widespread ph abnormalities in patients with malformations of cortical development and epilepsy: a...

www.elsevier.com/locate/braindev

Brain & Development xxx (2014) xxx–xxx

Original article

Widespread pH abnormalities in patients with malformationsof cortical development and epilepsy: A phosphorus-31 brain

MR spectroscopy study

Celi Santos Andrade a,⇑, Maria Concepcion Garcıa Otaduy a,Kette Dualibi Ramos Valente b, Eun Joo Park a, Alexandre Fligelman Kanas a,

Mauricio Ricardo Moreira da Silva Filho a, Miriam Harumi Tsunemi c,Claudia Costa Leite a

a Department of Radiology, Faculdade de Medicina da Universidade de Sao Paulo, Sao Paulo, Brazilb Department of Psychiatry, Faculdade de Medicina da Universidade de Sao Paulo, Sao Paulo, Brazilc Department of Statistics, Universidade Estadual Paulista Julio de Mesquita Filho, Sao Paulo, Brazil

Received 8 October 2013; received in revised form 26 December 2013; accepted 27 December 2013

Abstract

Introduction: Neuroimaging studies demonstrate that not only the lesions of malformations of cortical development (MCD) butalso the normal-appearing parenchyma (NAP) present metabolic impairments, as revealed with 1H-MRS. We have previouslydetected biochemical disturbances in MCD lesions with phosphorus-31 magnetic resonance spectroscopy (31P-MRS). Our hypoth-esis is that pH abnormalities extend beyond the visible lesions. Methods: Three-dimensional 31P-MRS at 3.0 T was performed in 37patients with epilepsy and MCD, and in 31 matched-control subjects. The patients were assigned into three main MCD subgroups:cortical dysplasia (n = 10); heterotopia (n = 14); schizencephaly/polymicrogyria (n = 13). Voxels (12.5 cm3) were selected in fivehomologous regions containing NAP: right putamen; left putamen; frontoparietal parasagittal cortex; right centrum semiovale;and left centrum semiovale. Robust methods of quantification were applied, and the intracellular pH was calculated with the chem-ical shifts of inorganic phosphate (Pi) relative to phosphocreatine (PCr). Results: In comparison to controls and considering a Bon-ferroni adjusted p-value <0.01, MCD patients presented significant reduction in intracellular pH in the frontoparietal parasagittalcortex (6.985 ± 0.022), right centrum semiovale (7.004 ± 0.029), and left centrum semiovale (6.995 ± 0.030), compared to controls(mean values ± standard deviations of 7.087 ± 0.048, 7.096 ± 0.042, 7.088 ± 0.045, respectively). Dunnet and Dunn tests demon-strated that the differences in pH values remained statistically significant in all MCD subgroups. No significant differences werefound for the putamina. Conclusion: The present study demonstrates widespread acidosis in the NAP, and reinforces the idea thatMCD visible lesions are only the tip of the iceberg.� 2014 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Keywords: Malformations of cortical development; Epilepsy; Neurometabolism; Phosphorus spectroscopy; Magnetic resonance imaging

0387-7604/$ - see front matter � 2014 The Japanese Society of Child Neuro

http://dx.doi.org/10.1016/j.braindev.2013.12.010

⇑ Corresponding author. Address: Av. Dr. Altino Arantes, 370, ap.32, Vila Clementino, 04042-002 Sao Paulo, Brazil. Tel./fax: +55 113069 7095.

E-mail address: [email protected] (C.S. Andrade).

Please cite this article in press as: Andrade CS et al. Widespread pH abnorepilepsy: A phosphorus-31 brain MR spectroscopy study. Brain Dev (201

1. Introduction

Malformations of cortical development (MCD) are aheterogeneous group of lesions that share in common astrong association with epilepsy. The severity of epilepsyvaries between patients, but it is estimated that MCD

logy. Published by Elsevier B.V. All rights reserved.

malities in patients with malformations of cortical development and4), http://dx.doi.org/10.1016/j.braindev.2013.12.010


mailto:[email protected]



2 C.S. Andrade et al. / Brain & Development xxx (2014) xxx–xxx

are responsible for up to 40% of intractable epilepsy inchildren [1]. Patients may also exhibit motor dysfunc-tion, speech disability, neurodevelopmental delay, andcognitive impairments, with profound compromise inthe life quality of the affected individuals [2].

