Stanislas DEHAENE,
Les neurones de la lecture, Odile Jacob, 2007.
For a short introduction to this book
click on « intro.lecture »
Contents :
Introduction : The science of reading
1.How we read
2.The literal brain
3.The reading neurones
4.The invention of reading
5.Learning how to read
6.The dyslexic brain
7.Reading and symmetry
8.Towards cultivating neurones
Conclusion : The future of reading
And now, let’s take a quick look at the reading process
from a neurophysiological point of view.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
How do we read?
Everything begins in the retina, where the photons
which are reflected by the page are projected (p.36).
The retina is the membrane which covers the back of the
eyeball : it is made up of several layers of cells which have
different functions.
It receives the images which enter through the eye, transforms
them into signals or electronic impuses and transmits them to the
brain through the optic nerve.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
The retina isn’t uniform, in the sense that only the central region of
the retina, called the fovea, contains numerous high resolution
photoreceptor cells, the cones. This region, which takes up only
about 15 degrees of the visual field, is also the only area of the
retina which is really useful for reading. It is the only area which
can perceive letters with sufficient detail to allow them to be
recognized (p.36).
This can be illustrated as follows :
The human eye and some of its components
cristalline lens
retina
Message read
«If the months and the days are eternal
passengers, the years which follow on
likewise travel. Whether one navigates
on a skiff all one’s life or one pulls on the
horse’s bit until the threshold of old age,
every day is a voyage, and one makes
one’s home of this voyage. I no longer
know in which year a solitary cloud
invited me into the wind, (...).
cornea
vitreous
humour
pupil
(Bashô, 17th century Japanese poet.)
iris
fovea
optic
nerve
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
The fovea’s narrowness forces us to move our eyes
constantly while we are reading.
Moreover, we don’t read through a text in a single continuous
process. Our eyes move in saccades (short rapid
movements). This is due to the fact that in the centre of the
fovea, visual information isn’t perceived with the same precision
everywhere. (…) Precision is maximum in the center and
diminishes towards the exterior. (p.37)
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
We identify 10 or 12 lettres per saccade : 3 or 4 to the left
of the centre of fixation and 7 or 8 to the right. This is
what is called the span of visual perception of letters.
This asymmetry comes from the direction in which
we read. A person who is reading Arabic or Hebrew,
whose fixation goes from the right to the left, has an
inverted span of visual perception (p.41).
Our eye imposes enormous and irremovable
constraints on reading. It has been proven that the
ocular saccades are actually what limit our reading
speed (p.42).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
When a word enters the retina, it breaks up into
a thousand pieces : each part of the image on the
page is recognized by a distinct photoreceptor.
The difficulty resides in re-assembling the
fragments in order to decode the letters which are
involved, the order they are presented in, and the
word which is in question. (p.35)
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Where does this take place ?
In a certain part of the brain, there is the visual word
recognition system : in each individual, in every culture in the
world, the same cerebral region, give or take a few milimeters,
intervenes to decode written words. Whether one reads in French
or in Chinese, learning how to read always goes through an
identical circuit (p. 27).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Our visual word recognition system perceives what doesn’t change
despite the variety of forms which words can take (size, fonts,
capital/small letters, bold type or not, underlined or not, …) : this is what
is called unvariable word recognition.
Examples : TWO, two, two, TWO, Two, …
R, R, R, r, r, r
Thus the system learns to ignore all variations which aren’t pertinent
for reading and on the other hand, to locate and amplify pertinent
differences, even tiny ones.
Example : the difference between « live » et « love »,
between « seat » et « sate », …
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
How does this take place?
Our visual system automatically breaks words down into their
basic components. The nature of these components remains
an ongoing subject of research (p.51).
Example :
unbuttonned
It’s likely that multiple levels
of analysis coexist (p. 51) :
un
button
un
Letters  Graphemes 
Syllables  Morphemes 
Words
u n
but
b
ut
ned
ton
t
ned
on n ed
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Broken down in this way, these elements can then be used
by the brain to give sound and sense.
Two channels exist :
- the phonological channel or the sound channel (= oralisation or silent
reading : this doesn’t involve either articulating or moving one’s lips, but rather
transforming the letters into sounds to reach the pronunciation of words). Also
called grapheme-phoneme conversion.
