Monday, June 16, 2008

Riitta Salmelin at Paris

Let me continue my summaries of the Paris conference workshop talks. The next speaker is Riitta Salmelin from Helsinki University. She is a professor and leader of the Language Perception and Production group. Check out her web page. She is actually a physicist by training like myself. I only briefly spoke to her at Paris. She is mainly interested in language perception and production per se, and studies stuttering as an interesting example to understand the general language processes better. That's fair enough, as long as we get interesting results from her lab!

She is best known for her MEG work on the stuttering brain. In fact, I believe that she is the only scientist who used Magnetoencephalography to study stuttering. MEG is a very interesting imaging method because it opens up a different world to the experimentalists: the world of millisecond changes of electrical activity of neurons as opposed to the fMRI world of "slow and stretched over time" pysiological changes in blood flow enduced by neuronal activity. I asked her why no-one else has done MEG, and she said: It is difficult to do! In the last years, progress on MEG has been considerable mainly due to new hardware and algorithms. MEG is especially useful to create movies of neuronal activity unlike fMRI. So you can see how activity goes from one region into another. Such timing studies are very important to establish causal relationships between different brain regions. For example with fMRI, you can only say that there is a correlation between two regions, but due to the low time resolution you cannot easily say which one was first.

Her talk was divided in two parts: first she gave a summary of her work over the last years, and secondly, she presented a newly developed MEG method that allows to extract relative timing directly. In a 2000 MEG study (see their Brain article abstract), they found abnormal functional activation in people who stutter even during fluent speech: earlier activation in left motor and pre-motor cortex but delayed activation in left inferior frontal cortex when preparing an utterance, and stronger activation in right motor and pre-motor cortex during speech production. From studies that contrasted silent to overt reading (which allows them to see which additional regions are involved in overt speech production), they interpret their findings as pointing to abnormalities in processes specifically involved in overt speech production rather than core linguistic analysis by stutterers. Moreover, she spoke about abnormal auditory processing in people who stutter. I cannot remember what her interpretations of these findings are. Then, she also spoke about the interesting observation that hand regions are strongly involved in overt reading, and speculates that stutterers do not have the same motor cortical specialisation.

Finally, she explicitly spoke about a new method of analysis called Dynamic Imaging of Coherent Sources (DICS) that they have spent years to develop, and which is now ready to go. Check out the abstracts here and here. The papers are quite mathematical, and I doubt any biologist or neuroligist will completely understand them. During her talk, she showed some graphs of real-time neural connectivity during reading, I think. They will apply (are applying?) their new method to stuttering, and I am very keen to see the results. We would hopefully be able to see what is really going on during stuttering and fluent speech in people who stutter.

In fact, I am getting more and more enthousiastic about MEG as opposed to fMRI and MRI which feels more and more like getting at the crime scene just after the murder or watching the crime far away rather than during the murder and close up, which is of couse much more fun! So be prepared to hear more about MEG in the future on this blog, and I will bet my money on this new approach to move the field forward.

No comments: