Summary
To develop novel therapeutic interventions for the prevention and management of back pain, animal models are required to examine the mechanisms and effectiveness of these therapies from a translational perspective. The present protocol describes the BMS test, a standardized method to assess back mechanical sensitivity in the rat.
Abstract
Low back pain is the leading cause of disability worldwide, with dramatic personal, economic, and social consequences. To develop novel therapeutics, animal models are needed to examine the mechanisms and effectiveness of novel therapies from a translational perspective. Several rodent models of back pain are used in current investigations. Surprisingly, however, no standardized behavioral test was validated to assess mechanical sensitivity in back pain models. This is critical to confirm that animals with presumed back pain present local hypersensitivity to nociceptive stimuli, and to monitor sensitivity during interventions designed to relieve back pain. The objective of this study is to lay down a simple and accessible test to assess mechanical sensitivity in the back of rats. A test cage was fabricated specifically for this method; length x width x height: 50 x 20 x 7 cm, having a stainless-steel mesh on the top. This test cage allows the application of mechanical stimuli to the back. To perform the test, the back of the animal is shaved in the region of interest, and the test area is marked to repeat the test on different days, as needed. The mechanical threshold is determined with Von Frey filaments applied to the paraspinal muscles, utilizing the up-down method described previously. The positive responses include (1) muscle twitching, (2) arching (back extension), (3) rotation of the neck (4) scratching or licking the back, and (5) escaping. This behavioral test (Back Mechanical Sensitivity (BMS) test) is useful for mechanistic research with rodent models of back pain for the development of therapeutic interventions for the prevention and management of back pain.
Introduction
Low back pain (LBP) is the leading cause of disability worldwide, which has dramatic personal, economic, and social consequences1,2,3,4. Every year, approximately 37% of the population is affected by LBP5. LBP usually resolves within a few weeks but recurs in 24%-33% of individuals, becoming chronic in 5%-10% of cases2. To understand the mechanisms and impacts of LBP as well as the effects of different therapeutic interventions, several animal models of LBP have been used, mimicking clinical conditions or some components of LBP6. These mouse and rat models can be classified in one or more of the following categories: (1) discogenic LBP7,8,, (2) radicular LBP8,9,10,11, (3) facet joint osteoarthritis12, and (4) muscle-induced LBP13,14. Since the pain cannot be measured directly in non-human species, numerous tests have been developed to quantify pain-like behaviors in these models8. These tests assess behaviors evoked by a noxious stimulus (mechanical force15,16,17, thermal stimulation18,19,20,21,22,23,24,25) or produced spontaneously26,27,28,29.
The methods using mechanical stimuli include the Von Frey test15,16 and the Randall-Selitto Test17. Methods using heat stimuli include the tail flick test18, hot plate test19, Hargreaves test20, and thermal probe test21. Methods using cold stimuli include the cold plate test22, acetone evaporation test23, and cold plantar assay24. Methods for spontaneous behaviors include the grimace scales26, burrowing27, weight-bearing and gait analysis28, as well as an automated behavioral analysis29. Despite these numerous available tests, none of them is designed specifically for back pain models.
The objective of this study is to lay down a simple and accessible test to assess mechanical sensitivity in the back of rats. The technique is largely based on the Von Frey test applied to the plantar surface of the hind paw15,16. The basic principle of the Von Frey test is to use a series of monofilaments to the region of interest, delivering constant pre-determined forces. A response is considered positive if the rat shows a nocifensive behavior. The mechanical threshold can then be calculated based on the filaments that evoked responses. In the present study, a simple and accessible method adapted from the Von Frey test is provided to determine mechanical sensitivity in the back of rats.
Subscription Required. Please recommend JoVE to your librarian.
Protocol
The experimental protocol was approved by the animal care committee of Université du Québec à Trois-Rivières and conformed to the Guidelines of the Canadian Council on Animal Care and the Guidelines of the Committee for Research and Ethical Issues of the International Association for the Study of Pain (IASP). The present study used six male Wistar rats (body weight: 320-450 g; age: 18-22 weeks). The animals were obtained from a commercial source (see Table of Materials). Data from these rats are from the larger sample of a previous study30.
1. Experimental preparation
- House the animals in a temperature-controlled room in standard animal facilities with access to food and water ad libitum and a light-dark cycle of 14 h-10 h. Ensure that all animals are in good health on the day of the experiments.