Phosphorus-31 magnetic resonance spectroscopy(31P-MRS) is an advanced method that provides non-invasive measurement of intracellular pH in humanbrain, both in physiological and pathological states[3–4]. This technique has been used in the evaluationof a variety of brain diseases, such as tumors, hyp-oxic-ischemic insults, migraine, and demyelinating dis-eases [5–7]. In the field of epilepsy, most studies havefocused in the evaluation of mesial temporal sclerosis[8–9].

Our group previously explored the visible lesions ofMCD with 31P-MRS. The findings indicated higherlevels of magnesium, reduction in intracellular pHand abnormalities in membrane phospholipids in themagnetic resonance imaging (MRI) visible lesions,pointing to bioenergetics dysfunctions and membraneturnover impairments [10]. Other MRI-based examina-tions, such as 1H-MRS, functional MRI, MRI-perfu-sion, and diffusion-weighted imaging (DWI), as wellas nuclear medicine tests have also shown metabolicdisturbances in these highly epileptogenic lesions [11–14].

In addition to the evidences of structural and met-abolic abnormalities in the aberrant lesions of MCD,recent studies pointed to the presence of associatedabnormalities in the normal-appearing parenchyma(NAP) of these patients. Studies with 1H-MRSshowed reductions in N-acetyl aspartate levels notonly in the MCD lesions, but also in the NAP [15–18]. There are also evidences of functional distur-bances in NAP with other multimodal neuroimagingtools, such as diffusion tensor imaging (DTI) [13],besides the findings of electrophysiological and histo-logical abnormalities outside the MRI visible malfor-mations [19].

Based on these findings, we postulate that intracel-lular pH, a parameter of brain bioenergetics and elec-trophysiological status, is abnormal in the NAP ofpatients with MCD. The purpose of this study wasto determine the intracellular pH in the NAP in fivecerebral regions of patients with MCD: the bilateralwhite matter (centrum semiovale), midline corticalgray matter (frontoparietal parasagittal cortex), andbilateral basal ganglia (mainly putamina). For thispurpose, we have performed structural and metabolicstudies in a large series of MCD patients and in anage- and sex-matched control group. A secondarygoal was to establish the pH patterns in individualsubgroups of MCD.


2. Methods

2.1. Study design

We have conducted a transversal, observational, con-trolled and non-randomized study. This investigationwas approved by the institution’s ethics committee. Allindividuals enrolled in the study or their legal guardianssigned a written informed consent.

2.2. Patients

Thirty-seven patients with previous known diagnosisof MCD were included in the study during one-year per-iod (20 female, 17 male). Patients’ mean age was29.5 years (range 19–55), and the mean years of educa-tion was 9.4 (range 1–17).

The patients were evaluated as a single whole groupof MCD and as individual subgroups of MCD. Thisclassification was based on the evaluation of the MRIstructural images by two experienced neuroradiologists(C.S.A., E.J.P), who defined, in agreement, which wasthe most representative and main lesion in each patient,according to the scheme proposed by Barkovich et al.[20]. Only patients with macroscopic MCD lesions evi-dent on MR images were included. These subgroupsof patients were assigned into three main subcategoriesof cortical malformations (Fig. 1): (1) focal cortical dys-plasia (n = 10; 27%); (2) subcortical and subependymalheterotopia (n = 14; 37.9%); (3) schizencephaly andpolymicrogyria (n = 13; 35.1%).

The majority of patients presented lesions in the fron-tal, parietal and temporal lobes (72.8%, 70.3% and51.3%, respectively). The insula and occipital lobes wereaffected in 29.7% and 37.8% of cases, respectively. Whilemost patients had diffuse disease, only eight patients(21.6%) had focal lesions that involved just one cerebrallobe. Bilateral MCD lesions were found in 48.7% ofpatients (n = 18), while 24.3% (n = 9) had unilateralMCD lesions in the right cerebral hemisphere, and27% (n = 10) presented unilateral lesions in the left cere-bral hemisphere.

In order to evaluate the effect of recent seizures,patients were further separated in two subgroups. Eightpatients (21.62%) reported seizures within 24 h prior toscan, while the remaining (n = 29, 78.38%) did not pres-ent recent seizures.

