3.4.4 CALCIUM EFFLUX

Adey, Bawin, and their colleagues have reported extensively on studies of changes in calcium efflux from preparations of neonate chick brain and cat cortex due to exposure of such preparations to very specific frequencies and intensities of unmodulated sub-ELF fields or to specific levels of VHF and UHF RFR amplitude-modulated at very specific sub-ELF frequencies, using the radioactive-calcium ion Ca-45++ as the tracer. Confirmation of the basic phenomenon was obtained by some investigators, and others reported negative or contradictory results. Representative papers selected from the literature on the subject are discussed below.

Bawin et al. (1975), in an early paper on calcium efflux, described the experimental protocol for chick brains in detail. Five hundred neonatal chicks in the age range 2-7 days were used. After decapitating each chick, its forebrain was quickly excised and each cerebral hemisphere was incubated for 30 min in 1.0 ml of physiologic medium and 0.2 ml of saline containing 0.2 microcurie of Ca-45++, within a polyallomer test tube. Following incubation, the samples were washed three times in nonradioactive solution. They were then immersed in 1 ml of physiologic medium for 20 min, during which one hemisphere of each chick was exposed to RFR with the other hemisphere as its control.

The exposures were to a 147-MHz field between two 4100-sq-cm triangular aluminum plates within an environmental chamber adapted for use of VHF fields and maintained at 37 deg C and 35% relative humidity. The field was modulated at 0.5, 3, 6, 9, 16, 20, 25, or 35 Hz at depths between 80% and 90% or was unmodulated. The RF from the source was applied to the narrow apices of the plates via an antenna coupler. The power was monitored with two inline wattmeters and an RF ammeter and was adjusted to yield intensities in the range 1-2 mW/sq cm.

Sets of 10 brains (10 each exposed and control hemispheres) were treated concurrently for each field condition and control. Each field condition was tested at least three times, to provide large enough populations for statistical analysis. In each experiment, four series of 10 samples each were treated, i.e., one series each was exposed to fields modulated at three different modulation frequencies and one series of unexposed samples served as controls. On treatment completion, a 0.2-ml aliquot of each bathing solution diluted in 11 ml of Packard "Instagel" was assayed for radioactivity by liquid scintillation counting. Regarding treatment of the data, the authors stated: "The radioactivities (cpm) of all samples were related to the mean value of the counts obtained with the 10 control samples, taken as 100. All normalized data within each condition were then statistically compared with matched samples of control values." (No unnormalized data were presented.)

The results for the unmodulated field and each modulation frequency were plotted as bar graphs of mean percentage increases (with SEs) of Ca-45++ concentration (effluxes) relative to the mean value in the absence of the field. The unmodulated field and those modulated at 0.5, 3, 6, 9, and 16 Hz yielded progressively larger increases in the percentage of calcium efflux, with a maximum of about 19% (to 119% of control) for 16 Hz. The results for 0 (unmodulated), 0.5, and 3 Hz were statistically not significant (p>0.05); those for 6 and 9 Hz were significant at the 5% level, and those for 11 and 16 Hz were significant at the 1% level. Above 16 Hz, calcium efflux percentage declined progressively with increasing modulation frequency, with the mean efflux significant (p<0.05) for 20 Hz and not significant (p>0.05) for 25 or 35 Hz.

Forty other chick brains were used to compare the effects of exposure on calcium efflux from brains in physiologic medium with those in brains poisoned by incubation with 0.0001-M sodium cyanide. Four sets, each consisting of five poisoned samples and five normal samples, were assayed, with the samples of each set treated simultaneously: One set each was tested for the effects of exposure to fields modulated at 0 (unmodulated), 0.5, and 16 Hz and one set was tested without exposure. The results are shown in Table 36, adapted from Table 1 of the paper.

TABLE 36: PERCENTAGE CALCIUM EFFLUX FROM NORMAL AND POISONED BRAINS

       Exposure          Condition % Efflux +/- SE (N)     Normal Condition   % Efflux +/- SE (N)       Poisned Brain
       Control     100.0 +/- 4.0 (15)       96.7 +/- 3.6 (15)
       Unmod.     103.7 +/- 6.0 (10)      102.3 +/- 5.3 (10)
       0.5-Hz Mod.     100.5 +/- 4.6 (10)       98.7 +/- 5.0 (10)
       16-Hz Mod.     114.2 +/- 6.4 (10)*      118.9 +/- 7.7 (10)*

----------

*p<0.05

 

As seen above, the difference in mean percentages of calcium efflux between poisoned and normal brains for each exposure condition was not significant. The authors stated: "The field effects observed previously were not altered by the cyanide treatment, which strongly suggests that the Ca-45++ effluxes from the cerebral tissues are independent of any ongoing metabolism."

The authors did not state what type of statistical analysis they used; their statements regarding data treatment, quoted above, do not indicate whether the paired t-test was used on the assays for each exposed brain hemisphere and its contralateral control hemisphere, which would seem to be an appropriate statistic. Regarding the 16-Hz value in Table 36 for normal brains (114.2%), which was labeled significant by them, it is interesting that by unpaired t-test, t=1.99, which is significant under the 1-tailed criterion (0.025<p<0.05), but is just outside the level of significance under the 2-tailed criterion (0.05<p<0.1). Also, that mean was smaller than the 119% obtained previously, but the difference presumably was not significant.

Also performed were preliminary experiments on possible calcium efflux from skeletal muscles (lateral head of the gastrocnemius). Muscle and cerebrum excised from the same chicks were tested concurrently for 20-min exposures to fields modulated at 6 or 16 Hz and no field. The results are shown in Table 37 (from Table 2 of the paper):

TABLE 37: PERCENTAGE CALCIUM EFFLUX FROM BRAIN AND MUSCLE

        Exposure            Condition        % Efflux +/- SE (N)         Brain      % Efflux +/- SE (N)             Muscle
Control 100.0 +/- 5.4 (10) 100.0 +/- 5.8 (10)
6-Hz Mod. 115.0 +/- 4.2 (10) 105.1 +/- 9.8 (10)
16-Hz Mod. 119.1 +/- 6.2 (10) 100.1 +/- 3.9 (15)

---------

*p<0.05

 

As seen above, the results for muscle exhibited much greater variability than those for brain tissue, but neither exposure condition had any significant effect on mean percentage calcium efflux therefrom.

The cerebral-calcium-efflux results of this study constitute evidence for the existence of a quite subtle RFR effect. The absence of the effect for unmodulated RFR and for modulation frequencies outside the range reported is an indication that heating of the preparation by the RFR is unlikely to be involved. Rather, some nonlinear mechanism may exist in the preparation that is capable of recovering and/or responding to one polarity of the modulation envelope. Such a mechanism would also have to have a time constant comparable to that for 16 Hz, at which the effect is near maximum.

The authors speculated that such slow undulations of the extracellular electric field could affect the binding of calcium to the neuronal membrane, and that a small displacement of calcium ions would result in cooperative interaction between modified adjacent binding sites, thus producing propagation and amplification of local electrical events.

Bawin and Adey (1976a,b), to ascertain whether calcium efflux induced by amplitude-modulated 147-MHz fields was due to the carrier frequency or the modulation per se, performed similar experiments with excised tissue from chicks (cerebral and muscle) and cats (cerebral only), but exposed the preparations to sinusoidal fields at discrete frequencies in the extremely-low-frequency (ELF) and sub-ELF ranges (1-75 Hz). Exposures were done between two parallel metal plates 1 square meter in area and spaced 50 cm apart in a chamber held at 36 deg C. Brain samples in 1 ml of physiologic medium were exposed for 20 min to 1, 6, 16, or 32 Hz at peak fields (in air) of 5, 10, 56, and 100 V/m. The muscle preparations were exposed for 20 min to a 16-Hz field at 20 V/m.

Cerebral and muscle samples from the chick were prepared and the 0.2-ml aliquots of bathing solutions were assayed for radioactivity following treatment as in the previous study. In addition, each brain sample was dissolved overnight in a digestive medium and assayed for radioactivity as a check. Cortical samples were dissected from the visual, auditory, somatosensory, and suprasylvian areas of anesthetized cats. Each sample was bisected and the pia mater removed before weighing. Half of each bisected sample for each field condition was exposed after incubation with Ca-45++, with the other half kept concurrently in a bath at 36 deg C as the control. Experiments were also done in which samples were sham-exposed, and the results were compared with those of controls.

About statistical treatment of the data, Bawin and Adey (1976b) stated: "Our pilot studies with radio frequency fields demonstrated that sample counts more than 40% above or below the mean of any set of 10 samples can be discarded as aberrations due to experimental errors in washing the tissue after incubation or in collecting the supernatant following the test period. In the present study, two statistical criteria applied to each set of data confirmed our previous observations. Extreme counts in any set more than 1.5 standard deviations away from the mean also satisfied the probability levels (0.1 to 0.01) in a statistical method based on the range of values [maximum ratio of extreme ranking observations (Dixon, 1951)]. Therefore, such extreme values were eliminated from the sets (fields as well as controls) before final analysis of the data. The radioactivities (cpm/g) of all samples (supernatant and digested tissues) were referred to the mean value of the counts obtained in control effluxes. This allows direct comparison between the amounts of Ca-45++ taken up by the tissues and subsequently released during the experimental conditions. All normalized data were statistically compared (t test) with matched samples of control values.

"The results are expressed in terms of the mean of all samples within a condition, plus or minus the standard error of the mean (m +/- sem). 340 neonate chicks and 39 adult cats were used in this study."

The mean chick-brain ratio m +/- SE for each of the four frequencies and field intensities (F) and for the controls (C) were presented in Table 1 of Bawin and Adey (1976b), shown adapted as Table 38. Also given were the numbers of samples (n) and the values of t. (These means and SEs were also displayed in Fig. 1 of the paper as bar graphs, but the SE bars were erroneously half the values given in Table 38.)

TABLE 38: MEAN CALCIUM-EFFLUX RATIOS FROM CHICK FOREBRAIN

Mod. m +/- SE (F) m +/- SE (C) n            t

                                                     5 V/m                                            

6Hz 0.932 +/- 0.036 1.000 +/- 0.038 30             1.450
16Hz 0.933 +/- 0.041 1.000 +/- 0.041 27             1.144
32Hz 0.948 +/- 0.038 1.000 +/- 0.041 27             0.974

                                                   10 V/m

1 Hz 0.943 +/- 0.041 1.0000 +/- 0.038 26            1.021
6 Hz 0.866 +/- 0.029 1.0000 +/- 0.037   26            3.069**
16 Hz 0.849 +/- 0.026 1.0000 +/- 0.031   38            3.726**
32 Hz 0.913 +/- 0.038 1.0000 +/- 0.037 27            1.633
                                                   56 V/m
1 1.028 +/- 0.042 1.0000 +/- 0.038    26          0.515          
6 0.882 +/- 0.032 1.0000 +/- 0.030   37          2.681*
16 0.889 +/- 0.035 1.0000 +/- 0.028   36          2.489*
32 0.942 +/- 0.031 1.0000 +/- 0.038

26         1.518

                                                  100 V/m
6 Hz 0.928 +/- 0.028 1.0000 +/- 0.029 36          1.735
16 Hz 0.995 +/- 0.037 1.0000 +/- 0.037 28          0.092

*p<0.05; **p<0.01

The results above indicate that the effect with ELF and sub-ELF fields was opposite to that with amplitude-modulated 147-MHz RFR (Bawin et al., 1975), i.e., mean ratios of less than unity (decreases rather than increases of calcium efflux) were obtained for all conditions except with 1 Hz at 10 V/m, with maximum effect with 6 and 16 Hz at 10 V/m. In addition, the data indicate the existence of a field-amplitude "window" (as well as a frequency window), i.e., with 6 and 16 Hz, the decreases were statistically significant at 10 V/m (p<0.01) and 56 V/m (p<0.05) but not at 5 or 100 V/m. The results for the chick-muscle preparations indicated no effect of field exposure, in consonance with the previous negative finding with amplitude-modulated 147-MHz RFR.

About the brain-tissue assays, the authors stated: "Tissue counts from exposed brain tissues were not statistically different from the control values. Each field condition was tested against the corresponding no field control to insure that the decrease seen in the Ca-45++ release was not due to an accidentally low tissue uptake. The mean uptake obtained for all brain samples (across all field conditions), expressed as a ratio with respect to the mean efflux of all controls, was 3.076 with a standard error of 0.104; the control values were 3.167 and 0.090. Thus the ratio of Ca-45++ uptake in the brain tissues versus the Ca-45++ released in the bathing fluid was three to one."

The results for the cat-cortex preparations are displayed in Table 39:

TABLE 39: MEAN CALCIUM-EFFLUX RATIOS FROM CAT CORTEX

 

Mod. m +/- SE (F) m +/- SE (C) n            t
                                                10 V/m
6 Hz 0.948 +/- 0.023 1.0000 +/- 0.032 23          1.296
16 Hz 0.982 +/- 0.037 1.0000 +/- 0.034 29          0.302
32 Hz 1.006 +/- 0.054 1.0000 +/- 0.036 16          0.108
                                                56 V/m
1 Hz 0.974 +/- 0.054 1.0000 +/- 0.036 23          0.386
6 Hz 0.855 +/- 0.034 1.0000 +/- 0.043 21          2.600*
16 Hz 0.874 +/- 0.025 1.0000 +/- 0.026   24          3.402**
32 Hz 0.909 +/- 0.034 1.0000 +/- 0.040 21          1.704
75 Hz 0.932 +/- 0.026 1.0000 +/- 0.033 22          1.600
6 Hz 1.000 +/- 0.025 1.0000 +/- 0.032 21          0.016
16 Hz 0.965 +/- 0.033 1.0000 +/- 0.025 29          0.830

 

*p<0.05; **p<0.01

 

As seen above, decreases in calcium efflux were obtained under almost all exposure conditions, but the only significant changes were at 56 V/m for 6 Hz (p<0.05) and 16 Hz (p<0.01), again indicating the existence of an amplitude window. The authors noted that no significant differences were found between different cortical regions, and that removal of the pia mater was essential for consistent effects in neocortical samples. They also found no differences in results for samples sham-exposed and samples placed in the heated bath, or between samples tested at 30 and 36 deg C.

The field amplitudes cited were those in air between the two exposure plates. Regarding the magnitudes of the induced internal field, the authors stated: "Tissue gradients were not directly measured, but in related studies, gradients of the order of 0.1 microvolt/cm were induced in a phantom monkey head by fields of similar geometry (Miller et al., 1974)." However, the description of the work on the phantom monkey head was obscure; also, it is difficult to understand the extrapolation from the results for the phantom monkey head to the brain preparations used in these experiments.

On the other hand, as with the study by Bawin et al. (1975), heating of the preparations by the ELF or sub-ELF fields was most unlikely to be involved. Instead, some nonlinear mechanism capable of rectifying the sinusoidal field with a time constant within those for the frequency range where the effect occurs may exist in the preparations. However, why decreases of calcium efflux were found for sub-ELF fields and increases were found for sub-ELF-modulated 147-MHz RFR is not readily explainable. The authors speculated on possible mechanisms for the basic phenomenon, including triggering of cooperative interactions at the cell membrane surface. In Bawin et al. (1978), a more complete discussion of possible mechanisms was presented. Also included therein were data on increases of calcium efflux from chick-brains exposed to 16-Hz-modulated 450-MHz RFR; these results indicated the existence of a power-density window in the range 0.1-1.0 mW/sq cm.

Regarding the statistical treatment of the numerical results, the validity of discarding extreme values a posteriori rather than because of foreknowledge that an experimental error may have occurred in any specific sample is questionable, and weakens the credibility of the results. It is noteworthy that since the opposite effect (decreases rather than increases in calcium efflux) was presumably not predicted before the results were obtained, use of the 2-tailed (rather than the 1-tailed) significance criterion was appropriate.

Sheppard et al. (1979) similarly prepared and exposed chick-brain halves to 450-MHz RFR sinusoidally modulated at 16 Hz only, this being the frequency for maximum effect in prior studies with 147-MHz RFR (Bawin et al., 1975). The exposures were for 20 min at 36 deg C and about 40% relative humidity to a TEM field within a tapered chamber fitted with RFR-absorbing material on the back and side walls. The incident average power densities were 0.05, 0.10, 1.0, 2.0, and 5.0 mW/sq cm. Following exposure, 0.2-ml aliquots of supernatant were assayed for Ca-45++ by liquid scintillation counting; the brains were dissolved overnight and also assayed for Ca-45++. Eight runs involving five each exposed and control half-brains per run were performed at each power density except 5 mW/sq cm, at which only six runs were done, for a total of 190 chick brains. Mean supernatant radioactivities of each exposure group were normalized to the mean values for the corresponding control groups, and the 2-tailed t-test was used. The normalized mean (m) and SE, number (N) of paired samples, and value of t for each RFR level are shown in Table 40, adapted from Table 1 of the paper, the asterisk indicating significance (p<0.05):

TABLE 40: CALCIUM EFFLUX FROM CHICK BRAINS

         RFR LEVEL            nW/sq cm

        EXPOSED          m +/- SE

        CONTROL          m +/- SE n         t
O.O5 0.937 +/- 0.025 1.0 +/- 0.030    40          1.615
0.10 1.080 +/- 0.026* 1.0 +/- 0.026   40          2.13
1.0 1.112 +/- 0.032* 1.0 +/- 0.026   40          2.68
2.0 1.035 +/- 0.030 1.0 +/- 0.023    40          0.339
5.0 1.022 +/- 0.030 1.0 +/- 0.034    30          0.484

 

The authors stated: "With attention to the important experimental steps (consistency in the pH and osmolarity of the physiological solution, and careful rinsing of the brains after incubation with the radioactive solution) it was possible to reduce the occurrence and magnitude of extreme values so that no data points were discarded. Significant results were found at the P<0.05 level, with statistical significance estimated by the two-tailed t-test."

Tissue counts for dissolved brains showed that the proportion of Ca-45++ remaining in the tissue was greater than in the supernatant; the ratio of the normalized counts in tissue to the counts per ml in supernatant was 3.5:1. The difference between mean tissue counts for exposed and control brains was not significant.

The authors proposed a model for the interaction of weak electric fields with neuronal membranes, which describes amplification of weak signals by the cooperative behavior of glycoprotein-bound ions in the membrane under the influence of field enhancement by polarization of surrounding unbound ions. They suggested that field-triggered cooperative behavior of the bound ions could change the calcium binding and give rise to the effects found and stated: "Adey has suggested that the sub-ELF envelope impressed on the RF carrier is detected at the polyanionic surface due to a strong asymmetry in charge distribution with respect to that surface. The charge asymmetry at the borders of glycoprotein-dense regions may allow demodulation much as in the case of a semiconductor diode but further details of the mechanism and its coupling to the cooperative system have not been developed."

Allis and Fromme (1979) pointed out that calcium is also transported across biological membranes by energy-dependent processes, suggesting specifically that amplitude-modulated RFR may directly stimulate the membrane to transport calcium (or stimulate calcium-sequestering sites within the cell to release or bind calcium), thus altering the quantity of calcium ions transferred across the membrane. With this hypothesis, they performed experiments in which specially prepared membrane-bound enzyme systems were exposed to almost 100% amplitude-modulated 2.45-GHz RFR in a spectrophotometric apparatus that was equipped with a slotted-waveguide applicator, thus permitting measurement of enzyme activity during exposure (Allis et al., 1975).

Two enzyme systems were selected for study: sodium-potassium adenosine triphosphatase (ATPase) from the erythrocyte's plasma membrane, which transports sodium and potassium ions across the membrane as adenosine triphosphate (ATP) is hydrolyzed (and may be responsible for maintaining the sodium-potassium balance in the cell); cytochrome oxidase, which is contained in the inner membrane of mitochondria (and is the final enzyme in the electron-transport chain of oxidative phosphorylization). Both preparations were derived from the rat.

Exposure was started after initiating the specific reaction for each enzyme system, and the reaction and exposure were continued for 15 min, during which the reaction rate was measured spectrophotometrically at the appropriate wavelengths. Three RFR-exposure and three control assays were made per day, with each set performed first on alternate days. From heating and cooling curves, the SAR was 26 W/kg (Allis et al., 1977). The sinusoidal modulation frequencies used were 16, 30, 90, and 120 Hz.

The membrane preparations for the ATPase system were found to vary in enzyme activity per gram of total protein, so membranes from a single preparation were used on each day (presumably for one frequency), but more than one preparation was necessary to complete several runs at each frequency. For this reason, the activities of the samples exposed to RFR each day were averaged and compared with the mean activity of that day's control samples. Paired t-tests of the results showed that the differences in enzyme activities were nonsignificant (p>0.1) at any modulation frequency. An analysis of variance was also performed on the mean values of each day's results. A nested design was used because the daily variation in results was nested within the effect of modulation frequency. The effects of treatment vs control, variation of modulation frequency, and frequency-treatment interaction were all nonsignificant.

The reaction rate of the cytochrome-oxidase system proved to be too high to permit direct measurements of its specific activity. Therefore, this assay was run under conditions for which the apparent first-order rate constant could be measured. Use of statistical techniques similar to those for the ATPase system yielded no significant differences between RFR-exposed and control samples.

However, as Allis and Fromme (1979) pointed out, the results of this study were not definitive, because: the membranes were tested for two functions under highly artificial conditions, e.g., the resting electrical potential across the membrane was not maintained in the in-vitro preparations used; the exposure levels were much higher than the power-density window found in the previously cited research; and of the four modulation frequencies used, only 16 Hz was within the frequency window found for nerve-tissue effects.

Blackman et al. (1979) performed experiments toward reproducing the 147-MHz chick-brain results of Bawin et al. (1975). After decapitating each chick, its forebrain was removed and divided at the midline to provide an exposure-control pair. After four brain-tissue pairs were prepared, each tissue sample was weighed. Each specimen was then immersed in 1 ml of physiologic medium labeled with Ca-45++ and agitated for 30 min at 37 deg C. The radioactive solution was then aspirated and the tissues were rinsed as follows: Two ml of medium was added to each tube and then was poured off with the tissues into small plastic sieves. Each sample was rinsed successively in two 250-ml volumes of nonradioactive medium, to render it free of any loosely associated Ca-45++, and was placed in a clear plastic tube containing 1 ml of medium for exposure.

Samples were exposed to 147-MHz RFR for 20 min in a rectangular TEM cell (Crawford, 1974; Weil, 1978) that was enclosed in a foamed polystyrene chamber maintained at 37 +/- 0.2 deg C. The sample-containing tubes in Lucite racks were arranged symmetrically on each side of the center conductor of the TEM cell while similar racks were placed on shelves within the chamber alongside the cell in a region where the field strength was more than 30 dB lower than in the cell. Two series of exposures were done. In one series, the power density was held constant at 0.75 mW/sq cm and the modulation frequencies were 0, 3, 9, 16, and 30 Hz. In the other series, the modulation frequency was held constant at 16 Hz and the power densities were 0 (sham), 0.5, 0.75, 1.0, 1.5, and 2.0 mW/sq cm.

Following RFR- or sham-exposure, 0.2-ml aliquots of bathing medium were assayed for radioactivity by liquid scintillation counting. The counts-per-min (CPM) value for each specimen was normalized to the weight of the specimen, and the difference, V, in normalized values between the exposed half-brain and its control was treated as the variable for statistical analysis. Preparation and exposure of four brain-tissue pairs were repeated nine times for each condition of frequency and power density, and a nested one-way analysis of variance was used for each series. For those values of the F statistic that were significant, a multiple-comparison procedure (sequential Newman-Keuls) was used to determine where significant differences existed between means.

The mean values of V for the constant-power-density series (0.75 mW/sq cm) and for sham-exposure were displayed in Fig. 2 of the paper as bar graphs with SEs. All mean values were positive (increases in calcium efflux); those for sham-exposure and 30 Hz were close to zero, which was well within their SEs; the values of (V-SE) for the other frequencies were all positive, but only the mean value of V for 16 Hz was labeled significant (p<0.05). The analysis of variance yielded a significant effect of frequency [F(5,48)=3.32, p<0.02]. The multiple-comparison procedure showed that V for 16 Hz was significantly higher than for sham-exposure or for 30 Hz, but that all the other values of V were nonsignificant (p>0.05), including the difference between the means for those exposed to the unmodulated RFR and those sham-exposed.

The mean values of V for the constant-modulation-frequency series (16 Hz) were similarly displayed in Fig. 3 of the paper. All mean values of V except for 0.75 mW/sq cm were close to zero. (Those for 0, 0.5, and 2.0 mW/sq cm were slightly positive and those for 1.0 and 1.5 mW/sq cm were slightly negative.) For 0.75 mW/sq cm, however, V was positive and labeled significant (p<0.01). Analysis of variance showed a significant power-density effect [F(5,48)=5.77, p<0.001]. The multiple-comparison procedure indicated that the mean V for 0.75 mW/sq cm was significantly higher (p<0.01) than the values for the other five power densities.

The authors stated: "Our results indicate that the modulation-frequency window in which calcium-ion flux is enhanced only occurs within a restricted range of power densities. Our initial attempts to reproduce the modulation-frequency window were hampered by the dramatic character of the power-density effect, of which we were initially unaware. These results indicate a maximal power-density effect at 0.75 mW/sq cm and no enhancement at levels plus or minus 0.25 mW/sq cm of this value. The narrow width of this window was not observed in the preliminary study in which intermediate values for enhanced efflux were found at 0.5 and 1.0 mW/sq cm."

They also noted: "In the experiments performed at 147 MHz, no difference could be detected in the measured values of incident and transmitted power [in the TEM cell]. This means that the rate of absorption of RF energy by the samples is low; i.e., less than the approximately 0.4 mW/g that can be resolved within the +/- 1% of full-scale uncertainty specified for our power meters. In measurements performed at a higher frequency of 500 MHz, where the ratio of sample size to wavelength is considerably greater than that existing at 147 MHz, an SAR of 1 mW/g was obtained in a 10 mW/sq cm field. Thus, at 500 MHz, an incident field of 0.75 mW/sq cm would be expected to produce an SAR of 0.075 mW/g, which doubtless represents an upper limit of the SAR in our experiments. This rate of energy deposition is too low to produce measurable heating by any readily available method (Allis et al., 1977; Blackman and Black, 1977; Kinn, 1977)."

The results of Blackman et al. (1979) appear to confirm the existence of the calcium-efflux phenomenon in excised chick brains, at least for 147-MHz RFR amplitude-modulated at 16 Hz. As noted by the authors, however, the power density window found is narrower than the windows found by Bawin and coworkers for modulated 450-MHz RFR (Sheppard et al., 1979) and sinusoidal sub-ELF fields (Bawin et al., 1976a,b).

Blackman et al. (1980a) extended this work with amplitude-modulated 147-MHz RFR, to determine whether the width of the power-density window depends on the number of brain samples exposed simultaneously in their system (a possible artifact) and whether the modulation frequency used affects the location or width of the power-density window. For this purpose, the experimental design was improved by obtaining, for each exposure condition, a companion set of results for sham-exposure under otherwise identical conditions.

Brain tissues were obtained from chicks 1 to 7 days old and processed as in the previous study. Lucite racks holding the tissue samples to be exposed were arranged symmetrically on each side of the center conductor of the TEM cell in the calculated uniform-field region and similar racks containing the control samples were placed on shelves outside the cell. Exposures were for 20 min to 147-MHz RFR sinusoidally modulated with 0, 9, or 16 Hz at power densities of 0.11, 0.55, 0.83, 1.1, 1.38, and 1.66 mW/sq cm. The authors noted that the 0.83-mW/sq-cm level represented a correction to the 0.75 mW/sq cm used in the previous study and was based on a more accurate calibration. Half of each chick brain was exposed to the RFR, and the other half, which was neither exposed nor sham-exposed, served as its control. Halves of other brain pairs were sham-exposed in the TEM cell, and the corresponding halves served as controls.

In one series of runs, four brain samples were treated concurrently per run in the TEM cell (with the corresponding controls outside the cell). In another series, six tubes containing equivalent amounts of unlabeled medium only (dummy loads) were placed in the Lucite racks together with four tubes containing brain samples (10 tubes total per run). In a third series, conducted to determine the influence of temperature only, Lucite racks holding four tubes containing brain halves were placed in water baths for 20 min at 32, 37, or 41 deg C while the complementary halves were held for 20 min at 37 deg C.

The relative quantity of Ca-45 ions released by each tissue pair was defined as the ratio, Vt/Vc, of the CPMs for the treated and control samples. Time and replication effects, sought by analysis of variance for a partially nested design, were not found, so analysis of variance for a one-way design was done for each combination.

Statistically significant differences in values of Vt/Vc were found between RFR- and sham-exposed tissues for eight combinations of power density, modulation frequency, and number of tubes/run (brains only or brains plus dummy loads). One combination that yielded a significant difference at the p<0.001 level was for 4 tubes/run with 16 Hz at 0.83 mW/sq cm; the corresponding differences for 4 tubes/run with 16 Hz at 0.55 and 1.11 mW/sq cm were nonsignificant (p>0.05), indicating the existence of a narrow power-density window. These results were similar to those found in the previous study. For 10 tubes/run with 16 Hz at 0.83 mW/sq cm, the difference in mean values for RFR- and sham-exposed tissues was significant (p<0.001), but so were the differences for 0.55, 1.11, and 1.38 mW/sq cm (0.01<p<0.05), implying that the power-density window was broadened by the presence of the dummy loads. This effect was provisionally ascribed by the authors to greater interaction among samples that distorted the fields in the vicinity of the brain samples.

The difference for 10 tubes/run with 9 Hz at 0.83 mW/sq cm was also significant (p<0.001), as were those for 10 tubes/run with 9 Hz at 0.55 and 1.11 mW/sq cm (0.01<p<0.05), thus indicating that the modulation-frequency window was relatively broad. (A series for 9 Hz with 4 tubes per run was not done.) The authors noted that although the differences in Vt/Vc between the RFR- and sham-exposed samples for the remaining combinations of power density, modulation frequency, and number of tubes/run were nonsignificant, the values of Vt/Vc for the RFR-exposed samples were all inexplicably larger than for their corresponding sham-exposed samples.

For the exposures to unmodulated RFR at 0.83 mW/sq cm (10 tubes/run), the differences in Vt/Vc between RFR- and sham-exposed samples were nonsignificant (p>0.05). For the tissues held at other temperatures than 37 deg C, the efflux was 9% lower at 32 deg C and 15% higher at 41 deg C than at 37 deg C. However, the authors excluded the possibility of RFR-induced temperature increases as a factor in the positive results above because the maximum SAR was undoubtedly less than 0.075 W/kg.

Blackman et al. (1980b) also conducted experiments with 16-Hz-modulated 50-MHz RFR in a TEM cell, to determine whether changes in carrier frequency altered the range of power densities effective in producing statistically significant increases in calcium efflux. The exposure conditions and protocol were similar to those previously used by these investigators. Ten tubes were treated concurrently per run (4 brain samples and 6 dummy loads), to broaden the power-density window as for 147 Mhz. The results are given in Table 41 (Table 2 of the paper):

TABLE 41: MEAN CALCIUM-EFFLUX RATIOS (Vt/Vc) VERSUS POWER DENSITY (PD)

         PD           (mW/sq cm) n     SHAM EXPOSURE   Vt/Vc       SE    RFR EXPOSURE    Vt/Vc       SE  P
0.37 32 1.063     0.044 1.053     0.036 0.863
0.72 47 1.148     0.047 1.089     0.032 0.301
1.44 32 1.014     0.035 1.152     0.044 0.013
1.67 64 1.026     0.023 1.182     0.034 <0.001
2.17 36 1.119     0.036 1.138     0.040 0.730
3.64 64 1.054     0.029 1.165     0.033 0.014
4.32 28 1.058     0.050 1.079     0.051 0.778

 

As seen above, the maximum value of Vt/Vc (1.182) occurred at 1.67 mW/sq cm and was significantly higher (p<0.001) than the corresponding Vt/Vc for sham-exposure (1.026). Significant increases in calcium efflux (about 16%) were also obtained at 1.44 and 3.64 mW/sq cm (p=0.013 and 0.014). Exposure at the intermediate level 2.17 mW/sq cm yielded 1.138, only marginally higher than 1.119 for sham-exposure, a nonsignificant difference (p=0.738). The value of Vt/Vc for RFR-exposure at 0.72 mW/sq cm, though greater than unity (1.089), was nonsignificantly lower (p=0.301) than for sham-exposure (1.148). For the remaining power densities, the differences, some of which were negative (decreases in calcium efflux), were also nonsignificant (p>0.05).

The authors stated: "Instead of a power-density window centered at 0.83 mW/sq cm [for 147-MHz], the results at 50 MHz indicated that there are at least two ranges of power densities that can induce enhanced calcium-ion efflux; one range spans 1.44 to 1.67 mW/sq cm and is bounded by 'no-enhancement' results at 0.72 and 2.17 mW/sq cm; the other range includes 3.64 mW/sq cm and is bounded by 2.17 and 4.32 mW/sq cm." This argument appears specious, however, in view of the two relatively large values of Vt/Vc for sham-exposure noted above (possibly indicative of the presence of uncontrolled non-RFR factors). If those values had been as close to unity as the other sham-exposure values, perhaps the increase in calcium efflux for exposure at 2.17 mW/sq cm also would have been significant, thus yielding a single relatively broad power-density window centered at 1.67 mW/sq cm.

Joines and Blackman (1980) modeled the chick-brain hemisphere bathed in buffer solution as a sphere within the field of uniform incident RFR, in an endeavor to account for the dependence of calcium-efflux increase on carrier frequency in the results for 16-Hz amplitude-modulated RFR at 450 MHz by Bawin et al. (1978) and Sheppard et al. (1979) and at 147 and 50 MHz by Blackman et al. (1980a,b). This theoretical analysis showed that the average electric-field intensity within such a spherical model can be made the same at the three carrier frequencies by adjusting the incident power density to compensate for the frequency dependence of the complex permittivity and internal wavelength of the sample.

Examination of the experimental data on this basis showed that all of the positive and negative results obtained at these three frequencies, when compared by average electric-field intensity within the sample, were in basic agreement and that no result, positive or negative, was contradicted by a corresponding experimental result at a different carrier frequency. However, the model did not take into account the amplitude-modulation frequencies per se. Because not all modulation frequencies are effective, comparisons among average electric-field intensities within the samples cannot be extended to other modulation frequencies.

Athey (1981) challenged the analysis above. He pointed out that there were uncertainties in the values of the electrical properties (complex permittivity) of brain material, that actual samples were predominantly saline and therefore different from the brain material assumed in the model, and that the simple homgeneous spherical geometry assumed may have been too unrealistic, all of which led to uncertainties that were too large to permit meaningful conclusions. He recommended that further work be based on experimental dosimetry to diminish such uncertainties.

In rebuttal, Joines and Blackman (1981) reported results on an improved model, that of a layered sphere. They performed calculations with the new model for various worst-case situations, which showed that the relationship of incident power density to internal field is relatively insensitive to small uncertainties in permittivity. A key aspect of the calcium-efflux effect, however, its dependence on modulation frequency (and its absence with unmodulated RFR), remained unanswered. Also among the observations not yet accounted for are why calcium efflux increases for modulated 147- and 450-MHz RFR and decreases for sinusoidal sub-ELF fields, and why the phenomenon only occurs within an amplitude or power-density window.

In two recent studies, Blackman et al. (1985a,b) reviewed prior work by Blackman and coworkers on the effects of fields at frequencies in the ELF range (defined by them for convenience as 1-300 Hz) on calcium efflux from chick brains in vitro. Among the findings were windows of frequency and field intensity within which calcium efflux was enhanced, and outside of which alterations of calcium efflux were nonsignificant. However, Blackman et al. (1985b) noted that their calcium-efflux changes (enhancements) were opposite in direction to the reductions found by Bawin and Adey (1976b) at frequencies in the same range.

Amajor difference between the two studies was the use, by Bawin and Adey (1976b), of AC electric fields only, whereas Blackman et al. (1985a) used AC fields having a magnetic component as well as an electric component. Therefore, Blackman et al. (1985b) hypothesized that the AC magnetic component could significantly influence changes in calcium efflux and that a DC magnetic field such as the local geomagnetic field (LGF) might also have a role.

The exposure apparatus used by Blackman et al. (1985b) consisted of a transmission line terminated with a 50-ohm load, a function generator to provide an AC electromagnetic field, and instrumentation (Bell Model SAB4-1808 probe, Bell Model 640 gaussmeter, and Hewlett-Packard Model 3582A spectrum analyzer) to measure the frequency and intensities of the field components. In consonance with the authors' usage, values cited below of AC electric field are peak-to-peak and those of AC magnetic field are rms. To obtain an electric field only, the 50-ohm load was removed. Specifically, with the load present and the field components adjusted to 40 V/m and 59.5 nanoteslas (nT), removal of the load reduced the magnetic component to a residual level of about 15 picoteslas (pT).

For exposure of samples to either type of AC field under altered LGF conditions, a pair of 18-inch-radius Helmholtz coils spaced 18 inches apart was oriented to produce a DC magnetic field parallel to the LGF (which was inclined at 85 deg), and the transmission line was placed within the coils. The gaussmeter was used to set the desired level of DC field within the transmission line in the absence of samples, and introduction of samples did not change the level. The authors assumed that sample magnetization was negligible, so the magnetic-field levels within and outside tissue were taken to be the same.

Sample preparation was similar to that used in previous studies. Also, in an endeavor to keep the procedures invariant, four tubes containing chick-brain-halves were treated concurrently with six tubes containing equivalent dummy loads (10 tubes total) as in Blackman et al. (1980b). Field exposures (including sham-exposures) were for 20 min at 37 deg C, with the corresponding half-brains held at the same temperature within a water bath. The following series of exposures were performed and the post-treatment assay for Ca-45++ in each field- or sham-exposed half-brain was normalized to the value for its control half-brain:

(1) Ten tubes were exposed to a 16-Hz electric field alone at 6, 10, or 40 V/m (peak-to-peak in air) and 10 other tubes were sham-exposed. The exposure at each level and the corresponding sham-exposure comprised a set. Exposure sets for 6 and 40 V/m were done eight times and sets for 10 V/m seven times. In addition, eight sets for 40 V/m were repeated eight months later.

