Copyright © 2003 American Psychological Society

1

CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE

Abstract

Although the brain was once

seen as a rather static organ, it isnow clear that the organizationof brain circuitry is constantlychanging as a function of expe-rience. These changes are re-ferred to as brain plasticity,and they are associated withfunctional changes that includephenomena such as memory,addiction, and recovery offunction. Recent research hasshown that brain plasticity andbehavior can be influenced bya myriad of factors, includingboth pre- and postnatal experi-ence, drugs, hormones, matu-ration, aging, diet, disease, andstress. Understanding howthese factors influence brain

organization and function isimportant not only for under-standing both normal and ab-normal behavior, but also fordesigning treatments for be-havioral and psychological dis-orders ranging from addictionto stroke.

Keywords

addiction; recovery; experi-ence; brain plasticity

One of the most intriguing ques-tions in behavioral neuroscienceconcerns the manner in which thenervous system can modify its or-ganization and ultimately its func-tion throughout an individual’slifetime, a property that is often re-ferred to as

plasticity.

The capacityto change is a fundamental charac-teristic of nervous systems and canbe seen in even the simplest of or-

ganisms, such as the tiny worm

C.elegans

, whose nervous system hasonly 302 cells. When the nervoussystem changes, there is often acorrelated change in behavior orpsychological function. This behav-ioral change is known by namessuch as learning, memory, addiction,maturation, and recovery. Thus, forexample, when people learn newmotor skills, such as in playing amusical instrument, there are plas-tic changes in the structure of cellsin the nervous system that underliethe motor skills. If the plasticchanges are somehow preventedfrom occurring, the motor learningdoes not occur. Although psychol-ogists have assumed that the ner-vous system is especially sensitiveto experience during develop-ment, it is only recently that theyhave begun to appreciate the po-tential for plastic changes in theadult brain. Understanding brainplasticity is obviously of consider-able interest both because it pro-vides a window to understandingthe development of the brain andbehavior and because it allows in-sight into the causes of normal andabnormal behavior.

THE NATURE OF BRAIN PLASTICITY

The underlying assumption of

studies of brain and behavioral plas-ticity is that if behavior changes,there must be some change in orga-nization or properties of the neuralcircuitry that produces the behav-ior. Conversely, if neural networks

are changed by experience, there

must be some corresponding changein the functions mediated by those

networks. For the investigator inter-ested in understanding the factorsthat can change brain circuits, andultimately behavior, a major chal-lenge is to find and to quantify thechanges. In principle, plastic changesin neuronal circuits are likely to re-flect either modifications of exist-ing circuits or the generation ofnew circuits. But how can research-ers measure changes in neural cir-cuitry? Because neural networksare composed of individual neu-rons, each of which connects with asubset of other neurons to form in-terconnected networks, the logicalplace to look for plastic changes isat the junctions between neurons,that is, at synapses. However, it is adaunting task to determine if syn-apses have been added or lost in aparticular region, given that thehuman brain has something like100 billion neurons and each neuronmakes on average several thousandsynapses. It is clearly impractical toscan the brain looking for alteredsynapses, so a small subset must beidentified and examined in detail.But which synapses should bestudied? Given that neuroscientistshave a pretty good idea of what re-gions of the brain are involved inparticular behaviors, they can nar-row their search to the likely areas,but are still left with an extraordi-narily complex system to examine.There is, however, a procedure thatmakes the job easier.

In the late 1800s, Camillo Golgiinvented a technique for staining arandom subset of neurons (1–5%)so that the cell bodies and the den-dritic trees of individual cells canbe visualized (Fig. 1). The den-drites of a cell function as the scaf-folding for synapses, much as treebranches provide a location forleaves to grow and be exposed tosunlight. The usefulness of Golgi’stechnique can be understood bypursuing this arboreal metaphor.There are a number of ways one

Brain Plasticity and Behavior

Bryan Kolb,

1

Robbin Gibb, and Terry E. Robinson

Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, Alberta, Canada (B.K., RG.), and Department of Psychology, University of Michigan,

Ann Arbor, Michigan (T.E.R.)

