Online Issues

<< All Back-issues

<< This Issue's Table of Contents

ILAR Journal V38(2) 1997
The Role of Computational Models in Animal Research

Educational Simulation Models in the Biomedical Sciences
Adrian Smith, Richard Fosse, David Dewhurst, and Karina Smith
Adrian J. Smith, Ph.D., is Professor in Laboratory Animal Health, Laboratory Animal Unit, Norwegian College of Veterinary Medicine, Oslo, Norway. David G. Dewhurst, Ph.D., is Professor of Health Sciences, Faculty of Health and Social Care, Leeds Metropolitan University, Leeds, United Kingdom. Richard T. Fosse, Dr. Sci. Vet., is Director, Laboratory Animal Veterinary Services, Vivarium, Medical Research Centre, University of Bergen, Haukeland Hospital, Norway. Karina Smith is a registered nurse and a consultant to the Laboratory Animal Unit, Norwegian College of Veterinary Medicine, Oslo, Norway.

INTRODUCTION

The prime goal of education in the biomedical sciences is to pass on prior knowledge, often obtained from animal experimentation, from teacher to student. This includes the development of practical skills. Indeed, this was often how our understanding of biological processes was originally attained: by designing and carrying out experiments on sentient animals, and then analyzing and discussing the results.

For the students, however, this knowledge is new, often phrased in unfamiliar technical terms, and may have ethical overtones. They frequently question the use of animals in class to demonstrate facts that have been proven by previous animal research and can be easily gleaned from a textbook. Furthermore, not all students require training in manual dexterity; many courses are designed to teach principles rather than to give practical experience in animal handling.

Many centers of learning are moving towards alternatives to animal experiments, for the ethical reasons mentioned above and for reasons of economy such as the high costs of technical support, laboratory equipment, and dedicated laboratory space, as well as the recurrent expenditure on consumables (such as animals, reagents, and disposable apparatus).

Microcomputers have been widely used for some time in biomedical education for word-processing, spreadsheet programs, databases, and statistical packages. Since the mid-1980s, a large number of computer models have been launched as potential alternatives to the use of animal experiments. The aim of the present paper is to assess the ability of computers to model real-life situations and thereby replace or reduce the use of animals or animal tissue in practical labs. Rather than presenting an exhaustive review of previous literature (for example, Model and others [1983], Dubin and others [1994]), we have chosen to illustrate the issues with examples from our own experience with computer modeling in biomedical education.

WHICH PRODUCTS ARE AVAILABLE?

Computer models for education in the biosciences may be broadly divided into 3 categories: pure computer software, integrated courseware, and virtual reality.

Computer Software

These products may be presented in any combination of text, graphics, sound, animations, and digitalized video. They may basically be divided into 2 types: emulations and simulations.

Emulations concentrate more on teaching principles, rather than the absolute response to a given stimulus. They have often a high degree of realism but lack quantitative accuracy. One of the earliest, and still one of the most elegant computer programs containing emulations is Pharmatutorİ1, devised by Daniel Keller (Keller 1987). Pharmatutor emulates the effects of several drugs on the cardiovascular system. The qualitative effects of epinephrine (adrenaline) on heart rate and blood pressure can be demonstrated (Figure 1).

Simulations aim to depict scenarios that are as close to the real-life situation as possible, with quantitatively correct data. Examples of these include demonstrations of pharmacokinetics, where the student can adjust the parameters in the model (such as body weight, circulating blood volume, site of administration, dose, and dosing interval) to simulate a given animal species. Computer simulations may be broadly divided into 2 categories.

1. Models of animal experiments. These programs use data previously obtained from actual in vivo experiments to generate realistic simulated tissue responses. The biological variation seen in the real experiment, from which the data were derived, is precisely reproduced in an interactive software program. Sections describing the preparation, the apparatus, and experimental method are frequently included, often with high-resolution graphics.

One example of this type of computer model is the Rat Blood Pressureİ program2 (Dewhurst and others 1996), which demonstrates the effects of a variety of pharmacological agents and procedures on blood pressure and heart rate in the anesthetized rat (Figure 2). Test substances are administered via the jugular vein while carotid artery pressure and heart rate are recorded. There are 16 experiments available, each of which is accompanied by a short protocol giving details of drug doses, order of drug administration, and student tasks.