- Generate the chronic back pain animal model following the steps below.
- To induce chronic back pain, perform an intramuscular injection of Complete Freund Adjuvant (CFA) into the back muscles following the previous reports14,30,31.
- Anesthetize the animal using isoflurane (4% for induction and 2%-2.5% for maintenance).
- Using a 27 G needle, inject 150 µL of a ready-to-use water-in-oil emulsion of CFA (see Table of Materials) into the paraspinal muscles unilaterally or bilaterally, depending on the protocol needs.
- Keep the injection needle in place for at least 3 min after completing the injection. For animals in the control group, use the same procedures30, but inject a solution of sterile physiological saline solution (150 µL, 0.9%) instead of CFA.
- Fabricate the test cage.
- Make a test cage for two animals that comprises one chamber for each animal.
NOTE: For the present study, each chamber has the following dimensions: length x width x height: 50 x 20 x 7 cm (see Table of Materials). - Mount the two contiguous chambers on four 33 cm long Plexiglass legs. Use transparent Plexiglass for the walls of the chambers, but use black Plexiglass to separate chambers to prevent animals from seeing each other.
- Use stainless-steel mesh made of 1 mm wire with an 8 mm inter-wire distance to make the floor and ceiling of the test cage (Figure 1).
- Make a test cage for two animals that comprises one chamber for each animal.
2. Back Mechanical Sensitivity (BMS) test
- Familiarize the animal with the test cage 30 min/day for 5-7 consecutive days prior to the first test. Repeat the test as needed.
- Anesthetize the animals using 2% isoflurane31 (see Table of Materials).
- In a prone position under isoflurane anesthesia, shave the back hair of the area of interest (from T6 to L6 vertebral levels) using an animal hair trimmer (see Table of Materials). For repeated measures, shave the back hair every 3 days on a day without behavioral assessment to ensure that stimuli are always applied directly to the skin. Draw a black mark on the skin with a permanent marker to ensure that filaments are always applied to the same area when repeating the test on different days.
- On the testing day, put the animals in the test cage for 15-30 min prior to the test until the animal is calm.
- During the test, apply Von Frey filaments (0.07, 0.16, 0.4, 0.6, 1, 2, 4, 6, 10, 15, and 26 g) perpendicularly to the back, always beginning with the 2 g filament and using the up-down method15 (see Table of Materials). Approach the back of the animal slowly with the filament from behind the animal.
- Apply the filament only when the animal is awake, standing on its four paws, and not moving. Apply the filament for 2 s bilaterally to the area of interest, 10 mm from the spinous process (Figure 2), every 15-30 s.
NOTE: A response is considered positive if the animal exhibits one or more of the following behaviors during or immediately after the filament is applied: (1) muscle twitching, (2) arching (back extension), (3) rotation of the neck to look at the back, (4) scratching or licking the back, and (5) escaping.
- Apply the filament only when the animal is awake, standing on its four paws, and not moving. Apply the filament for 2 s bilaterally to the area of interest, 10 mm from the spinous process (Figure 2), every 15-30 s.
- As described previously15, if no response is observed with the application of a filament, apply the next filament with a higher force in the series. If a response is observed, use the next filament with a lower force in the series. Continue this procedure until four readings are obtained after the first behavioral change (response after a series of "no response" or no response after a series of "response").
- Once data collection is completed, calculate the value representing 50% of the mechanical threshold, as described by Chaplan et al.15, using this formula:
50% threshold (g) = 10(Xf+kδ)/10 000
NOTE: In this formula, "Xf" is the Handle Mark of the last von Frey filament that was used. "k" is the tabular value based on the animal's response pattern15, and "δ" is the mean of Handle Mark's increments between Von Frey filaments. Depending on the experimental design and experimental needs, only one side of the spine may be assessed to report one threshold, or two sides may be evaluated, and thresholds are reported separately or as a mean. Refer to Supplementary Table 1 for the calculation template32.
3. Animal recovery
- After the intramuscular injection is completed, discontinue anesthesia and place the animal alone in a standard housing cage for recovery.
- During the recovery period, examine the animal's behavior and do not leave it unattended.
- Confirm that the animal recovers from anesthesia and moves normally within 5 min. Then, return the animal to its usual housing cage with the other animals.
