I found this interesting. Perhaps some of you will as well.
djc
-------
To see this story with its related links on the EducationGuardian.co.uk
site, go to http://education.guardian.co.uk
Elementary, my dear Watson, the clue is in the genes - or is it?
A summary of the lecture delivered yesterday by Annette Karmiloff-Smith,
from the Neurocognitive Development Unit at the Institute of Child Health,
UCL, at the British Academy.
Monday November 05 2001
The Guardian
The lecture celebrated the centenary of the British Psychological Society.
Let me begin by explaining my unusual title. When asked to give the British
Psychological Society's centennial lecture, hosted by the British Academy, I
pondered on how broad an audience the talk would have to reach. The British
Academy covers the humanities (archeology, classics, literature, history,
languages, philosophy and such like) and the social sciences (anthropology,
economics, geography, law, politics, psychology and such like). But even
within psychology, we cover a very wide range of topics: social psychology,
health psychology, emotional psychology, cognitive psychology. And my domain
- that of developmental cognitive neuroscience - is probably the furthest
away from the humanities. How could I link my interest in genes and human
development to the interests of these other disciplines? And, then, the name
"Watson" popped into my mind. On the science side: James Watson contributed
to the discovery of the structure of DNA in the 1950s and to the sequencing
of the !
human genome half a century later. On the humanities side: although Arthur
Conon Doyle may not have been the greatest of Britain's literary figures,
the sidekick of his Sherlock Holmes was also called Watson. Here was my
link. And the more I considered my title, the more convinced I was that the
link wasn't entirely tenuous, because understanding the complex pathways
from gene-to-brain-to-cognitive processes-to-behaviour is like a Holmes and
Watson detective story, in which tiny, seemingly unimportant clues early in
development play a vital role in the final outcome.
As we learn more about genes and the human brain, there is a temptation, not
only in the press but even in scientific literature, to seek one-to-one
relationships between complex behaviours like altruism, aggression,
intelligence, mathematics or language, on the one hand, and specific genes
and/or specific locations in the brain, on the other. In a series of popular
books, Steven Pinker has repeatedly supported such assumptions by using data
from adult neuropsychology and genetic disorders. Most recently, in an
important article in Nature [October 4], a group of researchers led by Tony
Monaco, from Oxford, and Faraneh Varga-Khadem, from the Institute of Child
Health in London, identified a point mutation in a gene in a family with
speech and language impairments. Now, these scientists made it clear that
they had identified a gene which affects fine oro-facial movement planning
which subsequently has cascading effects on speech and language, as well as
on intelligence and ge!
neral motor skills, and that there was a homologous gene in the mouse. But
Steven Pinker's gloss on the article makes it sound as if this gene were
only involved in language, and his commentary ran as follows:
"The discovery of a gene implicated in speech and language is among the
first fruits of the Human Genome Project for the cognitive sciences. Just as
the 1990s are remembered as the decade of the brain and the dawn of
cognitive neuroscience, the first decade of the 21st century may well be
thought of as the decade of the gene and the dawn of cognitive genetics."
The very notion of "cognitive genetics" is based on the assumption that
genes code for cognitive outcomes and, more generally, on the nativist claim
that human babies are born with genetically specified brains that contain
specialised modules. These modules are not only considered to operate for
low-level perceptual processes, but also for higher-level cognitive modules
like language, mathematics, spatial cognition, face processing and the like.
In other words, the infant brain is claimed to be like a Swiss Army knife.
Data used to bolster such claims come from adult neuropsychological patients
and children with genetic disorders. It is indeed the case that adults who
suffer a stroke or a road accident can damage a specific part of the brain
and this can result in an isolated impairment. Patients with prosopagnosia,
for instance, may have perfect language, be able to recognise all categories
of objects, yet present with an isolated impairment in recognising faces.
Other pat!
ients may process faces well and have fluent access to vocabulary, but
present with serious difficulties with grammar, and so forth. But these
neuropsychological cases concern adults who had developed normally until
their insult. Impairment after development cannot be equated to impairment
at the outset of development which affects the system throughout its growth.
