Letter-sound Training in Children Causes Brain Specialization for Letters

My research focuses on the left occipitotemporal region. One area in this region, also commonly referred to as the visual word form area, has been shown to activate selectively for letters. Presumably, since reading is too recent a phenomenon to have evolved a specialized brain region, the area develops as a result of  experience with words and letters.

To verify this, some studies have trained adults on a new writing system and scanned them pre and post training to see the effects on the occipitotemporal region. The results have been mixed, and complicated by the fact that adults already know a writing system. It would be simpler and more relevant to look at a training effect in children, and that is what Brem and colleagues did. They trained prereading kindergarteners on letters and found that sensitivity to words developed in the occipitotemporal cortex.



The children in this experiment trained on a computerized grapheme-phoneme correspondence game that taught them the sounds associated with individual letters. As a control, they also trained on a nonlinguistic number-knowledge game. The participants did eight weeks on each game, with half the group doing the grapheme training first and the other half doing the number training first. This resulted in a nice within-subject control.

The authors evaluated the children with fMRI and EEG at three time points: 1) Before training, 2) After training with the first game, and 3) after training with the second game. During the fMRI and EEG sessions, the children performed a simple modality judgment task. They were presented with either spoken or written words, false fonts, or unintelligible speech and simply had to say whether the stimulus was in the visual or auditory modality.

After grapheme-phoneme training, kids showed increased activation to words (as compared to false fonts) in the left occipitotemporal region.* The authors then looked more closely along the length of the fusiform gyrus (located in the occipitotemporal region) and found that there was an increase in activation to words in a posterior region a (MNI coordinates 46,-78, -12)**. This region is posterior to what is usually reported as the adult visual word form area. It would be interesting to see if the region shifts with age.

The EEG results supported the fMRI findings. One of the ERP components, the N1 peak, was stronger in response to words after training. The source of the N1 localized to the left occipitotemporal region, right cuneus, and posterior cingulate.

This is a nice study because we can see word expertise development in action, in the age group in which it presumably happens in real life. The authors argue based on previous literature that it’s the visual-phonological mapping that increases specialization in the fusiform, not just visual training. Apparently, previous studies with primarily visual training have not increased activation in the fusiform gyrus, while training adults on phoneme grapheme mapping did. I haven’t looked at those papers recently, but perhaps I’ll investigate them next.

* The posterior fusiform, right inferior temporal gyrus, and cuenus showed this effect.
**More specifically, the authors did an ROI analysis where they picked 5 ROIS along the length of the fusiform gyrus.  The 4th ROI from the front showed this effect.


Brem S, Bach S, Kucian K, Guttorm TK, Martin E, Lyytinen H, Brandeis D, & Richardson U (2010). Brain sensitivity to print emerges when children learn letter-speech sound correspondences. Proceedings of the National Academy of Sciences of the United States of America PMID: 20395549

Read more...

Posterior Brain Differences in Children with Dyslexia

Accessibility:  Intermediate-Advanced

I realized after the last post that we haven’t actually spent much time discussing brain differences between dyslexic and nonimpaired readers. So today, I’m covering an earlier experiment by the Shaywitz’s.



In a 2002 paper, Shaywitz and colleagues reported an experiment with 144 children aged 7-17, half dyslexic and half nonimpaired. The children performed several tasks in the scanner, but the paper focuses on two: nonword rhyme (NWR) (Does [PEAT] rhyme with [LEAT]?) and semantic categorization (CAT)(Are [CORN] and [RICE] in the same category?). A line match task was used as a baseline.

During fMRI, the nonimpaired readers showed more activation than dyslexic readers in a large number of left and right hemisphere brain regions.*

The authors also looked at brain regions where reading skill correlated with activation. The left occipitotemporal (OT) region correlated with skill in both tasks, while bilateral parietotemporal regions showed a correlation with skill in the categorization task only.

This isn’t the first time activation in the OT has been linked to reading skill. Specht 2009 found that OT activation during a categorization task correlates with reading score even before formal reading instruction. Shaywitz 2004 found activation increases in the left OT region a year after completion of a phonological intervention. Also, this paper reported negative correlation between reading skill and activation in right OT gyrus during a categorization task, a correlation that was also reported in Turkeltaub 2003**.

Finally, the authors looked at brain regions where activation correlates with age and found striking differences between dyslexic and nonimpaired readers. Dyslexic readers had many regions that increased in activation with age***. In normal readers, there were few correlations with increasing age, and age correlated negatively with superior frontal and middle frontal regions. One possible explanation is that dyslexics learn to compensate with other brain regions as they grow older. The normal readers, on the other hand, get more efficient in their reading.****

The age result also highlights the variability in the sample. Children with dyslexia change greatly in brain activation as they grow older. You can imagine the variability that this would produce in an random experimental sample of 15 kids. I wonder if there’s been much work on relative variability in dyslexic children vs. nonimpaired children.

*NWR in left hemisphere sites (inferior frontal gyrus, superior temporal sulcus, superior temporal gyrus, middle temporal gyrus, middle occipital gyrus) and right hemisphere (Inferior frontal, superior temporal sulcus, middle temporal gyrus, medial orbital.) CAT in left (angular gyrus, middle temporal gyrus, middle occipital) and in right (middle temporal gyrus, middle occipital)

**Turkeltaub didn’t find a positive correlation in left OT, and also used a lower level task (tall letter detection.

*** IN NWR in DYS, increased age correlated with activation in bilateral IFG, basal ganglia, posterior cingulate, cuneus, middle occipital gyri and left STG.

****Correlations with age in dyslexics and normal readers are also explored in Shaywitz 2007. In that paper, they do report regions in nonimpaired readers that increase activation with age. It might be the same dataset, but I’m not sure.



Shaywitz, B. (2002). Disruption of posterior brain systems for reading in children with developmental dyslexia Biological Psychiatry, 52 (2), 101-110 DOI: 10.1016/S0006-3223(02)01365-3

Read more...

Phonological Training Changes Brain Activation in Dyslexic Children

Note: Online Universities has included me in their list of top 50 female science bloggers. It’s not actually for this blog, but for my Brain Science and Creative Writing blog. Anyways, check out the list if you get a chance. There are lot of interesting bloggers.

 Accessibility:  Intermediate-Advanced

We’ve looked at the neuroscience of dyslexia and how the dyslexic brain processes words. Our ultimate goal, however, is treatment. Therefore, we’d like to see whether reading interventions cause brain changes in reading-impaired children. In a 2004 paper in Biological Psychiatry, Shaywitz and colleagues investigated this question.



The study focused on kids aged 6-9, divided into three groups. The experimental group consisted of reading-disabled students who went through an eight month experimental intervention that focused on  phonology: letter-sound associations, combining sounds, etc.. Another group of reading-impaired children were put in community intervention control group that participated in a variety of reading interventions, including remedial reading and tutoring. However, there was no specific focus on phonology. A third group, community control, consisted of normal reading children.*

All groups improved in their reading measures after 8 months (not surprising, since they continued to attend school). The experimental group showed more improvement than the community intervention group in one reading measure.

Shaywitz and colleagues were interested in brain differences before and after intervention. They scanned the kids pre/post intervention in a letter identification task.** Their main analysis was a second order comparion. They first determined the pre/post intervention changes within each group. Then they compared the changes between groups.

Compared to the community intervention control, both the experimental intervention and normal-reading control group showed a greater increase in left inferior frontal gyrus (often involved in phonological processing) activation. The experimental intervention group showed more increase in left middle temporal gyrus activation compared to the community intervention group.

In addition to comparing pre/post intervention differences between groups, Shaywitz also scanned the experimental group a year after finishing the intervention. The group showed continued increases in several left hemisphere areas, including left inferior frontal, superior temporal, and left occipitotemporal regions***. Also, they showed a decrease in right MTG and right caudate activation. This falls in line with the increase in left lateralization we saw in Turkeltaub 2003.

What’s the take home message? This study shows us that phonological intervention results in measurable brain changes in the left inferior frontal gyrus, a phonological region. This is encouraging. However, how does this actually impact reading performance? The experimental group only performed significantly better than the community intervention group in one reading measure, although it looks like they performed slightly better (but not statistically significant) in other measures. So there is a hint that phonological interventions might be more valuable than other interventions, but we’d have to get more data on this.

The study also shows that brain regions in the experimental group continue to develop during the year after the intervention. When did these changes start – during intervention or afterwards? It's hard to tell because they don't report the changes in the experimental intervention group right after intervention. They only report on the difference in changes between groups.

Also are these changes jumpstarted by the intervention, or would they have occurred anyway? Unfortunately, we can’t answer that question either. While the authors had hoped to also scan the two other groups a year afterwards, they were unable to.

Anyways, it's kinda cool to see brain differences as a result of training. It will be interesting to see in future studies what is going on in more detail.

