Seizure: European Journal of Epilepsy
Volume 16, Issue 1 , Pages 74-80, January 2007

Microstructural white matter abnormalities in nodular heterotopia with overlying polymicrogyria

  • Leonardo Bonilha

      Affiliations

    • Department of Neuropsychiatry, University of South Carolina, USA
    • Department of Communication Sciences and Disorders, University of South Carolina, USA
    • Corresponding Author InformationCorresponding author at: 1621 Greene Street, 6th Floor, Williams Brice Building, Columbia, SC 29209, USA. Tel.: +1 803 777 9242; fax: +1 803 777 3081.
    • ,
  • Jonathan Halford

      Affiliations

    • Department of Neurology, Medical University of South Carolina, USA
    • ,
  • Chris Rorden

      Affiliations

    • Department of Communication Sciences and Disorders, University of South Carolina, USA
    • ,
  • Li Min Li

      Affiliations

    • Department of Neurology State University of Campinas, Brazil
    • ,
  • Ashok Patel

      Affiliations

    • Department of Neurology, Medical University of South Carolina, USA
    • ,
  • Zoran Rumbolt

      Affiliations

    • Department of Radiology, Medical University of South Carolina, USA
    • ,
  • Paul Morgan

      Affiliations

    • Division of Academic Radiology, University of Nottingham, UK

Received 22 June 2006; received in revised form 17 August 2006; accepted 29 August 2006. published online 15 November 2006.

Article Outline

Summary 

Nodular heterotopia (NH) with overlying polymicrogyria can result in medically uncontrolled seizures. Most patients also exhibit deficits of function related to the location of the abnormal cortex. However, functional imaging studies show that the abnormal cortex can retain some function, making surgical planning difficult. It is not known if the connectivity of the abnormal cortex is normal. In this article, we performed an evaluation of molecular diffusion within the white matter in a patient with refractory epilepsy due to NH with overlying polymicrogyria. We observed that the white matter underlying the polymicrogyric area shows signs of microstructural abnormalities. This result suggests that the deficit of function from polymicrogyria result from both the structurally abnormal cortex and from its impaired connectivity.

Keywords: Polymicrogyria, White matter, Diffusion tensor imaging

 

Back to Article Outline

Introduction 

Polymicrogyria is a disruption in the cortical development process defined by the presence of multiple aberrant small gyri. Intrauterine insults causing laminar necrosis of cortical layers are possibly the cause of the majority of congenital cases of sporadic polymicrogyria.1 Polymicrogyria is usually detectable by conventional structural magnetic resonance imaging (MRI)2 and it can be associated with nodular heterotopia (NH).3

The clinical picture of NH depends on the location of the abnormal cortex,4 and it can involve different degrees of physical and cognitive impairment. Seizure control in patients with unilateral polymicrogyria is variable,5 and, when polymicrogyria is associated with NH, patients usually experience medically uncontrolled seizures.6

Surgery aiming to excise the abnormal cortex is typically an alternative form of treatment. Nevertheless, it is often difficult to plan surgery for periventricular NH or polymicrogyria. Even though abnormal, the polymicrogyric cortex can retain eloquent function.7, 8 Therefore, success rates of the surgical treatment of polymicrogyria are hindered by incapacity to remove the entire extension of the dysplastic tissue.

It is not yet completely understood, however, how much of a functional connection there is between the dysplastic tissue and other brain areas. There is evidence that patients with polymicrogyria do not exhibit extensive cortical functional re-organization,7 suggesting some preserved function, in contrast with other forms of cortical malformations.9 However, the evident impairment observed in the clinical picture suggests that both function and connectivity of the abnormal cortex are significantly disrupted.

In this article, we aimed to investigate microstructural abnormalities of the white matter in NH with overlying polymicrogyric cortex. We evaluated the spatial pattern of molecular diffusion using diffusion tensor imaging (DTI) of the white matter of the brain of a subject with NH with overlying polymicrogyria and clinical intractable seizures.

