From Scanner to Sound Bite:Issues in Interpreting and Reporting Sex Differences in the Brain by Cordelia Fine - HTML preview

Current Directions in Psychological

Science

http://cdp.sagepub.com/

From Scanner to Sound Bite : Issues in Interpreting and Reporting Sex Differences in the Brain Cordelia Fine

2010 19: 280

Current Directions in Psychological Science

DOI: 10.1177/0963721410383248

The online version of this article can be found at:

http://cdp.sagepub.com/content/19/5/280

Published by:

http://www.sagepublications.com

On behalf of:

Association for Psychological Science

can be found at:

Current Directions in Psychological Science

Additional services and information for

http://cdp.sagepub.com/cgi/alerts

Email Alerts:

http://cdp.sagepub.com/subscriptions

Subscriptions:

http://www.sagepub.com/journalsReprints.nav

Reprints:

http://www.sagepub.com/journalsPermissions.nav

Permissions:

at University of Melbourne Library on October 12, 2010

cdp.sagepub.com

Downloaded from

Current Directions in Psychological

From Scanner to Sound Bite: Issues in

Science

19(5) 280-283

ª The Author(s) 2010

Interpreting and Reporting Sex

Reprints and permission:

sagepub.com/journalsPermissions.nav

Differences in the Brain

DOI: 10.1177/0963721410383248

http://cdps.sagepub.com

Cordelia Fine

Centre for Agency, Values and Ethics, Macquarie University, and Department of Psychological Sciences, University of Melbourne

Abstract

Neuroimaging research is yielding reports of sex differences in the brain. Yet the likelihood of spurious findings of sex differences, the teething problems of new technology, the obscurity of the relation between brain structure and psychological function, and difficulties inferring mental states from neuroimaging data all require us to be considerably cautious in interpreting such results.

Unfortunately, these issues are often overlooked in popular accounts. Together with a tendency for people to regard neuroscientific information as more scientific than behavioral data, and as indicative of male and female ‘‘nature,’’ these issues point to the worrisome possibility of public misunderstanding of what contemporary neuroscience tells us about gender.

Keywords

neuroimaging, gender, science communication, neuroethics

For as long as there has been brain science there has been both

and, in keeping with the female brain’s supposedly more

scientific and popular interest in male–female differences and

interhemispheric functioning, are claimed to have a relatively

their psychological consequences. For example, 19th-century

larger corpus callosum (the bundle of neurons that connects the

scientific opinion held that women’s intellectual inferiority

two hemispheres). In both academic and especially popular

could be attributed to their smaller and lighter brains—a fact

work, these structural differences are proposed to have

that was widely known among the Victorian public as the

psychological implications.

‘‘missing five ounces’’ of female brain (see Russett, 1989).

Today, the weighing scales and other crude methodologies of

the Victorian brain scientists have been supplanted by sophis-

Production: The Problem of

ticated neuroimaging techniques that give unprecedented

Spurious Results

access to structural details of the brain and patterns of neural

When neuroscientists, in a single experiment, establish a ‘‘sig-

activity. Yet there remains cause for skepticism regarding

nificant difference’’ between the sexes, does this reflect a real

neuroscientific claims about sex differences and concern over

and reliable sex difference? Because sex is a primary and

the way such information is reported to, and interpreted by, the

ubiquitous social category, classifying participants by sex is

public. In this article, I lay out four scientific issues arising

obvious, easy, and may be done by default (Kaiser, Haller,

from the production and interpretation of ‘‘facts’’ about sex

Schmitz, & Nitsch, 2009). However, since by convention

differences in the brain, then discuss how these issues are

researchers declare a difference to be ‘‘significant’’ if there is overlooked and exacerbated when neuroscience findings are

no more than a one in 20 probability that it occurred by chance,

disseminated in the popular media and digested by the public.

if 20 researchers routinely test for sex differences, then even if To illustrate these points, I use as an example a long-there is no real difference between the populations, one

standing and influential claim about male–female brain

difference. The greater male lateralization (GML) hypothesis

proposes that males, compared with females, are more strongly

Corresponding Author:

left hemisphere dominant for language processing and right

Cordelia Fine, Department of Psychological Sciences, University of Melbourne, hemisphere dominant for visuospatial processing. Females,

Victoria 3010, Australia

by contrast, tend to engage both hemispheres for these tasks

E-mail: cfine@unimelb.edu.au

at University of Melbourne Library on October 12, 2010

cdp.sagepub.com

Downloaded from

Sex Differences in the Brain

281

researcher will find a statistically significant difference. The

Interpretation: The Obscurity of

concern is that, given that the publication process is geared

Structure–Function Relations

toward emphasizing difference rather than similarity, this

1-in-20 finding of difference will be reported while the 19 fail-

Further difficulties arise when it comes to understanding what,

ures to find a difference will not. Indeed, Kaiser et al. provide

if anything, brain differences might imply for psychological

examples of how, in language research, even marginal sex

function. Despite the extraordinary progress made in neu-

differences in the brain are given prominence in the published

roscience, we still have minimal understanding of how neural

paper.

structures contribute to complex psychological phenomena.