(2) To compare effects of an AC electric field alone with those of an AC electromagnetic field, nine sets, each comprised of an exposure to a 16-Hz electric field at 40 V/m and an exposure to a 16-Hz electromagnetic field at 40 V/m, 59.5 nT, were performed (sham-exposures omitted).

(3) A similar design was used for determining the possible influence of the LGF (given as 38 microteslas, presumably at the experimental site). Eight sets were performed, each consisting of two exposures to a 15-Hz, 40-V/m, 59.5-nT field, one with the normal LGF present and the other with the LGF reduced to 19 microteslas by the Helmholtz coils. Also, the eight sets were repeated one week later.

(4) Sets of exposures to a 30-Hz, 40-V/m, 59.5-nT field, but for several integral and fractional multiples of the LGF, including negative values representing field reversal. In each set, comparison was made of the results for the altered LGF with those of the normal LGF, and some sets were replicated one week later.

The normalized data collected for each series above were subjected to a two-way analysis of variance for a replication main effect and for a replication-by-exposure-group interaction. Neither replication effect was found to be significant, so a one-way analysis of variance was used to compare the combined normalized results for the different conditions in each series.

The results for series (1) yielded nonsignificant differences (p>0.2) between field- and sham-exposed sets for any level or for the combined initial and repeated 40-V/m set. By contrast, Blackman and coworkers had found that 16-Hz electromagnetic fields at 6 V/m, 8.9 nT and at 40 V/m, 59.5 nT enhanced calcium efflux significantly (p<0.05), whereas Bawin and Adey (1976b) had reported significant reduction (p<0.05) for a 16-Hz electric field at 10 V/m.

Direct comparison of the results of series (2) for an AC electric field only vs an AC electromagnetic field showed that calcium efflux for the latter was significantly larger than for the former (p=0.011). The authors stated: "The data demonstrate that the AC magnetic component [59.5 nT] must be present in the 6- and 40-V/m fields to induce an enhanced efflux."

The results for series (3) and (4) are shown in Table 42, adapted from Table 3 of the paper:

TABLE 42: MEAN CALCIUM-EFFLUX RATIOS

FOR VARIOUS MULTIPLES OF THE NORMAL LGF

For 15 Hz:
Net LGF N Mean Efflux Ratio SE P
1.0 32 1.199 0.054 *****************
0.5 32 1.045 0.049 0.043*
1.0 32 1.236 0.046 *****************
0.5 32 0.997 0.042

<0.0001*

For 30 Hz:
Net LGF N Mean Efflux Ratio SE P
1.0 32 1.040 .038 *****************
2.0 32 1.309 0.078 0.003
1.0 32 1.048 0.037 *****************
1.33 32 1.058 0.049 0.870
1.0 40 1.020 0.038 *****************
0.67 40 1.248 0.054 <0.001*
1.0 32 0.981 0.042 *****************
0.67 28 1.208 0.033 <0.001*
1.0 32 1.041 0.032 *****************
-0.67 32 1.143 0.048 0.085
1.0 32 1.065 0.040 *****************
-0.67 32 1.323 0.068 0.002*
1.0 32 1.028 0.035 *****************
-2.0 32 1.300 0.089 0.006*
1.0 32 1.002 0.049 *****************
-2.185 32 1.076 0.036 0.226

 

----------------

*Significantly different at the indicated level from the mean for 1.0

 

The authors stated: "The results demonstrate that a normally effective 15-Hz signal is ineffective when the net density of the LGF is reduced to half the ambient, and an ineffective 30-Hz signal is effective when the net density is changed to .67x or 2.0x ambient but not when it is increased to 1.33x or to 2.185x ambient...increased efflux resulted at LGFs +0.67x ambient and of +/- 2.0x ambient when compared to the lack of enhancement for ambient conditions...A slight increase in the density of the LGF from -2.0x to -2.185x the ambient was sufficient to remove the enhancement conditions."

The results above with changes in the LGF appear to indicate that the DC geomagnetic field is fundamentally involved in alterations of calcium efflux, but as is evident in the quotation above, the relation between the magnitudes of the LGF changes and the presence or absence of effect, if any, is obscure. Also not clear is the absence of effect in series (1), in which the AC magnetic field was absent but the LGF was present, since the magnitude of the normal LGF (38 microteslas) is about 600-fold larger than the AC magnetic component (59.5 nT rms) used in all series except (1). It seems plausible that the latter would have been swamped by the former. Moreover, the negative findings for series (1) do not gainsay the positive findings of Bawin and Adey (1976b) in the absence of an AC magnetic component.

Dutta et al. (1984) exposed monolayer cultures of human neuroblastoma cells in T-flasks to 16-Hz-amplitude-modulated (80%) 915-MHz RFR in a TEM exposure chamber at SARs ranging from 0.01 to 5 W/kg, and studied the effects of such exposure on calcium efflux from the cells. The SARs were determined by measurements of forward, reflected, and transmitted powers with: the TEM cell empty, empty flasks in the cell, and samples in flasks within the cell (Dutta et al., 1982).

Human neuroblastoma cells were grown to confluent monolayers within T-flasks (25 sq cm of growth surface), each containing 5 ml of minimum essential medium (MEM) supplemented with other necessary elements. The monolayer in each flask was then incubated in 5 ml of MEM containing 2 microliters of Ca-45++ (specific activity 5 Ci/l) for 1 hr to establish steady-state calcium-ion equilibrium within the cells. This incubation period was selected on the basis of preliminary experiments showing (in Fig. 2 of the paper) that Ca-45++ uptake rises linearly during the first hour to a plateau, with no difference in uptake between the first and second hours.) The labeled cells were washed with nonradioactive medium three times without disturbing the monolayer, to remove excess Ca-45++. After washing, 6 ml of MEM was added to the culture and mixed, and 1 ml was removed for determining the initial radioactivity (400-600 cpm/ml).

One flask was placed on each side of the center plane of the TEM cell within an incubator at 37 deg C and exposed for 30 min at 0.01, 0.05, 0.075, 0.1, 0.5, 0.75, 1.0, 1.5, 2, or 5 W/kg. Two control flasks for each exposure were kept concurrently at 37 deg C within the incubator, but outside the TEM cell. Sham-exposures were also conducted, but no significant differences in calcium efflux were found between control and sham-exposed cultures. At exposure end, a 1-ml sample of the medium was withdrawn from each flask. For calcium-efflux assay, the 1-ml pre- and post-exposure samples were centrifuged at 500g for 1 min to remove any cells or large debris, and 0.5 ml of each supernatant was added to 4 ml of scintillation cocktail and counted. A series of exposures was also conducted at 0.05 W/kg with modulation frequencies between 3 and 30 Hz.

RFR-induced changes of calcium influx into cells were also investigated. Monolayers were grown in medium that contained Ca-45++ instead of normal medium, and immediately exposed for 30 min to 16-Hz-modulated RFR at 0.05 or 0.1 W/kg, with similarly grown controls outside the TEM cell. After such treatment, the cells were washed three times as before and suspended in 0.5 ml of 0.25 M ethylenediaminetetracetic acid (EDTA) to detach the cells from the flask without affecting cell integrity. The cell suspensions were then assayed for radioactivity.

The results for the series of exposures at 0.05 W/kg with various modulation frequencies were graphed (in Fig. 3 of the paper) as the percentage increases of calcium efflux vs modulation frequency. This graph showed a sharp maximum of about 50% at 16 Hz, compared with about 35% at 14 Hz and 28% at 18 Hz, with much lower percentages above and below these frequencies.

The calcium-efflux results with 16-Hz modulated RFR are shown in Table 43, adapted from Table 1 of the paper. The differences between treated and control samples were assessed for significance by the two-tailed Student t-test.

TABLE 43: EFFECTS OF 16-HZ MODULATION ON CALCIUM EFFLUX

SAR (W/kg) N(a) CONTROL (cpm/ml)  Mean Efflux +/- SE  Exposed (cpm/ml)  Mean Efflux +/- SE % Increase
0.00(b) 6 691 +/- 25 689 +/- 20 -0.3
0.01 4 703 +/- 11 704 +/- 24 0.1
0.05 6 694 +/- 33 1074 +/- 76** 54.7
0.075 6 627 +/- 14 695 +/- 43 10.8
0.10 4 596 +/- 23 650 +/- 37 9.1
0.50 4 653 +/- 22 619 +/- 14 -5.2
0.75 4 653 +/- 21 729 +/- 16* 11.6
1.0 8 684 +/- 16 866 +/- 28** 16.6
1.5 4 606 +/- 23 621 +/- 26 2.5
2.0 6 630 +/- 33 733 +/- 36 16.3
5.0 6 665 +/- 36 641 +/- 14 -3.6

(a) N = number of flasks; (b) sham-exposure; *p<0.05; **p<0.001

 

As seen above, significant increases in calcium efflux were obtained at 0.05 and 1.0 W/kg (both p<0.001) and at 0.75 W/kg (p<0.05) but not for intermediate SARs or those outside the range 0.05-1.0 W/kg. For comparison, results for exposure to unmodulated (CW) 915-MHz RFR at 0.05 and 1.0 W/kg are shown in Table 44 (Table 2 of the paper):

TABLE 44: EFFECTS OF UNMODULATED (CW) RFR ON CALCIUM EFFLUX

    SAR     (W/kg) N   Control (cpm/ml)   Mean Efflux +/- SE Exposed (cpm/ml) Mean Efflux +/- SE % Increase
0.05 6 633 +/- 16 655 +/- 19 3.5
1.0 4 695 +/- 20 833 +/- 38

    19.8      (p<0.01)

 

Regarding the differences among the control values above, the authors stated: "The variations in calcium ion efflux in the control cultures shown in Tables 1 and 2 [43 and 44] could be due to the minor variation in cell numbers and uptake of Ca-45++ by the cells. However, since the exposed and control cultures used for experiments at any given SAR come from the same parental stock, there should not be a consistent bias."