Back to Article Outline

Case report 

A 34-year-old right-handed white female was referred to the epilepsy outpatient clinic of the Medical University of South Carolina with a 12-year-old history of seizures refractory to drug treatment. Her seizures were described as staring straight ahead, smacking her lips, rubbing her fingers and losing track of her train of thought. In general, each seizure would last between 15s and 2min, and would be followed by postictal confusion for about 5min. Typically, the patient has four to five complex partial seizures per week with rare secondary generalization. She reports having had one generalized seizure and denies a history of status epilepticus. Since her seizures started, several trails of anti-epileptic drugs were unsuccessful in controlling her seizures, including drugs such as carbamazepine, topiramate, and lamotrigine. Currently, she is being treated with the oxcarbazepine and levetiracetam, without significant improvement on her seizure frequency.

Previous medical history is unremarkable except for the seizure disorder. There is no history of head trauma, brain infections or childhood seizures. Past medical history is unremarkable. There is a family history of seizures in a maternal uncle and maternal cousin possibly suggesting a genetic component for her malformation of cortical development. She smokes one pack of cigarettes per day and denies use of alcohol or recreational drugs.

General physical exam and neurological exam are normal.

Further investigation of the patient's epilepsy revealed as follows.

Electroencephalography (EEG) 

During the patient's seizures, based on video/EEG monitoring, she shifts her position in bed and smacks her lips for about 45–60s. There is clear left temporal onset of 1–2s−1 sharp wave discharges at seizure onset (Fig. 1). Interictally, there are frequent left temporal spike and sharp-wave discharges (Fig. 1).

  • View full-size image.
  • Figure 1. 

    (Top panel) The video/EEG interictal recording, with frequent left temporal spike and sharp-wave discharges. (Bottom panel) Ictal recording with left temporal onset of 1–2s−1 sharp wave discharges at seizure onset.

Magnetic resonance imaging 

Scanning 

The patient was submitted to MRI scanning in a Philips 3T Intera scanner (Best, The Netherlands) using an eight-channel head coil at the Medical University of South Carolina. Routine T1 scans were acquired to investigate brain structure. Diffusion tensor images were acquired using single shot EPI with cubic voxels of 2.5mm×2.5mm×2.5mm, employing Philips 32 diffusion directions with b value of 1000s/mm2.

Structural analysis 

A T1 weighted MRI revealed NH with overlying polymicrogyria in the left temporal lobe, throughout the entire extension of the left lateral ventricle. The heterotopia extends from the temporal horn of the left lateral ventricle to the inferior temporal sulcus, encompassing the lateral and medial occipito-temporal and parahippocampal gyri. There is also dilation of the ipsilateral lateral ventricle (Fig. 2). The remaining brain structure is apparently normal. There are no signs of cortical malformation elsewhere, except for the left temporal polymicrogyria. The hippocampi are normal sized and not visually atrophied.

  • View full-size image.
  • Figure 2. 

    Top row is composed by three different orthogonal planes of the structural T1 weighted MRI, showing the unilateral temporal polymicrogyria. Middle row shows the FA map, on the left, followed by the histogram showing the distribution (in frequency of voxels) of FA between the normal and polymicrogyric temporal lobes. Bottom row shows the MD map followed by the histogram showing the distribution of the MD values between temporal lobes. Images are in neurological convention.

Diffusion tensor imaging (DTI) 

Image processing 

We investigated axonal integrity using a voxel-wise analysis of the magnitude (mean diffusivity: MD) and the directionality (fractional anisotropy: FA) of water diffusion. DTI images were submitted to pre-processing steps for voxel-wise evaluation of quantifiable MD and FA. FSL's (FMRIB's Software Library, http://www.fmrib.ox.ac.uk/fsl) diffusion toolkit (FDT) was used to pre-process the diffusion weighted images and to construct DTI data. Raw images were transformed into analyze format using MRIcro,10 then images underwent eddy current correction by way of affine transform of each diffusion weighted image (DWI) to the base b=0 T2 weighted image. This step removes the majority of the additional spatial distortion in the DWIs due to application of diffusion gradients in various directions. Following this correction, diffusion images acquired of the same slice are in alignment and so a pixel-wise calculation of the diffusion tensor may be performed. Variations in acquisition geometry were corrected and gradients were updated using the Java applet from the F.M. Kirby Center (http://www.mri.jhmi.edu/∼craig/protocols/dti.html). Pixel-wise calculation of the diffusion tensor was performed using FDT, including application of a binary brain mask extracted using FSL's brain extraction tool (BET) with fractional threshold of 0.3 to prevent erroneous DTI calculation in the noise background outside the head. This results in calculation of both the MD and FA maps from the DTI data, in the same space as the b=0 image volume of the original DTI acquisition. The b=0 image volume, in effect a T2 weighted spin-echo planar image, was linearly normalized to a T2 template in stereotaxic space using FLIRT (FMRIB's linear image registration tool-http://www.fmrib.ox.ac.uk/fsl/flirt/). The same normalization matrix is then applied to FA and MD maps obtained from the diffusion tensor reconstruction step.

