open access


A commercial research and development feasibility study was conducted using a critical bone gap of 1.0 cm, surgically created in the radial bone of the left forelimb, to determine the viability of PEMF for generating bone to refill the gap in an otherwise non-healing tissue injury site. The full duration of the feasibility study was not completed due to the subsequent financial insolvency of the sponsoring company, however, the critical experiments through week four were sufficiently completed to draw the following conclusions for commercial PEMF technology development purposes: PEMF did generate bone in critical gaps with a success rate of 100% for at least partial bone gap closure, and 40% for full bone gap closure, in the four weeks when PEMF was applied with a slew rate => 100 kG/s. Healing rates of 67% for partial gap closure, but 0% for full gap closure, was observed when PEMF slew rate was half that value, ~ 50 kG/s. Healing in the absence of PEMF was less than 10%, and only for very small amounts of bone gap closure on only one specimen, which may have been due to surgical error. Post-surgical pain was also greatly reduced when higher-slew rate PEMF was applied, compared to lower slew rates or control (no PEMF). The optimal magnetic waveform slew rate for PEMF when applied to orthopedic injuries, both for tissue regeneration and pain reduction, was therefore determined to be => 100 kG/s.


The TVEMF experiments at NASA-JSC (2, 3) were designed and conducted to develop the core technology for growing cells in culture in an orbital research laboratory (Space Shuttle mid-deck locker and ultimately the ISS). However, the findings were such that the technology attracted attention for other uses, including Earth-based industrial processing of biomolecules and clinical therapeutic applications. This report details the early developmental work on the technology for Earth-based clinical orthopedic applications.

TVEMF was the original designation at NASA-JSC for what was to be later designated PEMF, therefore the two terms may be considered to be equivalent. Following the TVEMF (PEMF) experiments on human neuronal cells in culture at NASA-JSC (2, 3), research on derivatives of this technology was privatized and carried out by licensed sub-contractors to the primary license holder of the NASA patents (16, 17), Regenetech, Inc., located in Houston, TX. The basic PEMF technology was further developed by sub-contract to the author (RG Dennis) into a second-generation PEMF system, developed for basic research on laboratory animals.

As described in this report, the technology was developed into the form of a small cuff that could be placed around the forelimb of small animals to determine the effects of PEMF on bone regeneration in a surgically-induced critical bone gap (1.0 cm) in the radial bone of the left forelimb. These studies were carried out by the Veterinary College at TAMU, in full compliance with their institutional animal care and use policies.

This work was intended for private research and development, and was not initially designed as a typical academic hypothesis test for publication in peer-review. Rather, as is the case for most research and development outside of academia, the process was adaptive, and allowed for changes to be made to the experimental conditions based upon observations, experience gained, and design improvements made during the experiments. So it is important to consider that while some experimental design features and conclusions are drawn based upon statistical significance of collected data sets, most conclusions from this study are drawn based upon direct measurement and observation done during the execution of the experiment compared to a priori thresholds established to define the anticipated clinical and commercial value of the technology under test.

PEMF had been shown in the 1970’s (1) to be of potential benefit for bone regeneration. The objective of this study was to determine whether the PEMF technology derived from the initial NASA-TVEMF experiments (2, 3) had potential clinical value, specifically for orthopedic injury and recovery, and also to further develop and refine the basic PEMF technology with the intent of developing commercial clinical systems based on this technology. The scientific objective was to test the hypothesis that application of this specific form of PEMF (ICES) would induce bone regrowth under conditions where bone loss was considered too extensive to expect regrowth under any normal conditions. Bone regeneration in 20% or more of the critical bone defects was established as the a priori threshold for successful demonstration of feasibility of the PEMF technology for clinical use.


The animal surgery and assessment portion of this study was carried out under contract at TAMU-MDI (College Station, TX). Surgeries were performed by David Nelson, DVM and his staff in the Small Animal Surgery section. Design, assembly, and characterization of the PEMF devices was performed by the author (RG Dennis) in his private laboratory at Micro-Pulse LLC (Chapel Hill, NC).

PEMF technology:

The PEMF systems for this experiment are referred to as “cuffs”. These consist of a copper coil wrapped around the animal leg to provide magnetic pulses to the injured tissue (osteotomy), an electronic PEMF pulse generator, and a lithium battery. The PEMF generator was based on a previously-published implantable muscle stimulator (4 - 9), and on the fact that the delta-wave was shown to have a significant effect on cell proliferation, metabolism, and gene expression in the initial NASA studies (2, 3).

