Read an Excerpt
Mouse Behavioral Testing
How to use Mice in Behavioral Neuroscience
By Douglas Wahlsten
Academic Press
Copyright © 2011 Elsevier Inc.
All right reserved.
ISBN: 978-0-12-375675-6
Chapter One
Introduction to the Research Process CHAPTER OUTLINE
The Research Process: Scientific Aspects 1 Adapting the approach to the stage of an investigation 2 Mice, mice, and more mice 4 Measuring instruments: Tests 4 The formal research design 4 Sample size 4 Ethics approval 5 Logistics 5 Obtaining subjects 5 The test day 6 Prelude to data analysis 6 Data analysis 7 Publishing results 7
The Institutional Context of Research 7 People 8 Facilities 9 Policies 11
Testing mouse behavior is done as part of a scientific investigation of some larger question, such as the genetic bases for neuropsychiatric disorders or mechanisms of memory formation. Those who begin their career in science with behavioral testing naturally focus on the fine details of the task set for them by a more experienced person. Only gradually does a larger picture form of how their work relates to other parts of the research process. This chapter sketches some of those parts and puts the tests into a context.
THE RESEARCH PROCESS: SCIENTIFIC ASPECTS
Whereas the research environment varies greatly in different countries and institutions, the actual process of doing research with mice is or should be the same everywhere. The main steps are illustrated in Figure 1.1, emphasizing things that are addressed in depth in this book. Conceptualization and literature review are governed by the existing knowledge and theories in a specific field of study. These are usually introduced to the student in an entire course such as behavioral neuroscience, behavioral genetics, psychopharmacology, or animal behavior. Through that kind of advanced study and extensive reading of the literature, the investigator becomes familiar with specialized methods of experimentation and learns what kinds of factors need to be controlled and measured in a good study. Little is said in this book about the first two steps in the process because they are thoroughly dealt with in many graduate training programs and postdoctoral settings. Several key reference works are listed at the end of this chapter for the benefit of those new to the field.
Most researchers working with mice have taken at least an introductory course in statistical data analysis. Because not everyone will have a strong background in these procedures, there is some coverage of statistical methods in Chapter 5, but no attempt is made to condense a one-term statistics course into one chapter. Statistical principles will not be grasped firmly until the investigator needs to apply them to real data, at which time a review of the basics can be helpful. This is especially true for the planning phase of a project. Many texts on statistics devote chapter after chapter to what should be done with data that have already been collected, while little is said about critical matters such as choosing the proper sample size before the experiment begins. When analysis of data is discussed here, attention is focused on issues that are common in work with mouse behavior.
Adapting the approach to the stage of an investigation
Most long-term research programs involve two stages: an exploratory stage followed by a formal test of well-specified hypotheses. The latter methodology is elaborated in many courses, whereas the former receives little attention. The exploratory stage is perhaps the most important because it can be the time when a discovery is first made or a decisive clue is found. More often than not, an important discovery occurs by accident during the course of some other experiment. Students and technicians are supposed to follow a protocol rigidly and everything is supposed to happen according to well-laid plans, but nobody tells the mice about this, and some of the mice behave unexpectedly. Good researchers take note of the exceptions and surprises. An experienced researcher may know when a surprising result is potentially very important, but this same experience can also make the investigator jaded and too willing to attribute every exception to sampling error. There is considerable benefit in having people new to the field take a fresh look at the procedures used in a behavioral test and also at the mice as they are run through their paces.
It is good practice to conduct the exploratory stage of a project without rigid protocols and controls. The pilot study allows different test parameters and procedures to be evaluated informally using just a few animals in any one condition or perhaps using repeated testing of the same animals in practice runs. This kind of study scans a wide range of possibilities in a way that will detect those with very large effects. Later, when the most important variables have been brought under control and optimized for that lab, factors with more subtle effects can be studied in a formal experiment with the stricter controls and larger samples that are needed to detect them. The principal investigator (PI) faces a special challenge in convincing an animal ethics committee to approve a pilot study. Suggestions about how to do this are offered in Chapter 6. The purpose of a pilot study is not to circumvent the ethical principles of research with animals. Those principles should be respected in any kind of study with mice, which imposes limits on what range of test parameters can be tried in a pilot study.