The problem of spurious results for all sex-differences

Again, the GML hypothesis is an instructive example: There

research has been long-noted. However, the inevitable teething

is no a priori reason to think that a more lateralized brain would problems of the new neuroimaging technologies may exacer-be advantageous for visuospatial processing but disadvanta-

bate the situation. Nuisance variables like breathing rate and

geous for language function. (Nor, to my knowledge, have any

caffeine intake can influence the imaging signal and give rise

such relations been demonstrated.) Bishop and Wahlsten

to spurious results; this is particularly the case when sample

(1997) questioned the assumption that a modest size difference

sizes are small, as they often are in the studies that have

in a structure as complex and massively interconnected as the

reported sex differences (Wallentin, 2009). Researchers have

corpus callosum would have tangible implications for a spe-

found that sex differences in language lateralization fail to

cific psychological construct.

generalize to a distinct but similar task within a second group

Nonetheless, functional speculations arising from the GML

of men and women and that identical analyses of the same par-

hypothesis have recently taken new form in suggestions that a

ticipants can also ‘‘discover’’ brain-activation differences

male brain skewed toward more lateralized, intrahemispheric

between randomly created groups matched on sex, perfor-

processing may be advantageous for scientific disciplines that

mance, and obvious demographic characteristics (Ihnen,

supposedly require focused scrutiny of details rather than

Church, Petersen, & Schlaggar, 2009). As controversies over

integration of information (Baron-Cohen, Knickmeyer, &

statistical procedures are resolved, researchers may turn out

Belmonte, 2005; Gur & Gur, 2007), and that a female brain

to have used inadequate or inappropriate techniques: It has

skewed toward more long-range processing may be advanta-

recently been argued that some reported sex differences in

geous for empathizing (Baron-Cohen et al., 2005).

language lateralization have not been put to adequate

While speculation is an important part of the scientific pro-

statistical test or that they can come and go depending on how

cess, as Fausto-Sterling (2000, p. 118) has observed, the prob-

the analysis is done (Kaiser et al., 2009). There is also an

lem is that ‘‘despite the many recent insights of brain research,

as-yet-unresolved

controversy

regarding

how—or

even

this organ remains a vast unknown, a perfect medium on which

whether—neuroscientists should control for the sex difference

to project, even unwittingly, assumptions about gender.’’ As

in average brain size when attempting to establish whether the

noted earlier, it was once readily assumed that brain weight

sexes differ in the volumes of particular brain regions (see, for

correlated with intelligence, thus explaining women’s sup-

example, Fausto-Sterling, 2000). A change in methodology can

posed intellectual inferiority. With that history in mind, we

transform what appears to be a sex difference into a difference

should be wary of suggestions that the typical female brain is

between people with smaller and larger brains (e.g., Im et al.,

suboptimally designed for currently male-dominated pursuits

2008).

like science—suggestions made in the absence of adequate

These difficulties point to the importance of not placing too

knowledge of how the brain enables scientific thinking and

much confidence in any single functional or structural neuro-

practice. There is no neuroscientific reason, for example, to

imaging study that seems to demonstrate a sex difference. The

think that the shorter circuits of an intrahemispheric brain will

wisdom of such caution is demonstrated by meta-analyses of

enable narrower focus in the mind.

tests of the GML hypothesis. A recent meta-analysis of 26 func-

Inferring a mental process from significant activation in a par-

tional neuroimaging studies of language lateralization found no

ticular brain region (for example, inferring that the amygdala was significant sex difference (Sommer, Aleman, Somers, Boks, &

activated, therefore participants were anxious) is known as

Kahn, 2008). Similarly, a meta-analysis of 49 postmortem and

reverse inference and is also fraught with difficulty. The statisti-structural neuroimaging studies of the corpus callosum found

cal procedures of functional neuroimaging identify regions that

no support for the hypothesis that this structure is larger in

are differentially activated by the experimental task, compared

females, even allowing for their relatively smaller brains

with a control task. However, while brain function involves spe-

(Bishop & Wahlsten, 1997).

