Neurogenetics And Behavior Research Paper

1. Historical Varieties Of Behavioral Neurogenetic Studies

There are three identifiably different paradigms that emerged in prior behavioral neurogenetic work. The first documented use of ‘neurogenetic’ was by the World Federation of Neurology, which hosted a series of international congresses ‘for Neuro-genetics and Neuro-ophthalmology’ in the 1960s. Here, as in most present-day use by the medical research community, the term refers to discovering relationships between gene sequence variants and ‘inherited’ neurological diseases:

Just as our understanding of pathology transformed neurology in the nineteenth century, new genetic knowledge has permeated many aspects of neurology today. Indeed, genetic methods have changed our basic understanding of neurologic disease processes, and the definition of diseases themselves has undergone changes as DNA based testing is substituted for clinical, morphological or biochemical definitions as a gold standard for analysis (Pulst 2000, p. ix).

Although this medical usage of ‘neurogenetic’ is fairly recent, the work it refers to has continuity with pre-1940s research in both biology and medicine. The prominent involvement of hereditarian views (and prewar researchers such as Ernst Rudin, a pioneer of genetic psychiatry) in eugenics movements and Nazi racial and genocidal policies made the subject of genetics, the nervous system, and behavior a rather uneasy one in medical circles for a period of time after 1945. Neurogenetic approaches to behavioral pathologies experienced a resurgence in the late 1950s and early 1960s, and developed an individual identity within the larger study of hereditary disorders.

A second neurogenetic paradigm was articulated outside of medicine in the late 1940s when the developmental biologist Paul Weiss (1898–1989) organized a meeting for the new discipline of ‘Genetic Neurology.’ In the preface to an edited volume of the same name, he explained the purview of this field:

The days of March 21–25, 1949 … may at some future time be remembered as the birth date of a new biological discipline—‘genetic neurology’—birthdate in the sense that an organism, which had for some time been in the making, came to light, gave evidence of its vitality, acquired a name, and set out for a career of its own. Neuroanatomy, neurophysiology, and neuropsychology deal essentially with the nervous system in its final mature state. Genetic neurology encompasses all those processes that lead up to that state (‘neurogenesis’), maintain it in its integrity, and restore it after disruption. The attribute ‘genetic,’ accordingly, refers to ‘genesis,’ i.e., development in the widest sense, and not to ‘genetics’ in the narrower modern sense of ‘inheritance,’ although it naturally includes the latter (Weiss 1950, p. v).

This developmental paradigm has a different emphasis from the ‘medical’ one: genes are not unifying explanations, disease predictors, or gold standards for classification, but rather tools for learning about development. Behavior becomes incorporated in the ‘developmental’ paradigm in the following way. Since all genetic changes to the nervous system do not necessarily lead to readily observable anatomical, biochemical, or physiological differences in individuals, it is necessary to have some way of assessing ‘invisible’ aspects of neural function. Behavior provides one convenient assay of this type, making it possible to identify subtle changes in development and/or ongoing function of the nervous system correlated with genetic changes. Even when behavioral assays are used, the goal of the developmental agenda is typically to understand the developmental processes particular gene products participate in rather than the genesis of the ‘assay’ behaviors themselves.

The third, ‘behavioral’ paradigm resembles the first, and shares a common history with it. Here the focus is on understanding the contribution of genetic differences to non-pathological individual differences in behavior:

More recently, it has become increasingly accepted that genetic variation makes an important contribution to differences among individuals in the normal range of behavior as well as for abnormal behavior. Moreover, many behavioral disorders, especially common ones, may represent variation at the extremes of the same genetic and environmental factors that are responsible for variation within the normal range (Plomin 1999).

As in the ‘medical’ paradigm, the emphasis of the ‘behavioral’ paradigm is on genes as unifying explanations for individual differences, as predictors for such differences, and as potential avenues for manipulating the characteristics of individuals. In spite of their procedural and explanatory similarities, it is important to note that the ‘medical’ and ‘behavioral’ paradigms have different aims. The ‘medical’ paradigm is fundamentally about understanding and alleviating dysfunction. The ‘behavioral’ paradigm is about explaining, predicting (and potentially manipulating) any differences in behavior between individuals or groups. These both contrast with but do not necessarily conflict with the ‘developmental’ paradigm, which is about understanding developmental processes within individuals that have an impact on behaviors.

