138 • Alternatives to Animal Use in Research, Testing, and Education 
multiple, relatively simple operations. On the other 
hand, there are major differences: 
• computers have larger and many fewer com- 
ponents than the brain; 
• computer operations occur with much greater 
speed than neural operations; 
• computers operate through sequential proc- 
essing; and 
• computers attend to all input, while the brain 
is selectively attentive. 
Such differences do not, however, preclude the 
use of computers as functional models (170). More- 
over, one advantage of computers is that they de- 
mand rigorous and exact description. And an in- 
vestigator need not invoke a variety of hypothetical 
or mentalistic variables (e.g., hope, fear, desire, 
and intention) to describe their functioning (171). 
It is important to note that any predictions gen- 
erated by a computer simulation must be tested 
and verified using the system the computer was 
designed to replace (149). In this sense, the use 
of animals in behavioral research is likely to con- 
tinue in lockstep with the development of com- 
puter simulation software. 
One commentator summarized the use of com- 
puters as an alternative to animals in behavioral 
research in this way (114): 
At the present time and for the foreseeable fu- 
ture it seems clear that the computer will not be 
a feasible substitute for experiments on animals. 
The fundamental reason is that a computer can- 
not acquire data other than those that are gener- 
ated by carefully designed experimental studies 
in animals. What the computer does provide is 
a superb technique for processing vast amounts 
of data with great speed and accuracy and for 
presenting them in almost any manner the inves- 
tigator desires. To suggest that enough data are 
already available from previous work, so that 
from them programs can be generated and sub- 
jected to a variety of permutations that would lead 
to new insights, overlooks an important fact. In 
any animal experiment there are numerous vari- 
ables over which we have little control, and there 
are virtually always as many more about which 
we as yet know nothing but which may have very 
significant influences on the phenomenon under 
investigation. In real life, which after all is what 
matters in biologic research, these variables may 
be crucial and may give important clues to en- 
tirely unsuspected phenomena that are some- 
times far more important than the original sub- 
ject of the study. In a word, computers do not 
generate new concepts or acquire new data. They 
process data and permit the investigator to view 
it in more manageable or novel ways, and this 
may facilitate new hypotheses or insights. 
In summary, computer simulation can serve to 
facilitate behavioral research. The need for cer- 
tain protocols may be precluded, or protocols may 
be refocused by computer simulation before they 
commence. Modeling techniques using computer 
simulation lead to the refinement of experimental 
protocols to be conducted on animals (36). Yet as 
the preceding quote implies, in facilitating be- 
havioral research computer simulation may actu- 
ally increase, rather than decrease, the use of ani- 
mals because data can be analyzed more quickly 
and in much greater detail, leading to proportion- 
ately more hypotheses to investigate (87,160). 
SUMMARY AND CONCLUSIONS 
Animal use in research can be modified in a num- 
ber of ways, including strengthening experimental 
design to minimize the number of animals used, 
reducing the degree of experimental insult, and 
substituting one species for another. The outright 
replacement of animals with nonanimal methods 
in research is not at hand, and, because of the na- 
ture of biomedical and behavioral research, in 
many instances it is not likely to become feasible. 
Advances in instrumentation are critical to the 
more refined or reduced use of live animals or liv- 
ing material. In the past decade, practically every 
piece of instrumentation in biomedical laboratories 
has been adapted to handle “micro” samples or 
has been replaced by new microtechnology. The 
use of small samples for analysis by mass spec- 
trometry and by gas or liquid chromatography 
leads to less invasive technology. Fiber optics, for 