2 VOLUME 12, NUMBER 1, FEBRUARY 2003

Published by Blackwell Publishing Inc.

could estimate how many leaves areon a tree without counting every leaf.Thus, one could measure the totallength of the tree’s branches as wellas the density of the leaves on arepresentative branch. Then, by sim-ply multiplying branch length byleaf density, one could estimate to-tal leafage. A similar procedure isused to estimate synapse number.About 95% of a cell’s synapses areon its dendrites (the neuron’sbranches). Furthermore, there is aroughly linear relationship be-tween the space available for syn-apses (dendritic surface) and thenumber of synapses, so researcherscan presume that increases or de-creases in dendritic surface reflectchanges in synaptic organization.

FACTORS AFFECTING BRAIN PLASTICITY

By using Golgi-staining proce-dures, various investigators haveshown that housing animals incomplex versus simple environ-ments produces widespread differ-ences in the number of synapses inspecific brain regions. In general,such experiments show that partic-ular experiences embellish cir-cuitry, whereas the absence ofthose experiences fails to do so(e.g., Greenough & Chang, 1989).Until recently, the impact of theseneuropsychological experimentswas surprisingly limited, in partbecause the environmental treat-

ments were perceived as extremeand thus not characteristic ofevents experienced by the normalbrain. It has become clear, how-ever, not only that synaptic organi-zation is changed by experience,but also that the scope of factorsthat can do this is much more ex-tensive than anyone had antici-pated. Factors that are now knownto affect neuronal structure and be-havior include the following:

experience (both leading pre-and post-natal)psychoactive drugs (e.g., amphet-amine, morphine)gonadal hormones (e.g., estrogen,testosterone)anti-inflammatory agents (e.g.,COX-2 inhibitors)growth factors (e.g., nerve growthfactor)dietary factors (e.g., vitamin andmineral supplements)genetic factors (e.g., strain differ-ences, genetically modified mice)disease (e.g., Parkinson’s disease,schizophrenia, epilepsy, stroke)stressbrain injury and leading disease

We discuss two examples to illus-trate.

Early Experience

It is generally assumed that ex-periences early in life have differ-ent effects on behavior than similarexperiences later in life. The reasonfor this difference is not under-stood, however. To investigate thisquestion, we placed animals incomplex environments either as ju-veniles, in adulthood, or in senes-cence (Kolb, Gibb, & Gorny, 2003).It was our expectation that therewould be quantitative differences inthe effects of experience on synapticorganization, but to our surprise, wealso found

qualitative

differences.Thus, like many investigators be-fore us, we found that the length ofdendrites and the density of syn-

•

•

•

•

•

•

•

•

••

Fig. 1. Photograph of a neuron. In the view on the left, the dendritic field with theextensive dendritic network is visible. On the right are higher-power views of den-dritic branches showing the spines, where most synapses are located. If there is anincrease in dendritic length, spine density, or both, there are presumed to be moresynapses in the neuron.

Copyright © 2003 American Psychological Society

CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE 3

apses were increased in neurons inthe motor and sensory cortical re-gions in adult and aged animalshoused in a complex environment(relative to a standard lab cage). Incontrast, animals placed in thesame environment as juvenilesshowed an increase in dendriticlength but a decrease in spine den-sity. In other words, the same envi-ronmental manipulation had quali-tatively different effects on theorganization of neuronal circuitryin juveniles than in adults.

To pursue this finding, we latergave infant animals 45 min of dailytactile stimulation with a littlepaintbrush (15 min three times perday) for the first 3 weeks of life.Our behavioral studies showedthat this seemingly benign earlyexperience enhanced motor andcognitive skills in adulthood. Theanatomical studies showed, in ad-dition, that in these animals therewas a decrease in spine density butno change in dendritic length incortical neurons—yet another pat-tern of experience-dependent neu-ronal change. (Parallel studies haveshown other changes, too, includingneurochemical changes, but theseare beyond the current discussion.)Armed with these findings, we thenasked whether prenatal experiencemight also change the structure ofthe brain months later in adulthood.Indeed, it does. For example, the off-spring of a rat housed in a complexenvironment during the term of herpregnancy have increased synapticspace on neurons in the cerebral cor-tex in adulthood. Although we donot know how prenatal experiencesalter the brain, it seems likely thatsome chemical response by themother, be it hormonal or otherwise,can cross the placental barrier and al-ter the genetic signals in the develop-ing brain.

