Monday, June 04, 2012

G r e a t e r / l e t t e r / s p a c i n g / helps reading in dyslexia

Simply increasing the spacing between letters improves the reading ability of children with developmental dyslexia, according to a group of Italian and French researchers (Zorzi et al., 2012). Dyslexic children were 20% faster and twice as accurate when reading the altered text. This impressive result was obtained without any prior training whatsoever.

The study was based on the phenomenon of crowding, where the recognition of individual letters is impaired by the close proximity of surrounding letters. Children with dyslexia are disproportionately affected by crowding, compared to normally developing children (Martelli et al., 2009). Other aspects of the printed word are known to affect reading ability, but surprisingly little is known about letter spacing. The recommendations of the British Dyslexia Association include optimizing the size and type of font, page layout, headings, type of paper, and line spacing but not letter spacing.1

The collaborative effort was a deliberate attempt to compare two languages that have different types of spelling-to-sound translation. Italian has completely regular spelling rules (a transparent orthography), meaning there are no exception words. Each combination of printed letters is always pronounced in a consistent way. By contrast, written French is orthographically opaque, meaning that pesky irregular spellings can trip you up. This is true in English as well: compare the pronunciation of the word "pint" to "hint", "mint", and "lint". The /i/ sound wins out over the /ī/ sound, in terms of regularity.

In the study, 34 Italian and 40 French children with dyslexia were tested on two separate occasions at least two weeks apart. They read 24 short sentences, which were written in standard text in one session and highly spaced text in the other. The order of sessions was counterbalanced to control for practice effects,2 with half assigned to read the spaced text at T1 and the other half at T2. Reading accuracy (number of errors) and reading speed (number of syllables per second) both interacted with test session (p<.0001), indicating a drastic improvement with the highly spaced text. This was true for both the Italian and the French children with dyslexia.

Fig. 2 (Zorzi et al., 2012). (C) Reading accuracy (number of errors) in the normal and spaced text conditions for Italian dyslexics, French dyslexics, and a younger group of Italian control children matched for reading level (RL) to the Italian dyslexic sample.

It came as quite a surprise to me that no one had demonstrated this letter spacing effect before. But then again, I'm not familiar with the literature on developmental reading disorders, so perhaps Professor Dorothy Bishop or Livia Blackburne can provide a more critical take on an [apparently] amazing finding.

Finally, the authors have developed DYS, a free iPhone/iPad application. You can test out the spacing effect for yourself and submit your results anonymously, in the name of science!

For more information, see the WSJ Health Blog.


1 Also note that bold is preferable to italic, as the latter induces crowding.

2 A control experiment in a different group of children presented the normal and spaced text within a single session, again in counterbalanced order. The critical difference here was that different sentences were used in each condition, so practice effects wouldn't be an issue.


Martelli M, Di Filippo G, Spinelli D, Zoccolotti P (2009). Crowding, reading, and developmental dyslexia. J Vis 9: 14, 1–18.

Marco Zorzi, Chiara Barbiero, Andrea Facoetti, Isabella Lonciari, Marco Carrozzi, Marcella Montico, Laura Bravar, Florence George, Catherine Pech-Georgel, and Johannes C. Ziegler (2012). Extra-large letter spacing improves reading in dyslexia. PNAS. doi:10.1073/pnas.1205566109.

Subscribe to Post Comments [Atom]


At June 05, 2012 7:34 AM, Blogger acorniv said...

Letter spacing is important for anyone with a visual perception disorder, which is caused by the eyes failing to team properly. When eyes don't team, both eyes focus independantly and causes blurring and movement - experienced as vibration and sliding to the viewer. When someone with this reads a line, the eye tends to slip down between words to the line below. You can imagine how hard that would make it for the reader to comprehend what they read. Spacing letters allows each eye to see the letter and interpret it. The font and slickness and color of the paper all contribute to the ease or difficulty of reading for anyone with this. Pulp paper makes letters bleed and fuzz., and thin fonts strobe more than sturdier ones, for example.

I wish every child over age 7 (the age at which eyes should have teamed) were screened for a visual processing disorder. It's a simple process ( my own Behavioral Optometrist has a test that can be done online for free in a few minutes). It would certainly spare children of the frustration of not being able to read, and spare both the child and parents from the ridiculous sorts of things teachers and school testers say and do to children with this. My son and I both test with very high IQ's but because of VP's we struggled with letter reversals and slow reading. Our schools both informed the parents the child was "mentally retarded" (the actual term used) but after testing, they decided this was nothing more than acting out and both of us were punished for it. I suffered for years from ongoing punishments, which my mother,trusting the school, supported and increased at home. I pulled my son out and home schooled him to spare him the same fate. I never heard of VP until my son was about 9, when we were both diagnosed, By then my son had taught him self to speed read, Evelyn Wood's style. I recommend bright students be introduced to that, and also, that anyone with VP see a specialist ( called either a Behavioral or a Developmental Optometrist). It is a travesty that researchers haven't done a better job of investigating causes for reading difficulties ( really? they never considered spacing the letters?) and the abuse students routinely recieve in schools because of the ignorance of school employees is also a travesty. This is nothing but a vision problem, but it is not covered by medical insurance, and treatment can easily run into thousands of dollars. Another travesty.

At June 05, 2012 8:06 AM, Anonymous Richard Kunert said...

Really interesting paper. Thanks for raising it in your blog.

I am not sure about one thing in the blog post and indeed the paper:

You write that Italian and Frech differ in terms of 'types of spelling-to-sound translation.' with Italian having 'completely regular spelling rules (a transparent orthography)', i.e 'Each combination of printed letters is always pronounced in a consistent way.' French, on the other hand, 'is orthographically opaque, meaning that pesky irregular spellings can trip you up.' You offer the following analogy from English: 'compare the pronunciation of the word "pint" to "hint", "mint", and "lint". The /i/ sound wins out over the /ī/ sound, in terms of regularity.'

What you refer to is called feedforward consistency (also called phonological consistency) at the rhyme level, referring to the difficulty of knowing how to pronounce a spelling (compare mint/pint and king/sing). However, in this respect French is actually quite regular. According to Ziegler et al. (1997) 31% of English one-syllable words are inconsistent. French one-syllable words are only 12% inconsistent.

The picture is quite different when looking at feedback consistency (also called spelling consistency), referring to the difficulty of how to write a sound (compare dutch/such with king/sing). In this respect 79% of French one-syllable words are inconsistent. For English it is 72% (Ziegler et al., 1997).

I tried to find similar values for Italian, but was unable to.

My point is this: French's spelling-to-sound translation is actually quite regular. Its sound-to-spelling translation, on the other hand, is very irregular. Given that it is the former one that appears intuitively more important for reading out loud (as in the Zorzi paper), I wonder whether the French replication is really all that surprising. An English replication, however, could turn out quite different, especially if consistency is manipulated rather than broadly assumed to be language inherent.

I am working on a paper about consistency in dyslexic word recognition at the moment. As soon as it is out, I will blog about it here: .

At June 05, 2012 11:38 AM, Anonymous Anonymous said...

"It's a simple process ( my own Behavioral Optometrist has a test that can be done online for free in a few minutes)"

acorniv, could you link to it please?

Letters spaced too closely together make captcha a nuisance even without dyslexia.

At June 06, 2012 12:34 PM, Blogger Allison M. Hermann, Ph.D said...

This is a significant study - terrific design. One can imagine how this can be included as a feature in screen readers. Thank you for your post!

At June 06, 2012 3:26 PM, Blogger Eric Charles said...


At June 06, 2012 4:16 PM, Anonymous Anonymous said...

I never knew EXACTLY how children and adults had such a hard time reading, but with me not being dyslexic I found it easier to read these words (spaced out). According to this study shows that it really works! I am just happy to see that there are ways to help or cure someone without medication. This was a very effective and smart way to help those that have trouble reading!

At June 06, 2012 11:46 PM, Blogger deevybee said...

Thanks for drawing attention to this paper.
As background: many years ago, everyone assumed dyslexia was a visual disorder; around the 1970s-80s, there was growing recognition that for most children with dyslexia, there were major problems in phonological awareness - i.e., recognising the individual speech sounds in words. The ‘phonological core deficit’ account now dominates, but there are still unanswered questions. And the pendulum has swung back recently, with some researchers arguing that problems with visual processing are important for at least a subset of poor readers.
To my mind, there are two questions raised by this paper. First, does increased letter spacing make it easier for dyslexics to read? Although there are some niggly issues about the design and stats, esp. for experiment 1, the overall results do seem convincing on that point. In fact, this result fits with anecdotal reports by dyslexics, who have argued that various aspects of the layout of text can make a big difference to them.
But there is another important question, which is whether dyslexics differ from other people in this regard. Maybe *everyone* finds spaced text easier - and the improvement would be particularly evident in anyone who doesn’t read all that well. Consider this analogy: suppose people were asked to read a text while balancing a box on their heads. Poor readers may show a bigger deficit from the box-balancing than good readers, but this wouldn’t mean that box-balancing could tell us anything about dyslexia, other than that dyslexics struggle to read, so if you make the task harder, they will be particularly impaired. Dyslexics could be sensitive to any manipulation that makes reading easier or harder, just because reading is challenging and consumes more of their attention and energy. The authors attempted to address this question using the ‘reading-level matched’ control design, with results illustrated above in Figure 2. This design is widely used as a way of asking whether research findings could just be a consequence of poor reading, rather than relating to its cause. If anyone who is not so good at reading - either because they are young, or because they are dyslexic - showed the same spacing effect, this would suggest it’s nothing specific to dyslexia.
But looking at the results in Figure 2, there’s something not quite right. The control group are supposed to be younger children whose reading efficiency is the same as dyslexics on a measure of words accurately read per second. But their reading accuracy looks much better than the dyslexic groups when standard (unspaced) text is used. The statistical interaction between group and spacing arises because the groups converge when spaced text is used. What we want to see is results for a group who make the same number of errors as dyslexics with regular text. If they failed to improve as much as dyslexics with spaced text, this would clinch it. As it stands, I don’t find the evidence convincing that the spacing effect is something special to dyslexics. The ‘reading level matched’ group just aren’t very well matched. This weakens the theoretical impact of the findings.
If you’re dyslexic, though, you may not care! Anything that can help you read will be welcome, and the app that the authors have produced sounds as if it would seem to be worth a try.

At June 07, 2012 12:47 AM, Blogger Neuroskeptic said...

It would be interesting to see whether it's increasing the spaces w i t h i n each word that matter, or the spaces between the words.

If it's the latter, it would be good because it would take up less space.

At June 07, 2012 2:37 PM, Anonymous Rhiannon Barrington said...

For my undergraduate dissertation I looked into the effects of inter-letter spacing on children's eye movements during reading. Through the use of an eye tracker I found that increasing inter-letter spacing facilitates reading in children, these children were typically developing. The children read the sentences presented to them faster when the spacing was larger than usual.

To investigate this further I manipulated the size of the increase of spacing, using 1.2pt or 2pt (double spacing). Children performed best in the 1.2pt spacing condition compared to the increased of 2pt. The increase of 1.2pt, probably, was more facilitatory as it was able to reduce crowding effects of the letters, yet not disrupt the visual word unit as a whole.

Furthermore, research has been published finding facilitatory effects of inter-letter spacing in adults (Paterson & Jordan, 2010).

It appears that increasing inter-letter spacing, even slightly, can help a range of people during reading. It is an extremely interesting finding which highlights the possible benefits in changing the way in which we present written text.

At June 08, 2012 1:43 AM, Blogger The Neurocritic said...

Thanks for the informative comments. There has been some degree of skepticism about this finding, primarily from those who support the dominant role of phonological processing deficits in dyslexia (and hence remediation thereof). On the other hand, some novel evidence has come to light, namely that letter spacing helps typical readers too. I have a few specific thoughts below.

Richard Kunert - You raised an interesting point about feedback consistency effects. {Not knowing any French, I was forced to use English examples.} A recent paper by Ziegler et al. (2008) reviewed the contradictory evidence about whether feedback consistency affects visual word recognition. Although a number of earlier papers supported this (for both lexical decision and naming), new experiments led them to conclude that it does not. I look forward to your paper on consistency effects in dyslexic word recognition.