cialization—the entire brain is not involved in all of its func-

The failure of meta-analyses to support predictions of the

tions—there is no simple one-to-one mapping between brain

GML hypothesis, to which much research attention has been

regions and psychological processes. Mental processes arise from

devoted, highlights the importance of remaining skeptical

the complex interaction of multiple areas, and any one region will about other reported sex differences. Clearly, isolated reports

be involved in any number of mental processes. The anterior cin-

of sex differences in brain activation or regional brain volume

gulate, for example, is activated by so many tasks that one cogni-

require replication and generalization before they can be

tive neuroscientist known to the author refers to this region as

assumed to be reliable.

‘‘the on button’’ (Geoffrey Boynton, personal communication).

at University of Melbourne Library on October 12, 2010

cdp.sagepub.com

Downloaded from

282

Fine

Moreover, the psychological implications of significant

(as much as 25%) larger in females, furnishing them not just

differences in the amount of brain activity in particular regions

with superior language skills but also greater multitasking abil-

are ambiguous. ‘‘More’’ activity does not necessarily imply

ity, a more intuitive leadership style, better emotion processing, psychologically ‘‘more’’ or ‘better,’’ or even that that region

and even greater capacity to remember to buy milk (see Fine,

is critically involved in that particular task (Poldrack, 2008).

2010).

Furthermore, data acquisition in functional magnetic resonance

Even books written by apparently authoritative and well-

imaging (fMRI) is slow: At its most sensitive, it averages over a

credentialed authors are rife with invalid structure–function

few seconds the activity of millions of neurons that can fire up

claims and reverse inferences, as well as factual errors, that

to one hundred impulses a second. (Positron emission tomogra-

go well beyond the GML hypothesis. To cite just a few exam-

phy, PET, is even slower.) This massively limits the interpreta-

ples from a multitude, functional neuroimaging findings may

tions that can be made about brief psychological events.

be used as evidence of sex differences in intrinsic interest in

Thus it becomes clear that observing, say, significantly

mathematics, in ‘‘hard-wired’’ ability to talk about feelings,

greater female anterior cingulate activity over the time course

or in capacity to empathize with the feelings of others. As

of the performance of a complex task is unlikely to indicate

I have discussed elsewhere (Fine, 2010), these particular claims

what mental process, if any, differs between males or females,

were made in part on the basis of studies that, respectively,

or in what direction that difference lies. It has been argued that didn’t involve mathematics, didn’t involve talking about feel-reverse inferences can be valuable when the underlying pro-

ings, and didn’t involve male participants. Furthermore, it

cesses involved in the participants’ task are well known, and

should by now be clear that currently no neuroimaging data

when they are used to generate hypotheses that drive further

could be cited as compelling support for such claims.

experimental work rather than to interpret neuroimaging results

(Poldrack, 2008). However, so far these conditions are rarely, if

Popular Digestion

ever, met in neuroimaging studies of sex differences. These

obstacles to valid reverse inference severely constrain the pos-

In addition to overinterpretation, misinterpretation, and misrep-

sibilities for making inferences about psychological differences

resentation, a number of characteristics of neuroscientific

between males and females from brain activation differences,

information and the way it is often communicated may further

even supposing these differences reported in studies are statis-

contribute to public misunderstanding. With its expensive,

tically valid and reliable.

complex machinery, the data yielded by neuroscience may

seem somehow more scientific and real than data collected in

less high-tech fashion. A consequence of this ‘‘neuro-realism’’

Popular Dissemination

(Racine, Bar-Ilan, & Illes, 2005) is that substantial behavioral While neuroscientists may routinely test for sex differences in

evidence of gender similarity may be overshadowed by a single

the brain and emphasize those they find, or engage in post hoc

finding of a sex difference in the brain. There may also be a ten-

speculations about such differences’ functional implications,

dency to equate ‘‘in the brain’’ with ‘‘innate.’’ Although the

they are presumably aware of the issues outlined previously

effects of gender socialization must manifest in the brain

and are engaging with peers who share this knowledge. The

(where else?), some popular authors promote the idea that brain

public, however, rarely gain its knowledge from neuroscientists

differences constitute evidence that the sexes are ‘‘hard-wired’’

or the neuroscientific literature. Instead, information is pre-

to be different. Lastly, neuroscientific data have been shown to

sented to the public by popular writers. That such writers are

have a ‘‘seductive allure’’: For instance, people find circular

either not aware of the critical issues of production and inter-

explanations of psychological phenomena more satisfying

pretation I have outlined or think them unimportant may be

when accompanied by information about brain responses

inferred from the fact that their books confidently purport to

(Weisberg, Keil, Goodstein, Rawson, & Gray, 2008). Thus,

offer practical applications for life, love, and learning on the

popular neuroscience is well placed to entice people to over-

basis of sex differences in the brain.