Design of the PEMF system:

Systems were designated “Sham” (no pulses, no PEMF) or “PEMF”. The difference was that the microcontrollers generating the pulses were programmed to output a steady zero voltage (no magnetic waveform) in the case of Sham, and 10 pulses per second in the case of PEMF. Pulses were alternating bipolar waveforms. Without a magnetic field detector, the systems were visually identical and indistinguishable to the technicians performing the animal portion of the experiment. The PEMF cuffs presented no detectable stimulus to the animal: the pulses are too low in intensity to elicit any stimulation that could be detected.

Cuff design:

Each cuff (Figure 1) was made from a cylinder of white 500 μm nylon mesh, formed around a machined Teflon mandrel and glued using silicone RTV to form a hollow cylinder. Each cylinder was supported on a Teflon mandrel and copper coil was wound (single-layer) onto it, then coated with several layers of silicone RTV. The cuffs used in the first half of the experiment are designated “small cuffs”. Half way through the experiment, cuff size was changed to “large cuffs”, which had a differently sized PEMF coil than the small cuffs used in the first half of the experiment. The PEMF pulse generators were then adjusted to provide the same peak Gauss (10 G) for both the small and large cuffs, which involved adjustment of the series resistance (Rc) as well as doubling the electronic pulse width from 100 μs to 200 μs.

Figure 1.PEMF “cuff” with copper coil shown to right and encapsulated lithium coin battery, shown to left and above. The PEMF pulse generator circuit (not shown) is attached by the 10-pin ribbon connector shown to the lower left.

Cuff Size Pulse Width Inner Diameter Coil Ohms Wire Gauge Coil Turns Coil Length Coil Impedance
Small 100 μs 25 mm 10.2 Ω 35 AWG 210 28 mm 0.30 mH
Large 200 μs 28 mm 25.1 Ω 34 AWG 300 46 mm 1.79 mH
Table 1.Cuff design parameters. Pulse width is the duration of the electrical pulse sent by the PEMF pulse generator to excite the PEMF coils.

Coils were tuned to generate 10 Gauss peak magnetic field pulses by addition of small-value series resistors (Rc). Magnetic waveforms were measured using two high-speed analog Hall effect sensors (Allegro A3515EUA, 5 mV/G sensitivity), configured with the sensing elements aligned coaxially and antiparallel, to give a combined sensitivity of 10 mV/G with double the small-signal bandwidth, zero voltage offset, and allowing precise, high-speed measurements of low-level, rapidly-changing magnetic waveforms.

The PEMF pulse generator was based directly on a design reported earlier (4, 5), using a microcontroller (Microchip PIC16F84A) and two dual N-P channel FET transistors (International Rectifier IRF7105), as shown in schematic in Figure 2. The PEMF pulse generator circuit and cuffs were operated at 3.1 Volts, powered by a single lithium battery, CR2477, 3.1 V, 1000 mAh 24.5 mm diameter coin with tabs. The microcontrollers were programmed in PICC, and clocked with a 40.000 kHz external crystal oscillator for micropower operation and to achieve highly precise pulse timing.

Electrical output pulses of defined pulse width were designed to generate corresponding magnetic waveforms of a skewed triangular shape with a defined slew rate (rising slope) and decay slope, as shown schematically in Figure 3.

The PEMF generator circuit was encapsulated in epoxy (Figure 4), and the cuff, circuit, and battery were encapsulated in silicone RTV (flowable viscosity). Coils were connected electrically to the PEMF pulse generator through 6 inches of thick-wall silicone tubing, glued at each end to the coil cuff and encapsulated pulse generator circuit connector.

Figure 2.Schematic of PEMF generator circuit. The value of Rc is selected to set the peak magnetic field to 10 Gauss.

Figure 3. Idealized schematic showing the relationship between electrical pulses in the coil and the resulting magnetic waveforms. Electrical pulses are applied to the coils in each cuff (top, solid line) with the resulting magnetic waveforms (bottom, dashed line). Note that the pulses are alternating bipolar, and that the magnetic waveform rises from zero to the peak value while the electrical current is applied to the coil, and falls at a different rate when the electrical current is allowed to passively return to ground state (0.0 volts). Pulse amplitudes and durations are not to scale.

Figure 4.Photograph of the encapsulated PEMF generator circuit, which connects to the cuff and battery shown in Figure 1. The circuit components are encapsulated in epoxy, leaving only the gold connector pins exposed. This arrangement allows damaged cuffs and drained batteries to be easily and quickly replaced while using the same PEMF generator. Batteries were replaced at ½ discharge to ensure full-power function of the PEMF device throughout the experiment.

