Jonathan Levy, Unraveler of the Unconscious

Recent research in PLoS ONE explored the neurological dynamics of reading in relation to functional anatomy.  It was found that words are routed through one distinct channel of the brain while non-words, or strange new words, are instead routed through their own distinct channel.  A summary of this research was published Monday in the article “Channeling grimps, yamps, and blin“.  What follows is the transcript of a text-based interview with the principal author of the study, Jonathan Levy.  Mr. Levy is affiliated with INSERM UMR 825, Université de Toulouse, the Donders Institute for Brain, Cognition and Behaviour and Radboud University Nijmegen.  He will defend his doctoral thesis within a couple months.

1)    What initially attracted you to the study of the neuroscience of reading?

I am primarily interested in unraveling whatever is unconscious, thus making it conscious. Reading offers a particular opportunity since it operates so fast (hundreds of words per minute), automatically and in such an irrepressible manner, that we are completely overlooking the numerous processes underlying it. By looking at brain activation underlying such subtle processes, we mirror people the remarkable quality of the brain, that is, its automatic and unconscious mechanism, like a computer. However, we are more than just a computer, since we also have the possibility of watching this animate computer. I think that neuroscience allows us to take distance and to watch what’s happening in our brain.

2)    Are there any concepts within the field of neuroscience that you find particularly captivating?

Yes, as said above, I think that neuroscience offers now many tools to help us understand how we operate in daily life, which is in other words, gaining consciousness or if you prefer, understanding of our body-mind. I am also attracted to brain-computer-interface investigations, which little by little offer a direct support and application to the advances in neuroscientific knowledge.

3)    Was there some odd wrinkle in the literature of this science that led you to formulate the hypothesis that you tested in this literature?

Not really. I was just curious about how our behavior (reading skill) can be predicted by our brain dynamics. So what we did was to take a group of participants and examine the major areas that they activated during reading. However, these were not just simple areas, but rather reflecting some of the most prominent reading processes (visual, orthographic and pre-lexical, phonological processes). We did so by testing brain activations induced by many possible linguistic and linguistic-like stimuli, from pseudo-letters (letter segments were visually rotated and manipulated so that they do not remind letters, but at the same time consist in the same visual complexity as letters do) to words and pseudo-words (illegal words, yet maintaining orthographic rules of real words, e.g. ‘scneer’ which I made from ‘screen’). The next step was that we checked information trafficking between these areas. We were intrigued to find out such a strong correlation between such trafficking and reading skill.

4)    One of the points that was stressed in this paper was that the functional processes involved in word processing were located in anatomically distinct areas.  Although we have known of specific functional areas, such as Broca’s and Wernicke’s Areas, for quite some time, popular culture continues to view much of the brain as an ephemeral computer.  Is the general field of quantitative neuroscience building up enough data to replace the popular conception of the brain with a more discretely compartmentalized view?

Of course. I actually thought that everyone thinks of the brain as very compartmentalized; I guess I am too long in the field … I actually recall that when I was teenager and I first came across the work of Antonio Damasio, and that I was so much fascinated by the accuracy between an isolated region and its emotional or cognitive function. I remember myself imagining how just by “altering” in some way (without cutting though!) one brain area, my whole daily life behavior would change, from a shy guy I would be heroic and brave…

In the field of cognitive neuroscience, we are often attempting to assign a function to a region. However, this is not that easy since there is a lot of variability between individual’s brain anatomy and also their functioning, but there is also variance within one individual, depending on the task to operate, on the characteristics of the stimulus, and even depending on the day or on the time of the day! So it is really not that simple, especially that in my field, laboratories all over the world also conduct experiments in different languages! This is why in my opinion we are often surprised to find different, or even reversed results, obtained by another group of researchers.

Nevertheless, there are some areas that area consistently appearing and reappearing, even if with variations of anatomical localization or response amplitude; such was the case also in our study where the areas which were the most prominent for the participants while they were reading, were very famous in their linguistic role (e.g. Broca’s area or the left occipito-temporal junction).

5)    It was noted in this paper that patients with acquired phonological dyslexia may have impairments in their left inferior-parietal and left infero-frontal regions, which handle grapheme to phoneme conversion.  If I understand correctly, grapheme to phoneme conversion is the process of translating the shapes of written letters into the sounds of words.  Why do we need to convert written words to auditory memories in order to comprehend text?