In other words, researchers cannot use the end state of development to make
claims about the start state. It could well be that the specialisations for
face processing, grammar and the like are not the starting point in infancy,
but the result in adulthood of a developmental process of learning which
gives rise over time to specialisations. Thus, isolated impairments in
adulthood tell us nothing necessary about the infant brain.
At first blush, there are a number of genetic disorders which seem to fit
the nativist/cognitive genetics model. Dyslexia is a disorder with a clear
genetic component and seems to present solely with impairments in reading.
Likewise for Specific Language Impairment (SLI) which by its very name
suggests that only language is impaired, with the rest of the child's skills
argued to be intact. Discussions of autism suggest a core deficit in
attributing mental states to others, such that researchers have argued for a
damaged theory-of-mind module. Finally, Williams syndrome, which I shall
discuss in detail in this talk, has been hailed by many, including Pinker,
as the prime example of a neat, compartmentalised package of impaired and
intact modules.
In my presentation, I challenge the notion of cognitive genetics, arguing
that there is no one-to-one, direct mapping between a specific gene (or
specific set of genes) and a cognitive outcome. Rather, there are
many-to-many very indirect mappings, with the regulation of gene expression
likely to contribute to very broad differences in developmental timing,
neuronal type, neuronal density, neuronal firing, neurotransmitter types,
etc. Instead of the cognitive genetics model, I argue for a
neuroconstructivist framework. In this framework, gene/gene interaction,
gene/environment interaction and, crucially, the process of ontogeny (pre-
and post-natal development) are all considered to play a vital role in how
the brain progressively sculpts itself and how it gradually becomes
specialised over developmental time. I take Williams syndrome as an example
of the neuroconstructivist approach.
Williams syndrome (WS) is a rare, genetic disorder occurring in 1 in 20,000
live births. A lot is already known about both the genotype and the
phenotype (the behavioural outcome). Yet despite this knowledge, the
relationship between genotype and phenotype is not at all obvious. WS
involves the deletion of 16 genes on one copy of chromosome 7. People with
WS have atypical brain anatomy and atypical brain chemistry. They present
with heart abnormalities, in particular supravalvular aortic stenosis
(SVAS), a facial dysmorphology (sometimes called "elfin faces"), and are
small in stature, have hoarse voices and an awkward gait. Their low IQs are
in the 50-65 range, with an uneven cognitive profile in which language
scores usually outstrip scores on spatial tasks and face processing is very
proficient. People with WS seem to be very sociable, sensitive to others'
emotional states and use erudite-sounding words. My favourite example of the
discrepancy between language and intell!
igence comes from an 18-year-old girl with WS, with a full IQ of 59. Her
favourite topic is vampires and the conversation runs as follows:
Exp: What do vampires do? WS: They break into women's bedrooms in the
middle of the night and sink their teeth into their necks. Exp: Why do they
do that? WS: (Clearly never having asked herself the question) Maybe they
are inordinately fond of necks.
This sophisticated-sounding language, yet shallow understanding of the
concept of a vampire, comes from a typical adolescent with WS who can
neither tie her shoe laces nor match the simple orientation of lines in a
spatial display.
The difference between very impaired spatial skills and seemingly proficient
face processing skills in Williams syndrome is particularly striking.
Children and adults with WS score in the normal range on a number of face
processing tasks, yet they score in the severely impaired range on spatial
tasks. This led a number of psychologists of a nativist persuasion to claim
that the syndrome presents with an intact face processing module and an
impaired face processing module. Geneticists working on the syndrome found
one non-WS family with deletions of two of the genes in the same region as
people with WS who displayed some spatial impairments. One of the two
deleted genes - Limkinase1 - is expressed in the brain. The geneticists
leapt to the conclusion that this gene, LIMK1, was directly linked to the
spatial impairment alone. It took little time for the press then to herald
the discovery of "a gene for spatial cognition" or even "a gene for
intelligence".
Could the heroes of my talk, Holmes and Watson, have found their first clue:
LIMK1 = a gene for spatial cognition?
Well, there are, alas, several problems with the direct mapping of LIMK1 to
spatial cognition. Firstly, as mentioned earlier, simple one-to-one, direct
mappings between specific genes and specific higher-level cognitive outcomes
like spatial cognition are extremely unlikely, given all we know about
many-to-many mappings and how the low-level effects of genes are widely
expressed during development. Secondly, drawing such strong conclusions from
the study of one family who may have other genetic impairments is
questionable. Thirdly, using the adult outcome to draw such conclusions
completely negates the role of development.