The percentage reduction of Ca-45++ uptake by cells exposed to the 16-Hz-modulated RFR at 0.05 and 0.1 W/kg are given in Table 45 (Table 3 of the paper):

TABLE 45: EFFECTS OF 16-HZ MODULATED RFR ON CALCIUM UPTAKE

     SAR          (W/kg) N Control (cpm/ml)   Mean Efflux +/- SE Exposed (cpm/ml) Mean Efflux +/- SE % Reduction
0.05 4 5768 +/- 048 5382 +/- 027 7.0 p<0.01)
0.10 4 4666 +/- 139 4875 +/- 172 ---

 

In summary, the authors stated: "Human neuroblastoma cells in culture show similar patterns of calcium ion efflux at 16 Hz AM as has been found in freshly isolated brain tissues (Bawin et al, 1975). Studies reported here indicate the presence of a narrow range of absorbed powers at which a significant increase in the efflux of calcium ions occurs from human neuroblastoma cells in culture. Similar intensity regions for radiofrequency fields were first noted in freshly removed chicken brain by Blackman and his colleagues...In the present studies, although we have actually measured SARs of samples (medium + cells), the energy absorbed by the neuroblastoma cells alone is not known. Thus it is difficult to compare our results with the effective radiofrequency power densities obtained for chicken brains."

They also noted: "The present study reveals that an unmodulated [915-MHz] carrier wave can induce calcium ion efflux at an SAR of 1 mW/g." (This finding is contrary to those of the chick-brain studies).

Lastly, they stated: "The uptake studies suggest that specific microwave intensities known to cause enhanced net efflux of calcium ions from prelabeled cells can also reduce the net uptake of calcium ions when the cells are exposed during the labeling period. Because the exposure time during the uptake studies [0.5 hr] was not long enough for equilibrium to be reached [at least 1 hr], there can be no comparison between the uptake and efflux results."

Shelton and Merritt (1981) investigated whether RFR pulses at repetition rates comparable to the modulation frequencies used in the chick-brain studies would elicit alterations in calcium efflux from the rat brain. After rats were euthanized by cervical dislocation, their brains were excised rapidly, the olfactory bulbs were removed, a vertical cut was made to extract about one-third of the cerebrum, and the pia mater was removed. In each experiment, three pairs of cerebral-hemisphere samples were processed together for RFR- and sham-exposure. Each tissue sample (about 100 mg) was immersed in 4 ml of medium containing 2 microcuries of Ca-45++ within a 50-ml glass beaker, in which it was incubated for 20 min at 37 deg C. The bicarbonate concentration in the medium was 26 mM (that of mammalian cerebrospinal fluid instead of 2.4 mM in the chick-brain studies).

Following incubation, samples were washed once with 1-ml portions of medium that was free of Ca-45++, and were transferred to clean 50-ml beakers containing 2-ml aliquots of the radioactivity-free medium for RFR- or sham-exposure. Because the Ca-45++ concentration could be affected by the extent of tissue washing, the samples in one experiment were washed with 1-ml portions of radioactivity-free medium five times instead of once.

Six beakers containing cerebral samples thus treated were placed in two parallel rows of three each below, and in the far field of, a standard-gain horn within an anechoic chamber. The exposures in four experiments were for 20 min to 20-ms pulses of 1-GHz RFR, 16 pps (0.32 duty factor), a repetition rate analogous to the 16-Hz amplitude modulation used in the chick-brain experiments; the average power densities were 0.5, 1.0, 2.0, or 15 mW/sq cm, selected to search for the reported power-density window. In two other experiments, samples were exposed for 20 min to 10-ms pulses, 32 pps pulses (same duty factor), at 1.0 or 2.0 mW/sq cm. Sham-exposures were conducted in similar fashion for each experiment. Four groups each of RFR-exposed and sham-exposed samples comprised the population for each experiment.

In a seventh experiment, samples washed five times instead of once were exposed to 20-ms pulses, 16 pps, at 1.0 mW/sq cm (or sham-exposed) as in the second experiment. In still another experiment (with samples washed only once), the exposure parameters used were the same as in the second experiment again, but the exposure was interrupted for about 1 min each after 4, 8, and 12 min of exposure, and 0.5-ml aliquots of incubation medium were taken and assayed for possible time-dependent effects and replaced with fresh medium.

On completion of exposure, a 0.5-ml aliquot of each incubation medium was transferred to a counting vial. The tissue samples were washed once with 1-ml portions of medium free of Ca-45++ and also placed in counting vials, to which 2 ml of Soluene was added for overnight digestion at 37 deg C. Dimilume was added to all vials, and the radioactivities of the media and tissue samples were assayed by liquid scintillation counting.

The authors defined "efflux value" as the ratio (in percent) of the CPM of the medium to the sum of the CPMs in tissue and medium, a measure of efflux that differed from that used by other investigators, and the mean values (and SDs) were compared by Student's t-test. The results for cerebral tissue from 143 rats used in the first six experiments showed statistically nonsignificant (p>0.05) differences between any of the mean Ca-45++ effluxes when RFR-exposed samples were compared with their corresponding sham-exposed control samples. The mean efflux values for RFR- and sham-exposure in the seventh experiment (24 rats, five washings instead of one) also did not differ significantly from one another, but both means were higher than those obtained in the first six experiments. The results of the time-dependence experiment (presumably 24 rats also) showed successive rises in efflux at the three epochs, but again the differences in means for RFR- and sham-exposed samples at corresponding times were nonsignificant.

Thus, the findings of Shelton and Merritt (1981) were negative, but, as pointed out by the authors, no direct comparisons can be made between their results and those of the investigators on chick brains discussed above because in addition to the difference in species, the samples were actually exposed (with a duty factor of 0.32) for only a third of the 20-min period (at correspondingly higher peak levels), and because the spectral distribution of energy for the 16-pps repetition rate differed markedly from that in the sidebands for 16-Hz amplitude-modulated RFR.

Merritt et al. (1982) performed experiments in vivo as well as in vitro on possible pulsed-RFR-induced alterations of calcium efflux from the rat brain. For both types of experiment, brain tissue was loaded with Ca-45++ by injection directly into the right lateral ventricle of ether-anesthetized male Sprague-Dawley rats (175-225 g).

In the in-vitro experiments, cervical dislocation was used to euthanize the rats after intraventricular injection of Ca-45++. Six brain-tissue samples excised as in the previous study were exposed concurrently for 20 min to 20-ms pulses, 16 pps, of 1-GHz RFR at 1 or 10 mW/sq cm (SAR 0.29 or 2.9 W/kg) or of 2.45-GHz RFR at 1 mW/sq cm (SAR 0.3 W/kg), with a like number of samples sham-exposed as controls. The incident power densities at the sample sites were measured with a Narda Microwave Corp. Model 8316B RF monitor and 8323 probe, but the method for measuring the SARs of the samples was not described.

Although the radioactivities of the incubation media and digested tissue samples were assayed by liquid scintillation counting, results were not given in terms of the efflux values as defined in Shelton and Merritt (1981); instead, the mean disintegrations per min per gram of tissue (DPM/g) and SDs were presented. By 2-tailed t-test, the difference in mean DPM/g values between the RFR- and sham-exposed samples was not statistically significant for any of the three exposure conditions.

For the whole-animal exposures, two hours after the rats were injected intraventricularly with Ca-45++, each rat was gently squeezed between two Styrofoam sides of a Plexiglass holder to maintain its body axis constant during exposure, and groups of 12 rats each were exposed for 20 min to 2.06-GHz RFR with long axes parallel to the E-field. One group each was exposed to CW at 0.5, 1.0, 5.0, or 10.0 mW/sq cm and one group each to 10-ms pulses at 8, 16, or 32 pps and average power density of 0.5, 1.0, 5.0, or 10.0 mW/sq cm (16 groups total). For controls, a 17th group was sham-exposed. Each rat was assigned randomly to an exposure condition, and exposure conditions were presented randomly to eliminate time-of-day effects. The normalized SAR, measured in muscle-equivalent rat models by calorimetry (Allen and Hurt, 1979), was 0.24 W/kg per mW/sq cm or 0.12, 0.24, 1.2, and 2.4 W/kg for the four power densities.

After exposure, the rats were euthanized by cervical dislocation and the brains were quickly removed, rinsed in physiologic medium, blotted dry, weighed, and solubilized in Soluene. An aliquot of solubilized tissue of each rat was transferred to a counting vial containing Dimilume and assayed for 45-Ca++ by liquid scintillation counting. Statistical tests performed on the 17 treatment combinations (4x4 RFR-exposures, 1 sham-exposure) showed that difference between the sham group and the combined RFR groups and the differences between the sham group and the individual RFR groups were statistically nonsignificant.

Acontrol series of experiments was conducted on 12 rats to obtain baseline Ca-45++ levels. These rats were treated in the same manner as the sham-exposed rats, i.e., they were injected intraventricularly with Ca-45++ and allowed to stabilize for 2 hr. However, their brains were assayed immediately after the 2-hr period to ascertain the amount of Ca-45++ at the beginning of the 20-min period used for the sham- and RFR-exposed rats. This experiment yielded a mean concentration of 0.944 +/- 0.033 (SD) ng-atoms/g of brain tissue, whereas for the 12 sham-exposed rats, the mean value at the end of sham-exposure was 0.659 +/- 0.156 ng-atoms/g, indicating that there was a mean net efflux of 0.285 ng-atoms/g from the brain during the 20-min period.