A magnetic pulse detector (not shown) was fashioned to verify function of the PEMF cuffs during the experiment, using two similarly-configured Hall effect sensors, and an instrumentation amplifier with op-amp buffered outputs to drive two green LEDs, which flashed when placed over the cuff on the bone defect.

Prior to execution of the animal study, each PEMF device to be used in the study was built, characterized, and endurance tested for 480 hours to detect any devices that would be subject to premature failure or accelerated battery drain, indicating an over-current condition in the device. One defective device was detected and removed prior to animal testing. At the conclusion of the endurance testing, new batteries were installed before placement on the animals.

Animal model:

Rabbits; New Zealand White, male, skeletally mature (> 12 months old) with weight ranging from 2.5 to 4.0 kg. Rabbits were sourced, handled, fed, housed, and anaesthetized according to TAMU Study Protocol MDI-0106. All animals received humane care in compliance with “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research, and “Guide for the Care and Use of Laboratory Animals” (NIH Publication No. 85-23, revised 1996), and the USDA Animal Welfare Act CFR Parts 1, 2, and 3, and the specific animal use protocol approved by the IACUC at TAMU-MDI.

Surgical Procedure:

A critical bone gap (osteotomy) of 1.0 cm length was surgically created at the junction of the middle and distal thirds of the radius using an oscillating micro-sagittal saw, perpendicular to the long axis of the bone, using saline irrigation to control thermal damage to the tissue. Initial experimental groups consisted of ten (10) control (Sham device with no PEMF waveform) and ten (10) experimental (active PEMF device). The PEMF cuff was then placed on the forelimb and positioned over the osteotomy site.


The operated forelimb (left) was imaged by radiograph as well as computer tomography (CT), every two weeks, and healing was to be evaluated by 2 independent veterinary radiologists.

Bone regeneration was to be scaled according to the following categories (Table 2):

Score Description
0 No signs of healing
1 Blurring of osteotomy edges (early healing)
2 Beginning of callus formation
3 External callus well visualized and mineralized
4 Reestablishment of cortical margins
5 Medullary cavity reestablished
Table 2.Score descriptions for bone regeneration scaling

Post-operative observations: To monitor recovery and post-operative pain, animals were closely observed for the following symptoms (Table 3):

1 Decreased appetite or water consumption
2 Weight Loss
3 Hypo activity/lethargy
4 Restlessness
5 Vocalization/grunting
6 Alert but reluctant to move
7 Sudden aggression
8 Ataxia
9 Protruded third eyelid (nictitating membrane)
10 Hunched posture
11 Cyanosis
12 Pale mucous membranes
13 Fecal stained fur/lack of grooming
14 Respiratory distress/labored or rapid breathing
15 Dehydration
16 Tooth grinding
17 Licking, rubbing, or scratching at a certain area
Table 3.Category descriptions for post operative observations

Initial design:

Twenty-two rabbits were used in total: ten controls, ten experimental, plus two animals held in reserve for the experiment. Animals that developed infections or other severe health complications were removed from the study and were to be replaced with animals from the reserve as necessary.

The experimental protocol was double-blind: surgeons and small animal care technicians did not know which animals were given active PEMF or sham cuffs. Animals were given surgical bone defects (osteotomy) and were then fitted with cuffs on the surgical area. The cuffs were randomly either active (PEMF) or sham (control, no PEMF) by design and were identical in appearance. The cuffs were randomly serialized to identify each cuff as “PEMF” or “Sham”, and this key list was maintained by the designer/builder RG Dennis in North Carolina, without disclosure to the experiment technicians at TAMU in Texas. The animals received post-operative care without knowledge of the experimental group of each animal. Observations were made by small animal veterinary staff to observe and record each animal during recovery. The PEMF/Sham key was only to be disclosed when all data collection was complete. The animals were to be imaged every two weeks for the duration of the study, which was to be 16 weeks. Final assessment of the PEMF generator function, samples for tissue pathology, bone density measures, and radiologist assessment of images was to be done at the conclusion of the experiment.

The key feature of this study was the use of a well-established and reliable critical bone defect model, which is not expected to regenerate bone under normal circumstances. Any amount of healing or closure of this gap with new bone tissue would indicate that the PEMF devices under test were eliciting bone tissue healing greater than would normally be expected. Therefore, assessment of the effectiveness of the PEMF devices to enhance bone healing was not to be determined by the standard statistical method of comparison of means, rather, the presence of any healing would indicate the effectiveness of the PEMF devices. Sham controls served to ensure that the study had no systematic errors with the critical bone defect model surgery.