An open-ended exploratory study can be done in some situations. Crowcroft (1973), in his charming book Mice All Over, described a study of mice free to run loose in an entire building. At the Institute for Behavioral Genetics in Colorado, Gerald McClearn had a device constructed that he dubbed the "hypothesis generating apparatus." It was a large and complex environment populated with mice of both sexes from several inbred strains. The students and postdoctoral fellows sat in chairs and watched the antics of 5 to 10 mice for hours at a time at different phases of the light—dark cycle (dim light for the dark phase). This was an excellent way to become familiar with the range of behaviors that mice commonly express, and hypotheses flowed freely from this device (e.g., Yanai & McClearn, 1972). David Sternthal, a graduate student at the University of Waterloo, went one step further by filling a small lab room with topsoil covered by grass turf with places to build deep burrows and a hut where an outcast could find shelter away from the main group of mice. Those inbred mice that sometimes seem so dull and listless in a "shoebox" colony cage or boring behavioral tests came to life and utilized every portion of an environment that was vastly larger than the typical colony cage. A complex mouse society emerged quickly and soon the entire environment was transformed with tunnels, nests, and home bases. Hans-Peter Lipp and co-workers (Lipp, Amrein, Slomankia, Wolfer, 2007) at the University of Zürich and Moscow State University took inbred mice to the outer limits of practical research by releasing them into large outdoor pens in a rural research station north of Moscow, where they were exposed to the rigors of natural selection. To the amazement of almost everyone, inbred mice survived and some of them even thrived in the Russian winter . This was high-risk, exploratory research at its finest, because nobody knew if the study would provide any useful data.
Another source of ideas about testing mouse behavior is the Internet site YouTube. There are fascinating videos of "the smartest mouse in the world" that rapidly navigates an astoundingly complex sequence (probably aided by odor trails) over barriers and through a wide variety of devices that put anything we do in the lab to shame. There are mice on surf boards that exhibit the notorious aversion of mice to water, even at the beach. At the beginning of scientific work with mice early in the twentieth century, laboratory researchers went to mouse fanciers to obtain their subjects and learn about how to breed and maintain mice (Chapter 2). Today there is still a thriving community of mouse fanciers with fascinating Web sites who keep these animals as pets and have a greater interest in behavior than fanciers 100 years ago, who were interested mainly in coat color variants.
One perfectly legal approach to exploratory work is to buy mice at a local pet shop and keep them in a basement or garage. This approach needs to respect the needs of the mice and adhere to regulations that proscribe animal cruelty. Never should a pet be made to suffer just in case it might give a scientist some prize winning idea. At the same time, all the things that are done to train cats and dogs for the show ring are reasonable for training mice and testing the limits of what they can achieve. Animal trainers have been able to induce bears to play ice hockey and porpoises to serve cocktails on silver trays at poolside. There is no good reason why a professional scientist who studies mouse behavior in the lab cannot also be an animal trainer or observer after hours, provided friends and family can live with the mouse odors.
Mice, mice, and more mice
Pets are fun to watch, but research is done with special mice developed for use in laboratories around the world. Their surprising origins are now fairly well understood. Modern lab mice are a mixture of genes from four wild Mus subspecies. Mouse fanciers in Europe, the United States, China, and Japan kept mice as pets, and those animals were convenient sources of research subjects in the twentieth century when the scientific study of mouse behavior commenced. Before long, standard inbred strains were created and mouse genetics progressed rapidly. With the advent of molecular genetic technology to sequence the genome and introduce mutations into known genes, mice became more popular than ever. Today, mouse models of human genetic disorders have become prime subjects for neuroscience research. Chapter 2 reviews mouse history and the many kinds of mice that inhabit labs today.
Measuring instruments: Tests
Behavior is a measurable phenotype (Wahlsten & Crabbe, 2007). The process of measuring it is termed a test. The test involves a physical apparatus and a protocol instructing how to conduct the trials. Virtually every kind of behavior, with the possible exception of some simple reflexes, is complex and can be measured in several ways. The test chosen for a study must pertain to the kinds of behavioral or neural processes that are the subjects of the study. The researcher needs to decide on the specific behavioral tests to employ in a study quite early in the process before there can be sample size estimation or ethics approval. Chapter 3 introduces more than 50 kinds of tests. The entire second half of the book (Chapters 10 to 14) is devoted to an in-depth discussion of many kinds of tests. Chapters on general features of behavioral tests cover a wide range of topics from ways to construct a test battery to methods for assessing reliability. Specialized techniques to provide adequate motivation and obtain good results from computer-based video tracking are explored in considerable depth. Finally, the challenge of obtaining comparable results in different labs is described, and possibilities for standardizing the tests as well as the conditions in a lab are discussed.