2. Models of animal preparations. This type of program allows the student to design investigative experiments on a simulated animal preparation. In contrast to models of animal experiments, the emphasis in these simulations is often to let the student learn by trial and error. Since the experiment is performed on a computer, they are free to "administer'' drug doses and combinations that would kill a real animal. The constraints normally imposed by the passing of time have also been removed: the student can "wash out" drugs instantaneously, leaving them with a new simulation, or the experiment can be paused, and the experimental variables changed if desired.

An example of this type of computer model is the Guinea Pig Ileum3 (Dewhurst and Meehan 1989), which may be used to investigate the pharmacology of the enteric nervous system (Figure 3). The program offers little help to the students, which leads to a desirable freedom of experimental design, but also can be counterproductive if students are disheartened by early failures (if, for example, they select unrealistic drug doses). These programs are more successful if students are guided by a schedule designed by the teacher.

Integrated Courseware

Many computer programs are criticized for their lack of realism. This criticism has been met by the development of interactions between computers and other audiovisual media such as video, laser disks, and CD-ROM.

The medium that has proved to be of most use so far is the storage of video film, or still photographs such as slides, on laserdisks (also known as videodisks). Laserdisks resemble large CDs. Each image or film frame is numbered, which, coupled with the inherent accuracy and speed of a laser, enables extremely rapid searching and accurate retrieval of information The video signal can either be presented on a separate screen or, if an overlay card is installed in the computer, it can be combined with computer-generated text and graphics on one monitor. The program is inter-active--students participate actively by answering questions or making decisions that affect the outcome--and the presentations are realistic because they show color images from real life. Film sequences or individual images can be accessed repeatedly and as quickly as desired, with virtually no waiting time (access time is usually 1 to 2 seconds, compared with the many minutes it may take to spool a conventional video film).

Playback from the videodisk is usually controlled by the computer. However, to keep costs down, the videodisk player can be steered by barcodes that are read by a light pen. Barcodes can be easily and quickly generated prior to the tutorial session using computer software.

The main problem with this technology is that interactive video has proved to be an expensive and rather limited market, due to the high hardware costs. Its future as an extension of computational modeling in animal classes is therefore uncertain. The role of videodisks will in all probability be taken over by CD-ROM technology in the future.

Virtual Reality

Virtual reality, as its name suggests, creates images that delude the observer into believing that he or she is in "another world". Usually, the operator wears a special helmet in which the computer images are projected, using small monitors at eye level, into the field of vision. Sound is provided by stereo headphones. All sensory input from the "real world" is effectively cut off. The feeling of participating within the virtual world can be further enhanced by the use of gloves carrying sensors that analyze the position of the hands and fingers and in turn use this information to adjust the visual image.

In another variation, used to train laparoscopy techniques, the operator manipulates surgical instruments attached to a body simulator, while observing the results on a monitor. A feeling of physical resistance can be achieved by mechanical devices within the simulator, and the effects of operator actions, such as hemorrhage, can be displayed.

Progress is also being made in the development of programmable mannequins. An example of these is "Harvey", from the University of Florida (Waugh and others 1995). "Harvey" can be used to practice chest palpation, pulse measurement, auscultation of heart sounds, and EKG recording. The software can be programmed for a variety of normal and pathologic cardiovascular conditions.

Needless to say, the technology for virtual reality is still prohibitively expensive for most situations, but there are, however, several areas of bioscience where the technique could have great value.

AVAILABILITY OF COMPUTATIONAL MODELS

Even the most realistic techniques are of little value if, as is often the case, knowledge of their existence is limited. To make matters more difficult, personal computers have made it relatively easy to develop courseware so that, in addition to traditional software companies, individuals and academic institutions are creating computer models. The standard of the program, availability of technical support, reliability of the program, freedom from programming errors, production of upgrades, and degree of advertising therefore varies markedly from product to product.

Clearly, there is a need for systematic efforts to collect and disseminate knowledge about these products as widely as possible. One such project is the NORINA database (Smith and Smith 1996). NORINA (A Norwegian Inventory of Alternatives) provides information on several thousand audiovisual aids and computer programs from around the world that may be used as alternatives or supplements to animals in biomedical teaching. The whole of NORINA is available online on the WorldWideWeb (http://oslovet.veths.no/NORINA), enabling searches to be carried out easily and without charge by anyone with access to the Internet. Individual copies for installation on personal computers or institutional networks are also available.4 Information on other similar resources can be found at <http://oslovet.veths.no/ databases.html>.