The threshold for demonstrating a useful clinical effect was set a priori to 20%. That is, if 20% or more of the animals tested showed evidence of regeneration of bone tissue into the critical bone gap, the PEMF device was to be considered clinically effective and would be further developed. This criterion of success was established before the study was conducted, and served as the justification for the number of animals used in the experimental group.

Deviations from initial experimental design:

As an adaptive research and development study, experimental conditions were expected to be modified as deemed necessary and appropriate during the course of the study. Incorporated into the initial contract, deviations were expressly permitted as part of the study design as it was to be a technology development instrument, provided the changes were submitted in writing and approved by the corporate sponsor.

After half of the surgeries had been performed (ten surgeries), and careful observations were made, the decision was taken to modify the PEMF cuffs with larger cuffs. This resulted in changes to the magnetic waveform as described in the RESULTS section. This effectively resulted in two, smaller experimental groups as described below. Several animals were excluded from the study due to health complications, and several PEMF device failures, due mainly to animal chewing, resulted in smaller than anticipated experimental groups. Also, the study was terminated prematurely due to financial problems at the sponsor (Regenetech, Inc.), which were unrelated to this study but which ultimately resulted in insolvency and failure of the company. The technology and data then fell to the stewardship of the author (RG Dennis).

By the termination point in the study, three of the animals had been excluded on the basis of health issues, and four of the PEMF systems had been destroyed by the animal chewing on the device. These latter failures happened almost immediately after surgical recovery and device placement, and the animals were deemed otherwise healthy by the veterinary technicians, so for the purposes of analysis those animals were moved to the control (Sham, no PEMF group). The two reserve animals were added into the randomized process. This ended up yielding the following group sizes by week 4, the final week of the study that will be analyzed in this report:

  1. 5 animals in the Small Cuff PEMF group
  2. 3 animals in the Large Cuff PEMF group
  3. 11 animals in the Sham (control) group

As a result of early termination of the study, a number of assessments were not made that had been planned for the end of the study, such as bone density measures, detailed radiologist assessments of radiographs, and final tissue pathology reports. Also, the data sets for images beyond 6 weeks were incomplete. These limitations should be considered when assessing the value of the results and conclusions reported in this study.

However, images for all test animals up to 4 weeks were complete, as were behavioral assessments and reports from veterinary technicians that were caring for and observing the animals. And data related to the PEMF generator performance and waveform characterization was comprehensive and complete. Taken together, this provided ample data for the intended PEMF technology research and development purposes, and allowed several important conclusions to be drawn.


Some of the most important results from this experiment arose because careful observations were made throughout the study, and during unexpected experimental problems. This is very often the case in the conduct of real science and real-world research and development, although many, perhaps most, academic scientists, attempting to maintain the charade of ideal scientific performance and flawless execution, and needing to curry the favor of federal funding agencies, often distort, conceal, or ignore the real-world events that invariably occur during the execution of any experiment, and report only those aspects that can be presented in an “ideal” light (20, 21). However, unfettered by these concerns, we are free to learn much from this imperfect but realistic and very enlightening experiment, with unanticipated positive results qualifying it as a felix culpa rather than a waste of resources. Therefore, first we consider the lessons learned from the unanticipated imperfections in the experiment:

Concerns and unanticipated events observed during the experiment:Small amounts of periosteum was suspected to have remained following the surgical creation of the critical gap defect in some cases. As a result, some amount of bone regrowth was observed for control (no PEMF or Sham), although this was considered insignificant enough to permit the use of the resulting images in a comparative assessment of the effectiveness of PEMF to promote bone regrowth, since the suspected traces of periosteum were presumed to be infrequent and equally distributed between the experimental groups due to randomization of the application of PEMF or Sham (control) for each surgery.

  1. Corrective action: after noting this error early in the study, subsequent surgical procedures were modified to assure removal of traces of periosteal tissue from the bone defect area.
    1. Corrective action: after noting this error early in the study, subsequent surgical procedures were modified to assure removal of traces of periosteal tissue from the bone defect area.
    2. Lessons learned: the surgeries could be improved, but also, real-life injuries will be quite different, and may offer unexpected opportunities for recovery, when compared to highly refined surgical defects with specific biological recovery mechanisms eliminated. Of note also was the clinical observation that the healing rate and completeness was much faster for injuries treated with PEMF, even accounting for the possible presence of residual periosteum in some of the surgical defects. This suggests that, under more realistic conditions of injury (bone fracture, not bone removal, with periosteum still in place), the application of PEMF appears to accelerate and improve healing, even with severe bone tissue injury and loss.
  2. Several animals chewed through and disabled their active PEMF cuffs.
    1. Corrective action: the cuff design was modified to resist chewing, and a test device was provided to allow a technician to verify the function of each cuff every day during feeding and care. This presented the problem that one designated individual among the animal care staff broke with the double-blind design. This person maintained a record of which devices were active, which had failed, and communicated this to the PEMF designer and the Chief Scientist at Regenetech.
    2. Lessons learned: unanticipated damage and failure mechanisms should be more carefully considered when future clinical PEMF systems are designed for human use (though failures due to chewing of the device are unlikely when used clinically with humans).
  3. Animal care technicians quickly observed that all animals fell into two groups based upon their post-surgical behavior: one group of animals that recovered quickly and resumed normal activity, and a second group of animals that initially were observed to be largely immobile, inactive, and appeared to recover more slowly. Further, the technicians voiced their opinion that these two groups corresponded to the experimental status of their PEMF cuffs, and suggested that animals with active cuffs were recovering swiftly, whereas animals in the control group with sham cuffs were not recovering well. Later analysis of these observations confirmed an exact correspondence between animals that recovered swiftly and regained normal activity with those animals that had active PEMF cuffs. Animals with sham cuffs or those that had damaged them by chewing exhibited very different behavior: much slower recovery. While this suspicion by the animal care staff did break the double-blind nature of the study, it also provided unexpected and very valuable observations about the effectiveness of PEMF for the reduction of pain during post-surgical recovery.
    1. Corrective action: technicians were asked to treat animals as if they did not know or suspect the experimental group they were in. Imperfect as this corrective action may be, it is not unreasonable to expect unbiased compliance with this request from veterinary professionals.
    2. Lessons learned: This is one of the most important observations made during the course of this study and could, by itself, form the basis for an entirely separate experiment if it were conducted strictly as a clinical study of post-surgical recovery based on the assessment of animal activity. Many animal studies report only animal activity with careful observations and quantification made. For example, the efficacy of many drugs and devices are assessed on the basis of how active an animal is, their ability to learn, their balance, social, feeding, and sleeping behavior, just as examples.
  4. The original cuffs did not fit well because they were too small: they needed to have a slightly larger diameter to fit on the animal leg properly, and the coil itself needed to be made slightly longer to assure that the bone gap injury was located inside the cuff as the animal recovered.
    1. Corrective action: exactly half way through the experiment, the cuff size was changed. This had the expected and initially unwanted result of changing the PEMF magnetic waveform, for reasons that are given in detail elsewhere. However, the outcome was that animals with larger cuffs recovered less well, and this is reflected in both the post-surgical observations of activity as well as in the images of bone regeneration in the defect gaps. Effectively, this created three experimental groups: (1) control, (2) small cuffs, and (3) large cuffs with some of the PEMF parameters being different (lower slew rate and longer pulse width), while other parameters were maintained the same between small and large cuffs (for example, peak Gauss and frequency were unchanged).
    2. Lessons learned: an important range for one PEMF parameter (magnetic waveform slew rate) was discovered, quite by accident, which improved the consistency and reliability of subsequent ICES-PEMF® technology markedly, while the importance of other parameters (peak Gauss and frequency) were shown to be of only secondary importance.

Characterization and analysis of magnetic waveforms:

The first major experimental result derived from concern (4) above: the change of the size of the cuffs themselves half way through the experiment. The magnetic waveform was measured using a high-speed Hall-effect analog sensor, and recorded on a digital oscilloscope (Fig. 5). This allowed a detailed analysis of the actual shape of the magnetic pulses of each cuff.

Figure 5.Narrow “square” electrical pulses from the PEMF pulse generator circuit yielded skewed triangular magnetic waveforms, as shown. This magnetic waveform is typical, resulting from a 100 micro-second current pulse applied to the PEMF coil in a short cuff, and measured using an analog Hall effect sensor. Note that the rise time corresponds to the electrical current pulse applied the coils (0 μs to 100 μs), while the fall-off time occurs while the coil current drains to ground potential, generating an asymmetric triangle (sawtooth) magnetic pulse.