The structural image was linearly normalized to a T1 template in stereotaxic space using FLIRT. Two masks were created in MRIcro, each involving one temporal lobe, using the structural image as a background to ensure anatomical accuracy. The boundaries for creation of temporal lobe masks were the same for both sides, were defined based on segmentation protocols of the medial temporal lobe11 and can be described as follows: temporal pole as the anterior boundary, the slice corresponding to the posterior end of the occipital horn of the lateral ventricle in the polymicrogyria side as the posterior boundary, the superior temporal gyrus as super-lateral boundary, and the endorhinal sulcus as the supero-medial border.

These masks were transposed to the normalized FA and MD maps, with visual check of anatomical location authenticity. Data for each voxel contained in both masks regarding FA or MD values were extracted. This way, MD and FA values were computed for the temporal lobe containing the polymicrogyria, and for the healthy temporal lobe.

Statistical analysis 

Typically, FA maps are not normally distributed. Therefore, FA values between the healthy and polymicrogyric temporal lobes were compared using the Wilcoxon signed-rank test. MD values, in turn, usually follow a normal distribution, so a paired t-test was used to compare both sides. We used a statistical threshold of p<0.05.

We obtained 15387 samples from the normal temporal lobe, and 19982 from the polymicrogyric side. We observed that there was a significant difference between both temporal lobes for FA values (Z=−19.9, p<0.001) and MD values (t=6.7, p<0.001) between both temporal lobes.

The distribution of MD and FA values is shown in Fig. 2.

It is noticeable that there is a large difference in the FA distribution in the range of FA between 0.4 and 0.8, which largely corresponds to white matter. These results show that the white matter of the temporal lobe containing the polymicrogyria exhibits reduced FA. It is possible that reduced FA in the polymicrogyric temporal lobe reflects less organization of molecular movement within axonal fibers,12 which are not as organized and homogeneous, therefore corresponding to disrupted connectivity.

Conversely, the distribution of MD between the polymicrogyric side and the healthy side is similar, and possibly the significant difference between sides is largely due to the comparison between samples, rather than reflecting intrinsic tissue structure differences.

Tractography 

In order to further explore the presence of disrupted connectivity of the polymicrogyric temporal lobe, we also employed a three-dimensional spatial reconstruction of the tractography of the DTI data. The raw data from the Philips scanner was processed with the software DtiStudioV2.4 (http://lbam.med.jhmi.edu/DTIuser/DTIuser.asp). Tensor calculation was defined with a background noise level of 20. Fiber tracking calculation parameters were set so as the fiber tracking would start if FA were greater than 0.25, or would stop if FA were less 0.25, or if tract turning angle would be higher than 70°. Spherical seeds (15mm×15mm) were defined in order to symmetrically involve the stem white matter of temporal lobes. Results are shown in Fig. 3. It is possible to observe that the white matter of polymicrogyric side is less cohesive and organized than the white matter in the healthy side. The process of tracing white matter within the healthy side is able to detect robust and bundled fibers. Conversely, the same process, with an equivalently placed same sized seed on the polymicrogyric side, is not able to trace the same number and density of fibers. This indicates that the white matter underlying the polymicrogyric side is microstructurally disrupted.

  • View full-size image.
  • Figure 3. 

