CSIRO - June 1994





People with normal hearing are able to perceive pulse-modulated RF and microwave radiation between about 200 MHz and 6.5 GHz. The sound perception probably results from the thermoelastic expansion of brain tissue caused by a small but rapid increase in temperature. The perception threshold for pulses shorter than 30 Ás depends on the specific energy density in the head and has been estimated to be as low as 30 mJ/kg. Receptors in the skin are sensitive and absorb RF and microwave radiation at power densities of approximately 300 W/m2 at 3 GHz during exposure for 10 s. Meanwhile infrared radiation applied for 10 s is detected at power densities an order of magnitude lower due to its greater absorption (therefore greater SAR) in the skin. Thresholds of perception depend on frequency and exposure duration as well as the locality on the body and the exposed area. The perception of skin warming by microwave and RF frequencies in the rage 0.5-100 GHz does not afford a reliable means of protection against potentially harmful exposure from heating (Elder 1984a).

Healthy subjects at rest in light clothing and in comfortable ambient conditions (21 - 22°C, 50% RH and adequate ventilation) are able to dissipate RF power at SARs of 1 W/kg, and to up to 4 W/kg for short periods, although sweating and increased heart rate was observed in the upper part of this range after 20 min exposure. The total heat load of an exposed person represents the sum of the SAR from RF or microwave heating and the rate of metabolic heat production and must be compensated by heat loss. The limit of tolerable SAR is affected by adverse conditions (high temperatures or humidity), moderate physical exercise, some medication, or conditions which impair thermoregulation (including pregnancy). The relationship between local SAR and body temperature increase is not well established. Neither is the rise in local temperature in response to high, localised SARs elsewhere within the body. Further dosimetric research is required to determine whether local heating, rather than whole-body heating, could become a limiting factor in some circumstances.

7.1 Perception

Auditory perception

It is well established that humans with normal hearing are able to perceive pulse-modulated RF and microwave radiation as buzzing, clicking, hissing, or popping noise, depending on the modulation characteristics (NCRP 1986). First reports (Frey 1961, 1962, 1963) of this phenomenon described the perception of pulsed radiation frequencies between 216 MHz and 2.98 GHz. Pulse widths varied between 1 and 1000 Ás and the average threshold power density was 4 W/m2. Sensitivity was increased by lowering the ambient audible noise levels. Since then the average power density threshold for RF hearing has been reported as low as 0.01 W/m2 in people with normal hearing (Cain & Rissman 1978). Perception of different pulse-modulated frequencies has been reported (Constant 1967) at 3 and 6.5 GHz but not at 9 GHz.

In a study of the threshold conditions for this effect in one human subject exposed to pulsed 2.45 GHz radiation (Guy et al 1975) it was determined that, for pulse widths less than 30 Ás, the perception threshold depended on the energy density per pulse. A threshold value of 280 mJ/m2 (estimated 10 mJ/kg) was measured when the subject wore earplugs. An animal study measured brain stem auditory evoked potentials in guinea-pigs exposed to pulsed 918 MHz radiation (Chou & Guy 1979) and concluded that the threshold for microwave hearing is related to the incident energy density per pulse for pulses shorter than 30 Ás and is related to the peak power for longer pulses (up to 500 Ás).

It is generally agreed (NCRP 1986; Foster & Finch 1974) that the mechanism of acoustic perception of short pulses of RF and microwave radiation is due to thermoelastic expansion of brain tissue following a small but rapid temperature increase (<10-5°C). As the effect must depend on absorption of some incident energy it would be limited to frequencies which penetrate the skull and are significantly absorbed by brain tissue. The effect of the expansion is an acoustic pressure wave which is transmitted through the skull to the cochlea where vibration-sensitive hair cell receptors respond as they would to acoustically generated pressure stimuli. It has been calculated that the frequency of the induced sound is related to head size and the acoustic properties of brain tissue, regardless of the RF frequency (Lin 1977).

Cutaneous perception

The absorption of RF and microwave radiation can be detected by receptors in the skin, although variability in the sensitivity has been reported. Most studies have involved exposures in the frequency range 2.45 - 10 GHz (NCRP 1986; Elder 1984; Adair 1983). Microwave radiation was found to be ten to fifteen times less effective than infrared in heating the skin (Justesen et al 1982). The difference is attributed to the scattering of two-thirds of the microwave energy and the relatively small proportion of microwave energy, estimated to be one-fifth of the value for infrared radiation, absorbed in the skin. Subjective awareness of warmth is not a reliable indicator of microwave hazard. Perception threshold values are frequency dependent. Threshold response by different parts of the body is variable, at low levels of irradiation the face is the most sensitive region, the trunk intermediate, and the limbs least sensitive (by a factor of two or three). Threshold power density must be qualified by the duration of the stimulus and the area of the exposed skin; e.g. infrared power density thresholds decrease as the duration of the stimulus increases (up to a critical value), and as the size of the area of exposed skin increases. The response latency is also an important variable that depends on stimulus duration and area. In general, skin sensory receptors respond to transient rather than constant stimuli, although the effect of adaptation on the perception of microwave radiation is not known.

