Neuroanatomy of developmental dyslexia. Franck Ramus CNRS Institut d Etude de la Cognition, Ecole Normale Supérieure, Paris

Transcription

1 Neuroanatomy of developmental dyslexia Franck Ramus CNRS Institut d Etude de la Cognition, Ecole Normale Supérieure, Paris

2 Environment Causal modelling of dyslexia Genetic difference Brain structure difference Biology Brain function difference Normal intelligence Normal perception Cognitive deficit Cognition Normal environment Expected reading level >> Observed reading level Behaviour

3 Aims of Genedys project Investigate neuroanatomical correlates of dyslexia with: More participants (than previous studies). Better quality images. More fine-grained analyses, in each individual s native space, respecting individual anatomy. Comprehensive behavioural battery DNA With C. Billard, G. Dehaene and the Neurospin team.

4 Genedys study Dyslexics Controls Number Sex (M/F) 18/14 18/14 Age (years) 11.7 (1.3) 11.5 (1.4) Sequences: T1: 0.9 mm resolution 3-D DTI: 1.7x1.7x1.7 mm, 60 directions, b=1400 s/mm². T2, T2*, flair fmri: mapping of the ventral visual pathway.

5 Received knowledge on the neuroanatomy of dyslexia Less grey matter in left perisylvian (language) regions Less-well connected white matter in the left arcuate fasciculus (connecting language regions). Less leftward asymmetry of planum temporale? Corpus callosum? Smaller brains?

6 Richlan, Kronbichler, & Wimmer (2012) HBM Received knowledge on gray matter volumes in dyslexia Eckert (2004) Neuroscientist 9 studies, 266 participants

7 Richlan, Kronbichler, & Wimmer (2012) HBM

8 Our VBM results (32/32)

9 Complementary/replication studies 23 French dyslexics and 23 controls from Monzalvo et al. (2012) younger, reading-matched controls 46 Polish dyslexics and 35 controls (Jednorog et al.) 45 German dyslexics and 26 controls (Heim et al.)

10 L thalamus VBM study of dyslexia across 3 countries Jednorog, Marchewka, Gawron, Altarelli, Ramus, & Heim (in prep.) 84 French, 71 German, and 81 Polish children (n=236) control dyslexic Group differences in gray matter volume overlaid on the Anatomy Atlas, Julich. Region in the left thalamus shown in red yellow exhibited reduced volume in the dyslexic group French t= 2.28; p=0.025 German t= 2.75; p=0.008 Polish t= 1.86; p= French German Polish Consistent with VBM study by Brown et al. (2001), and with meta-analysis of fmri by Maisog et al (2008).

11 GM volume = cortical thickness x surface T1 anatomical sequence processing (Freesurfer software) original T1 surfaces Previous results on dyslexia (Frye et al Cereb. Cortex): Greater overall GMV and cortical surface in dyslexics Greater surface in rsmg in dyslexics Lower surface in IFG and fusiform gyrus in dyslexics

12 Our results on cortical thickness and surface (Altarelli + Monzalvo) Lower whole brain volume and surface in dyslexics. No consistent local difference. Genedys dataset Monzalvo et al. (2012) dataset ctr>dys dys>ctr

13 Functional sequence 9.7 years old Controls Ctr > Dys Dyslexics Monzalvo et al., 2012 Neuroimage

14 Cortical thickness of VWFA in dyslexic children Altarelli, I., Monzalvo, K., Iannuzzi, S., Fluss, J., Billard, C., Ramus, F., & Dehaene-Lambertz, G. (2013) J. of Neuroscience WORDS Anterior Medial Lateral Posterior

15 Diffusion Tensor Imaging Meta-analysis (Vandermosten et al. 2012) Dyslexics: lower FA in left long segment of arcuate fasciculus

16 FA Dissection of the arcuate fasciculus (32/31) Zhao, Thiebaut de Schotten, Altarelli, Dubois, & Ramus (in prep) * Anterior Posterior Long Control Dyslexia Group X Segment: F(2, 82)=3.446, p=0.037 We find the same trend in both hemispheres And the opposite effect in the anterior segment And no correlation with behavioural measures

17 Problems with standard DTI analysis methods Fiber-crossing regions not well handled Arcuate fasciculus confounded with superior longitudinal fasciculus. Our sequence parameters (1.7x1.7x1.7 mm, 60 non-collinear directions, b=1400 s/mm²) allow us to try and do better. New measure: Hindrance modulated orientational anisotropy (HMOA or OA).