Actually we don’t. There are two ways of reading a word, by converting its letter forms to sounds and then extracting its meaning, or by directly extracting its meaning. This is broadly the principle behind the hypothesis of the dual-route cascade model. Acquired phonological dyslexics have trouble with reading stimuli that should be read while following grapheme to phoneme conversion (pseudo-words), while they have no problem in reading irregular words, which do not follow these conversion rules (e.g. yacht, sword, Lincoln); their neuronal pathway for non-lexical reading is impaired. Acquired surface dyslexics have the reversed pattern of impairment, namely, they easily read normal (regular) words but they are unable to read those irregular words. Their neuronal pathway for lexical reading is impaired. So to return to your question, we don’t need to read words in a ‘non-lexical’ way, which is more cumbersome and slow, unless unfortunately, our lexical pathway is damaged. In our paper we showed that the pathways between the prominent areas that we obtained could account for these two ways of reading, although we are sure that other different pathways are also involved.

6)    Similarly, do disparate languages with different transcription rules involve different requirements of grapheme to phoneme conversion in native speakers?  For example, the transcription of German is very literal and exacting whereas spoken French does not sound much like it is written.

Actually the conversion rules in French are quite straightforward… although I think that this statement may be difficult to be agreed by non-French speakers

Yes, different languages result in different linguistic awareness and abilities, and furthermore, it even involves different brain regions. In our work we discussed this point when we compared our results to those obtained in English which is very non-straightforward (opaque) in terms of grapheme to phoneme conversion rules.

7)    Is it possible that all the experimental subjects used in this study were native French speakers introduced a confounding variable into the results?

Yes of course, as said above, it involved variations in the anatomy of the functional results.

8)    Would you anticipate that data from a different patient sample (all university students in this study) might yield different conclusions?  For example, would an adult who became literate later in life process written words in comparably novel ways relative to individuals who have been literate since a very young age?

Certainly. We know that reading acquisition starts very early at school time, and as a result, neuronal specialization for reading is enhanced at that time. An important point to retain is that our brain is particularly malleable at early age. Later on, plasticity is always present, but probably to lesser a degree. Actually we addressed this point in our work since we also checked whether the age and reading practice of the participants could predict their reading skill, and if it could even predict variations in information routing within the brain. We found that age (nor gender) did not exert any influence, but that reading practice predicted reading skill. More interestingly in my opinion, we found that non of these factors could predict the dynamics within the neuronal network of reading. So this comes back to your question by suggesting that the neuronal dynamics are shaped early (early school), and that later practice (university) cannot change them, but it can change their reading skill. Interesting to know hah?

9)    With this data in hand, where do you hope to go next?  Are there any promising data or new tools coming in that you can disclose?

I am now using different tools, magneto-encephalography, which is more apt for measuring neuronal oscillations, which are a good predictor for consciousness studies, in which I am more involved now.

10)    I note in your credentials that you are formally affiliated with four different universities in three different countries.  How does one manage their time between these institutions and has the formation of the EU enabled this kind of collaborative spreading out?

I am affiliated to the INSERM UMR 825 in Toulouse (France) and to the Donders Institute for Brain, Cognition and Behaviour in Nijmegen (the Netherlands). At the same time I am doing my PhD degree (finished in a couple of months!) in a co-tutelle (joint) program under that academic umbrella of the Université de Toulouse (UPS) and Radboud University Nijmegen (RU). Yes, the EU helped a lot to this international exchange by creating such cooperative agreements, and my directors Dr. Jean-François Démonet and Prof. Pascal Fries were fantastic in accepting to coordinate me in this exchange.

11)    Lastly, just for fun, what music do you listen to most often when you are working on science?

I don’t listen to music while working, instead, I go often for walks in the park, it helps me to reflect…

Levy J, Pernet C, Treserras S, Boulanouar K, Aubry F, et al. (2009) Testing for the Dual-Route Cascade Reading Model in the Brain: An fMRI Effective Connectivity Account of an Efficient Reading Style. PLoS ONE 4(8): e6675. doi:10.1371/journal.pone.0006675