Anonymous of June 05, 2012 11:38 AM - I had never heard of Behavioral Optometrists, but I did find this NYT article, which had a lot of critical voices.

Neuroskeptic - I think both are needed, because you want to enhance the recognition of individual letters while avoiding what was difficult about the t i t l e / o f / t h i s / p o s t -- not enough spacing between words (due to b l o g .... f o r m a t t i n g .... l i m i t a t i o n s).

deevybee - Thanks so much for your insights. Although the authors claimed that spacing didn't significantly improve the performance of 'reading-level matched' controls, your point about the less-than-optimal matching here is well taken. Which brings us to the most recent comment...

Rhiannon Barrington - Very interesting results! I hope you'll publish them. And thanks for pointing out the Paterson & Jordan paper - Effects of increased letter spacing on word identification and eye guidance during reading.

At June 08, 2012 6:22 AM, Blogger Eric Charles said...

To add to the "it is not just phonological" theme... I recall their being a lot of research a while ago showing that using colored paper could help dyslexic's significantly. Given that black on white is not the easiest color pairing for normal people (in projections white on black is better, and on paper there are some other combinations that work better), that further suggests there might be a continuum of reading ability, where "normal people" can simply handle a little more non-idealness than "dyslexics".

At June 08, 2012 11:47 AM, Blogger Alexander Riccio said...

Hahahaha even the URL is spaced

At June 11, 2012 10:45 AM, Anonymous Anonymous said...

Ive heard that a person with dyslexia sees words as jumbled and floating letters. So the idea to space the letters apart isnt such a bad idea.Visual perseption is a huge part of dyslexia and if that could be fixed just by changing the way things are written would almost seem to easy. What i want to know is if there are any disadvantages to changing the spacing and or tools to the lettering.

At June 11, 2012 11:17 AM, Blogger Eric Charles said...

In the digital age, it might make little difference. In a print age, it makes a huge difference. 20% more spacing would have required 20% more paper.

At June 20, 2012 1:17 AM, Anonymous Anonymous said...

But but... How about all that Merzenich/Tallal auditory training program to fix dyslexia?

At July 04, 2012 4:15 PM, Blogger PeggyL said...

I was EXTREMELY unhappy when the online and print standard for blank spaces after a period changed from two spaces to one. Never did understand how or why that happened. I hope this kind of research will help to put it back the way it was, which made it easier for ANYONE to read a dense paragraph with understanding. (And I add a second space every chance I get.)

At July 04, 2012 4:19 PM, Blogger PeggyL said...

It would also help EVERYONE if we changed the convention for the number of blank spaces after a period back to what it used to be -2 spaces.

At September 19, 2012 12:51 AM, Blogger Chris Phoenix said...

Anonymous of June 20: (I'm writing for the blog author as well. I would like to discuss this comment with you, or anyone. Email me at cphoenix at gmail dot com.)

There are lots of researchers who have found lots of interventions that work in carefully-controlled studies. Geiger/Lettvin at MIT: training reading direction (purely visual). Tallal/Merzenich/Shaywitz: Purely auditory training. Then there are the colored lenses, etc, all the way back to Delacato and movement training.

It seems unlikely that they're all wrong, or all-but-one wrong. The alternative is that dyslexia must be capable of influence through multiple sensory modalities. And in fact, this is also suggested by the number of "dys" conditions - dysgraphia, dyspraxia, dyscalculia, etc, etc - involving very diverse tasks.

From my own reading, talking with various researchers, observations providing Davis Dyslexia Correction for several years, and experiences as a mild dyslexic/dyscalculeic/dyspraxic, it seems clear that dyslexia is a sensory scramble involving multiple senses - furthermore, it can vary over time, quite rapidly, dependent on task and emotional state. "Conditional dyslexia" is recognized but not widely talked about. I know a man who can't read at even individual words - unless he's totally relaxed - and then he can read whole phrases until he realizes what he's doing.