This paper, like the previous one, described an attempt to determine whether changes in calcium efflux reported to be induced in chick brains by in-vitro exposure to amplitude-modulated RFR might also be seen in rats from in-vitro or in-vivo exposure to pulse-modulated waveforms. No RFR-induced calcium-efflux changes were found. However, in addition to the previously noted differences in species, carrier frequency, and waveform, there were others:

1) Loading of the brain with Ca-45++ was done by injection into the right ventricle of the brain of the intact animal, whereas external bathing media containing Ca-45++ were used in the other studies. The dynamics of regional uptake of Ca-45++ by this intraventricular method were not described, but measurements of calcium diffusion in the other studies indicated that their reported RFR-induced changes were probably obtained from the outer mm or so of cortical tissue. Not reported was whether the intraventricular technique loaded cortical tissue; indeed, it is not clear whether this technique may have resulted in passage of Ca-45++ from the cerebrospinal fluid directly into the bloodstream in addition to the assumed perfusion from cerebrospinal fluid into brain tissue. No measurements were given of Ca-45++ levels in the blood.

2) Also not stated was whether only the right cerebral hemisphere or both hemispheres were used to provide tissue samples for the in-vitro exposures. Injection only into the right hemisphere might have provided asymmetrical loading. For the in-vivo exposures, apparently the whole brain was solubilized. It is unclear what the effect of any Ca-45++ still remaining in the ventricular fluid might have had on the liquid-scintillation-counting results. Additional characterization of the intraventricular technique, originally developed for injecting tritiated norepinephrine into the lateral ventricle of the rat, was required for the present application.

3) As in the other in-vitro studies, great variability was found in the liquid-scintillation-counting results. In those other studies, the technique of normalizing the results of an RFR-exposed half-brain to the results of a matched, sham-exposed half-brain was used to reduce the overall variability. The Ca-45++ brain-loading technique used in the present study did not permit such normalization.

4) In other studies (both in vivo and in vitro), alterations in calcium efflux have been detected as alterations in the concentration of Ca-45++ in the incubating medium, but not in solubilized brain tissue. In the in-vitro part of the present study, measurements were made on samples of bathing medium as well as for solubilized brain case, but only results for the latter were given. However, the control series of experiments in vivo did show that the Ca-45++ concentration in whole brain decreased by about 30% during the 20-min exposure period. Under the experimental conditions described, such a decrease could have occurred by uptake into the bloodstream and distribution to the rest of the body. The presence of Ca-45++ in the blood could have yielded erroneous readings in whole-brain liquid-scintillation counts, a possibility that could be studied by perfusion of the animals with saline following euthanasia.

Adey et al. (1982) presented results of a study of Ca-45++ efflux in vivo from the cortex of the paralyzed but awake cat. A 12-mm-diameter trephine hole centered over the right suprasylvian cortex was drilled under ether anesthesia. The dura mater was carefully removed and a plastic cylindrical well was fitted in the aperture so as to make gentle contact with the surface of the pia. Nonradioactive physiologic medium was added to the well and all skin incisions and pressure points were infiltrated with local anesthetic. Use of ether was then discontinued, the cat was paralyzed with gallamine triethiodide IV, and artificial respiration was maintained with a tracheotomy. The end-tidal level of CO2, a measure of the CO2 level in the blood, was monitored and normally maintained at 4%, but was changed during episodes of hypoventilation and hyperventilation imposed to alter the state of arousal.

During recovery from ether anesthesia, the fluid level in the well was replaced at 10-min intervals for 30 min, to ensure that the fluid was clear. Incubation was then begun with 1.0 ml of medium that contained 20 microcuries of Ca-45++, and was continued for 90 min, at the end of which the fluid was replaced with nonradioactive medium. At 10-min intervals throughout the remainder of the experiment, the fluid was exchanged completely with fresh medium by pipetting, and 0.2-ml aliquots of each solution removed were diluted with Dimilume and assayed for Ca-45++. All samples were counted twice, once during the experiment and again to a stated accuracy of 1% within 24 hr.

Twenty-three female cats (2.8-3.6 kg) were used. Starting at different times following completion of incubation of the cortex with Ca-45++, the cats were individually sham-exposed or exposed for 60 min to 450-MHz RFR 85% amplitude-modulated at 16 Hz, and intervals ranging from 80 to 120 min were tested for possible differences in efflux patterns. Exposures were done with a horn about seven wavelengths long within an anechoic chamber maintained at 28 deg C and 30-40% relative humidity. Each cat was placed in a plastic stereotaxic headholder, with body axis at right angles to the incident field and with the right cerebral cortex nearest the RFR source. The power density was 3.0 mW/sq cm, for which the electric field in the interhemispheric fissure was found by Adey et al. (1981) to be 33 V/m, which corresponded to an SAR of 0.29 W/kg.

Description of the data-analysis methods used was rather obscure. The authors indicated that the relative Ca-45++ efflux data taken at 10-min intervals in the absence of RFR could be fitted by a time-dependence equation involving the sum of two exponential terms (but not clear was how the actual radioactivity data were normalized to obtain relative Ca-45++ effluxes). Accordingly, they log-transformed such data and used linear regression to obtain an idealized curve of relative efflux vs time. Though not displayed, this idealized curve was said to consist of two linear branches, one of large negative slope indicating that the rate of decrease in relative efflux during approximately the first hour of successive 10-min measurements is rapid, and the other of smaller negative slope showing that the rate of relative efflux decrease is less rapid during the remainder of the measurement period.

The authors then quantified the results for RFR-exposed cats in terms of the means of the relative squared deviations of experimental data from the idealized curve at the sampling points, and similarly with the results for sham-exposed cats. By this analysis, the total variance for the preexposure periods was found to be the same for data from the RFR- and sham-exposed animals, but the total variance in the data from the RFR-exposed cats for the exposure- and post-exposure period was higher than for the sham-exposed cats. A 1-tailed binomial probability test was used to show that this exposure- and post-exposure increase in variance was significant (p<0.05), i.e., that the results for the RFR- and sham-exposed cats were from statistically different populations. Experiments with imposed hypoventilation and hyperventilation showed that the effect was not a consequence of alteration of end-tidal CO2 levels, i.e., not secondary to raised cerebral CO2 levels.

Next, they paired experimental curves of relative Ca-45++ efflux vs time from RFR- and sham-exposed cats. (The pairing criteria used were rather involved, but were designed to ensure the best matching of the available data.) The data for a representative pair of cats, for which RFR- and sham-exposure were initiated 90 min into the measurement period, were exhibited in Fig. 1 of the paper. By eye, the average slopes of the two curves were approximately the same up to about the first 60 min of the measurement period (corresponding to the period of rapid decrease noted above). Beyond this interval (during and after RFR- or sham-exposure), both slopes were smaller (corresponding to the lower rate of decrease above), but the average slope for the RFR-exposed cat was less negative than for the sham-exposed cat and exhibited cyclic variations ("waves of increased Ca-45++ efflux") having a periodicity of about 25 min.

No experimental data for RFR- and sham-exposed animals were presented (other than mention of their use in the curve-fitting method as noted above) or were directly compared for significant differences, and as noted previously, no information on how the radioactivity data were normalized was given. Absence of such data seemed to imply that the important difference between the calcium efflux curves for the RFR- and sham-exposed cats was in the oscillations about the idealized straight-line fit to the data, and not between values themselves at corresponding measurement times or their means. Only amplitude modulation at 16 Hz was used in this study. Whether CW or other modulation frequencies would also be effective in causing the Ca-45++ efflux oscillations noted above is unknown.

In overall summary, there is a large volume of work by Bawin, Adey, and coworkers, and by Blackman and coworkers on changes of calcium efflux from chick brain resulting from in-vitro exposure to various regimens of electromagnetic fields at specific frequencies in the sub-ELF range and at 50-, 147- and 450-MHz RFR carriers sinusoidally amplitude-modulated at specific sub-ELF frequencies. Only a few studies involving in-vivo exposure were performed, notably those of Adey et al. (1982) with cats and of Merritt et al. (1982) with rats. The findings of some of the studies are corroborative, those of others appear to be contradictory, and those of still others negative (no effect). Noteworthy was the study by Dutta et al. (1984), which yielded a significant increase of calcium efflux from human neuroblastoma cells from in-vitro exposure to unmodulated 915-MHz RFR (as well as for 16-Hz modulated RFR at specific SARs). Thus, the occurrence of RFR-induced calcium efflux changes may depend on a complex function of incident power density and modulation frequency.

Calcium is known to play a major role in the regulation of not only secretion of neurotransmitters by nerve cells, but also of proteins by endocrine and exocrine glandular cells (Schramm, 1967). Myers and Ross (1981) performed a comprehensive and detailed critical analysis of the studies (to that date) on the RFR-induced calcium-efflux phenomenon, in which they raised some doubts about certain aspects of the methods used, statistical treatments of the data, and interpretations of the results. Disageement with their conclusions has been expressed by some of the researchers who studied the effect experimentally, with little if any likelihood that the controversy will be resolved in the near future.

In spite of the differences in findings among the various investigators, there appears to be substantial common agreement that the calcium-efflux phenomenon may be worthy of further study to further elicit the basic mechanisms of interaction of electromagnetic fields with biological entities at the microscopic level, but that the effect is not of concern with regard to possible hazards to humans from in-vivo exposure to RFR.