2. Behavior As An Object Of Neurogenetic Studies

Whatever their aims and assumptions, all neurogenetic paradigms use behavior as a measurable object to which neurobiological and genetic variation can be related. However, behavioral variation does not manifest itself in predefined units that have a direct relationship with underlying biological processes. Researchers must make decisions about how they are going to classify and measure behavior in order to make it suitable for subsequent biological analyses. Approaches to behavioral biology that grew out of different research traditions (such as medicine, experimental psychology, and zoology) have divided behavioral variation differently to suit their particular interests, and these categorization schemes continue to change over time. The most important attributes of behavioral categorization with respect to biology are whether the behavioral phenomena of interest can be measured directly or not, and whether these phenomena are measured at the level of individuals or groups of individuals.

2.1 Direct Output vs. Inferential Linkage

Two broadly defined methods of objectifying behavior for biological studies can be identified. The first, direct output, grew out of an interest in mechanistically explaining externally observable behavioral acts. What events go on in a brain when a particular behavior such as a yawn or the production of fluent speech is generated? Notice that such acts can either be narrowly or loosely defined (a particular type of yawn versus all yawning, or the generation of a particular word or class of words versus any vocal emission), so that these measures can either deal with individual acts or some population of acts. The thing to be explained is always the thing that is physically being measured. Yet there are many behavioral phenomena that are more abstract than a concrete motor act, for instance, behavioral states such as hunger and thirst, or emotional states such as happiness or sadness, or finding stimuli pleasant or distasteful. While it is sometimes possible to find a combination of externally measurable variables that will uniquely characterize a state (as in the case of sleep), it is more usual that there is no combination of behavioral measures that will mark one of these phenomena off from all others. Rather, it is a combination of both a subject’s behavior and the context in which that behavior occurs that indirectly defines this type of behavioral variation.

The second method for objectifying behavior, inferential linkage, takes its starting point from these facts. Investigators have an underlying model (not always made explicit) that defines the relationship between externally observable variables and the underlying phenomena of interest. For instance, operant conditioning paradigms use motor responses like bar pressing to study learning in non-human animals. A neurogeneticist who adopts this paradigm is not interested in the neurobiology of bar pressing per se, but in brain processes that subserve learning (particularly the influence of prior history on the brain, and what happens in the brain up to the time a decision is made to press or not press the bar). Implicitly or explicitly, such an investigator has a schema for how these learning processes are linked to bar pressing, allowing bar pressing to be a proxy for learning. The same applies to measurements of human behavior from written or oral questionnaires or tests. The thing that is to be explained is not what is directly measured. Another way of stating this is that there are intervening variables between the object being measured and the object of interest.

The distinction between these two methods is important because of the readiness with which biological variables can be causally related to behaviors. Neurogenetic studies using direct outputs are studying gene effects on the object of interest. Unfortunately, there are many behavioral phenomena that cannot be studied this way. Studies using inferential linkages are thus a necessity, but they do not provide an immediate and direct causal link between the behavioral phenomenon of interest and biological variables. Biological effects on processes that intervene between the object of study and what is actually measured must first be ruled out.

2.2 Individual Behavior vs. Group Behavior

Most behavioral scientists are used to thinking of behavior as a product of a single individual’s nervous system that can be validly measured and analyzed within that individual. For instance, one can measure the extent to which a behavioral trait occurs repeatedly in the same individual by presenting him or her with the same context multiple times, and using this set of measures as a basis for documenting differences in behavior among different individuals. In this case, the performance of each individual is preserved and individual differences can be directly assessed.

However, for the purpose of assessing population expectations in behavioral patterns, it may be desirable to measure a behavior with respect to groups rather than single individuals. In this case, individuals are classified according to some predetermined schema and the behavioral responses of different individuals in each group are measured in the same context. Individuals are typically assessed only once or a few times, and the data of interest are differences in the distribution of group scores. Individual deviations from group means (or from the grand mean of all groups) have been suggested as measures indicating individual differences. Such group measures have been very popular in neurogenetic studies, where individuals can be classified according to the identity of particular genetic variants to look for behavioral differences, or sorted according to behavior to look for genetic differences.