NOTE: At the end of the experiment, the animal is perfused through the heart with a 10% formalin solution, under deep isoflurane anesthesia (5%). The back muscles in the injected area are then extracted for histology and confirmation of inflammatory changes.
Subscription Required. Please recommend JoVE to your librarian.
Representative Results
The method was used in a previous study, in which full data and statistics were presented to compare back mechanical sensitivity between CFA and control rats30. Representative individual data (mean of left and right thresholds) from six rats included in the previous study are presented in Figure 3 and Table 1. At baseline, mechanical sensitivity was similar between groups. Intramuscular injection of CFA in the lumbar muscles caused a marked increase in mechanical sensitivity (decreased threshold) from 7 days to 28 days after CFA injection. In contrast, the control (CTL) rats did not show this change. As shown in Figure 3, variability was observed within and between animals, as expected with this type of behavioral assessment. However, hypersensitive CFA rats showed decreased variability. Based on the previous study30, 16 animals (8 CFA and 8 CTL) is sufficient to detect a significant effect between groups over time (η2p = 0.38) for 5 time points.
In this study, the presence of chronic inflammatory changes in the muscles injected with CFA was confirmed by histological examination (Figure 4)30. Also, mechanical hypersensitivity was observed at the hind paw with a standard Von Frey test, in addition to the back (Figure 5)30. In previous studies with the same back pain model, we showed increased spontaneous pain behavior and neuroinflammatory and neurophysiological changes14,31. Indeed, licking behaviors were increased in CFA compared with control rats during the formalin test, and single-unit responses to noxious stimulation of the sciatic nerve were altered in the right amygdala31. In addition, NF-kB protein expression was increased in the spinal cord of CFA compared with control rats14. Together, the results of these studies validate this chronic back pain model, and the present study visually demonstrates how to confirm the presence of mechanical hypersensitivity in the back of this rat model.
Figure 1: Back Mechanical Sensitivity (BMS) test cage. (A) Schematic drawing of the test cage. (B) Custom-made test cage comprising two chambers, one for each animal. (C) Lateral view of the test cage with a rat in one of the chambers. Please click here to view a larger version of this figure.
Figure 2: Back Mechanical Sensitivity assessment. The experimenter approaches the animal from behind and applies the Von Frey filament to the area of interest, 10 mm laterally from the spinous process. Please click here to view a larger version of this figure.
Figure 3: Individual examples of back mechanical sensitivity. Back mechanical sensitivity in CFA and control (CTL) rats, at Baseline and 7, 14, 21, and 28 days after the intramuscular injection of CFA or saline, respectively. Individual data are shown by gray (CTL) and black (CFA) filled circles. Horizontal bars indicate the means. Error bars denote the standard error of the mean. Please click here to view a larger version of this figure.
Figure 4: Histological confirmation of chronic muscle inflammation. Individual examples of back muscles from CFA rats and controls30. (A) Healthy back muscle from a control rat 14 days after intramuscular injection of saline. (B-C) Back muscles from two CFA-treated rats showed chronic inflammation 14 days after intramuscular CFA injection, with a clear leukocyte infiltration. Hematoxylin-eosin coloration was used for staining the muscle slices. Scale bar = 250 µm. Please click here to view a larger version of this figure.
Figure 5: Mechanical hypersensitivity in CFA rats30. Time course of mechanical sensitivity over 4 weeks, following either CFA (n = 8) or saline (n = 8) injection into the back muscles (L5-L6 level). (A-B) Mechanical sensitivity under the left and right hind paws. Mechanical thresholds significantly reduced in CFA compared with control rats (P < 0.01) over time. This effect was not significantly different between left and right hind paws (P = 0.7). For both hind paws combined, the Tukey HSD test revealed lower mechanical thresholds in CFA compared with control rats, from 1 week to 4 weeks after injection (all P's < 0.03). Time courses for separate hind paws are shown for illustration purposes only (interaction not significant, see results for details). (C-D) Mechanical sensitivity in the back. Mechanical thresholds significantly reduced in CFA compared with control rats (P < 0.001) over time. This effect was not significantly different between left and right assessment sites (P = 0.3). For left and right assessment sites combined, the Tukey HSD test revealed a lower mechanical threshold in CFA compared with control rats, from 1 week to 4 weeks after injection (all P's < 0.05). Time courses for separate hind paws are shown for illustration purposes only (interaction not significant, see results for details). In panel (D), the individual data of one CFA rat is not shown at the baseline (9.6 g) for illustration purposes. Shaded areas represent baseline assessment. Please click here to view a larger version of this figure.