The coordination of work on information databases was recently the topic of an ECVAM workshop. ECVAM (European Centre for the Validation of Alternative Methods) was established by the European Commission to coordinate efforts to validate alternative test methods in the biosciences. The workshop results were presented at the 2nd World Congress on Alternatives and Animal Use in the Life Sciences, held in Utrecht, The Netherlands in October 1996. The workshop's main result was to recommend the establishment of a reference center for dissemination of information on alternatives to animal experiments.

WHEN SHOULD COMPUTER MODELS BE USED?

Computer models do not have to be used merely as replacements of animal experiments. They are also valuable as illustrations in discussing animal experiments step-by-step. Pitfalls may be discussed prior to the experiment, thus sparing unnecessary suffering or use of animals. Equally, lessons learned by individual students or the class as a whole may be illustrated in class discussions. In both instances, such an audiovisual demonstration may well be enhanced by frequent pauses to print out screen images depicting events on which the tutor or students wish to focus.

PEDAGOGICAL ASPECTS OF COMPUTER MODELS

Computers have an inherent tendency either to fascinate or repel newcomers, depending upon the user's prior interest in modern technology. Knowing that their use clearly is saving animal lives will, fortunately, assist some skeptics in adopting computers in the classroom. There is, however, a great deal that can be done to help smoothly introduce technology into a teaching situation:

· Computer models should have a simple, user-friendly interface and they should be intuitive. Both the tutor and student must feel at ease using the entire computer system (both hardware and software). If this is not the case, the mechanics of using the computer will assume proportions that at best overshadow and frequently even erase the true message of the exercise, namely to use models merely as a means to better understand the biological process. Although many students will invariably be highly motivated since they recognize that they are saving animal lives by participating in a computer lab, this interest will soon be extinguished if they run into a succession of problems.
POTENTIAL PROBLEMS ASSOCIATED WITH THE USE OF COMPUTER-BASED TEACHING METHODS

The rapid development of faster and more effective hardware by computer manufacturers may act as an incentive to update computer labs faster than is strictly necessary for the purpose of using animal alternatives technology. Likewise there is a temptation when designing courseware to incorporate special effects that, although they improve the realism of the model, are counterproductive because they delay courseware use until it is possible to finance more powerful hardware. This should be borne in mind by those designing software intended to replace animal experiments.

Those institutions that are most in need of computational models as replacements for animal labs have often very limited resources to develop their own courseware. This can result in the need to use material from other centers that does not match the exact needs of the institution or may even show procedures that are no longer considered acceptable. In addition, there are national differences in standards for video signals and electric voltages that need to be taken into account.

A disadvantage of the use of computers that is less frequently recognized is that work with animals introduces the student to an additional range of issues not generally apparent in computer emulations or simulations. These include the housing and management of laboratory animals; handling and restraint; drug administration techniques; and anesthesia, analgesia, and euthanasia. Within all these areas there is opportunity for discussion of ethical and legislative principles, and these discussions are undoubtedly of great benefit in increasing the student's awareness of the discipline of modern laboratory animal science as a vital adjunct to biomedical research.

Finally, students seem to have a greater tendency to believe the results of a computer model, whereas unexpected findings in an animal class are often investigated to eliminate causes such as equipment failure or poor technique.

ADVANTAGES OF COMPUTER-BASED TEACHING METHODS

In their classical work, Russell and Burch (1959) described the three Rs of animal use (reduction, replacement, and refinement), which have become the cornerstone of modern laboratory animal science. Computer modeling can significantly reduce and replace the number of animals used in biomedical education. Refinement can also be achieved by using computer software as a tool before and after laboratory practicaIs to discuss potential pitfalls that could cause the animal to suffer thus avoiding these procedures with the actual animal.

Computer simulations can successfully achieve many of the learning objectives of laboratory animal classes at a fraction of the cost. Guy and Frisby (1992) found no difference in test marks between groups of medical students learning anatomy either by using traditional cadaver dissections or interactive video. Furthermore, Dewhurst and others (1994) demonstrated that use of a computer simulation was approximately 10 times cheaper than the equivalent animal experiment. The students achieved the majority of the learning objectives of the lab class. The resource analysis took into account consumables (animals, reagents, disposable apparatus) and staff resources (academic staff time, technician time, demonstrator time, and glass-washing time). The capital cost of equipment for both groups was ignored in this study.