Our studies showing that expe-rience can uniquely affect the de-veloping brain led us to wonder ifthe injured infant brain might berepaired by environmental treat-

ments. We were not surprised tofind that postinjury experience,such as tactile stroking, could mod-ify both brain plasticity and behav-ior because we had come to believethat such experiences were power-ful modulators of brain develop-ment (Kolb, Gibb, & Gorny, 2000).What was surprising, however,was that prenatal experience, suchas housing the pregnant mother ina complex environment, could af-fect how the brain responded to aninjury that it would not receive un-til after birth. In other words, pre-natal experience altered the brain’sresponse to injury later in life. Thistype of study has profound impli-cations for preemptive treatmentsof children at risk for a variety ofneurological disorders.

Psychoactive Drugs

Many people who take stimu-lant drugs like nicotine, amphet-amine, or cocaine do so for theirpotent psychoactive effects. Thelong-term behavioral consequencesof abusing such psychoactivedrugs are now well documented,but much less is known about howrepeated exposure to these drugsalters the nervous system. One ex-perimental demonstration of a verypersistent form of drug experience-dependent plasticity is known asbehavioral sensitization. For exam-ple, if a rat is given a small dose ofamphetamine, it initially will showa small increase in motor activity(e.g., locomotion, rearing). Whenthe rat is given the same dose onsubsequent occasions, however,the increase in motor activity in-creases, or sensitizes, and the ani-mal may remain sensitized forweeks, months, or even years, evenif drug treatment is discontinued.

Changes in behavior that occur asa consequence of past experience,and can persist for months or years,like memories, are thought to be dueto changes in patterns of synaptic or-

ganization. The parallels betweendrug-induced sensitization andmemory led us to ask whether theneurons of animals sensitized todrugs of abuse exhibit long-lastingchanges similar to those associatedwith memory (e.g., Robinson &Kolb, 1999). A comparison of the ef-fects of amphetamine and salinetreatments on the structure of neu-rons showed that neurons in am-phetamine-treated brains had greaterdendritic material, as well as moredensely organized spines. Theseplastic changes were not foundthroughout the brain, however, butrather were localized to regions suchas the prefrontal cortex and nu-cleus accumbens, both of which arethought to play a role in the reward-ing properties of these drugs. Laterstudies have shown that these drug-induced changes are found not onlywhen animals are given injections byan experimenter, but also when ani-mals are trained to self-administerdrugs, leading us to speculate thatsimilar changes in synaptic organi-zation will be found in human drugaddicts.

Other Factors

All of the factors we listed earlierhave effects that are conceptuallysimilar to the two examples thatwe just discussed. For instance,brain injury disrupts the synapticorganization of the brain, andwhen there is functional improve-ment after the injury, there is a cor-related reorganization of neuralcircuits (e.g., Kolb, 1995). But notall factors act the same way acrossthe brain. For instance, estrogenstimulates synapse formation insome structures but reduces syn-apse number in other structures(e.g., Kolb, Forgie, Gibb, Gorny, &Rowntree, 1998), a pattern ofchange that can also be seen withsome psychoactive drugs, such asmorphine. In sum, it now appearsthat virtually any manipulation

4 VOLUME 12, NUMBER 1, FEBRUARY 2003

Published by Blackwell Publishing Inc.

that produces an enduring changein behavior leaves an anatomicalfootprint in the brain.

CONCLUSIONS AND ISSUES

There are several conclusions todraw from our studies. First, expe-rience alters the brain, and it doesso in an age-related manner. Second,both pre- and postnatal experiencehave such effects, and these effectsare long-lasting and can influencenot only brain structure but alsoadult behavior. Third, seeminglysimilar experiences can alter neu-ronal circuits in different ways, al-though each of the alterations ismanifest in behavioral change.Fourth, a variety of behavioral con-ditions, ranging from addiction toneurological and psychiatric disor-ders, are correlated with localizedchanges in neural circuits. Finally,therapies that are intended to alterbehavior, such as treatment for ad-diction, stroke, or schizophrenia,are likely to be most effective ifthey are able to further reorganizerelevant brain circuitry. Further-more, studies of neuronal structureprovide a simple method of screen-ing for treatments that are likely tobe effective in treating disorderssuch as dementia. Indeed, ourstudies show that the new genera-tion of antiarthritic drugs (knownas COX-2 inhibitors), which act to re-duce inflammation, can reverse age-related synaptic loss and thus oughtto be considered as useful treatmentsfor age-related cognitive loss.