From the point of view of biology, individual and group measures are not equivalent. The vast majority of behaviors that are studied biologically are not collectively produced by the linked physiological and biochemical activity of many nervous systems simultaneously. Neurogeneticists who are interested in disease, in development, or in the causation of behavior all need to find out how gene variants in individual bodies make differences in developmental decisions. While accepting that there are many arenas that group measures of behavior are good and appropriate for, their application to neurogenetic research is limited. They can be useful in determining whether a correlation exists between variation in a particular gene sequence and behavioral variation at a population level, but they cannot be directly converted into the currency of developmental events that go on in individuals.

2.3 Limits On Behavioral Measurement Constrain Causal Attribution

To many scientists and non-scientists alike, it may appear that behaviors are relatively simple and straightforward to measure, while neurochemistry, neuroanatomy, gene sequences, and gene expression patterns are vastly more complicated. It is frequently argued that causal attributions in neurogenetic work are limited by an ability to quantify genetic or neurobiological variation, or limitations in the statistical techniques used to relate behavioral variation to neurobiology and genetics. In reality, just the opposite is the case. Measures of neurobiological and genetic variables are typically subjected to closer critical scrutiny than behavioral measures. They are typically also direct measures. Since so many of the behaviors people study neurogenetically can only be measured via inferential linkages, the characteristics of the behavioral variables themselves are typically the factors which most constrain causal attributions.

One example is provided by work on rodent learning that uses the amount of time it takes animals to find a platform hidden under the surface of a deep pool of water as a behavioral variable. During some number of training sessions, individual animals can presumably use a constellation of spatial cues in the room surrounding the pool to memorize where the platform is. However, there are many other factors that will determine whether animals can utilize these training opportunities in the same way, and also determine the speed with which an animal familiarized with this situation will find the platform.

To take a hypothetical illustration, animals which are more prone to panic attacks may find being tossed into a pool of water quite stressful, and may spend a good deal of time swimming around the edges of the pool trying to escape. Thus, they may not have the same learning opportunities as calmer animals, and, even if this difference in opportunity to learn is controlled for, they may still be slower because it takes them a while to calm down enough and remember that there is a platform to be looked for. Other animals may have motor problems with coordinated swimming and be hesitant to leave the side of the pool or simply be slower swimmers. There are many reasonable scenarios by which animals whose genotypes differ may show reliable differences in this behavioral task, all unrelated to their spatial learning ability. Researchers must conduct numerous tests to rule out these alternative explanations before accepting that a neurobiological or genetic linkage is with learning ability rather than with stress, motor coordination, or a host of other possible explanations.

Because there is currently no systematic way of advancing reasonable alternative explanations, or of deciding how many potential alternatives there are, this open-ended process can take a long time. When one is finally sure of a linkage between a behavioral process and genetic and/or neurobiological variation, one must then make sure that individual differences in behavior can be sufficiently well quantified to enable the biological linkages between genes, neurobiology, and behavior to be adequately explored.

3. Current Trends In Neurogenetic Studies

Study designs commonly utilized in neurogenetics can be divided according to whether they involve correlations between behavior and genetic variation based on classifications of subjects into groups, or study of the behavioral outcome of controlled genetic and/or neurobiological manipulations. Correlative studies can be further subdivided by whether they classify subjects by gene variants and examine behavioral differences, or whether they classify subjects by behavioral variants and examine genetic differences. Manipulative studies can directly alter single genes (gene deletions, gene insertions, and changes in gene expression), groups of genes (inactivation or loss of parts of, or whole, chromosomes), and/or manipulations on brain cell groups (substitutions, deletions, and chemical alterations of function).

3.1 Correlative Studies: Candidate Behaviors And Candidate Genes

Correlative designs currently dominate neurogenetic studies in both humans and animals. In the first of these study designs, individuals are classified according to some behavioral attribute and the genomes of individuals in the different groups are searched for sequence variation that correlates with group membership. For example, individuals are administered a psychological or IQ test, and a breakpoint is determined whereby the subjects are divided into ‘highscore’ and ‘low-score’ groups. The DNA sequences of group members are subsequently searched for sequence variants that statistically differ according to group membership. The sequence variants can either be part of a DNA sequence coding for a gene product, part of a DNA sequence that either controls the expression of a nearby gene or contributes to the local 3-dimensional geometry of a stretch of DNA, or part of a sequence that has no identifiable function. A common variant of this procedure is to take extremes of these two groups, say subjects who scored in the upper and lower 10% of the test distribution, and compare their DNA.