In short, dyslexia is context-dependent. Confusion or upset - anything disorienting - make it worse. Conversely, just like singing can bypass stuttering, changing some aspect of the task can bypass dyslexia... for a while, until the confusion accumulates again. I'd be very interested in a long-term study of whether the letter-spacing improvement persists after a few thousand pages of text. (Likewise for the colored lenses.)

People can be trained, over time, not to use the "sensory scramble" mode. This is what Merzenich aims explicitly to do. Geiger and Lettvin apparently accomplish it also, in at least some cases - they told me of a person who came to them for treatment, but stopped partway through because he was losing his ability to do interior design.

Let me make it explicit: the "sensory scramble" is not all bad. It can be very useful, when dealing with non-symbolic input. Pick up an object and turn it over in your hands; the more you look, the more you learn its shape. Now, try to read something printed on the object... that's the two faces of dyslexia. Some people, for some tasks, can use the sensory scramble. This may be the source of the belief that dyslexics are "visual thinkers." People with permanently well-integrated perceptions can only see what their eyes see. A lot of dyslexics can "move their mind's eye" and quickly imagine/synthesize a useful realistic view of what they're seeing from some other viewpoint.

(Wow, I wrote over 4096 characters; I'll have to break this into two comments.)

At September 19, 2012 12:51 AM, Blogger Chris Phoenix said...

If you've read this far, allow me to speculate. Suppose that dyslexia is based on a weakness in the magnocellular system, so that timing information on sensory input sometimes gets lost. This could, among other things, cause the saccade reassembly mechanism to stitch together fragments of visual input to make an apparently-coherent version of what is seen, with more flexibility than non-dyslexics have. There might also be some compensation in the parvocellular system, increasing sensitivity to visual detail. Apparently, there are analogs of the visual magnocellular system in the auditory system.

It's worth stressing how rapidly the perceptual distortion can come and go. I've seen kids who could not even copy geometric shapes correctly, after half an hour of training, make the whole alphabet correctly - and several days later, read much better. And I've seen many people go in and out of the sensory-scramble state in mere seconds, once they learned to control it.

When giving a Davis program, I could watch people get disoriented and start to stumble while reading - it happens quickly enough that I could frequently identify which symbol caused the confusion, and with practice, so could they. Then I could remind them to get "on point" and within seconds they'd be reading better again. (And we'd stop and learn the offending symbol. It's amazing how quickly a dyslexic can learn symbols, once they know how to straighten out their perceptions.)

As you can probably tell, I'm a big fan of the Davis approach - it's the only one I know of that gives lasting benefit _and_ lets dyslexics keep using their perceptual flexibility when they want it, rather than training it away entirely. But I want to repeat the point I started with - there are lots of dyslexia interventions that demonstrably work, using a variety of sensory modalities. (Davis has several modes of turning off the scramble - visual imagination, kinesthetic imagination, and sound.)

At September 22, 2012 11:18 AM, Blogger manuel perea said...

Nice blog. A quick note. There is recent work (earlier to the submission of the Zorzi et al. paper) on the effects of small increases in interletter spacing during visual-word recognition with skilled readers, and also with dyslexic children:

To cite just three examples:

Perea, M., Moret-Tatay, C., & Gómez, P. (2011). The effects of interletter spacing in visual-word recognition. Acta Psychologica, 137, 345–351. DOI: 10.1016/j.actpsy.2011.04.003

Perea, M., & Gómez, P. (2012). Increasing interletter spacing facilitates encoding of words. Psychonomic Bulletin and Review, 19, 332–338. DOI: 10.3758/s13423-011-0214-6

Perea, M., Panadero, V., Moret-Tatay, C., & Gómez, P. (2012). The effects of inter-letter spacing in visual-word recognition: Evidence with young normal readers and developmental dyslexics. Learning and Instruction, 22, 420-430. DOI: 10.1016/j.learninstruc.2012.04.001

At September 22, 2012 11:57 AM, Blogger The Neurocritic said...

Professor Perea - Thanks for providing the references. It seems the authors should have cited your work. But then their finding wouldn't have been as novel...

Sorry that happened.


Post a Comment

Links to this post:

Create a Link

<< Home

eXTReMe Tracker