Although much is now knownabout brain plasticity and behav-ior, many theoretical issues re-main. Knowing that a wide vari-ety of experiences and agents canalter synaptic organization andbehavior is important, but leads toa new question: How does thishappen? This is not an easy ques-tion to answer, and it is certainthat there is more than one an-

swer. We provide a single exam-ple to illustrate.

Neurotrophic factors are a classof chemicals that are known to af-fect synaptic organization. An ex-ample is fibroblast growth factor-2(FGF-2). The production of FGF-2 isincreased by various experiences,such as complex housing and tactilestroking, as well as by drugs such asamphetamine. Thus, it is possiblethat experience stimulates the pro-duction of FGF-2 and this, in turn,increases synapse production. Butagain, the question is how. Onehypothesis is that FGF-2 somehowalters the way different genes areexpressed by specific neurons andthis, in turn, affects the way synapsesare generated or lost. In other words,factors that alter behavior, includingexperience, can do so by altering geneexpression, a result that renders thetraditional gene-versus-environmentdiscussions meaningless.

Other issues revolve around thelimits and permanence of plasticchanges. After all, people encounterand learn new information daily. Isthere some limit to how much cellscan change? It seems unlikely thatcells could continue to enlarge andadd synapses indefinitely, but whatcontrols this? We saw in our studiesof experience-dependent changes ininfants, juveniles, and adults that ex-perience both adds and prunes syn-apses, but what are the rules govern-ing when one or the other mightoccur? This question leads to another,which is whether plastic changes inresponse to different experiencesmight interact. For example, does ex-posure to a drug like nicotine affecthow the brain changes in learning amotor skill like playing the piano?Consider, too, the issue of the perma-nence of plastic changes. If a personstops smoking, how long do the nico-tine-induced plastic changes persist,and do they affect later changes?

One additional issue surroundsthe role of plastic changes in disor-dered behavior. Thus, althoughmost studies of plasticity imply

that remodeling neural circuitry isa good thing, it is reasonable towonder if plastic changes mightalso be the basis of pathological be-havior. Less is known about thispossibility, but it does seem likely.For example, drug addicts oftenshow cognitive deficits, and itseems reasonable to propose that atleast some of these deficits couldarise from abnormal circuitry, es-pecially in the frontal lobe.

In sum, the structure of the brainis constantly changing in responseto an unexpectedly wide range ofexperiential factors. Understandinghow the brain changes and the rulesgoverning these changes is impor-tant not only for understandingboth normal and abnormal behav-ior, but also for designing treat-ments for behavioral and psycho-logical disorders ranging fromaddiction to stroke.

Recommended Reading

Kolb, B., & Whishaw, I.Q. (1998). Brainplasticity and behavior.

Annual Re-view of Psychology, 49

, 43–64.Robinson, T.E., & Berridge, K.C. (in

press). Addiction.

Annual Reviewof Psychology.

Shaw, C.A. , & McEachern, J .C.(2001).

Toward a theory of neuro-plasticity.

New York: Taylor andFrancis.

Acknowledgments—

This research wassupported by a Natural Sciences andEngineering Research Council grant toB.K. and a National Institute on DrugAbuse grant to T.E.R.

Note

1. Address correspondence to BryanKolb, CCBN, University of Lethbridge,Lethbridge, AB, Canada T1K 3M4.

References

Greenough, W.T., & Chang, F.F. (1989). Plasticityof synapse structure and pattern in the cerebralcortex. In A. Peters & E.G. Jones (Eds.),

Cere-

Copyright © 2003 American Psychological Society

CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE 5

bral cortex: Vol. 7

(pp. 391–440). New York: Ple-num Press.

Kolb, B. (1995).

Brain plasticity and behavior.

Mah-wah, NJ: Erlbaum.