2.3. Control subjects

Similar MR experiments were concurrently per-formed in 31 control subjects (19 female, 12 male) withno clinical complaints or previous neurological disor-ders. These healthy individuals were selected based on



Fig. 1. Representative graphic of patients’ distribution in the threesubcategories of MCD, along with illustrative examples: corticaldysplasia (green), heterotopia (purple), and polymicrogyria/schizen-cephaly (gray). The white arrows point to the MCD lesions on T1weighted coronal images. (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)

C.S. Andrade et al. / Brain & Development xxx (2014) xxx–xxx 3


demographic characteristics in order to match age, gen-der and level of education to the MCD patients group.Controls’ mean age was 28.8 years (range 11–57), andthe mean years of education was 11 (range 5–16).

The population evaluated in the present work is thesame enrolled in our previous study that focused onthe visible MCD lesions [10]. Although there is a smalldifference in the number of subjects in each group(patients and controls), we have chosen not to excludeany patients, maintaining the sample original and com-plete. Moreover, for the evaluation of smaller MCDsubgroups, a maximal sample size would be critical tokeep the statistical power of comparisons. There wereno significant differences of demographic characteristics(age, gender, level of education) between patients andcontrol subjects.

2.4. MRI

All the examinations were performed in a 3.0T MRIsystem (Intera Achieva, Philips Healthcare, Best, TheNetherlands). The structural images were performed inorder to evaluate and classify MCD lesions. These imagesets were obtained with an 8-channel head coil (PhilipsHealthcare, Best, The Netherlands). The first sequenceof the protocol consisted of a sagittal 3D T1 fast-fieldecho (TR = 7 ms; TE = 3.2 ms; TI = 900 ms; flipangle = 8�; isotropic 1-mm3 resolution). It was alsoobtained a FLAIR sequence in the axial plane(TR = 11,000 ms; TE = 130 ms; TI = 2800 ms; sectionthickness = 4.5 mm). The total scanning time for thisstructural imaging phase was 11 min.

2.5. 31P-MRS

The spectroscopic phase was performed with a dou-ble-tuned 31P/1H birdcage head coil (AIRI, Cleveland,USA). First we obtained T1-weighted fast-field echoimages in the axial plane (TR = 7.6 ms; TE = 3.7 ms;flip angle = 8�; isotropic 1-mm3 resolution) with multi-planar reconstructions, which were used for the uniformpositioning of the 31P-MRS grid in the mid-sagittalplane with an orbitomeatal angulation. A multivoxel31P-MRS acquisition was performed to include thewhole supratentorial brain with a 3D chemical shiftimaging (TR = 5090 ms; TE = 0.31 ms; FOV = 175 �200 � 120 mm; matrix = 6 � 7 � 8; BW = 6000 Hz;number of points = 1024). An adiabatic RF pulse wasapplied with a broadband decoupling technique (powerfactor = 0.4; offset = 100 Hz). An automatic shimmingprocedure was also performed, with an acceptance of amaximum value of 35 Hz. The individual voxels were25 � 25 � 20 mm in size (nominal volumes of12.5 cm3). Subsequently, T1-weighted axial images(TR = 9.4 ms; TE = 4.6 ms; flip angle = 8�; sectionthickness = 6 mm) were acquired to make sure that the




positioning of patients’ head was the same as the begin-ning of the exam. This second phase of the experimenthad a total scanning time of 45 min.

2.6. Voxel selection

A large number of voxels was obtained with thethree-dimensional spectroscopy protocol (6 � 7 � 8 =336). However, for practical purposes and to avoid mul-tiple comparisons errors, we have decided to analyze alimited number of voxels that were representative ofbilateral white matter, deep gray matter and corticalgray matter. Other investigators have mentioned thatdifferences in metabolites concentrations in distinct tis-sue types (e.g., gray � white matter) outweigh any differ-ence seen between the same tissue types on differentcerebral locations (e.g., temporal � frontal lobe) [21].Five homologous cerebral regions containing NAP inpatients and controls were selected, numbered as fol-lows: (1) right putamen, (2) left putamen, (3) frontopa-rietal parasagittal cortex, (4) right centrum semiovale,and (5) left centrum semiovale (Fig. 2). The voxels thatoverlapped MCD lesions or contained any MR signalchanges were discarded from the analysis. Thereforethe number of voxels included in the analysis variedamong the five cerebral regions in the patients’ group(n = 24, 26, 29, 21, and 17 for the voxels 1, 2, 3, 4,and 5, respectively). The number of voxels was the samefor the control group in these five regions (n = 31).

2.7. pH determination

31P-MRS raw data were processed offline using apublic-domain MR user-interface software for spectros-copy analysis (available at: www.mrui.uab.es/mrui) [22].The advanced method for accurate, robust and efficientspectral fitting (AMARES algorithm) [23] was appliedto fit the time-domain. Wide baseline distortions due

Fig. 2. T1 weighted axial images from a control subject exhibit the selected v(a), frontoparietal parasagittal cortex, right centrum semiovale, and left centfigure legend, the reader is referred to the web version of this article.)


to brain macromolecules were suppressed by omittingthe two first data points. Starting values for line-widthconstraints, chemical shifts and J-coupling were sup-plied as prior knowledge [24]. The manual steps forspectral convergence included zero- and first-orderphase corrections with subsequent identification of inor-ganic phosphate (Pi) and phosphocreatine (PCr) peaks(Fig. 3). For a more objective and conservativeapproach, we analyzed only the intracellular pH in thisstudy. The pH values were calculated using the chemicalshift differences between the Pi and PCr resonances,accordingly to derived mathematical algorithms [25–26].

2.8. Statistical analysis

A statistical expert carried out the analysis by usingthe Statistical Package for the Social Sciences, version17.0.1 (SPSS Inc., Chicago, IL, USA). A Bonferroniadjusted probability p-value <0.01 was adopted for sta-tistically significant differences, with confidence intervalsof 95%. For normal distributions, the parametric Stu-dent’s t-test was used for comparisons between groupsof patients and controls. On the other hand, if theassumption of normality was rejected for the distribu-tion of pH values, a non-parametric Mann–Whitney testwas used for the comparisons.

Analysis of variance (ANOVA) and the non-para-metric Kruskall–Wallis tests were used for the evalua-tion of pH values in individual subgroups of MCD,comparing each one with the control group. Dunnetand Dunn tests with Bonferroni corrections werealso applied for comparisons in the subcategories of cor-tical malformations (focal cortical dysplasia; subcorti-cal and subependymal heterotopia; schizencephaly andpolymicrogyria).

Dunnet and Dunn tests were also used to evaluate theeffect of recent seizures on pH values. Pearson’s coeffi-cients were applied to examine the correlations between

oxels (small boxes highlighted in yellow): right putamen, left putamenrum semiovale (b). (For interpretation of the references to color in this


http://www.mrui.uab.es/mrui


Fig. 3. Spectral convergence obtained with MRUI software. The curves obtained during the fitting process are displayed in the frequency domain inparts per million (ppm). Residual differences (top curve) between the original and estimated data are minimal. The Pi and PCr peaks are depicted inblue in the estimated curve. Note that PCr is the zero reference (ppm = 0), according to conventions.


intracellular pH and duration of epilepsy (while control-ling for age), and between pH and age of seizures onset(while controlling for epilepsy duration).

3. Results

There were no significant differences between patientsand controls in the right and left putamina (Fig. 4).Intracellular pH was significantly lower in the NAP ofthe frontoparietal parasagittal cortex (6.985 ± 0.022),right centrum semiovale (7.004 ± 0.029), and left cen-trum semiovale (6.995 ± 0.030). The pH values in thecontrol group were 7.087 ± 0.048, 7.096 ± 0.042 and

Fig. 4. Box-plot graphics. Comparisons of mean pH values in the NAPfrontoparietal parasagittal cortex, (d) right centrum semiovale, (e) left centru2 = heterotopia (purple); 3 = schizencephaly and polymicrogyria (gray). Odemonstrated statistically significant differences (p < 0.001) between the MCDreferences to color in this figure legend, the reader is referred to the web ver


7.088 ± 0.045 in these same regions, respectively, withp-values <0.001 for all comparisons (Table 1).

The intracellular pH values were also evaluated withANOVA and Kruskall–Wallis tests in these three areasof NAP (frontoparietal parassagital cortex, right cen-trum semiovale, and left centrum semiovale), consider-ing the three subcategories of MCD as individualsubgroups compared to the control group, with resul-tant p-values <0.001 (Fig. 4). In multiple comparisons(Dunnet and Dunn tests), the p-values also remainedstatistically significant different (p <0.001) after Bonfer-roni corrections. Metabolic abnormalities in the visiblelesions were presented and discussed elsewhere [10].

in the following regions: (a) right putamen, (b) left putamen, (c)m semiovale. 0 = control-group (blue); 1 = cortical dysplasia (green);nly in (c–e), Dunnet and Dunn tests with Bonferroni correctionssubgroups compared to the control-group. (For interpretation of the

sion of this article.)