*Children from the EI group were from Syracuse, NY, while the other two groups were recruited from New Haven.
**The kids heard a letter name and had to choose the correct letter from two options. This task was compared against baseline of hearing tone and specifying position of asterisk.
***LIFG, STG, left OT, left lingual, and left inferior occipital


Shaywitz BA, Shaywitz SE, Blachman BA, Pugh KR, Fulbright RK, Skudlarski P, Mencl WE, Constable RT, Holahan JM, Marchione KE, Fletcher JM, Lyon GR, & Gore JC (2004). Development of left occipitotemporal systems for skilled reading in children after a phonologically- based intervention. Biological psychiatry, 55 (9), 926-33 PMID: 15110736

Read more...

The Development of Visual Word Recognition

Accessibility:  Intermediate-Advanced

We’ve looked at brain regions and development during word related tasks (word generation, reading and repeating), but we haven’t yet looked at a straight up study of word recognition and development.



What’s the best task to use to study visual word recognition? You can have people read out loud, but that involves processes like speech generation. Likewise, reading sentences or paragraphs requires the reader to process meaning and grammar in addition to the words on the page.

One segment of the field has gravitated towards tasks of single word processing that don’t require reading at all. In this particular study, Turkeltaub and colleagues use a tall letter detection task. The subjects press a button if the word has a tall letter (like d or l). As a control condition, subjects perform the same task on false fonts. Even though you can do this task without reading the words, the assumption is that reading, being highly automatic, will occur anyways. This approach, focusing on the automatic, bottom up process, allows for a more tightly controlled study. However, it also limits the findings to that very thin slice of the reading process.

Turkeltaub and colleagues tested forty one subjects ranging from 8 to 20 years old. In the whole group, the words > symbols contrast gives activation in the left posterior temporal, left inferior frontal, and right inferior parietal regions.

The authors also looked at correlations between activation and reading ability. The trend here seems to be increasing lateralization (more reliance on left hemisphere regions and less reliance on right hemisphere regions), with reading skill.*   Interesting.  I wonder how this relates to lateralization of spoken language.

Finally, the authors looked for regions that correlated with other behavioral measures, including phonetic working memory (left intraparietal sulcus and left and right middle frontal gyri), phonological awareness (left hemisphere network, incluing posterior STS and ventral inferior frontal), and phonological naming (bilateral network, including right posterior superior temporal, right middle tempral, and left ventral inferior frontal.) Surprisingly (to me at least) there is almost no overlap between the regions for the three measures. This could either mean that these measures involve very different cognitive and neural processes, or that the automatic task used in this experiment was not suited for accurately tapping into these abilities.

*Reading ability correlated positively with activation left hemisphere frontal and temporal cortical areas, and negatively with right hemisphere posterior regions. There was no correlation in the left fusiform (visual word form area), but there is a negative correlation in right posterior fusiform. 

Turkeltaub, P., Gareau, L., Flowers, D., Zeffiro, T., & Eden, G. (2003). Development of neural mechanisms for reading Nature Neuroscience, 6 (7), 767-773 DOI: 10.1038/nn1065

Read more...

Rats Who Can't Read Good: A Rodent Model for Dyslexia

Accesibility:  Intermediate

Dyslexic rats? Really? Well, these rats can’t read, but they’re still used as an animal model for dyslexia.



First, some background. The underlying cause of dyslexia is still under debate, but it’s generally accepted that it involves deficits in auditory and phonological (language sounds) processing, with a possibility of visual deficits as well. Post mortem studies of dyslexic human brains have turned up brain anomalies, including cortical ectopias (nests of neurons in the wrong layer in the cortex) and focal microgyri (micro folding). Researchers have also found abnormalities in the thalamus and cerebellum.

Dyslexia rat models are created by inducing these same abnormalities, usually focal microgyria and molecular layer ectopias, in rats. Interestingly, some of these rats develop deficits in rapid auditory processing, which is important for phonological processing in humans. Introducing microgyria also causes thalamic changes in male rats, similar to dyslexic thalami in humans. The thalamic changes are also associated with auditory perceptual deficits in the males.

Another interesting observation:  boys are more at risk than girls for dyslexia, and the same trend occurs in rats. Young male rats have a higher risk for developing rapid auditory processing deficits from induced cortical malformations. There seems to be something about the male brain that increases risk for language related disorders.

Hmm, anyone want to start Hooked on Phonics for rodents?

[And kudos to anyone who got the Zoolander reference in the title]

Galaburda AM, LoTurco J, Ramus F, Fitch RH, & Rosen GD (2006). From genes to behavior in developmental dyslexia. Nature neuroscience, 9 (10), 1213-7 PMID: 17001339

Read more...