These approaches are sometimes referred to as ‘quantitative trait loci’ (QTL) approaches because when one takes group measures of the behavioral performances one can generate estimates of the amount of behavioral variation in the population that is removed (‘explained’) by variation in particular gene sequences. Such estimates are not necessarily very informative: different genes may affect behavioral traits through the same biological pathways, so one does not expect the variance contributions of all genes to add to 100%. The point of these approaches is not to generate tallies of explained variance, but rather to discover genes that putatively influence the behavioral trait of interest.

A more classical research design used in animal studies is to induce genetic changes in reproductive cells of parents, mate them, and use a particular behavior as a ‘screen’ for animals containing gene sequence changes that affect the criterion behavior. While formerly these approaches relied upon chemical mutagens or radiation-induced damage to chromosomes, it is now possible via molecular biological techniques to target changes to particular genes, insert these into cells in embryos, and implant the embryos in foster mothers (‘transgenic’ techniques). When such a design is applied with a particular behavior as the object of study, it becomes a variant of the ‘candidate behavior’ approach.

The second type of correlative study is quite literally the inverse of the first design. Individuals are classified by which sequence variants of particular genes they possess, and are measured for one or a collection of behavioral attributes. Group behavioral measures are then screened for covariation with group membership. In this method one typically knows something about the molecule whose gene sequence is used for classification, and where in the brain this molecule is typically found. Educated guesses can be made about the kinds of developmental and cellular processes this molecule might participate in, and hence the range of behaviors one might expect to be affected by structural variants in the molecule. This approach is potentially capable of providing faster reliable links between biology and behavior than the candidate behavior approach.

3.2 Caveats Governing The Interpretation Of Correlative Studies

Correlative studies have an appealingly simple design that belies the complexity of their interpretation. Three major interpretational difficulties are:

(a) The illusion of homogeneity in group composition. Correlation designs can give the illusion of a ‘one-to-one’ correspondence between trait and behavioral variation when this may not be the case. Because one has chosen variation in a limited number of gene sequences or behavioral traits as a basis for group classification, this does not mean that groups could not also systematically differ in other gene sequences and other behavioral traits. This problem would cause correlative studies to produce spurious positive correlations between a particular genetic variant and a particular behavioral variant. The same problem could also cause investigators to underestimate the complexity of the genetic system contributing to behavioral variation significantly, since an effect due to multiple co-varying genes would be attributed to a single gene. The candidate behavior approach is particularly prone to underestimating the complexity of a gene’s effects in development via some fundamental cellular process, because it focuses on linkages between a single gene and a single behavior. This may cause investigators to misconstrue seriously the developmental and functional pathways that a particular gene contributes to, if they fail to arrive at a true accounting of the multiple co-varying behavioral systems a gene product may affect.

(b) Non-genetic explanations for the correlation. There may also be correlations between non-genetic biological or experiential factors and group membership. It is frequently the case that correlational designs are applied to groups of subjects with uncontrolled (and unknown) prior developmental histories. Thus, while there may indeed be correlations between genetic and behavioral variables, the genetic variation may be causally unrelated to the behavioral variation.

(c) Misidentification mislabeling of the behavioral dimension that co-varies with gene sequences. As described in Sect. 2.3, when using behavioral measures that depend on inferential linkages in a candidate behavior approach, it is possible to misidentify the domain specificity of a behavioral effect completely if only one or a small number of similar behavioral measures are employed. This misidentification has serious consequences in a biological research program, where domain assignments (for instance, ‘learning’ as versus ‘fear responses’) have implications for which specific brain regions and biochemical pathways one implicates in a behavioral difference, based on the historical body of neurobiological work.

Correlational studies generate possible linkages between particular genes and particular behaviors, but cannot be said to constitute a demonstration of such a linkage. Because they are the only neurogenetic studies possible in humans, they will undoubtedly remain a staple of research in the medical and behavioral paradigms. It is important that their conclusions be constrained by their limitations.

3.3 Manipulative Studies In Genes And Cells And Their Limitations

Manipulative studies are only possible in the domain of animal work. They are the best means available for definitively linking changes in particular biological pathways with particular behavioral differences. Manipulative studies are especially important for mapping out the full range of biological processes that particular gene variants participate in, which is essential for understanding the linkage between gene variation and behavioral variation.

The most basic level at which such manipulations can be made is to single genes. As discussed in Sect. 3.1, it is now possible to make controlled changes to single genes and place these altered genes into animals. In addition to using this approach to correlate gene changes simply with behavioral changes, one can also systematically alter the structure of molecules, the timing of when they are produced, and the places where they are produced in an animal’s nervous system. Such manipulations make it possible to move into the realm of identifying and characterizing the fine structure of the developmental and cellular processes in which gene products participate.