Consensus of a panel of a Symposium on Microwaves and Thermoregulation, Connecticut, 1981 was that microwaves of 30 GHz and above would probably be similar to infrared in their perception threshold values and may be sensed at the limit recommended by the American National Standards Institute (ANSI 1982) which is 50 W/m2 in this part of the spectrum (Adair 1983). However, over much of the radiofrequency spectrum, the perception thresholds are higher than the ANSI standard, and the deeper penetration results in a larger mass of tissue heated by microwaves for a comparable rise in skin temperature. A further problem is that the threshold temperature (41-42°C) of cellular injury for sustained temperature elevations is below the threshold (_ 45°C) of pain. For the frequency range 0.5 MHz to 100 GHz, cutaneous perception of heat and thermal pain may be an unreliable sensory protection mechanism against RF radiation exposures.

7.2 Thermophysiological responses

Normal body temperature is maintained by a complex control system of heat loss or gain responses, including behaviour (Simon et al 1986), at a so-called ôset-pointö value of around 37°C. Body temperature is maintained within a narrow range (_0.5 °C fluctuation) by the processes of sweating or increasing metabolic heat production. Data on thermoregulatory responses to RF or microwave heating is obtained from experiments with passively heated volunteers. Ambient conditions such as high humidity can profoundly limit the thermoregulatory capabilities and exert a significant effect on the ability to tolerate different whole-body SARs. A change in ambient temperature of 1°C can produce a change in heat flux of about 0.15 W/kg in a clothed individual. It is clear then, that the thermal burden from a given SAR that can be tolerated in a cool, dry environment may pose a health hazard if the environment is either very hot or humid.

In an attempt to estimate the maximum SAR that could be tolerated by a fit healthy person, under strictly specified ambient conditions, Durney et al (1978) defined a rectal temperature of 39.2°C as an upper limit of physiologically tolerable body temperature. A SAR of approximately 3 W/kg was calculated to induce this rectal temperature within 1 h in a person in an environmental temperature of 40°C and a relatively humidity of 80%. Raising the ambient temperature to 41°C decreased the tolerable SAR to about 1 W/kg. The rectal temperature of 39.2°C is an estimate of the upper limit of tolerance and should not be considered as a safety limit as there is a wide range in physiological tolerance amongst different members of the population.

Other physical and physiological factors which reduce the ability to adapt to an extra heat load include old age, obesity, hypertension and effects of many drugs such as diuretics, antihistamines, tranquilizers B-blockers and amphetamines. The thermoregulatory ability of infants is not well developed while pregnant women have an extra circulatory load which may compromise their ability to dissipate heat. Heat loss from the embryo and fetus across the placental barrier may be less efficient than heat dissipation in other, well-vascularised tissues.

The American National Institute of Occupational Safety and Health (NIOSH 1972) and the American Conference of Governmental Industrial Hygienists (ACGIH 1983) recommend an upper threshold limit value of a rise in body temperature of 1°C. This is endorsed by the World Health Organisation (WHO 1980). It was recommended that rates of physical work and environmental factors should be such as to limit excursions of body temperature beyond 1°C.

Calculations relating whole-body SAR to increases in body temperature are generally supported by the results of studies of the thermoregulatory responses of patients and volunteers exposed to RF fields of up to 4 W/kg in magnetic resonance systems. However, the subjects are exposed at rest and in controlled environments and the whole-body SARs quoted result from much higher, localised, SARs.

Shellock et al (1989) exposed the abdomen of six volunteers to 64 MHz RF magnetic fields for 30 min at whole-body SARs from 2.7 to 4.0 W/kg. Although body temperature was reported to rise by an average of only 0.1°C, all of the subjects reported feeling warm and had visible signs of perspiration on their forehead, chest and abdomen during the procedure. Abart et al (1989) reported increases in rectal or sublingual temperature of up to 0.7°C, with a mean increase of 0.33°C, in 12 healthy volunteers exposed at a whole-body SAR of 3 W/kg for 20 min. Heart rate was also observed to increase by up to 45%. The magnitude of body temperature increase was similar to that reported in 12 healthy volunteers exposed at a whole-body SAR of 4 W/kg for 20 min (Schaefer et al 1985).

In contrast, differing thermoregulatory responses (to magnetic resonance imaging) have been reported in 50 patients with unspecified clinically impaired temperature regulation (Shellock & Crues 1987). Exposure to 64 MHz RF magnetic fields at whole-body SARs between 0.4 and 1.2 W/kg for 15 min increased body temperature by 0.5°C at low SARs. The mean skin temperatures of localised areas of the hand and trunk when imaged, were reported to increase by up to 1.2°C. Another study (Kido et al 1985) reported mean increases in body temperature of 0.5°C in volunteers exposed to a magnetic resonance abdominal scan for 17 min where the whole-body SAR was only 0.8 W/kg. Heart rate increased by 3 beats per minute.

Calculations of SAR distribution based on a heterogeneous model of the human body indicate that localised SARs in small tissue volumes could be 10 to 70 times greater than the whole-body average SAR during exposure to magnetic resonance imaging (Orcutt & Gandhi 1988).

The relationship between local SAR and temperature increase is not well established. Localised heating may occur in various parts of the body depending on the conditions of exposure, particularly antenna proximity and the radiation wave-length, and on the shape and variation in tissue conductivity and blood circulation in the exposed part of the body. The amount by which localised heating will exceed whole-body average is not known at present.