| Rat | Group | Baseline | Day 7 | Day 14 | Day 21 | Day 28 |
| 1 | 2.34 | 0.29 | 0.12 | 0.29 | 0.29 | |
| 2 | CFA | 1 | 0.48 | 0.05 | 0.48 | 0.08 |
| 3 | 1.26 | 0.05 | 0.05 | 0.05 | 0.19 | |
| Mean ± SD | 1.53 ± 0.58 | 0.27 ± 0.18 | 0.07 ± 0.03 | 0.27 ± 0.18 | 0.19 ± 0.09 | |
| 4 | 1.59 | 2.61 | 0.64 | 3.26 | 2.45 | |
| 5 | CTL | 1.15 | 0.63 | 3.41 | 2.3 | 1.29 |
| 6 | 0.43 | 1.26 | 0.77 | 0.32 | 2.09 | |
| Mean ± SD | 1.06 ± 0.48 | 1.50 ± 0.83 | 1.61 ± 1.28 | 1.96 ± 1.22 | 1.94 ± 0.48 |
Table 1: Individual examples of back mechanical sensitivity in CFA and control rats.
Supplementary Table 1: Determination of mechanical threshold. This template table is used to calculate the mechanical threshold. The pattern of responses (X/O) is noted, and the values needed for the calculation are entered for Xf and k only, corresponding to the Handle Marking of the last filament that was used for the test and to the k-value associated with the response pattern, in this case, XX followed by OOXXO. Please click here to download this File.
Subscription Required. Please recommend JoVE to your librarian.
Discussion
Critical steps
The BMS test is a simple method to assess mechanical sensitivity in the back of rats, either at one time point or repeatedly over days or weeks, when changes are expected to occur (pain models) or after pharmacological or non-pharmacological intervention. Critical issues of the method include the test cage, whose dimensions must ensure that the rat is comfortable but does not move too much. The animal's back must stay accessible through the mesh ceiling for reproducible mechanical stimulation. To limit variability in the threshold assessment, the back area under investigation must be shaved so that mechanical stimuli are applied directly to the skin. In addition, the skin needs to be marked for mechanical stimuli to be applied to the same area. Lastly, the experimenter must approach the filament to the skin from behind the animal to avoid being seen by the animal.
Compared with the Von Frey test used for assessing the mechanical sensitivity at the hind paw15,16, the mechanical force needed to produce a positive response in the BMS test is lower. The filaments used for the test should be selected carefully. Using the following filaments should cover most experimental needs (0.07, 0.16, 0.4, 0.6, 1, 2, 4, 6, 10, 15, and 26 g) and prevent facing a ceiling or a floor effect. In this case, the 2 g filament is used for the first application. This can be adapted to experimental needs as long as the calculation is adjusted accordingly.
Modifications and troubleshooting
During a pilot experiment, the ideal area for the test was determined. Because of the shape of the rat's body, the thoracolumbar region is the most accessible area in the test cage. If there is no reason for the test to be performed in other regions of the spine, this is the area of choice for applying mechanical stimuli. The lumbar area is also easily accessible. When deciding which area to test, it must be kept in mind that the filament must be applied perpendicularly to the surface and bend properly to deliver the pre-determined calibrated force.
Limitations
The experimenter must be trained to observe the behaviors associated with the test. The five positive responses include muscle twitching, arching, neck rotation to look at the back, licking or scratching the back, and escaping30. While most of these responses are easily observable, muscle twitching is sometimes subtle for stimuli of a lower force. Also, the rat may move spontaneously in the cage, so this must not be confounded with escaping, which occurs specifically when the filament is applied. To avoid confounding both behaviors, the experimenter must wait for the animal to be calm for at least a few seconds.