REFERENCES:

Adey, W.R., S.M. Bawin, B.F. Burge, H.I. Bassen, and K.E. Franke,
ELECTRIC FIELDS IN CAT BRAIN EXPOSED TO 450 MHZ CW FIELDS IN SEMI-FAR FIELDS
In Abstracts of Bioelectromagnetics Symposium, Washington, D. C., p. 35 (August 1981)

Adey, W.R., S.M. Bawin, and A.F. Lawrence
EFFECTS OF WEAK AMPLITUDE-MODULATED MICROWAVE FIELDS ON CALCIUM EFFLUX FROM AWAKE CAT CEREBRAL CORTEX
Bioelectromagnetics, Vol. 3, No. 3, pp. 295-307 (1982)

Allen, S. and W. Hurt
CALORIMETRIC MEASUREMENTS OF MICROWAVE ENERGY ABSORPTION BY MICE AFTER SIMULTANEOUS EXPOSURE OF 18 ANIMALS
Radio Sci., Vol. 14, No. 6S, pp. 1-4 (1979)

Allis, J.W., C.M. Weil, and D.E. Janes, Jr.
A CROSSED-BEAM APPARATUS FOR SIMULTANEOUS SPECTROPHOTOMETRIC OBSERVATION AND MICROWAVE EXPOSURE OF BIOCHEMICAL SAMPLES
Rev. Sci. Instrum., Vol. 46, pp. 1344-1349 (1975)

Allis, J.W., C.F. Blackman, M.L. Fromme, and S.G. Benane
MEASUREMENT OF MICROWAVE RADIATION ABSORBED BY BIOLOGICAL SYSTEMS: I. ANALYSIS OF HEATING AND COOLING DATA
Radio Sci., Vol 12, No. 6S, pp. 1-8 (1977)

Allis, J.W. and M.L. Fromme
ACTIVITY OF MEMBRANE-BOUND ENZYMES EXPOSED TO SINUSOIDALLY MODULATED 2450-MHZ MICROWAVE RADIATION
Radio Sci., Vol. 14, No. 6S, pp. 85-91 (1979)

Athey, T.W.
COMPARISON OF RF-INDUCED CALCIUM EFFLUX FROM CHICK BRAIN TISSUE AT DIFFERENT FREQUENCIES: DO THE SCALED POWER DENSITY WINDOWS ALIGN?
Bioelectromagnetics, Vol. 2, No. 4, pp. 407-409 (1981)

Bawin, S.M., L.K. Kaczmarek, and W.R. Adey
EFFECTS OF MODULATED VHF FIELDS ON THE CENTRAL NERVOUS SYSTEM
Ann. N.Y. Acad. Sci., Vol. 247, pp. 74-81 (1975)

Bawin, S.M. and W.R. Adey
INTERACTIONS BETWEEN NERVOUS TISSUES AND WEAK ENVIRONMENTAL ELECTRIC FIELDS
In C.C. Johnson and M. Shore (eds.), BIOLOGICAL EFFECTS OF ELECTROMAGNETIC WAVES, U.S. Dept. of Health, Education, and Welfare, Washington, D.C., HEW Publication (FDA) 77-8010, pp. 323-330 (1976a)

Bawin, S.M. and W.R. Adey
SENSITIVITY OF CALCIUM BINDING IN CEREBRAL TISSUE TO WEAK ENVIRONMENTAL ELECTRIC FIELDS OSCILLATING AT LOW FREQUENCIES
Proc. Nat. Acad. Sci., Vol. 73, No. 6, pp. 1999-2003 (1976b)

Bawin, S.M., A. Sheppard, and W.R. Adey
POSSIBLE MECHANISMS OF WEAK ELECTROMAGNETIC FIELD COUPLING IN BRAIN TISSUE
Biochemistry and Bioenergetics, Vol. 5, pp. 67-76 (1978)

Blackman, C.F. and J.A. Black
MEASUREMENT OF MICROWAVE RADIATION ABSORBED BY BIOLOGICAL SYSTEMS: II. ANALYSIS BY DEWAR-FLASK CALORIMETRY
Radio Sci., Vol 12, No. 6S, pp. 9-14 (1977)

Blackman, C.F., J.A. Elder, C.M. Weil, S.G. Benane, D.C. Eichinger, and D.E. House
INDUCTION OF CALCIUM-ION EFFLUX FROM BRAIN TISSUE BY RADIO-FREQUENCY RADIATION: EFFECTS OF MODULATION FREQUENCY AND FIELD STRENGTH
Radio Sci., Vol. 14, No. 6S, pp. 93-98 (1979)

Blackman, C.F., S.G. Benane, J.A. Elder, D.E. House, J.A. Lampe, and J.M. Faulk
INDUCTION OF CALCIUM-ION EFFLUX FROM BRAIN TISSUE BY RADIO-FREQUENCY RADIATION: EFFECT OF SAMPLE NUMBER AND MODULATION FREQUENCY ON THE POWER-DENSITY WINDOW
Bioelectromagnetics, Vol. 1, No. 1, pp. 35-43 (1980a)

Blackman, C.F., S.G. Benane, W.T. Joines, M.A. Hollis, and D.E. House
CALCIUM-ION EFFLUX FROM BRAIN TISSUE: POWER DENSITY VERSUS INTERNAL FIELD-INTENSITY DEPENDENCIES AT 50-MHZ RF RADIATION
Bioelectromagnetics, Vol. 1, No. 3, pp. 277-283 (1980b)

Blackman, C.F., S.G. Benane, D.E. House, and W.T. Joines
EFFECTS OF ELF (1-120 HZ) AND MODULATED (50 HZ) RF FIELDS ON THE EFFLUX OF CALCIUM IONS FROM BRAIN TISSUE
Bioelectromagnetics, Vol. 6, No. 1, pp. 1-11 (1985a)

Blackman, C.F., S.G. Benane, J.R. Rabinowitz, D.E. House, and W.T. Joines
A ROLE FOR THE MAGNETIC FIELD IN THE RADIATION-INDUCED EFFLUX OF CALCIUM IONS FROM BRAIN TISSUE IN VITRO
Bioelectromagnetics, Vol. 6, No. 4, pp. 327-337 (1985b)

Crawford, M.L.
GENERATION OF STANDARD EM FIELDS USING TEM TRANSMISSION CELLS
IEEE Trans. Electromagn. Compat., Vol. 16, No. 4, pp. 189-195 (1974)

Dixon, W.J.
RATIOS INVOLVING EXTREME VALUES
Ann. Math. Stat., Vol. 22, pp. 68-78 (1951)

Dutta, S.K., J. Choppalla, M.A. Hossain, T.N. Bhar, and H.S. Ho
DOSIMETRIC MEASUREMENT AND BIOEFFECT STUDIES OF LOW LEVEL 915 MHZ CW MICROWAVES USING TEM CRAWFORD CELL
J. Basic Appl. Sci., Vol. 2, pp. 43-52 (1982)

Dutta, S.K., A. Subramoniam, B. Ghosh, and R. Parshad
MICROWAVE RADIATION-INDUCED CALCIUM ION EFFLUX FROM HUMAN NEUROBLASTOMA CELLS IN CULTURE
Bioelectromagnetics, Vol. 5, No.1, pp. 71-78 (1984)

Joines, W.T. and C.F. Blackman
POWER DENSITY, FIELD INTENSITY, AND CARRIER FREQUENCY DETERMINANTS OF RF-ENERGY-INDUCED CALCIUM-ION EFFLUX FROM BRAIN TISSUE
Bioelectromagnetics, Vol. 1, No. 3, pp. 271-275 (1980)

Joines, W.T. and C.F. Blackman
EQUALIZING THE ELECTRIC FIELD INTENSITY WITHIN CHICK BRAIN IMMERSED IN BUFFER SOLUTION AT DIFFERENT CARRIER FREQUENCIES
Bioelectromagnetics, Vol. 2, No. 4, pp. 411-413 (1981)

Kinn, J.F.
WHOLE BODY DOSIMETRY OF SMALL ANIMALS: THE EFFECT OF WEIGHT AND EXPOSURE GEOMETRY
Radio Sci., Vol 12, No. 6S, pp. 61-64 (1977)

Merritt, J.H., W.W. Shelton, and A.F. Chamness
ATTEMPTS TO ALTER 45-CA++ BINDING TO BRAIN TISSUE WITH PULSE-MODULATED MICROWAVE ENERGY
Bioelectromagnetics, Vol. 3, No. 4, pp. 475-478 (1982)

Miller, D.A., A.R. Valentino, and M. Benedick
Tech. Memorandum No. 2, Project No. IIT RIE-6249, U. S. Naval Electronic Systems Command, Washington, D. C. (1974)

Myers, R.D. and D.H. Ross
RADIATION AND BRAIN CALCIUM: A REVIEW AND CRITIQUE
Neurosci. Biobehav. Rev., Vol. 5, No. 4, pp. 503-543 (1981)

Schramm, M.
SECRETION OF ENZYMES AND OTHER MACROMOLECULES
Ann. Rev. Biochem., Vol. 36, pp. 307-320 (1967)

Shelton, W.W., Jr. and J.H. Merritt
IN VITRO STUDY OF MICROWAVE EFFECTS ON CALCIUM EFFLUX IN RAT BRAIN TISSUE
Bioelectromagnetics, Vol. 2, No. 2, pp. 161-167 (1981)

Sheppard, A.R., S.M. Bawin, and W.R. Adey
MODELS OF LONG-RANGE ORDER IN CEREBRAL MACROMOLECULES: EFFECTS OF SUB-ELF AND OF MODULATED VHF AND UHF FIELDS
Radio Sci., Vol. 14, No. 6S, pp. 141-145 (1979)

Weil, C.M.
THE CHARACTERISTIC IMPEDANCE OF RECTANGULAR TRANSMISSION LINES WITH THIN CENTER CONDUCTOR AND AIR DIELECTRIC
IEEE Trans. Microwave Theory Tech., Vol. 26, pp. 238-242 (1978)