There have traditionally been genetic techniques available in animal species such as fruit flies for manipulating the expression of whole groups of genes in particular cells of the nervous system. This makes it possible to produce so-called ‘chimeric’ animals, where some cell types express one set of genes, while other cell types express another set. Chimeras provide an important tool for examining developmental interactions at the level of cells, and techniques for producing them in both mammals and birds have made it possible to examine developmental differences in cells that have many genetic differences. In this way, complex behavioral differences that may be due to many gene changes in many different cell types, such as behavioral differences among species of animals, become amenable to empirical investigation.

Manipulative studies can suffer from the same interpretational difficulties as correlational studies. The major difference is that in manipulative studies, one has no excuse for failing to control for non-genetic explanations, for failing to measure enough different organismal attributes to avoid mislabeling or misidentification of the relevant behavioral dimensions, and for failing to look carefully enough at genetic and behavioral variables to avoid any illusions of homogeneity in characterizing genotypic or behavioral groups. There are some known technical problems with particular transgenic manipulations in rodents that can compromise causal attributions of gene effects on behavior (see Gerlai 1996); some of these problems may be remedied by future technical developments. The major limitation to manipulative studies comes from our limited ability to measure many different biological and behavioral attributes in a short period of time in single individuals. This limits our ability to understand cellular and developmental processes that are important for brain development and brain function in general, and for cells that participate in the formation and function of particular behavioral circuits.

4. Probable Future Directions

Clearly, an important development for the field of behavioral neurogenetics will occur when researchers take appropriate account of the limitations imposed by their behavioral measurements, and make research decisions and interpretations in line with these limitations. New developments in automated recording and analysis of natural behaviors will not necessarily advance this agenda, because the majority of ‘behaviors’ that are studied must be indirectly inferred. It is rather an increasing sophistication in defining the domain specificity (or, just as importantly, the domain generality) of particular linkages to biology that will most clearly benefit the field.

Up to the present time, behavioral neurogenetic research has been dominated by paradigms that emphasize genes as unifying explanations or predictors for behavioral dysfunction or differences. Yet the greatest untapped potential contribution of behavioral neurogenetics lies in the developmental domain. While Paul Weiss’ Genetic Neurobiology has largely lain dormant for more than 40 years, it is starting to be realized in the work of developmentally oriented neurogeneticists. They believe that to understand evolutionary, developmental, and functional contributions of molecules and cells to behavior, one must, as Weiss first articulated, understand the developmental processes the molecules and cells participate in, as well as those processes which ‘maintain it (the nervous system) in its integrity, and restore it after disruption.’

Such ‘process-oriented’ views toward understanding both behavior and brain function have been slow to gain acceptance, because for a good part of this past century, scientists have thought of the brain as a stable structure with unvarying neuronal connections set down during a defined developmental period. Since the 1980s the brain has come to be appreciated as an entity in which apparent stasis at more global levels is produced by competing dynamic processes at the levels of cell groups and individual cells. This has fostered the perceived value of relating global products of nervous systems like brain structures and behaviors to the dynamic processes that give rise to and maintain them.

The change in the emphasis of behavioral neurogenetic work from finding molecules to fitting them into developmental and functional pathways also entails a shift in the technical emphasis of the ‘genetic’ part of the research, a change that is already underway. Investigators are beginning to focus more on where and when in nervous systems genes are expressed during development, and how gene expression in cells is modified in development and in adulthood.

4.1 Gene Expression Correlations And Manipulations

Many non-biologists envision genes purely as molecules of heredity whose reason for being disappears (everywhere but in the germ cells) once an organism develops. Yet genes are a fundamental part of the physiological machinery in cells, especially in nerve cells which must constantly manufacture the products they need to communicate with other cells, to receive communications from other cells, and to change their size, shape, and other aspects of their morphology. While analogies between nerve cells and electrical components can be terribly misleading, it would not be an error to regard nerve cells as akin to circuits containing tunable resistors, inductances, and capacitors, which enable them to alter both their sensitivity to incoming signals and the signals they send in return. To change any aspect of their physiology or morphology (changing their cellular ‘behavior’), nerve cells need to alter their patterns of gene expression. They need to be able to make these changes repeatedly throughout the entire life of the individual in whose brain they are found.