18 OA Oriented Anistrophy New method: spherical 420 deconvolution 380 tractography 360 DYS 340 Zhao, Thiebaut de Schotten, Altarelli, Dubois, & 320 Ramus (in 300 prep) SLF I SLF II SLF III No results Left on the arcuate fasciculus. Right CON DYS Superior longitudinal fasciculus: Group effect F(1, 56)=4.443, p=0.04 Group X hemisphere F(1, 56)=3.814, p=0.056 Group X Hemisphere X Segment F(2, 112)=3.982, p=0.021 CON Inferior frontal-occipital fasciculus Left CON Right 440 IFOF 400 SLF I SLF II SLF III DYS Group X Hemisphere F(1, 56)=6.805, p=0.012 ** CON DYS Left Right

19 Summary of tractography results Using standard DTI analysis of arcuate fasciculus: Partial replication of disruption (lower FA) of the long segment of arcuate. But bilateral Seems compensated (?) by higher FA in anterior segment (consistent with greater frontal involvement) Using spherical deconvolution tractography: No difference in arcuate. Higher OA in right SLF II in dyslexics (greater involvement of attentional networks?) More symmetrical IFOF (less specialised direct reading route?)

20 Asymmetry of the planum temporale A long, sad story Authors N Anatomical criteria Ctr asymm Group difference Green Heschl's sulcus (excluding H2) to end of PAR no larger PT+ area overall in dys; no diff in asymm Leonard 2002 (study 1) 14 Heschl's sulcus (excluding H2??) to intersection Leftward PT asymm in dys, symm in SLI (but comparison ns) Leonard Heschl's sulcus (excluding H2??) to intersection marked leftward asymm Schulz 1994 Heschl's sulcus (or end of insula) to end of syl 31 yes fis. Excluding H2? No diff between groups, trend for dys girls symm but not signif, age correlated with struct size/area Hynd 1990 Larsen 1990 Semrud-Clikeman 1991 Semrud-Clikeman 1996 ridge of Heschl's gyrus (excluding H2) to end 20 of syl fis ridge of Heschl's gyrus (excluding H2) to end 36 of syl fis ridge of Heschl's gyrus (excluding H2) to end 20 of syl fis ridge of Heschl's gyrus (excluding H2) to end 20 of syl fis yes yes yes yes Dyslexic symm Dyslexic symm or rightward length, left PT smaller L>R 63% of cases Geschwind & Levitsky 1968 Galaburda et al L R 37% of cases Best and Demb Heschl's sulcus to end of desc branch yes No diff in PT asymm, regardless of method Heiervang Heschl's sulcus to end of desc branch yes all leftward asymm, smaller left PT in dys Hugdahl Heschl's sulcus to end of desc branch yes Trend for smaller left PT in dys; reduced leftward asymm in dys Hugdahl Heschl's sulcus to end of desc branch yes No diff in asymm, smaller left PT in dys Preis Heschl's sulcus to end of desc branch yes No diff in PT or PP asymm Robichon Heschl's sulcus to end of syl fis yes No diff in PT asymm, more leftward asymm in PO in dys Eckert Heschl's sulcus to intersection No group nor group*gender effect Foster Heschl's sulcus to intersection? Larger right PT in dys Kibby 2004 Heschl's sulcus to intersection; desc branch 17 excluded yes? No diff in PT or pars triang asymm L R in 5 out 5 male dyslexic brains dissected by Galaburda et al Leonard 1993 Rumsey 1997 Leonard 2001 Heschl's sulcus to intersection; desc branch 21 excluded Heschl's sulcus to intersection; desc branch 30 excluded yes yes More marked leftward in dys but signif not reported; rightward asymm PP; greater par than temp tissue in the right hemi in dys No diff in PT or PP asymm Tendency towards more marked leftward but not Heschl's sulcus to intersection; desc branch excluded; right: if 28 signif; leftward asymm PT+ in all; no diff in no planum, asc branch instead ratio of PT to asc branch

21 Asymmetry of the planum temporale Altarelli, Leroy, Dehaene-Lambertz, & Ramus (in prep) Under guidance of Al Galaburda 81 subjects, all right-handed 25 control M 21 control F 20 dyslexic M 15 dyslexic F Heschl s gyrus Planum Temporale Posterior ramus

22 Surface area ±1 SE (mm 2 ) PT Hemi*gender*diagnosis F (1,71) =4.2 p=0.04 Boys Girls Left Right Control Dyslexic Control Dyslexic Boys Hemi*diagnosis F (1,39) =6.1 p=0.02 Girls Hemi*diagnosis n.s.

23

24 Thanks LSCP: I. Altarelli, K. Jednorog, J. J. Zhao, E. Cauvet, S. Iannuzi, C. Bogliotti, L. Scotto di Covella, N. Villiermet ICM: M. Thiebaut de Schotten Neurospin: G. Dehaene-Lambertz, K. Monzalvo, J. Dubois, F. Leroy, nurses and radiographers Bicêtre: C. Billard, F. Delteil-Pinton, J. Fluss Funding: Ville de Paris, Agence Nationale de la Recherche, European Commission, Ecole de Neurosciences de Paris.

25