Pulse Peak Rising Falling
Small Cuff 100 μs 10.1 G 101 kG/s -28 kG/s
Large Cuff 200 μs 9.6 G 48 kG/s -54 kG/s
Table 4.Average Gauss (B) and slew rate (dB/dt) during the rising and falling phases of each magnetic pulse, using small and large cuffs. Note that the pulses have nearly identical peak magnetic flux values (~10 Gauss) but the duration of the pulses and the rise and fall times differ by a factor of ~2. “Pulse” is the pulse width, defined by the duration of the electrical current pulse applied to the coils, measured in micro-seconds (μs). “Peak” is the peak magnetic field measured for each magnetic pulse, measured in Gauss (G). Slew rate (“Rising” and “Falling”) occurs during (rising), then begins immediately after (falling), the electrical current pulse has been applied and the coil drains to ground potential, both measured in kilo-Gauss/second. Note also that pulses were alternating bipolar, with the sign convention that + kG/s indicates the change in the magnetic field while current is applied, and – kG/s indicates the change in the magnetic field immediately after the electrical pulse to the coils is ended and the magnetic field returns to baseline, therefore the magnetic polarity of each pulse reverses with each pulse.

Post-operative pain relief and activity level:

Very early during the study, the veterinary technicians notified the study coordinator that the animals clearly fell into two groups based on post-surgical behavior: those that gave clear indications of post-operative pain, and those that did not. The vet techs were basing this mostly upon the general activity level, grooming and feeding activity of the animals. Some amount of post-operative distress (pain) was expected, and in this case was not considered to be sufficient to terminate the experiment. Upon checking against the double-blind group identification key, the vet techs had identified correctly, with 100% accuracy, which animals were in each group based on their post-operative behavior. Furthermore, in the cases where the animals caused damage to the PEMF cuffs, the vet techs detected this based on the change in behavior of the animals (from pain-free to showing some signs of pain). Upon testing, the PEMF devices of these animals was found to be damaged and not functioning. This occurred very early post-operatively, so these animals were moved into the Sham (control) group for further analysis.

Bone regeneration and healing:

This study was not designed as a standard A – B comparison test with a threshold p-value set to the standard value of 0.05. As a privately funded research and development effort, the experiment was designed to utilize a critical bone defect for which no bone regrowth was normally to be expected. The a priori value of 20% of samples showing evidence of bone regeneration was taken to indicate that the PEMF device would have clinical and commercial success. Experimental group PCT images, and representative PCT images for the Sham group, are shown below, four weeks post-surgically:

Figure 6.All five bone defects in the Small Cuff PEMF group, with the highest dB/dt (magnetic waveform slew rate) values, after four weeks:

Figure 7.All 3 bone defects in the Large Cuff PEMF group, with lower dB/dt (magnetic waveform slew rate) values, after 4 weeks.

Figure 8.Three typical examples of the Sham (control, no PEMF) group (n = 11), showing no evidence of bone gap healing after four weeks:

Group n n* n** %* %** peak Gauss Frequency dB/dt
Small Cuff 5 5 2 100% 40% 10.1 G 10.0 Hz 101 kG/s
Large Cuff 3 2 0 67% 0% 9.6 G 10.0 Hz 48 kG/s
Control 11 1 0 9% 0% no PEMF no PEMF no PEMF
Table 5.Summary of assessment of bone growth after 4 weeksn = total number in group, n* = number showing some bone regeneration, n** = number showing extensive bone regeneration that bridged the critical defect gap, %* = percent showing some bone regeneration, %** = percent showing extensive bone regeneration that bridged the critical defect gap.


While it may be argued that these results can be discounted due to the loss of double-blind conditions, it can equally well be argued that the difference in recovery between the two groups was so dramatic and unanticipated that it underscores the vast improvement in the recovery of animals when a specific type of PEMF (ICES) was applied versus when it was not, and that the double-blind conditions of the study were lost precisely because the beneficial effects of PEMF were so clearly and evident to any observer. Future experiments of this nature must take these this very powerful and easily observed effects into account if there is a need to maintain double-blind conditions, or a careful single-blind experimental design must be employed to assure equivalent post-operative animal care and assessment.

Although it might be argued that the study as designed was not fully completed and therefore should be repeated, the fact is that the results of the study, incomplete as it is, yielded useful information far beyond our initial expectations. Because this study was conducted in 2007, we have the benefit of 12 years of subsequent work built upon these findings, so its impact on further PEMF technology development is already known and can be clearly stated. It is the opinion of the author that, while replication of this study would be easily achievable, and equivalent results would be expected, and increased scientific credibility would result from each subsequent replication, that nonetheless, it is not necessary to replicate this study. Far better, more humane, and more useful results would be gained by focusing subsequent studies on a diverse human population with a range of refractory bone injuries, rather than subjecting more animals to severe and unrealistic surgically created bone defects. The small gain in statistical confidence would be far outweighed by the arbitrarily unrealistic experimental conditions that would need to be established, and the fact that ultimately our real interest is on the potential benefits in a diverse human population with varied but realistic injuries.