Table 1Comparisons of pH values between MCD patients and controls in five cerebral regions containing NAP.

Cerebral region Patients Controls(n = 37) (n = 31)

Mean SD Mean SD p-Value

Right putamen 6.914 0.033 7.007 0.046 p = 0.621Left putamen 7.011 0.040 6.995 0.043 p = 0.170Frontoparietal parasagittal cortex 6.985 0.022 7.087 0.048 p < 0.001*

Right centrum semiovale 7.004 0.029 7.096 0.042 p < 0.001

Left centrum semiovale 6.995 0.030 7.088 0.045 p < 0.001

Note: SD indicates standard deviation. p-values obtained with the Student’s t test, and the Mann–Whitney test*. Bonferroni adjusted p-values <0.01are printed in bold type.


Patients with recent seizures (within 24 h prior to MRscan, n = 8) as well as patients who did not presentrecent seizures (>24 h prior to MR scan, n = 29) exhib-ited lower pH values in the bilateral white matter andfrontoparietal cortex in comparison to controls(n = 31, p-values <0.001 for both comparisons withDunnet and Dunn tests).

There were no significant correlations between pHvalues and epilepsy duration, or between pH and ageof seizures onset for all voxels, as evaluated with Pear-son’s correlation coefficients at p <0.01.

4. Discussion

Herein, we confirm the hypothesis that metabolicabnormalities extend beyond the visible lesions inpatients with cortical malformations and epilepsy. Wehave demonstrated shifts toward acidosis in three areasof NAP in a large series of MCD patients, encompassingindividuals with focal cortical dysplasia, heterotopia,schizencephaly, and polymicrogyria. The differencesremained significant even when considering the MCDsubgroups individually, in comparison to level of educa-tion-, age- and sex-matched controls. Although MCDcomprise a heterogeneous group of lesions, they areoften presented together in some individuals. The simi-lar findings in all subgroups also reinforce the idea thatMCD are different spectrums of a more broad disease,and that distinct MCD may share similar overlappingpathogenic mechanisms.

The pH is a physicochemical measure of hydrogenions concentration, calculated in a logarithm scale ofbase 10. Thus, even a small decimal variation may indi-cate a major change in the concentration of hydrogenions. We have found lower intracellular pH in theputamina in comparison to the cortex and white matterof healthy subjects. However, all the normative pHvalues reported in this work are in agreement with pre-vious studies (range 6.96–7.10) [25]. The modulation ofintracellular pH in human brain results from a myriadof osmotic and metabolic mechanisms that are primarilyrelated to transport and diffusion of ions, buffering


systems, activity of carbonic anhydrases, and energyconsumption [27].

There is growing evidence linking brain bioenergeticsdisturbances to epileptogenesis. Mitochondria are keystructures in the living cells to keep homeostasis andto provide fuel for several interrelated physiologicalmechanisms, including the maintenance of the electro-physiological state [28]. The role of hydrogen ions inregulating several mitochondrial reactions is striking,and the intracellular pH is a very important factor inthe modulation of bioenergetics. Changes in intracellu-lar pH may modify the characteristics of the creatinekinase enzyme, which catalyzes the replenishment ofadenosine triphosphate (ATP). Neuronal and glial cellsare particularly dependent on energy, and therefore theyare susceptible to disturbances in the intracellular bio-chemical scenario and in the pH microenvironmentsteady-state [29].

The control of membrane potential is closely relatedto Na+ and K+. For instance, boosts in extracellularK+ trigger neuronal depolarization. Under normal cir-cumstances, the concentration of K+ is controlled byNa+/K+ pump and by ion channels present in astro-cytes. These voltage-dependent ion exchangers are alsomodulated by intracellular pH. Additionally, abnormal-ities in glutamate and calcium cycles may have recipro-cal effects on intracellular pH [27,30].

Although epilepsy might result from the networkeffect of disruptions in multiple molecular pathways, itis a general rule that the development of a seizure is dri-ven by an imbalance between inhibition and excitation inneuronal tissue. Ictogenesis may outcome from distur-bances in different levels in the central nervous system:abnormalities in ion concentration and membrane trans-port in neurons and glial cells, disruptions in signal trans-duction between neuronal paths and, ultimately,abnormal connections between major cerebral tracts.Conversely, in vivo studies with animal models showedthat seizures and oxidative stress might also lead to mito-chondrial damage and metabolic impairment [28,31].