My team and I decided that three approaches were needed to properly explore
the relationship between genotype and phenotype: i) We joined forces with
colleagues in clinical genetics at Manchester's St. Mary's Hospital and
examined a larger number of cases from different families, with partial gene
deletions in the WS critical region; ii) we carried out in-depth studies of
those areas which other research teams had deemed to be intact in WS, by
dissecting the phenotype in much greater detail; iii) we explored the start
state by studying infants and toddlers with WS and comparing them to infants
with other genetic disorders.
Here I will just briefly summarise the results from our three approaches.
Our genetic studies showed that non-WS patients with two, three or even 13
deleted genes in the WS critical region showed severe SVAS and other
physical problems, but normal and even above normal intelligence, and with
no linguistic/spatial imbalance.
The heroes of my talk, Holmes and Watson, have to conclude, finally, that
the search for the gene for spatial cognition was predicated on two
erroneous assumptions. The first false assumption is that genes code
directly for spatial cognition, and the second that scores in the normal
range imply intact genes, and scores in the impaired range imply mutated
genes. Holmes and Watson needed to recall the very indirect, many-to-many
mappings involved in gene expression across the developing system.
Our second approach was to examine in detail the so-called "intact" domains
of WS functioning. My team and I thus undertook in-depth studies of face
processing, language and social cognition in older children and adults with
Williams syndrome. These domains of seeming proficiency turned out to be far
from intact. Normal controls process faces configurally - they look at the
whole face and the relationships between the parts. By contrast, people with
Williams syndrome process faces featurally; they focus on single details in
a face. In other words, the WS scores in the normal range are arrived at via
a different cognitive process. Such differences were also revealed in our
studies of WS brains. Normal controls process faces predominantly with the
right hemisphere - the hemisphere involved in holistic processing. However,
people with WS show either bilateral processing or a predominance of the
left hemisphere - the one usually involved in more piecemeal, featural
processing. !
Interestingly, too, normal controls display clear-cut differences in brain
electrophysiology when processing human faces, monkey faces or cars, whereas
people with WS process all three in the same way. So, it is not the case
that face processing is intact and spatial processing is impaired in WS;
both are impaired compared to normal controls. Whatever the ultimate
function of the deleted genes, their effects are pervasive and influence far
more than spatial cognition. And our studies revealed the same picture for
social cognition and language in WS. Whatever we examined in language -
vocabulary, semantics, grammar, pragmatics, reading - all showed subtle
impairments, despite the superficial fluency in WS and the use of erudite
terms. Yet the claims of nativism and cognitive genetics, and the use of a
genetic disorder like Williams syndrome to back those claims, require a neat
pattern of intact and impaired modules. This is clearly not the case.
Our third line of experimental attack was to look at infants and toddlers
with Williams syndrome. Does the pattern of impairments found in adulthood
look the same in infants with WS? You may wonder how the psychologist can
study language, spatial cognition, number and face processing in young
babies who can't talk, can't move very well, can't draw, and can't count.
All young babies can do is suck, look and move their heads. But, in fact,
this is all the experimental psychologist needs them to do in order to
display their implicit knowledge. For example, we present infants
repeatedly with the same display on a wide computer screen until the infant
shows boredom (looks away). We then change the display in subtle ways to
see whether the infant renews interest (notices the change) and measure the
amount of time he or she looks at the changed display. For instance, we
might present a schematic face five or six times, and then present a choice
between an identical face or one !
that has its eyes changed (a featural change) or the spacing between its
features changed (a configural change). Does the infant notice either of
these changes and look longer? Or we might present infants with pairs of
objects many times and then suddenly change one of the pairs to three
objects. Does the infant look longer at the changed number? Or, for
language understanding, we might present a picture of a dog on one side of
the screen and a picture of a car on the other, and through a loudspeaker
exclaim: "Car, look at the car." and measure whether the infant looks longer
at the named object.
Our results with infants and toddlers with WS showed that for faces, they
notice both featural and configural changes but, unlike control infants, the
ones with WS prefer to focus on features if given a choice between the two.
For number, infants with WS notice small changes in numerosity, whereas
infants with Down's syndrome (DS) of the same chronological and mental age
do not. Yet in adulthood, people with DS are less impaired in arithmetic
tasks than those with WS. I will come back later in the talk to why I think
this obtains. With respect to language, surprisingly infants with WS are
just as impaired as those with DS during the early years. Yet, in this
case, it is adults with WS who easily outstrip those with DS in the language
domain. Clearly the infant profiles do not look the same as the adult
profiles. Yet again, the nativist/cognitive genetics view, based on adult
outcomes, would require that the infant profiles look similar to the adult
outcome.
In the third part of the talk I consider why language is so late to develop
in WS, given the proficiency of language in their adulthood, and use this to
examine increasingly lower-level mechanisms which might explain the overall
impairments in the cognitive profile of Williams syndrome. It turns out
that toddlers with WS show less hemispheric lateralisation than normal
controls when reaching for objects, and that they have difficulties
segmenting the speech stream into separate words even before language
understanding kicks in. Furthermore, toddlers with WS don't understand the
function of pointing which normal children use to learn new words. In
general, infants and toddlers with WS have difficulty disengaging from faces
and therefore do not follow pointing towards objects outside the dyadic
interaction. This leads the heroes of my talk, Holmes and Watson, to
re-examine the infancy number results. Perhaps it is the WS focus on detail
that explains why they seem to perfor!
m well on tasks involving tiny changes in numerosity but that their
behaviour actually has nothing to do with an understanding of number. Focus
on detail will, over developmental time, cause impairments in understanding
cardinality if children do not also focus on the whole. Again, development
itself is playing a vital role in the outcome.
Finally, in the talk, I examine low-level mechanisms like eye movement
planning and oscillatory neuronal firing. We find that infants with WS
cannot update the image on the retina with extra-retinal information about
where their eyes are focusing in the world - something we do so
automatically that we are unaware of it. The infants with WS stay fixated
on one point, leading to an impairment in their general exploration of the
world. And, when we reanalysed the brain data from the electrophysiological
studies, we found that people with WS do not show the normal synchronisation
of oscillatory activity in the brain. In normal brains over development,
neurons that fire together simultaneously wire together, leading to learning
and memory. This simultaneous oscillatory activity occurs as of six months
in normal infants, but adults with WS displayed a pattern that resembled the
firing activity of three month olds. So, across developmental time, infants,
toddlers, children and !
adults with WS have impaired neuronal activity, impaired eye movement
activity and impaired grouping of features of faces and objects. This
affects their spatial cognition seriously, but also has subtle effects on
their face processing, number processing, language and social cognition. The
genetic mutations in this clinical population do not have neat, single
effects on cognitive domains, as the cognitive genetics model would require.
My team and I are at present carrying out the same exercise at the brain,
cognitive and infancy levels with other genetic disorders. Noteworthy is the
fact that even in a syndrome with a single mutated gene, like Fragile X
syndrome, the same complex story holds; subtle impairments occur across the
multiple aspects of the developing system, because of the wide effects of
one-to-many mappings and the fact that this mutated gene constrains synaptic
plasticity throughout development.
The brain is an organ for learning, equipped with a number of different
learning mechanisms. We need to understand how the brain gradually sculpts
itself in interaction with the structures inherent in environmental input,
and how learning progressively takes place over developmental time. I
conclude that nativists and those who believe in cognitive genetics cannot
call on data from adult neuropsychology and genetic disorders to bolster
claims about genetically determined, modular specialisations of the human
brain. This is not, in my view, the dawn of cognitive genetics but the
dawn, I hope, of forging an understanding of how genes are expressed through
learning and development, because the major clue for the heroes of my talk,
Holmes and Watson, turns out to be the very process of development itself.
Copyright Guardian Newspapers Limited
--- Message from peirce-l forum to subscriber cunningh@indiana.edu To unsubscribe send a blank email to: leave-peirce-l-38161V@lyris.ttu.edu
This archive was generated by hypermail 2b29 : Sat Dec 01 2001 - 01:01:00 PST