The formal research design
Many brilliant "discoveries" in a pilot study, accidental findings in a project aimed at studying something else, or flashes of insight while watching pets at play arise from events that cannot be repeated. Many hunches turn out to be wrong. Good ideas need to survive a rigorous, formal test before the rest of the field needs to take notice. For this purpose, the investigator must adopt a research design that is capable of proving the truth of an idea to a skeptical audience of peers. Many elaborate designs with sophisticated control groups are used for genetic studies of mouse behavior. A few of the main ones are described in Chapter 4. Design of a study is critically important for sample size determination, ethics approval, and logistics of running the experiment. Once those hurdles have been crossed, the proper execution of an elegant design is essential for obtaining good data. It is fairly easy to design a study on a piece of paper so that logic appears to govern all else. Taking that design into a real laboratory, however, soon teaches the researcher to anticipate and respect the many pitfalls and complications that beset the practical implementation of well-laid plans. It is for this reason that most of the remaining chapters address the practical side of research on mouse behavior.
Sample size
Only after the design of the study and the specific behavioral tests are chosen can the researcher rationally decide the appropriate number of mice to be tested. This step is taken before the application for ethics approval and then grant funding, because the number of animals in the study is an important aspect of external review and consent. After the data are collected and the report is submitted for publication, the author may be asked by a skeptical reviewer to justify the sample size. If the number of mice is found to be too small at that stage of the process, the ship of science will break apart on the shoals of bad methodology. It is better by far to plan ahead and take great care to get this number right. Chapter 5 is devoted to this topic, and a series of statistical utilities is provided to make the process as quick and painless as possible.
Ethics approval
Having decided upon the design, the tests, and the number of mice, permission to conduct the study must be given. This is usually done through some kind of peer review by colleagues who also work with animals or have expertise pertinent to the study. The specific rules and procedures differ greatly in different countries, and within a single country an institution may have considerable latitude in implementing ethical principles. The journal to which an article is submitted for publication may also, through its peer review process, judge the ethical aspects of a study. Official bodies sanctioned by a government agency may be given powers to pass judgment on a research proposal, and those bodies provide instructions on how to pass over the hurdles and through the hoops of regulations. The discussion in Chapter 6 does not attempt to replace or reduce the need for an investigator to engage the official mechanism. Instead, it offers a researcher's perspective on the process and advice on how to get the job done well and done quickly. Specific examples of research procedures that are considered good practice or in some cases unethical in work with mice are presented, and a scheme for rating the degree of discomfort or invasiveness for a wide range of behavioral tests is proposed.
Logistics
After funding and permissions are obtained, it is time to plan the details of execution of the experiment. This step could be taken earlier in the process, but everything could change after the sample size calculation or the ethics review. Planning means a list of which mouse is to be tested at what time on which day under what conditions, and it requires other aspects of the lab to be arranged so that the testing can progress smoothly according to protocol. This stage is complete when the researchers have produced a stack of data sheets, one per mouse, in the correct order for the entire experiment. The study should not begin until every detail of every trial for every mouse in the entire experiment has been specified.
There are two major parts to the planning process. First, the throughput for the test, how much time on how many days is required to obtain measures on just one animal, must be determined. Second, the order in which animals from the various genotypic and treatment groups are to be tested must be decided. The order should be carefully balanced and randomized so there is no preponderance of mice from any one group at any particular period of testing. The effects of test order must always be considered. Some mouse must go first or last and early or late each day. A skillful execution of test order causes variance arising from time of day and sequential effects such as odor trails to end up in the within-groups variance, rather than creating spurious between-group effects. Chapter 7 explains how to achieve this and provides several elaborate examples from real experiments.
(Continues...)
Excerpted from Mouse Behavioral Testing by Douglas Wahlsten Copyright © 2011 by Elsevier Inc.. Excerpted by permission of Academic Press. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
