An MRI-based diagnostic framework for early diagnosis of dyslexia

  • A. El-Baz
  • M. Casanova
  • G. Gimel’farb
  • M. Mott
  • A. Switala
Original Article

Abstract

Purpose

A computer-aided diagnosis (CAD) system for early diagnosis of dyslexia was developed and tested. Dyslexia can severely impair the learning abilities of children so improved diagnostic methods are needed. Neuropathological studies show abnormal anatomy of the cerebral white matter (CWM) in dyslexic brains. We sought to develop an MRI-based macroscopic neuropathological correlate to the minicolumnopathy of dyslexia that relates to cortical connectivity: the gyral window. The brains of dyslexic patients often exhibit decreased gyrifications, so the thickness of gyral CWM for dyslexic subjects is greater than for normal subjects. We developed an MRI-based method for assessment of gyral CWM thickness with automated recognition of abnormal (e.g., dyslexic) brains.

Methods

In vivo data was collected from 16 right-handed dyslexic men aged 18–40 years, and a group of 14 controls matched for gender, age, educational level, socioeconomic background, handedness and general intelligence. All the subjects were physically healthy and free of history of neurological diseases and head injury. Images were acquired with the same 1.5T MRI scanner (GE, Milwaukee, WI, USA) with voxel resolution 0.9375  ×  0.9375  ×  1.5 mm using a T1-weighted imaging sequence protocol. The “ground truth” diagnosis to evaluate the classification accuracy for each patient was given by the clinicians. The accuracy of diagnosis/classification of both the training and test subjects was evaluated using the Chi-square test at the three confidence levels—85, 90 and 95%—in order to examine significant differences in the Levy distances.

Results

As expected, the 85% confidence level yielded the best results, the system correctly classified 16 out of 16 dyslexic subjects (a 100% accuracy) and 14 out of 14 control subjects (a 100% accuracy). At the 90% confidence level, 16 out of 16 dyslexic subjects were still classified correctly; however, only 13 out of 14 control subjects were correct, bringing the accuracy rate for the control group down to 92.86%. The 95% confidence level obviously gives the smaller accuracy rates for both the groups, namely, 14 out of 16 correct answers for dyslexic subjects (87.5%) and still 13 out of 14 control subjects (92.86%). The classification based on traditional volumetric approach is 7 out of 16 dyslexic subjects (a 43.75% accuracy), and 9 out of 14 control subjects (a 64.29% accuracy) at a 85 confidence interval. These results highlight the advantage of the proposed diagnostic approach.

Conclusion

We found that 3D texture analysis of MRI brain scans can accurately discriminate dyslexic and normal subjects in this feasibility trial. Our method for white matter segmentation and classification outperforms volumetric descriptions of brain structures and may be influenced less by age effects and segmentation errors. The proposed approach efficiently extracts quantitative features from 3D shapes of brain structures.

Keywords

Diagnosis Dyslexia MRI Levy distance Cerebral White Matter (CWM) segmentation CWM gyrifications 