The post-surgical behavior of the rabbits alone, originally included only for purposes of animal care and for identifying animals for exclusion from the study due to illness, turned out to yield invaluable data that lead to studies related to optimization and use of the PEMF technology for pain reduction, which was an unanticipated outcome at that time, but has become the most clinically important application of ICES-PEMF technology. The use of PEMF for many forms of pain, including those unrelated to orthopedic injury, has been reported (10, 12). And the unplanned two different levels of slew rate for the cuff coils allowed the identification of the key threshold value necessary to achieve reliable biological effects. Further, it was interesting to note that the PEMF parameter that was held constant, peak Gauss, was only found to be effective when adequate magnetic waveform slew rate was applied. Therefore, we conclude that peak Gauss is not a key fundamental parameter of PEMF for biological effectiveness. Rather, peak Gauss is a secondary, derived value, resulting simply from the multiplicative product of the key parameter values slew rate (Gauss/time) and pulse width (time). Further, the pulse width had been based upon earlier, and verified by subsequent, studies of orthopedic tissue development (6 – 9). Pulse rate (Hz) was not tested in this study, but was thought to be non-critical, based on the wide range of PEMF frequencies reported in the literature, now numbering approximately 1000 manuscripts published in peer-review, with no systematic indication that any one frequency is critical for biological effectiveness.

This study resulted in the determination of several key PEMF parameters. These established key PEMF parameter values have been engineered into all subsequently manufactured ICES-PEMF devices, and have consistently resulted in positive biological effects in 93% to 99% of subjects in every subsequent study, including humans with TBI (11), humans with a wide range of severe refractory and idiopathic pain conditions (12), and in the standard animal (rat) model used for drug discovery for identifying new NSAIDs (manuscript in preparation). Therefore, there is no practical or scientific need to repeat this study, as it would be wasteful of time, resources, and laboratory animals. Far more compelling is the fact that, in the 13 years since this study was carried out, the resulting key PEMF parameter values have been used time and again in subsequent studies, both published and unpublished, with consistent beneficial biological effects generally observed at a rate far exceeding 90%.

This work, and subsequent technology development, resulted in the issuance of three United States patents (13-15) and two trademarks (18-19).


The key parameter for biological effectiveness of PEMF was determined to be magnetic slew rate (dB/dt), and the minimum threshold of this parameter for clinical effectiveness for regeneration of bone tissue after orthopedic injury was found to be ~ 100 kG/s. This magnetic slew rate, when sustained for 100 μs at a pulse rate of 10 Hz, was found to be effective both for pain reduction as well as to induce bone regeneration in a critical defect gap.

Statement of Potential Conflict of Interest

The author, Robert Dennis, PhD, has a financial interest in this research, and while working on this technology has served as a paid consultant for NASA-JSC, and as a subcontractor for NASA-JSC and DARPA during the initial development of this technology, as a scientific advisor and stockholder of Regenetech Inc., licensee of the original NASA-TVEMF technology, and finally as the sole owner and proprietor of Micro-Pulse LLC that continues to develop ICES-PEMF® technology for research and commercial sale.

List of Abbreviations

NASA-JSC – National Aeronautics and Space Administration at Johnson Space Center, Houston, TX

ISS – International Space Station

DARPA – Defense Advanced Research Projects Agency

TVEMF – Time Varying Electro-Magnetic Fields

PEMF – Pulsed Electro-Magnetic Fields, equivalent in meaning to TVEMF

ICES – Inductively-Coupled Electrical Stimulation, a specific, patented and trademarked form of PEMF

TAMU – Texas A & M University, College Station, TX

MDI – Michael DeBakey Institute of Comparative Cardiovascular Sciences and Biomedical Devices

IACUC – Institutional Animal Care and Use Committee

CT or PCT – Peripheral Computer (Aided) Tomography


1. Bassett et al. (1974) Augmentation of bone repair by inductively coupled electromagnetic fields, Science, New Series, v. 184, No. 4136, pp. 575-77, May 3.

2. Dennis R. (2019) Inductively Coupled Electrical Stimulation Part I: Overview and First Observations. ProcACIM 1(1):20-35.