It is possible to simulate treatments that provide excellent illustrations of the topic of study, but which would be unacceptable or illegal to perform on real animals. For example, toxicological simulations can be demonstrated using drugs that could evoke pain, suffering, or death in real animals. Furthermore, simulations can be made of human experiments. For students in fields such as medicine, nursing, and nutrition, computer models thereby provide direct information on human function, rather than information from an animal which in turn is a model for humans.

The ability of computers to handle large amounts of data efficiently and rapidly takes the monotony out of many class situations and frees time and mental concentration for more fruitful work. Likewise, students can work their way through a computational model at their leisure, in a speed and location that suits them, and with ample opportunity to repeat an exercise until they feel they have grasped the principles involved.

EVALUATION OF THE IMPACT OF COMPUTER MODELING IN EDUCATION

Any alternative approach to education ought to be compared critically with existing teaching methods. This entails an evaluation of the ability of computer simulations or emulations to achieve the learning objectives of the practical class to be replaced. There have been few such evaluations within the biomedical sciences, a problem exacerbated by the fact that learning objectives are often poorly defined and rarely made explicit to students. Modern laboratory practical classes in biomedicine aim to achieve far more than a simple transfer of information about one aspect of, say, physiology or pharmacology. Such aims may be:

It is obvious that the 2nd of these aims cannot be achieved by means of computer models. However, for many students (the majority in many fields of study) it is questionable whether this is a major loss. The other aims can still be fulfilled with computational modeling. Indeed, some universities specifically use computer simulations to train students who will later perform animal experiments; they may try out drug or dose combinations or ascertain appropriate stimulus parameters.

Retrospective studies have demonstrated no difference in student performance, measured in terms of the quality of their practical reports, between similar groups of students taught either in practical classes using animals or using computer alternatives (Dewhurst and others 1988). One study has also shown that students using computer simulations learn better when working in groups than when working alone (Dewhurst and Meehan 1993). Other studies have attempted to compare the effectiveness of these 2 teaching forms by assessing:

Fawver and others (1990) found no difference in knowledge gain either among medical or veterinary students.

The impact of computer-based learning on the numbers of animals used in university teaching has been investigated in several countries by means of a questionnaire (Dewhurst and Jenkinson 1995). This study focused on academic staff perceptions of the usability of computer learning packages in physiology and pharmacology, and whether these programs have in fact led to a reduction in the numbers of animals used. Most packages were used in a staff-supervised learning situation, either to replace or support a practical class. Their use saved academic and nonacademic staff time, they were considered to be less expensive, and they were claimed to be an effective and enjoyable means of student learning. It was also evident that their use had resulted in a significant reduction in animal use.

CONCLUSIONS

Computer modeling can undoubtedly contribute to the implementation of the three Rs of Russell and Burch, when used intelligently. The mathematician Searle has said, "Learn what to compute and when to compute before you learn how to compute" (Searle 1989). Whether the educator already possesses computer skills or not, it is vital that instructors analyze their own needs and the actual potential for replacing animals or their products, before the latest computer techniques are employed.

Computer models, like all other audiovisual aids, are merely a means to an end, not the end in themselves. When that has been said, modern computer technology offers tremendous room for implementation of the three Rs that form the backbone of modern laboratory animal science.

1 Available from Fonds fur Versuchstierfreie Forschung, Biberlinsstrasse 5, 8032 Ziirich, Switzerland. Tel: (41) 1-42-33-70-70; Fax: (41) 1-42-22-80-10.
2 Available for 199 from Dr. David Dewhurst, Sheffield BioScience Programs, 5 Woodlands Green, Harrogate, North Yorkshire, HG2 8QD, England. Tel: (44)1-42-388-8514; E-mail: d.dewhurst@lmu.ac.uk.
3 Available for 120 from Dr. David Dewhurst, Sheffield BioScience Programs, 5 Woodlands Green, Harrogate, North Yorkshire, HG2 8QD, England. Tel: (44)1-42-388-8514; Emaih d. Dewhurst@lmu.ac.uk
4 For further details contact Professor Adrian Smith, Laboratory Animal Unit, Norwegian College of Veterinary Medicine, P.O. Box 8146 Dep., N-0033 Oslo, Norway. Tel: (47) 22 96 45 74; Fax: (47)22 96 45 35; E-mail: adrian.smith @veths.no.

REFERENCES

Dewhurst DG, Meehan AS. 1989. Computer simulation of the effects of drugs on neurotransmitter release in the enteric nervous system. Br J Pharmacol Proc Suppl 97:597P.