- the lexical channel or the direct channel which gives immediate access to
meaning.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
This is a very controversial subject, as scientists have differing
opinions :
- for some, it is impossible to reach meaning without first going
through the phonological channel
- for others, going through the phonological channel is a
characteristic of a beginning reader and not of a good reader.
The author’s position: Today a consensus is beginning to
appear : for the adult, both channels of reading exist and
are active simultaneously. (…) They therefore function on
a parallel basis, one supporting the other (p.53).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
The phonological channel : the only possible way to read new or rare words
with regular spelling, neologisms, …
Decoding letters, then searching for possible
pronunciation, then meaning.
The lexical channel : used for common words and indispensable first for irregular
words (numerous in French and even more so in English)
Decoding letters, then searching for
meaning, then pronunciation.
Languages which are
very rich in phonemes !
≠ for Italian where there are almost no irregular
words – each letter corresponds to a sound 
children’s reading scores are significantly better
than for French children and there is almost no
dyslexia!
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Neither of these channels, of itself, is sufficient for reading all words (p.70).
When we read aloud, the two channels
conspire together, as it were and collaborate
with each other (p.70).
Thus, most contemporary psychological models agree with
the idea that expert and fluent reading is the result of close
coordination between the two channels of reading (p. 71).
Please note : it would be more correct to speak of the multiple channels of
reading.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Please note with respect to the lexical channel :
It is based on the storage of tens of thousands of words in a
« mental glossary » or rather in several glossaries : orthographic,
phonological, grammatical, semantic.
And all of these glossaries act in parallel and
not at all sequentailly  great efficiency and
rapidity ! (see pages 74 and following for the
pandemonium metaphor).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Without the action of our mental glossary, the written word would remain « a
dead letter ». The identification of letters and words is an active decoding
process during which the brain adds information to the visual signal (p.80).
In order for a word to be recognized, the multiple
cerebral systems have to agree on an unequivocal
interpretation of the visual input. The time that it takes
us to read a word, therefore, depends less on its
intrinsic characteristics than on the conflicts or
coalitions that it induces within our cerebral
architecture (p. 82).
Our glossary is like an arena where there is
strong competition and where the advantage
goes to the « regulars », that is to say, the
words used most frequently (p. 82).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
In short, in the brain :
Each lobe specializes in
one or several sensorial
functions.
VISION
TASTE OLFACTION
HEARING
Left hemisphere : external view
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
The information (words, faces, objects, …)
perceived by the eyes, activates the visual areas
of the occipital lobe in each hemisphere.
L
R
L
R
These regions make a primary analysis
of the image, probably in order to
extract the elementary forms (lines,
curves, surfaces, …) . At this stage of
information processing, the brain
doesn’t yet know what kind of stimulus
it is dealing with (p. 115).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Then (50 milliseconds later), the information begins to be sorted and the words
give rise to the activation of the word recognition area which we have just
spoken about (especially in the left hemisphere, in the ventral occipitotemporal region).
All of this happens
automatically, in less than
fifth of a second !
a
And after visual recognition, what path does
reading take ? How do we have access to the
sense and the sound of words ?
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
The occipito-temporal region which we have just
spoken about then distributes information to
numerous cortical regions.
In two principal circuits : one which
converts the information into sounds,
the other which gives it meaning.
These regions are not specifically for reading.
And these two channels for reading (which give access
to the meaning and the sound of words) activate
distinct cerebral areas.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Conversion of letters into sounds :
The left temporal lobe is largely implicated, notably an
upper region of the lobe which is called the temporal plane.
Because it allows visual and auditive information
to meet, the temporal plane very likely plays a
role of essential junction in the process of
learning how to read (p. 152).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Means of access to meaning :
Several regions are activated, none of which
is specific for written words.
The complexity of these mechanisms is such that it
is impossible to summarize in just a few lines !
We are just at the beginning of the neurology of meaning. (…) In the area of
meaning, humility is necessary, because no one, for the moment, can claim to
have a precise neurological model of this mysterious flash of understanding
which means that the activity of a network of neurones, all of a sudden,
« makes sense » (p.155).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
We know at least one thing, that it would be naive to think that
meaning is limited to a small number of cerebral regions. On the
contrary, semantics calls on very vast populations of neurones
distributed in all of the regions of the cortex (p.156).