    White matter tractography generated by two 15mm×15mm seeds (red circles on axial slice demonstrated in the insert) shows asymmetry regarding the complexity of the connections between the side with polymicrogyria (left) and the normal side.

Another additional explanation for differences in FA maps between temporal lobes is an imbalance between the quantity of gray and white matter. This provides further evidence for the insufficient number of fibers linking the polymicrogyric cortex. To examine this hypothesis, we evaluated the difference of T1 weighted signal distribution between temporal lobes. We extracted the voxel-wise signal from the structural T1 image using the temporal masks. We observed a significant difference between T1 signals from both lobes, which is showed in Fig. 4 (Wilcoxon's Z=−27.8, p<0.001). The polymicrogyric lobe exhibited a significant reduction in the quantity of signal corresponding to the white matter.

  • View full-size image.
  • Figure 4. 

    Left panel shows the location of the temporal lobe masks, from which data was extracted for DTI and structural analyses. The histogram shows the distribution of T1 signal for both temporal lobes, demonstrating the reduced frequency of signal corresponding to the white matter on the abnormal side. The bar graph shows the ratio of gray and white matter for the normal and abnormal side. The abnormal side shows a greater ratio of gray over white matter than the normal side.

We also segmented the gray and white matter of the T1 weighted structural image using the segmentation routine in-built in the SPM5 software (http://www.fil.ion.ucl.ac.uk/spm/software/spm5/). We applied an intensity filter to the segmented white or gray matter images in order to exclude voxels with less than 80% of chance of being white or gray matter. We then calculated the number of white and gray matter voxels contained in each mask of each temporal lobes, and the ratio between gray and white matter from both lobes. The polymicrogyric lobe exhibited a larger ratio of gray/white matter (Fig. 4), indicating less white matter to support the cortex.

Summarizing the results, we observed reduced FA and less organized white matter tractography in the abnormal temporal lobe. The abnormal temporal lobe also exhibited an imbalance between gray and white matter, with a smaller percentage of white matter, compared to the normal lobe.

Back to Article Outline

Discussion 

Even though patients with very extensive polymicrogyria involving eloquent areas can perform language and motor tasks, it is usually clinically apparent that these functions are severely impaired. Function within the polymicrogyric cortex has been elegantly demonstrated by functional MRI studies,13 specifically targeting increased specificity of detecting activation coupled with simple language and motor paradigms.

Clinical evidence supports that, even though the polymicrogyric cortex can retain function, it is overall different and less effective than normal cortex. It is currently not possible to evaluate the reliability of fMRI activation in regards to how effective the activation is, or to how it can be readily translated to purposeful acts. In fact, most fMRI paradigms rely on fitting the normal canonical hemodynamic response function to observed activations. It is possible that within the abnormal cortex, the hemodynamics of activation is significantly different.

This problem is even more accentuated when complex and eloquent functions are taken into account.14 Therefore, the investigation of the pattern of connectivity of the abnormal cortex can be insightful concerning the capability of the abnormal cortex to transmit function.

We observed that the temporal lobe stem of a patient with temporal NH with overlying polymicrogyria is significantly different from the contralateral healthy temporal lobe stem. This finding indicates that even if the polymicrogyric cortex retains normal function, the white matter underlying it is different from the normal side and appears to be disrupted. This observation, extrapolated to other patients with polymicrogyria, can help explain why although function is apparently preserved within the abnormal cortex yet it is not readily translated into completely normal acts. Likely, the tissue damage in specific cortical layers and the hindered functionality of the abnormal cortex, combined with impaired connectivity, result in cognitive and motor deficits.

Our findings are in accordance with previous DTI studies of polymicrogyria that suggested that the long projection fibers were preferentially reduced in patients with polymicrogyria.15 We also observed that there is a range of FA differences between normal and abnormal white matter, suggesting that not only large but also small fibers are disrupted. These results suggest overall impaired connectivity of the polymicrogyric area.