Indeed, it is not possible to establish if the resultsdescribed in our study are primary causes or ultimate




consequences of the epileptogenic brain behavior ofMCD patients. However, because we have found thesame results in patients with and without recent seizures,one may speculate that the metabolic changes are notimmediate consequences of epileptic activity. Moreover,we believe that tissue plasticity that occurs in the brainof patients with epilepsy may result in a self-propagatecycle that reduces the excitability threshold and perpet-uates epileptogenic activity.

The idea that MCD visible lesions represent only the“tip of the iceberg” has been intensely debated. 1H-MRS, a more vastly employed spectroscopic modality,showed a reduction in N-acetyl aspartate levels not onlyin the MCD lesions, but also in the NAP, defined eitheras the perilesional parenchyma or the contralateral side[11,16,18].

Eriksson and coworkers demonstrated anisotropyand diffusivity changes in the NAP outside the MRI-vis-ible cortical malformations with DTI, suggesting thatmore widespread functional and microstructural abnor-malities are present [13]. Several others imaging tech-niques, as well as studies with electroencephalogramand histopathology have demonstrated abnormalitiesoutside the MRI visible malformations [12,14,19].

Different theories could explain the findings of meta-bolic impairments in the NAP of MCD patients. A firsthypothesis is that such abnormalities occur in tissueswith normal cytoarchitecture, but with intrinsic epilep-togenic activity. This is supported by a study that corre-lated positron emission tomography, MRI, andintraoperative electrocorticography. The authors foundaberrant electric pattern and abnormal metabolism inhistologically normal perilesional tissues [32]. A secondtheory is that the functional disruptions occur in tissueswith microscopic abnormalities, not detectable by con-ventional MRI sequences, such as heterotopic clustersof neurons in otherwise NAP. Finally, we may postulatethat these findings in NAP could ultimately reflect thepropagation of abnormal electric activity to functionallyand histologically normal tissues.

There are evidences of abnormal cortical connectionsinvolving the thalamus and putamen ipsilateral or con-tralateral in patients with epilepsy secondary to mesialtemporal sclerosis [28]. Nevertheless, we have not foundsignificant differences in mean pH values in the bilateralputamina in our study. Patients exhibited lower pH val-ues in the right putamen in comparison to their left side.However, differences in the proportion of MCD lesionsbetween cerebral hemispheres could not explain this find-ing. These voxels, although centered in the basal ganglia,might have suffered from partial volume effect from thesurrounding structures, such as internal and externalcapsules, as well as lower signal-to-noise ratio, that couldpossibly have underestimated subtle differences [33].

For the comparisons of NAP between patients andcontrols, we have chosen homologous voxels, paired


for each cerebral hemisphere and in similar topography,resulting in comparable gray and white matter concen-trations and alike signal-to-noise ratios between patientsand controls. Furthermore, we have not selected voxelslocated in the periphery of the spectroscopy matrix, inorder to avoid contamination of voxels with surround-ing skeletal muscle that would certainly bias the analysis[33,34].

One drawback of our study is that we could not ruleout potential effects of antiepileptic drugs in the results.It would be difficult to evaluate a large cohort of drug-naive patients with epilepsy secondary to MCD. Katoet al. have confirmed the findings of acidosis in the brainparenchyma of drug-free patients with bipolar disorder[35]. To our knowledge, so far, there are no studiesinvestigating the effects of antiepileptic drugs in intracel-lular pH.

31P-MRS poses several technical challenges that stilllimit its clinical usage nowadays. It is necessary a dedi-cated head coil, and scan times are usually much longer.However, the more widespread deployment and clinicaluse of MRI scanners with stronger magnetic fields gra-dients will progressively enable the use of this methodin the investigation of a variety of brain diseases.

The depiction of metabolic abnormalities in other-wise NAP may be of particular importance in MCDpatients who are candidates for lesion resection. Indeed,some MCD patients undergo seizure recurrence aftercomplete excision of the visible epileptogenic lesion,and others do not present complete cognitive recovery[36]. Our results suggest that clinical treatment for con-trolling seizures might also be extended for the controlof metabolic impairments in the NAP. The biochemicalabnormalities detected in NAP could be a secondary tar-get for medical interventions. The understanding ofthese biochemical abnormalities that occur beyond thevisible lesions may facilitate this approach.