look psychological and sociological data showing that gender

Consider, for example, how the GML is presented in the

difference is contingent on historical period, ethnicity, socio-

popular literature. It is not only regularly asserted as fact rather economic group, and social context and to instead conclude

than a hypothesis (a poorly supported one, as noted) but is used

that gender differences are immutable, inevitable, and the prod-

as a springboard for scientifically unwarranted claims about

uct of fixed differences between the ‘‘male brain’’ and the

men’s and women’s different psychological abilities. Thus one

‘‘female brain.’’

author (working from an implicit metaphor of the brain as pin-

ball machine) explains how men’s language lateralization

Conclusion

impairs their ability to talk about their feelings, describing how the ‘‘signal’’ of an emotional feeling, having made it to the

While neuroimaging has potential to contribute in novel ways

right hemisphere, ‘‘may well get stopped, disappearing into

to our understanding of gender, scientists, popular commenta-

neural oblivion because the signal found no access to a receptor

tors, and the public need to be alert to the problem of premature

in a language center in the left side of the brain’’ (Gurian, 2004, speculation

about

sex

differences.

The

possibility

of

p. 88). Similarly, the corpus callosum is regularly claimed to be

neuroimaging ‘‘facts’’ about male and female brains—that may

at University of Melbourne Library on October 12, 2010

cdp.sagepub.com

Downloaded from

Sex Differences in the Brain

283

be spurious, overinterpreted, misinterpreted, or even fabri-

References

cated—influencing public attitudes about gender raises ethical

Baron-Cohen, S., Knickmeyer, R.C., & Belmonte, M.K. (2005). Sex concerns. The imaginative reader will not have too much

differences in the brain: Implications for explaining autism. Sci-

difficulty envisaging how, by reinforcing stereotypes, such

ence, 310, 819–823.

claims may affect people’s social attitudes in ways that oppose

Bishop, K.M., & Wahlsten, D. (1997). Sex differences in the human progress toward greater gender equality, just as such claims did

corpus callosum: Myth or reality? Neuroscience and Biobehav-

in the past. The task now is to open a dialogue between the pro-

ioral Reviews, 21, 581–601.

ducers, communicators, and consumers of knowledge about

Fausto-Sterling, A. (2000). Sexing the body: Gender politics and the sex differences in the brain and to take seriously the issues that construction of sexuality. New York, NY: Basic Books.

arise from scanner to sound bite.

Fine, C. (2010). Delusions of gender: How our minds, society, and

neurosexism create difference. New York, NY: W.W. Norton.

Gur, R.C., & Gur, R.E. (2007). Neural substrates for sex differences in Recommended Reading

cognition. In S.J. Ceci and W.M. Williams (Eds.), Why aren’t

Fine, C. (2010). (See References). An accessible book that covers all more women in science? Top researchers debate the evidence.

the issues in more detail than the current article.

(pp.

189–198).

Washington,

DC:

American

Psychological

Lehrer, J. (2008, August 17). Of course I love you, and I have the brain Association.

scan to prove it – We’re looking for too much in brain scans. Bos-

Gurian, M. (2004). What could he be thinking? A guide to the mys-

ton Globe, K1. An accessible account of general interpretative

teries of a man’s mind. London, England: Element.

issues in fMRI research.

Ihnen, S.K.Z., Church, J.A., Petersen, S.E., & Schlaggar, B.L. (2009).

Miller, G. (2008). Growing Pains for fMRI. Science, 320, 1412–1414.

Lack of generalizability of sex differences in the fMRI BOLD

A useful and accessible overview of issues, controversies, and

activity associated with language processing in adults. Neuro-

potential solutions in the interpretation of fMRI data that provides Image, 45, 1020–1032.

concrete examples.

Im, K., Lee, J., Lyttelton, O., Kim, S.H., Evans, A.C., & Kim, S.I.

Russett, C. E. (1989). (See References). A historical account of Victor-

(2008). Brain size and cortical structure in the adult huma