Kolb, B., Forgie, M., Gibb, R., Gorny, G., & Rown-tree, S. (1998). Age, experience, and the chang-ing brain.

Neuroscience and Biobehavioral Reviews,22

, 143–159.

Kolb, B., Gibb, R., & Gorny, G. (2000). Corticalplasticity and the development of behavior af-ter early frontal cortical injury.

DevelopmentalNeuropsychology, 18

, 423–444.

Kolb, B., Gibb, R., & Gorny, G. (2003). Experi-ence-dependent changes in dendritic arborand spine density in neocortex vary with age

and sex.

Neurobiology of Learning and Memory,79

, 1–10.

Robinson, T.E., & Kolb, B. (1999). Alterations inthe morphology of dendrites and dendriticspines in the nucleus accumbens and prefron-tal cortex following repeated treatment withamphetamine or cocaine.

European Journal ofNeuroscience, 11

, 1598–1604.

The Malicious Serpent: Snakes as a Prototypical Stimulus for an Evolved Module of Fear

Arne Öhman

1

and Susan Mineka

Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden (A.Ö.),

and Department of Psychology, Northwestern University, Evanston, Illinois (S.M.)

Abstract

As repti les, snakes mayhave signified deadly threatsin the environment of earlymammals. We review findingssuggesting that snakes remainspecial stimuli for humans. In-tense snake fear is prevalent inboth humans and other pri-mates. Humans and monkeyslearn snake fear more easilythan fear of most other stimulithrough direct or vicariousconditioning. Neither the elici-tation nor the conditioning ofsnake fear in humans requiresthat snakes be consciously per-ceived; rather, both processescan occur with masked stimuli.Humans tend to perceive illu-sory correlations betweensnakes and aversive stimuli,and their attention is automati-cally captured by snakes incomplex visual displays. To-gether, these and other findingsdelineate an evolved fear mod-ule in the brain. This module isselectively and automaticallyactivated by once-threateningstimuli, is relatively encapsu-lated from cognition, and de-rives from specialized neuralcircuitry.

Keywords

evolution; snake fear; fearmodule

Snakes are commonly regardedas slimy, slithering creatures worthyof fear and disgust. If one were to be-lieve the Book of Genesis, humans’dislike for snakes resulted from adivine intervention: To avenge thesnake’s luring of Eve to taste the fruitof knowledge, God instituted eternalenmity between their descendants.Alternatively, the human dislike ofsnakes and the common appear-ances of reptiles as the embodimentof evil in myths and art might reflectan evolutionary heritage. Indeed,Sagan (1977) speculated that humanfear of snakes and other reptilesmay be a distant effect of the condi-tions under which early mammalsevolved. In the world they inhabited,the animal kingdom was dominatedby awesome reptiles, the dinosaurs,and so a prerequisite for early mam-mals to deliver genes to future gen-erations was to avoid getting caughtin the fangs of Tyrannosaurus rexand its relatives. Thus, fear and re-spect for reptiles is a likely coremammalian heritage. From thisperspective, snakes and other rep-

tiles may continue to have a specialpsychological significance even forhumans, and considerable evi-dence suggests this is indeed true.Furthermore, the pattern of find-ings appears consistent with theevolutionary premise.

THE PREVALENCE OF SNAKE FEARS IN PRIMATES

Snakes are obviously fearsomecreatures to many humans. Agras,Sylvester, and Oliveau (1969) inter-viewed a sample of New England-ers about fears, and found snakesto be clearly the most prevalent ob-ject of intense fear, reported by38% of females and 12% of males.

Fear of snakes is also commonamong other primates. Accordingto an exhaustive review of fielddata (King, 1997), 11 genera of pri-mates showed fear-related responses(alarm calls, avoidance, mobbing) invirtually all instances in which theywere observed confronting largesnakes. For studies of captive pri-mates, King did not find consistentevidence of snake fear. However,in direct comparisons, rhesus (andsquirrel) monkeys reared in thewild were far more likely than lab-reared monkeys to show strongphobiclike fear responses to snakes(e.g., Mineka, Keir, & Price, 1980).That this fear is adaptive in thewild is further supported by inde-pendent field reports of largesnakes attacking primates (M.Cook & Mineka, 1991).

This high prevalence of snakefear in humans as well as in our