Significance and potential applications
Several rodent models of back pain are used in current investigations8. Surprisingly, however, no standardized behavioral test was validated to assess mechanical sensitivity in back pain models. This is critical to confirm that animals with presumed back pain present local hypersensitivity to nociceptive stimuli, and to monitor sensitivity during interventions designed to relieve back pain. The BMS test presented here provides a simple and accessible solution for these purposes. Although it was developed for rats30, it may be adapted to other laboratory animals in the future.
Subscription Required. Please recommend JoVE to your librarian.
Disclosures
The authors declare no competing interests or relationships that may lead to any conflicts of interest.
Acknowledgments
This work was supported by a grant from the Fondation Chiropratique du Québec and the Natural Sciences and Engineering Research Council of Canada (MP: grant #06659). The contribution of HK was supported by the Université du Québec à Trois-Rivières (PAIR program). The contribution of BP was supported by the Fonds de recherche du Québec en Santé (FRQS) and the Fondation Chiropratique du Québec. The contribution of TP was supported by the Natural Sciences and Engineering Research Council of Canada. The contribution of NE and EK was supported by the Fondation Chiropratique du Québec. The contribution of MP was supported by the FRQS.
Materials
| Name | Company | Catalog Number | Comments |
| Aerrane (isoflurane, USP) - Veterinary Use Only | Baxter | NDC 10019-773-60 | Inhalation Anaesthetic ; DIN 02225875, for inducing anasthesia |
| Complete Freund Adjuvant (CFA) | Fisher Scientific | #77140 | Water-in-oil emulsion of Complete Freund Adjuvant (CFA) with killed cells of Mycobacterium butyricum. |
| Male Wistar Rats | Charles River Laboratories | body weight: 320–450 g; age: 18-22 weeks. | |
| Penlon Sigma Delta Vaporizer | Penlon | 990-VI5K-SVEEK | Penlon Sigma Delta Vaporizer used for anasthesia |
| Sharpie Permanent Marker | Sharpie | BC23636 | Permanent Marker, Fine Point, Black |
| Test cage | Custom-made | Width: 20 cm; Length: 50 cm; Height from the bottom to the top: 40 cm; Height from the bottom mesh to the top of the cage: 7 cm; Wall thickness: 5 mm; Mesh: 1 mm wire with an 8 mm inter-wire distance | |
| Von Frey Filaments | Aesthesio, Precise Tactile Sensory Evaluator | 514000-20C | Filaments from 0.07 g to 26 g |
| Wahl Professional Animal, ARCO Cordless Pet Clipper, Trimmer Grooming | Wahl | Kit #8786-1201 | Animal hair trimmer, for shaving purposes, zero blade |
References
- Hartvigsen, J., et al. What low back pain is and why we need to pay attention. Lancet. 391 (10137), 2356-2367 (2018).
- Manchikanti, L., Singh, V., Falco, F. J., Benyamin, R. M., Hirsch, J. A. Epidemiology of low back pain in adults. Neuromodulation. 17, Suppl 2 3-10 (2014).
- Urits, I., et al. Low back pain, a comprehensive review: Pathophysiology, diagnosis, and treatment. Current Pain and Headache Reports. 23 (3), 23 (2019).
- James, S. L., et al. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 392 (10159), 1789-1858 (2018).
- Hoy, D., et al. A systematic review of the global prevalence of low back pain. Arthritis & Rheumatology. 64 (6), 2028-2037 (2012).
- Shi, C., et al. Animal models for studying the etiology and treatment of low back pain. Journal of Orthopaedic Research. 36 (5), 1305-1312 (2018).
- Olmarker, K. Puncture of a lumbar intervertebral disc induces changes in spontaneous pain behavior: An experimental study in rats. Spine. 33 (8), 850-855 (2008).
- Deuis, J. R., Dvorakova, L. S., Vetter, I. Methods used to evaluate pain behaviors in rodents. Frontiers in Molecular Neuroscience. 10, 284 (2017).
- Kawakami, M., et al. Pathomechanism of pain-related behavior produced by allografts of intervertebral disc in the rat. Spine. 21 (18), 2101-2107 (1996).
- Hu, S. -J., Xing, J. -L. An experimental model for chronic compression of dorsal root ganglion produced by intervertebral foramen stenosis in the rat. Pain. 77 (1), 15-23 (1998).
- Xie, W. R., et al. Robust increase of cutaneous sensitivity, cytokine production and sympathetic sprouting in rats with localized inflammatory irritation of the spinal ganglia. Neuroscience. 142 (3), 809-822 (2006).
- Arthritis and Rheumatism. Characterization of a new animal model for evaluation and treatment of back pain due to lumbar facet joint osteoarthritis. Arthritis and Rheumatism. 63 (10), 2966-2973 (2011).
- Kobayashi, Y., Sekiguchi, M., Konno, S. -I., Kikuchi, S. -I. Increased intramuscular pressure in lumbar paraspinal muscles and low back pain: Model development and expression of substance P in the dorsal root ganglion. Spine. 35 (15), 1423-1428 (2010).
- Touj, S., et al. Sympathetic regulation and anterior cingulate cortex volume are altered in a rat model of chronic back pain. Neuroscience. 352, 9-18 (2017).
- Chaplan, S. R., Bach, F. W., Pogrel, J. W., Chung, J. M., Yaksh, T. L. Quantitative assessment of tactile allodynia in the rat paw. Journal of Neuroscience Methods. 53 (1), 55-63 (1994).
- Deuis, J. R., et al. Analgesic effects of clinically used compounds in novel mouse models of polyneuropathy induced by oxaliplatin and cisplatin. Neuro-Oncology. 16 (10), 1324-1332 (2014).
- Randall, L. O., Selitto, J. J. A method for measurement of analgesic activity on inflamed tissue. Archives Internationales de Pharmacodynamie et de Therapie. 111 (4), 409-419 (1957).
- D'Amour, F. E., Smith, D. L. A method for determining loss of pain sensation. Journal of Pharmacology and Experimental Therapeutics. 72 (1), 74-79 (1941).
- Woolfe, G. The evaluation of the analgesic actions of pethidine hydrochlodide (Demerol). Journal of Pharmacology and Experimental Therapeutics. 80 (3), 300-307 (1944).
- Hargreaves, K., Dubner, R., Brown, F., Flores, C., Joris, J. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain. 32 (1), 77-88 (1988).
- Deuis, J. R., Vetter, I. The thermal probe test: A novel behavioral assay to quantify thermal paw withdrawal thresholds in mice. Temperature. 3 (2), 199-207 (2016).
- Allchorne, A. J., Broom, D. C., Woolf, C. J. Detection of cold pain, cold allodynia and cold hyperalgesia in freely behaving rats. Molecular Pain. 1, 36 (2005).
- Carlton, S. M., Lekan, H. A., Kim, S. H., Chung, J. M. Behavioral manifestations of an experimental model for peripheral neuropathy produced by spinal nerve ligation in the primate. Pain. 56 (2), 155-166 (1994).
- Brenner, D. S., Golden, J. P., Gereau, R. W. I. V. A novel behavioral assay for measuring cold sensation in mice. PLoS One. 7 (6), 39765 (2012).
- Moqrich, A., et al. Impaired thermosensation in mice lacking TRPV3, a heat and camphor sensor in the skin. Science. 307 (5714), 1468-1472 (2005).
- Langford, D. J., et al. Coding of facial expressions of pain in the laboratory mouse. Nature Methods. 7 (6), 447-449 (2010).
- Deacon, R. M. J. Burrowing in rodents: a sensitive method for detecting behavioral dysfunction. Nature Protocols. 1 (1), 118-121 (2006).
- Griffioen, M. A., et al. Evaluation of dynamic weight bearing for measuring nonevoked inflammatory hyperalgesia in mice. Nursing Research. 64 (2), 81-87 (2015).
- Brodkin, J., et al. Validation and implementation of a novel high-throughput behavioral phenotyping instrument for mice. Journal of Neuroscience Methods. 224, 48-57 (2014).
- Paquette, T., Eskandari, N., Leblond, H., Piché, M. Spinal neurovascular coupling is preserved despite time dependent alterations of spinal cord blood flow responses in a rat model of chronic back pain: implications for functional spinal cord imaging. Pain. , (2022).
- Tokunaga, R., et al. Attenuation of widespread hypersensitivity to noxious mechanical stimuli by inhibition of GABAergic neurons of the right amygdala in a rat model of chronic back pain. European Journal of Pain. 26 (4), 911-928 (2022).
- Dixon, W. J. Efficient analysis of experimental observations. Annual Review of Pharmacology and Toxicology. 20, 441-462 (1980).