Behavioral neurogeneticists are just starting to do correlative and manipulative work with gene expression patterns. Such research is likely to provide important empirical and theoretical insights in three major areas: developmental and functional pathways that behaviorally important brain circuits utilize; how external stimuli and an individual’s own behavior can affect the development and function of his or her brains; and how ‘inborn’ behavioral influences are manifest in nervous systems and how these meld with ongoing behavioral development and function.

4.2 Developmental And Functional Pathways And Interactions

By examining gene expression pattern changes in response to manipulation of the structure or expression of ‘targeted’ genes, much new information should be gained about how gene products interact during development and ongoing adult function. This will facilitate the construction of functionally coherent biochemical and physiological ‘pathways’ that are vital for understanding cellular interactions and decision-making processes that help construct and run behavioral circuitry in the brain. Such work may also lead to an increased appreciation of the role that stochastic variation in cell interactions may play in generating behavioral variation among individuals. The ‘cause’ of such behavioral variety would not necessarily be identifiable with variation in any particular gene product involved in a pathway, but rather as a joint property of the structure of molecular constituents and the types of interactions in which they are involved. A better appreciation of the mechanics underlying stochastic contributions to development may also provide profound insights into dysfunction and radically alter the way we identify and assess developmental ‘risk’ factors.

4.3 How Stimuli And Behavior Feedback On Brain Development And Function

A second way that developmental and functional studies of gene expression may advance the field of behavioral neurogenetics is by providing a new arena in which to examine environmental and social influences on brain development. It would be especially enlightening to identify the pathways through which an organism’s own behaviors may provide feedback that alters the development and function of its own brain. There has been a long and controversial history of the idea that organisms are actively involved in the regulation of their own development, just as there has been a tendency in behavioral neurogenetic work to disregard the biological pathways that mediate environmental influences on development. By being able to study and to manipulate the gene expression responses of cells and cell groups, we have much more powerful and sensitive assays to bring to bear on these questions.

4.4 ‘Inborn’ Influences: How Development, Evolution, And Ongoing Function Are Integrated

The study of species differences has proven to be a powerful method for learning about developmental mechanisms in biology for well over a century. The application of gene expression studies (together with the other manipulative neurogenetic techniques described above) toward the problem of inborn species differences in behavior is likely to produce powerful insights into the developmental pathways that make different members of a species behaviorally similar. Of equal importance will be the light such research sheds on developmental pathways that make individuals in one species all different from members of another species in the same way.

Comparative work is also expected to provide unique information on the mechanisms evolution uses to change brains in order to change behavior. Such mechanisms are likely to involve changes to both the autonomous characteristics of cells and changes in cellular interactions. Many of these changes may depend not on changes to the structure of gene products, but rather on changes in ‘silent’ parts of DNA sequences that influence the context and timing of gene expression. Understanding how such changes are accomplished evolutionarily may provide new technologies for dealing with neural dysfunction caused by injury and disease.

Finally, behaviors that show inborn species differences typically also have individually distinctive components that are malleable in the lifetimes of individuals. Comparative work could produce major insights into the ways that dynamic developmental and functional processes within and between neural cells integrate development and evolution with ongoing activity.

Bibliography:

1. Breakefield X O (ed.) 1979 Neurogenetics: Genetic Approaches to the Nervous System. Elsevier, New York
2. Gerlai R 1996 Seminars in Neuroscience 8: 153–61
3. Greenspan R J, Kyriacou C P (eds.) 1994 Flexibility and Constraint in Behavioral Systems. Wiley, New York
4. Hall J C, Greenspan R J, Harris W A 1982 Genetic Neurobiology. MIT Press, Cambridge, MA
5. Kelner K, Benditt J (eds.) 1994 Genes and behavior: A special report. Science 264: 1685–739
6. Pfaff D W, Berrettini W H, Joh T H, Maxon S C (eds.) 2000 Genetic Influences on Neural and Behavioral Functions. CRC Press, New York
7. Plomin R 1999 Nature 402: C25–C29 , p. C25
8. Pulst S M (ed.) 2000 Neurogenetics. Oxford University Press, Oxford, UK
9. Wahlster D 1999 Single gene influences on brain and behavior. Annual Review of Psychology 50: 599–624
10. Weiss P (ed.) 1950 Genetic Neurology. The University of Chicago Press, Chicago