In conclusion, our findings suggest that patients withMCD have more extensive metabolic abnormalities thatextend beyond the MR-visible lesions. There are shiftstowards acidosis in the NAP of diverse subgroups ofMCD patients.

Disclosure of conflicts of interest

None of the authors has any conflict of interest todisclose.

Acknowledgments

Dr. Celi Santos Andrade is a recipient of a post-doc-toral Grant from FAPESP (Sao Paulo Research Foun-dation, Grants 2012/00398-1 and 2013/1552-9). Dr.Claudia Costa Leite is supported by CNPq (National




Council for Scientific and Technological Development,Grant 308267/008-7).

References

[1] Guerrini R, Marini C. Genetic malformations of cortical devel-opment. Exp Brain Res 2006;173:322–33.

[2] Tinuper P, D’Orsi G, Bisulli F, Zaniboni A, Piraccini A, BernardiA, et al. Malformation of cortical development in adult patients.Epileptic Disord 2003;5:85–90.

[3] Hanaoka S, Takashima S, Morooka K. Study of the maturationof the child’s brain using P-31-MRS. Pediatr Neurol1998;18:305–10.

[4] Menuel C, Guillevin R, Costalat R, Perrin M, Sahli-Amor M,Martin-Duverneuil N, et al. Phosphorus magnetic resonancespectroscopy: brain pathologies applications. J Neuroradiol2010;37:73–82.

[5] Maintz D, Heindel W, Kugel H, Jaeger R, Lackner KJ.Phosphorus-31 MR spectroscopy of normal adult human brainand brain tumours. NMR Biomed 2002;15:18–27.

[6] Lodi R, Lotti S, Cortelli P, Pierangeli G, Cevolli S, Clementi V,et al. Deficient energy metabolism is associated with low freemagnesium in the brains of patients with migraine and clusterheadache. Brain Res Bull 2001;54:437–41.

[7] Minderhoud JM, Mooyaart EL, Kamman RL, Teelken AW,Hoogstraten MC, Vencken LM, et al. In vivo phosphorusmagnetic-resonance spectroscopy in multiple sclerosis. ArchNeurol 1992;49:161–5.

[8] Hetherington HP, Pan JW, Spencer DD. H-1 and P-31 spectros-copy and bioenergetics in the lateralization of seizures in temporallobe epilepsy. J Magn Reson Imaging 2002;16:477–83.

[9] Pan JW, Kim JH, Cohen-Gadol A, Pan C, Spencer DD,Hetherington HP. Regional energetic dysfunction in hippocampalepilepsy. Acta Neurol Scand 2005;111:218–24.

[10] Andrade CS, Otaduy MCG, Valente KDR, Maia DF, Park EJ,Valerio RM, et al. Phosphorus magnetic resonance spectroscopyin malformations of cortical development. Epilepsia2011;52:2276–84.

[11] Leite CC, Lucato LT, Sato JR, Valente KDR, Otaduy MCG.Multivoxel proton MR spectroscopy in malformations of corticaldevelopment. Am J Neuroradiol 2007;28:1071–5.

[12] Cross JH. Functional neuroimaging of malformations of corticaldevelopment. Epileptic Disord 2003;5:S73–80.

[13] Eriksson SH, Rugg-Gunn EJ, Symms MR, Barker GJ, DuncanJS. Diffusion tensor imaging in patients with epilepsy andmalformations of cortical development. Brain 2001;124:617–26.

[14] Avery RA, Zubal IG, Stokking R, Studholme C, Corsi M, SeibylJP, et al. Decreased cerebral blood flow during seizures with ictalSPECT injections. Epilepsy Res 2000;40:53–61.

[15] Andrade CS, Leite CC. Malformations of cortical development:current concepts and advanced neuroimaging review. Arq Neu-ropsiquiatr 2011;69:130–8.

[16] Kuzniecky R, Hetherington H, Pan J, Hugg J, Palmer C, GilliamF, et al. Proton spectroscopic imaging at 4.1 Tesla in patients withmalformations of cortical development and epilepsy. Neurology1997;48:1018–24.

[17] Mueller SG, Laxer KD, Barakos JA, Cashdollar N, FlennikenDL, Vermathen P, et al. Metabolic characteristics of corticalmalformations causing epilepsy. J Neurol 2005;252:1082–92.

[18] Li LM, Cendes F, Bastos AC, Andermann F, Dubeau F, ArnoldDL. Neuronal metabolic dysfunction in patients with cortical


developmental malformations: a proton magnetic resonancespectroscopic imaging study. Neurology 1998;50:755–9.

[19] Tassi L, Pasquier B, Minotti L, Garbelli R, Kahane P, BenabidAL, et al. Cortical dysplasia: electroclinical, imaging, andneuropathologic study of 13 patients. Epilepsia 2001;42:1112–23.

[20] Barkovich AJ, Kuzniecky RI, Jackson GD, Guerrini R, DobynsWB. A developmental and genetic classification for malforma-tions of cortical development. Neurology 2005;65:1873–87.

[21] Hetherington HP, Spencer DD, Vaughan JT, Pan JW. Quanti-tative 31P spectroscopic imaging of human brain at 4 Tesla:assessment of gray and white matter differences of phosphocre-atine and ATP. Magn Reson Med 2001;45:46–52.

[22] Naressi A, Couturier C, Castang I, de Beer R, Graveron-DemillyD. Java-based graphical user interface for MRUI, a softwarepackage for quantitation of in vivo/medical magnetic resonancespectroscopy signals. Comput Biol Med 2001;31:269–86.

[23] Vanhamme L, van den Boogaart A, Van Huffel S. Improvedmethod for accurate and efficient quantification of MRS data withuse of prior knowledge. J Magn Reson 1997;129:35–43.

[24] Hamilton G, Patel N, Forton DM, Hajnal JV, Taylor-RobinsonSD. Prior knowledge for time domain quantification of in vivobrain or liver P-31 MR spectra. NMR Biomed 2003;16:168–76.

[25] Barker PB, Butterworth EJ, Boska MD, Nelson J, Welch KM.Magnesium and pH imaging of the human brain at 3.0 Tesla.Magn Reson Med 1999;41:400–6.

[26] Petroff OAC, Prichard JW, Behar KL, Alger JR, den HollanderJA, Shulman RG, et al. Cerebral intracellular pH by P-31 nuclearmagnetic-resonance spectroscopy. Neurology 1985;35:781–8.

[27] Chesler M. Regulation and modulation of pH in the brain.Physiol Rev 2003;83:1183–221.

[28] Pan JW, Williamson A, Cavus I, Hetherington HP, Zaveri H,Petroff OA, et al. Neurometabolism in human epilepsy. Epilepsia2008;49(Suppl. 3):31–41.

[29] Forester B, Berlow YA, Harper DG, Jensen JE, Lange N,Froimowitz MP, et al. Age-related changes in brain energetics andphospholipid metabolism. NMR Biomed 2010;23:242–50.

[30] Traub RD, Michelson-Law H, Bibbig AEJ, Buhl EH, Whitting-ton MA. Gap junctions, fast oscillations and the initiation ofseizures. Adv Exp Med Biol 2004;548:110–22.

[31] Liang LP, Patel M. Seizure-induced changes in mitochondrialredox status. Free Radic Biol Med 2006;40:316–22.

[32] Juhasz C, Chugani DC, Muzik O, Watson C, Shah J, Shah A,et al. Electroclinical correlates of flumazenil and fluorodeoxyglu-cose PET abnormalities in lesional epilepsy. Neurology2000;55:825–34.

[33] Kreis R. Issues of spectral quality in clinical 1H-magneticresonance spectroscopy and a gallery of artifacts. NMR Biomed2004;17:361–81.

[34] Mason GF, Chu WJ, Vaughan JT, Ponder SL, Twieg DB, AdamsD, et al. Evaluation of P-31 metabolite differences in humancerebral gray and white matter. Mag Reson Med 1998;39:346–53.

[35] Kato T, Murashita J, Kamiya A, Shiori T, Kato N, Inubushi T.Decreased intracellular pH measured by P31-MRS in bipolardisorder: a confirmation in drug-free patients and correlation withwhite matter hyperintensity. Eur Arch Psychiatry Clin Neurosci1998;248:301–6.

[36] Roulet-Perez E, Davidoff V, Mayor-Dubois C, Maeder-Ingvar M,Seeck M, Ruffieux C, et al. Impact of severe epilepsy ondevelopment: recovery potential after successful early epilepsysurgery. Epilepsia 2010;51:1266–76.


http://refhub.elsevier.com/S0387-7604(14)00003-5/h0005
























































































































widespread ph abnormalities in patients with malformations of cortical development and epilepsy: a...

Documents