3. Goodwin, T.J. (2003) Physiological and Molecular Genetic Effects of Time-Varying Electromagnetic Fields on Human Neuronal Cells. NASA/TP-2003-212054, 2003, available electronically from:

4. Dennis R.G. (1998) Bipolar implantable stimulator for long-term denervated muscle experiments. Med & Biol Eng & Comput, 36: 225-28, March.

5. Dennis, R.G., Dow, D.E., Faulkner, J.A. (2003) An implantable device for stimulation of denervated muscles in rats. Medical Engin & Physics, 25(3), pp. 239-253.

6. Dennis, R.G., Dow, D. E. (2007) Excitability of skeletal muscle during development, denervation, and tissue culture. Tissue Engineering, 13:10, 2395-2404, October.

7. Dennis, R.G., Paul E. Kosnik, Mark E. Gilbert, and John A. Faulkner. (2001) Excitability and contractility of skeletal muscle engineered from primary cultures and cell lines. Am J Physiol Cell Physiol 280: C288-C295.

8. Dennis R.G., Kosnik P.E. (2000) Excitability and isometric contractile properties of mammalian skeletal muscle constructs engineered in vitro. In Vitro Cell. Dev. Biol. Anim. 36(5): 327-335.

9. Kosnik P. Jr., Faulkner J.A., and Dennis R.G. (2001) Functional development of engineered skeletal muscle from adult and neonatal rats. Tissue Engineering, 7(5) 573-584.

10. Hubbard, D.K., Dennis, R.G. (2012) Pain relief and tissue healing using PEMF therapy: a review of stimulation waveform effects. Asia Health Care Journal 1(1), pp. 26-35, July.

11. Pawluk, W, Dennis, R, and Tommerdahl, M. (2017) Tracking the effects of pulsed electro-magnetic field (PEMF) on individuals with a history of traumatic brain injury (TBI) with the Brain Gauge. Cortical Metrics Magazine, 24 March 2017.

12. Ravid, S. (2019) Doctoral Dissertation: Descriptive Exploratory Study of Individuals’ Use of Pulsed Electromagnetic Fields, the Micro-Pulse, for Pain Relief. Florida Atlantic University School of Nursing, Boca Raton, FL, as presented at ACIM 2019, Orlando FL, 14 Nov 2019.

13. Dennis, R.G., Kosnik, P.E., Clark, J.R. (2011) Magnetic system for treatment of cellular dysfunction of a tissue or an extracellular matrix of a tissue. U.S. Patent Number 8,029,432 B2, issued October 4, 2011. Utility patent, Attorney docket no. 1814.001, EFS ID: 6552631, Application Number: 12628990. Filing date: 1 December 2009.

14. Dennis, R.G., Kosnik, P.E., Clark, J.R. (2012) Magnetic system for treatment of cellular dysfunction of a tissue or an extracellular matrix of a tissue. U.S. Patent Number 8,137,258 B1, issued March 20, 2012. Attorney docket no. 1814.002A, Application Number: 12952761. Filing date: 23 November 2010.

15. Dennis, R.G., Kosnik, P.E., Clark, J.R. (2012) Magnetic system for treatment of cellular dysfunction of a tissue or an extracellular matrix of a tissue. U.S. Patent Number 8,137,259 B2, issued March 20, 2012. Attorney docket no. 1814.003A, Application Number: 12952797. Filing date: 23 November 2010.

16. Wolf, David A., Goodwin, Thomas J., Dennis, Robert G. (2002) Growth stimulation of biological cells and tissue by electromagnetic fields and uses thereof. U.S. Patent No. 6,485,963, issued November 26, 2002. An investigation of inventorship was initiated by NASA-JSC on 5-22-2007 ref: MSC 22633-1 and MSC 22633-2, and it was determined that R.G. Dennis will be added as co-inventor.

17. Wolf, David A., Goodwin, Thomas J., Dennis, Robert G. (2004) Growth stimulation of biological cells and tissue by electromagnetic fields and uses thereof. U.S. Patent No. 6,673,597 B2, issued January 6, 2004. An investigation of inventorship was initiated by NASA-JSC on 5-22-2007 ref: MSC 22633-1 and MSC 22633-2, and it was determined that R.G. Dennis will be added as co-inventor.

18. “ICES” Registered Trademark, United States PTO

19. “DIGICEUTICAL” Registered Trademark, United States PTO

20. Fanelli D. (2010) Do Pressures to Publish Increase Scientists’ Bias? An Empirical Support from US States Data. PLoS ONE 5(4): e10271. doi:10.1371/ journal.pone.001027121.

21. Baker, M. (2016) Is there a reproducibility crisis? NATURE 452-454, VOL 533