References

  1. 1.
    Pugh K, Mencl W, Jenner A, Katz L, Frost S, Lee J, Shaywitz S, Shaywitz B (2000) Functional neuroimaging studies of reading and reading disability (developmental dyslexia). Mental Retardation And Developmental Disabilities Research Reviews 6: 207–213PubMedCrossRefGoogle Scholar
  2. 2.
    Brambilla P, Hardan A, Ucelli S, Perez J, Soares J, Barale F (2003) Brain anatomy and development in autism: review of structural MRI studies. Brain Res Bull 61: 557–569PubMedCrossRefGoogle Scholar
  3. 3.
    Vellutino F, Scanlon D (1987) Phonological coding, phonological awareness, and reading ability: evidence from a longitudinal and experimental study. Merril Palmer Quart 33(3): 321–363Google Scholar
  4. 4.
    Torgesen J (1994) Lessons learned from research on interventions for students who experience difficulty learning to read. In: McCardle P, Chabra V(eds) The voice of evidence in reading research. Brookes, Baltimore, pp 355–382Google Scholar
  5. 5.
    Brodmann K (1913) Neue forschungsergebnisse der grosshirnrindenanatomie mit besonderer berücksichtigung anthropologischer fragen. Gesselsch Deuts Naturf Artze 85: 200–240Google Scholar
  6. 6.
    El-Zehiry N, Casanova M, Hassan H, Farag A (2006) Effect of mnincolumnar disturbance on dyslexic brains: an MRI study. In: Proceedings of international symposuim on biomedical imaging, (ISBI’06), Arlington, Virginia, April 2006, pp 1336–1339Google Scholar
  7. 7.
    Casanova M, Araque J, Giedd J (2004) Reduced brain size and gyrification in the brains of dyslexic patients. J Child Neurol 9(4): 275–281CrossRefGoogle Scholar
  8. 8.
    Eliez S, Rumsey J, Giedd J, Schmitt J, Patwardhan A, Reiss A (2000) Morphological alteration of temporal lobe gray matter in dyslexia: an MRI study. J Child Psychol Psychiatry 41(5): 637–644PubMedCrossRefGoogle Scholar
  9. 9.
    Hynd G, Hall J, Novey E, Eliopulos D, Black K, Gonzalez J, Edmonds J, Riccio C, Cohen M (1995) Dyslexia and corpus callosum morphology. Arch Neurol 52(1): 32–38PubMedGoogle Scholar
  10. 10.
    Robichon F, Bouchard P, Démonet J, Habib M (2000) Developmental dyslexia: re-evaluation of the corpus callosum in male adults. Euro J Neurol 43(4): 233–237CrossRefGoogle Scholar
  11. 11.
    Herbert M, Ziegler D, Makris N, Filipek P, Kemper T, Normandin J, Sanders H, Kennedy D, Caviness V (2004) Localization of white matter volume increase in autism and developmental language disorder. Ann Neurol 55(4): 530–540PubMedCrossRefGoogle Scholar
  12. 12.
    Atkins M, Mackiewich B (1997) Fully automatic segmentation of the brain in MRI. IEEE Trans Med Imaging 16: 41–54CrossRefGoogle Scholar
  13. 13.
    Farag A, El-Baz A, Gimel’farb G (2006) Precise segmentation of multi-modal images. IEEE Trans Image Process 15(4): 952–968PubMedCrossRefGoogle Scholar
  14. 14.
    Collins D, Peters T, Dai W, Evans A (1994) Model based segmentation of individual brain structures from MRI Data. SPIE Vis Biomed Comput 1808: 10–19Google Scholar
  15. 15.
    Dawant B, Hartman S, Thirion J, Maes F, Vandermeulen D, Demaerel P (1999) Automatic 3-D segmentation of internal structures of the head in MR images using a combination of similarity and free-form transformations. IEEE Trans Med Imaging 18(10): 909–916PubMedCrossRefGoogle Scholar
  16. 16.
    Holden M, Schnabel J, Hill D (2001) Quantifying small changes in brain ventricular volume using non-rigid registration. In: Proceedings of international conference on medical image computing and computer-assisted intervention, (MICCAI’01), Utrecht, Netherlands, October 14–17, pp 49–56Google Scholar
  17. 17.
    Kass M, Witkin A, Terzopoulos D (1987) Snakes: active contour models. Int J Comput Vis 1: 321–331CrossRefGoogle Scholar
  18. 18.
    Rousson M, Paragios N, Deriche R (2004) Implicit active shape models for 3D segmentation in MR imaging. In: Medical image computing and computer assisted intervention (MICCAI’04), Saint-Malo, France, September 26–29, pp 209–216Google Scholar
  19. 19.
    Caselles V, Kimmel R, Sapiro G (1997) Geodesic active contours. Int J Comp Vis 22: 61–79CrossRefGoogle Scholar
  20. 20.
    Zeng X, Staib L, Duncan J (1998) Volumetric layer segmentation using coupled surface propagation. In: Proceedings of IEEE conference on computer vision and pattern recognition (CVPR’98), Santa Barbara, CA, USA, June 23–25, pp 708–715Google Scholar
  21. 21.
    Xu C, Pham D, Prince J (1999) Reconstruction of the human cerebral cortex from magnetic resonance images. IEEE Trans Med Imaging 18(6): 467–480PubMedCrossRefGoogle Scholar
  22. 22.
    El-Baz A (2006) Novel stochastic models for medical image analysis. PhD dissertation, University of Louisville, LouisvilleGoogle Scholar
  23. 23.
    Adalsteinsson D, Sethian J (1995) A fast level set method for propagating interfaces. J Comput Phys 118: 269–277CrossRefGoogle Scholar
  24. 24.
    Lamperti J (1996) Probability. Wiley, New YorkGoogle Scholar
  25. 25.
    Vese L, Chan T (2002) A multiphase level set framework for image segmentation using the Mumford and Shah model. Int J Comp Vis 50(3): 271–293CrossRefGoogle Scholar
  26. 26.
    Aylward E, Minshew N, Field K, Sparks B, Singh N (2002) Effects of age on brain volume and head circumference in autism. J Neurol 59(2): 175–183Google Scholar
  27. 27.
    Courchesne E, Karns C, Davis H, Ziccardi R, Carper R, Tigue Z, Chisum H, Moses P, Pierce K, Lord C, Lincoln A, Pizzo S, Schreibman L, Haas R, Akshoomoff N, Courchesne R (2001) Unusual brain growth patterns in early life in patients with autistic disorder: an MRI study. J Neurol 57(2): 245–254Google Scholar
  28. 28.
    Courchesne R, Carper R, Akshoomoff N (2003) Evidence of brain overgrowth in the first year of life in autism. J Am Med Assoc 290: 337–344CrossRefGoogle Scholar

Copyright information

© CARS 2008

Authors and Affiliations

  • A. El-Baz
    • 1
  • M. Casanova
    • 2
  • G. Gimel’farb
    • 3
  • M. Mott
    • 2
  • A. Switala
    • 2
  1. 1.Bioengineering DepartmentUniversity of LouisvilleLouisvilleUSA
  2. 2.Department of Psychiatry and Behavioral ScienceUniversity of LouisvilleLouisvilleUSA
  3. 3.Computer Science DepartmentUniversity of AucklandAucklandNew Zealand

Personalised recommendations