And in chapter 5, entitled « Learning how to read » the author
shows the ways in which learning how to read modifies a child’s
brain; he describes the phases of this learning process and
suggests directions to follow to optimize the teaching of reading.
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Notably, he demonstrates the inefficiency of the global method :
To sum up, today there can no longer be any doubt : the global
outline of words plays practically no role in reading. Visual
recognition of words is not based on a global perception of its
outline, but on breaking it down into simple elements, letters and
graphemes. The cortical region of the visual form of words handles
all of the letters of the word in parallel, which is responsible,
historically, for the impression of global reading. But the immediacy
of reading is merely an illusion, caused by the extreme
automatization of the different steps, which takes place outside of
our awareness (p. 297).
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
To conclude this overview, we would like to highlight in chapter 6 where
the author speaks about dyslexia :
In the majority of cases, dyslexia can be linked to a faulty
mental manipulation of phonemes. Dyslexic children’s brains
present several characteristic anomalies : (…).
Do these anomalies imply that dyslexia is incurable ?
Not at all. (…) (p.309)
Paper and Microsoft Power Point Presentation 2003
by Hélène Delvaux for the IF Belgiim
Images : clipart at
http://office.microsoft.com
For the European project Signesetsens
2009
Stanislas DEHAENE, Les neurones de la lecture, Odile Jacob, 2007.
Some introductory words :
P. 1/4
Behind each reader there is hidden a neuronal mechanism which is
admirable for its precision and efficiency, the organisation of which
we are beginning to understand. During the last twenty years an
authentic science of reading has been born (p.21).
In this fascinating and remarkably clear book, the author tries to
share this science of reading, as well as the experimental progress
which supports it. He hopes to see the appearance of a true
educational neuroscience which would allow for the optimization of
teaching strategies ( by permanently setting aside certain methods
used to teach reading, such as, for example, the method which is
known as the global method, which he has shown to be inefficient
because it isn’t adapted to a child’s cerebral organisation).
There are two opposing models used to « explain » the brain : the
old model, that of generalized plasticity and cultural relativity, and
the new one, which the author defends, that of neuronal recycling.
P. 2/4
The old model develops the following ideas :
• The brain is a completely flexible and malleable organ ; therefore
it poses no constraints on the full range of human activities (p.26).
• The human brain is a sort of clean slate on which natural and
cultural data from the environment are imprinted.
There is thus no biological human nature, but a progressive
construction of human nature through immersion in a given culture.
• It is only the capacity to learn which is characteristic of our human
nature (p.26-27)
This model has been challenged by recent data from cerebral
imagery and from neuropsychology. We shall see to what extent the
image of an infinitely malleable brain, like a clean slate, and which
would limit itself to absorbing information from its cultural
environment, is false (p.27).
P. 3/4
The author develops another model, that of neuronal recycling. Of
course our brain is capable of learning and shows plasticity and the
ability to adapt to the environment, he says, but its architecture is
strictly managed by strong genetic constraints. (p.27) This learning is
therefore limited. For example, in each individual, in every culture in
the world, the same cerebral region, give or take a few milimeters,
intervenes to decode written words. Whether one reads in French or
in Chinese, learning how to read always goes through an identical
circuit (p. 27).
This model is largely based on the idea that our cortical circuits,
inherited from our evolving past, adapt as best they can to reading :
learning how to read imposes profound modifications on brain circuits
(pp. 22-23). The brain is a highly structured organ which makes new
things out of old. To learn new competencies, we recycle our old
cerebral circuits of primates – to the extent that these tolerate a
minimum of change (p.28).
P. 4/4
The paradox of reading underlines the indubitable fact that our
genes haven’t evolved so as to allow us to learn how to read. I only
see one solution. If the brain hasn’t had the time to evolve under
the pressure of the constraints imposed by what is written, then it
is what is written which has evolved to take into account the
constraints of our brains (p.29). And a bit further along, the author
finds the traces of an unceasing, evolving construction which
relentlessly adapts the written objects to our brains’ constraints.
Please note : reading in this book is only directed towards reading words and
not images.
Descargar

Diapositive 1 - Signes et Sens