was registered continuously during the exposures to search for RF-EMF originated changes. The dosimetry was based on simulations with the anatomical heterogeneous numerical model of the pig head. The simulation results were validated by experimental measurements with the exposure dipole and a homogeneous liquid phantom resembling the pig head. The specific absorption rate (SAR), defined as a maximum average over 10 g tissue mass (SAR(10g)), was 7.3 W/kg for the first set of short bursts and 31 W/kg for the second set of short bursts. The SAR(10g) in the continuous 10-min exposure was 31 W/kg. The estimated uncertainty for the dosimetry was +/-25% (K = 2).     

The McClure and Weiss models of Fe–O2 bonding for oxyhemes, and the HbO2 + NO reaction

For the Fe–O₂(S = 0) linkages of oxyhemes, valence bond (VB) structures are re-presented for the McClure [Fe^II(S = 1) + O₂(S = 1)], Pauling–Coryell [Fe^II(S = 0) + O₂*(S = 0)], and Weiss [Fe^III(S = ½) + O₂ ^?(S = ½)] models of bonding. The VB structures for the McClure and Weiss models are of the increased-valence type, with more electrons participating in bonding than occur in their component Lewis structures. The Fe–O bond number and O–O bond order for the McClure structure are correlated with measured Fe–O and O–O bond lengths for oxymyoglobin. Back-bonding from O ₂ ^? to Fe^III of the Weiss structure gives a restricted form of the McClure structure. The McClure and Weiss increased-valence structures are used to provide VB formulations of mechanisms for the oxyhemoglobin + NO reaction. The products of these two formulations are Hb⁺ and NO₃ ^? (where Hb is hemoglobin) and Hb⁺ and OONO^?, respectively. Because Hb⁺ and NO₃ ^? are the observed products, they provide an experimental procedure for distinguishing the McClure and Weiss models. It is also shown that the same type of agreement between McClure-type theory and experiment occurs for oxycoboglobin + NO, cytochrome P450 monooxygenases, and related hydrogen atom transfer reactions. In the appendices, the results of density functional theory and multireference molecular orbital calculations for oxyhemes are related to one formulation of the increased-valence wavefunction for the McClure model, and theory is presented for the calculation of approximate weights for the Lewis structures that are components of the McClure increased-valence structure.     

The relation between the specific absorption rate and electromagnetic field intensity for heterogeneous exposure conditions at mobile communications frequencies

The relation between the incident electromagnetic field strength and both the whole‐body and the local specific absorption rate (SAR) was investigated for typical heterogeneous exposure scenarios for frequencies relevant for mobile communication. The results were compared to results from plane wave exposure. Heterogeneous exposure arises from multiple path propagation of the electromagnetic waves to the location of interest. It is shown that plane wave exposure does not represent worst‐case exposure conditions. When the electric field strength arising at plane wave exposure is compared to the electric field strength averaged over the volume of the human body occurring during multipath exposure, 12% of all heterogeneous cases examined represent worse exposure conditions than plane wave exposure for whole‐body exposure at 946 MHz, 15% at 1840 MHz, and 22% at 2140 MHz. The deviation between plane wave and heterogeneous whole‐body SAR ranges from ?54% to 54%. For partial‐body SAR averaged over 10 g of tissue, a range from ?93% to 209% was found when comparing multiple wave exposure to single incoming plane waves. The investigations performed using the Visible Human as phantom showed that the basic restrictions are met as long as the reference levels are not exceeded. However, this must not be necessarily the case when different phantoms are used to perform similar investigations because recent studies demonstrated that reference levels might not be conservative when phantoms of children are used. Therefore, the results of this work indicate the need to extend the investigations to numerical simulations with additional human phantoms representing parts of the human population having different anatomy and morphology compared to the phantom used within the frame of this project. This also applies to phantoms of children. Bioelectromagnetics 30:651–662, 2009. © 2009 Wiley‐Liss, Inc.     

Multiple proton relay routes in the reaction mechanism of RNAP II: Assessing the effect of structural model

RNA polymerase II catalyzes the nucleotidyl transfer reaction for messenger RNA synthesis in eukaryotes. Two crystal structures of this system have been resolved, each with its own defects in the coordination sphere of Mg²⁺(A) resulting from chemical modifications. We have used both structures and also a novel hybrid of the two that allows a better exploration of the parts of configuration space that reflect substrate–enzyme interactions. MD and QM/MM calculations have been performed, the latter with the semiempirical AM1/d‐PhoT method, calibrated against density functional theory. Reaction path scans in 1‐D provided insights about the role of Mg²⁺ (A) which turns out to be more structural than catalytic. In contrast, 1‐D scans of the incorporation of the nucleotidyl group yielded barriers that were much too high, necessitating the use of 2‐D reaction coordinates. Three different proton acceptors for the initial reaction step were examined. For those models based on the two PDB structures the 2‐D scans continued to yield very high barriers, indicating that the reaction is unlikely to proceed from these configurations. On the other hand, two hybrid models, chosen from the early and late parts of a 12ns molecular dynamics simulation yielded greatly reduced barriers in the range of ～17 to ～27 kcal/mol for the three proton acceptors, as compared to the experimental estimate of 18 kcal/mol. The final step, release of pyrophosphate, was found to be facile. Our overall mechanism involves only active site residues or water without the need for external reactive agents such as the hydroxide ion previously proposed. Proteins 2015; 83:268–281. © 2014 Wiley Periodicals, Inc.     

Sampling-Design Effects on Properties of Species-Area Relationships – A Case Study from Estonian Dry Grassland Communities

Despite widespread use of species-area relationships (SARs), dispute remains over the most representative SAR model. Using data of small-scale SARs of Estonian dry grassland communities, we address three questions: (1) Which model describes these SARs best when known artifacts are excluded? (2) How do deviating sampling procedures (marginal instead of central position of the smaller plots in relation to the largest plot; single values instead of average values; randomly located subplots instead of nested subplots) influence the properties of the SARs? (3) Are those effects likely to bias the selection of the best model? Our general dataset consisted of 16 series of nested-plots (1 cm²–100 m², any-part system), each of which comprised five series of subplots located in the four corners and the centre of the 100-m² plot. Data for the three pairs of compared sampling designs were generated from this dataset by subsampling. Five function types (power, quadratic power, logarithmic, Michaelis-Menten, Lomolino) were fitted with non-linear regression. In some of the communities, we found extremely high species densities (including bryophytes and lichens), namely up to eight species in 1 cm² and up to 140 species in 100 m², which appear to be the highest documented values on these scales. For SARs constructed from nested-plot average-value data, the regular power function generally was the best model, closely followed by the quadratic power function, while the logarithmic and Michaelis-Menten functions performed poorly throughout. However, the relative fit of the latter two models increased significantly relative to the respective best model when the single-value or random-sampling method was applied, however, the power function normally remained far superior. These results confirm the hypothesis that both single-value and random-sampling approaches cause artifacts by increasing stochasticity in the data, which can lead to the selection of inappropriate models.