Impaired diffusivity such as reduced FA has been demonstrated in patients with other types of malformation of cortical development(MCD), such as focal cortical dysplasia and band heterotopias.16 Taken together, the results from our paper and previous results support the notion that there is a common underlying reduction in diffusivity in patients with MCD, both when MCD is due to abnormal cortical migration or abnormal cortical proliferation. Our findings also support the notion that the combination of disrupted cortical architectural organization leads to impaired function and connectivity of the polymicrogyric cortex. As a consequence, even though the abnormal cortex can retain function, the transmission of information from the abnormal cortex to other brain areas is disrupted. Future work can improve the understanding of the extent of the functionality of the abnormal cortex that is crucial to function, helping to guide surgery.

Back to Article Outline

References 

  1. Harding BN, Copp AJ. Malformations. In:  Graham DI,  Lantos PL editor. Greenfield's neuropathology. New York: Edward Arnold Press; 2001;p. 357–483
  2. Barkovich AJ, Kuzniecky RI, Jackson GD, Guerrini R, Dobyns WB. Classification system for malformations of cortical development: update 2001. Neurology. 2001;57:2168–2178
  3. Wieck G, Leventer RJ, Squier WM, Jansen A, Andermann E, Dubeau F, et al. Periventricular nodular heterotopia with overlying polymicrogyria. Brain. 2005;128:2811–2821
  4. Jansen A, Andermann E. Genetics of the polymicrogyria syndromes. J Med Genet. 2005;42:369–378
  5. Ohtsuka Y, Tanaka A, Kobayashi K, Ohta H, Abiru K, Nakano K, et al. Childhood-onset epilepsy associated with polymicrogyria. Brain Dev. 2002;24:758–765
  6. Li LM, Dubeau F, Andermann F, Fish DR, Watson C, Cascino GD, et al. Periventricular nodular heterotopia and intractable temporal lobe epilepsy: poor outcome after temporal lobe resection. Ann Neurol. 1997;41:662–668
  7. Burneo JG, Kuzniecky RI, Bebin M, Knowlton RC. Cortical reorganization in malformations of cortical development: a magnetoencephalographic study. Neurology. 2004;63:1818–1824
  8. Janszky J, Ebner A, Kruse B, Mertens M, Jokeit H, Seitz RJ, et al. Functional organization of the brain with malformations of cortical development. Ann Neurol. 2003;53:759–767
  9. de Bode S, Firestine A, Mathern GW, Dobkin B. Residual motor control and cortical representations of function following hemispherectomy: effects of etiology. J Child Neurol. 2005;20:64–75
  10. Rorden C, Brett M. Stereotaxic display of brain lesions. Behav Neurol. 2000;12:191–200
  11. Bonilha L, Kobayashi E, Cendes F, Min LL. Protocol for volumetric segmentation of medial temporal structures using high-resolution 3-D magnetic resonance imaging. Hum Brain Mapp. 2004;22:145–154
  12. Beaulieu C. The basis of anisotropic water diffusion in the nervous system—a technical review. NMR Biomed. 2002;15:435–455
  13. Araujo D, de Araujo DB, Pontes-Neto OM, Escorsi-Rosset S, Simao GN, Wichert-Ana L, et al. Language and motor FMRI activation in polymicrogyric cortex. Epilepsia. 2006;47:589–592
  14. Fridriksson J, Ryalls J, Rorden C, Morgan PS, George MS, Baylis GC. Brain damage and cortical compensation in foreign accent syndrome. Neurocase. 2005;11:319–324
  15. Munakata M, Onuma A, Takeo K, Oishi T, Haginoya K, Iinuma K. Morphofunctional organization in three patients with unilateral polymicrogyria: combined use of diffusion tensor imaging and functional magnetic resonance imaging. Brain Dev. 2006;28:405–409
  16. Eriksson SH, Rugg-Gunn FJ, Symms MR, Barker GJ, Duncan JS. Diffusion tensor imaging in patients with epilepsy and malformations of cortical development. Brain. 2001, Mar;124:617–626

PII: S1059-1311(06)00160-9

doi:10.1016/j.seizure.2006.08.009

Seizure: European Journal of Epilepsy
Volume 16, Issue 1 , Pages 74-80, January 2007

Article Tools

* Image Usage & Resolution