Dewhurst DG, Meehan AS. 1993. Evaluation of the use of computer simulations of experiments in teaching undergraduate students. Br J Pharmacol Proc Suppl 108:238P.

Dewhurst DG, Jenkinson L. 1995. The impact of computer-based alternatives on the use of animals in undergraduate teaching. ATLA 23:521-530.

Dewhurst DG, Brown GJ, Meehan AS. 1988. Microcomputer simulations of laboratory experiments in physiology. ATLA 15:280-289.

Dewhurst DG, Hardcastle J, Hardcastle PT, Stuart E. 1994. Comparison of a computer simulation program with a traditional laboratory practical class for teaching thc principles of intestinal absorption. Am J Physiol (Adv Physiol Ed) 12(1):S95-S103.

Dewhurst DG, Hughes IE, Williams AD. 1996. An interactive computer program to replace in vivo experiments on rat blood pressure for teaching undergraduate students. ATLA 24:707-714.

Dubin S, Zietz S, Naim K. 1994. Computer models as an alternative in comparative medicine. Lab Animal 23:44-47.

ECVAM [European Centre for the Validation of Alternative Methods]. 1996. From a workshop, Neubiberg, Germany, September 12-15. Conclusions presented at the 2nd World Congress on Animal Alternatives, Utrecht, The Netherlands, October 20-24.

Fawver AL, Branch CE, Trentham L, Robertson BT, Beckeft SD. 1990. A comparison of interactive videodisc instruction with live animal laboratories. Am J Physiol (Adv Physiol Ed) 259(4):S11-S14.

Guy JF, Frisby AJ. 1992. Using interactive videodiscs to teach gross anatomy to undergraduates at The Ohio State University. Acad Med 67:132-133.

Keller D. 1987. Pharmakologie-Unterricht am Computer. In: Alternativen zu Tierexperimenten, December 1987, Fonds Fiir Versuchstierfeic Forschung (FFVFF), Zurich, Switzerland. Vol 4, p 5-11.

Modell HI, Olszowka AJ, Piewes JL, Farhi LE. 1983. Role of computer graphics in simulations for teaching physiology. Physiologist 26: 93-95.

Russell WMS, Burch RL. 1959. The Principles of Humane Experimental Technique. (Special Edition). Herts, England: Universities Federation for Animal Welfare.

Searle SR. 1989. Statistical computing packages: Some words of caution. Am Statist 43:189-190.

Smith AJ, Smith K. 1996. The NORINA database of audiovisual alternatives. Proceedings from the 2nd World Congress on Alternatives and Animal Use in the Life Sciences, Utrecht, The Netherlands, 20-24 October.

Waugh RA, Mayer JW, Ewy GA, Felner JM, Issenberg BS, Gessner LH, Rich S, Sajid AW, Safford RE. 1995. Multimedia computer-assisted instruction in cardiology. Arch Intern Med 155:197-203.



FIGURE 1 Screen display from the Pharmatutor program. The figure shows an emulation of an experiment where the effect of epinephrine (adrenaline) on blood pressure in an anesthetized rat is studied. The pressure recorder in the bottom right corner of the figure produces a trace that scrolls to the left across the screen. A menu bar (not shown) allows different combinations of drugs (including blockers) to be added, the experiment to be halted at any time, and the screen image to be printed.



FIGURE 2 Screen display from the Rat Blood Pressure program. The figure shows simulated data of heart rate (upper trace) and carotid artery pressure (lower trace), recorded from the anesthetized rat in response to administration of isoprenaline (1 nm/kg iv). Students may use a crosshair cursor to take accurate measurements from the traces. They can access the experiment schedule or 2 types of student task: a traditional lab report or a series of 5 multiple-choice questions with feedback related to the experiment.



FIGURE 3 Screen display from the Guinea Pig Ileum program1. The figure shows simulated responses of the isolated guinea pig ileum to different concentrations of acetylcholine added to the tissue bath. The dose shown produced the final response which is a maximum contraction. The display scrolls in real time to allow the student to see the characteristics of the response in more detail, that is, it shows the fast contraction and slower relaxation of the gut. The menu bar indicates the tools the student has available: an electrical stimulator, a selection of drugs that may be added at any concentration and in any combination, and a wash facility that instantly removes the drugs from the tissue bath.





Copyright © 2008. National Academy of Sciences.
All rights reserved.
500 Fifth St. N.W., Washington, D.C. 20001.
Terms of Use and Privacy Statement