Diagnostic Imaging Centre, Oncology Institute of Vojvodina, Sremska Kamenica, Serbia

INTRODUCTION

According to the theory of neuroplasticity,experience can actually change the brain's physical structure and functional organization from “top to bottom”, while certain approaches targeting neurite growth inhibitors may provide multifactorial protection against brain insults[1].Functional magnetic resonance imaging (fMRI) shows that various functional cortical changes occur after brain injuries of different etiologies, and these functions are thought to be represented in dynamic large-scale networks[2].Brain plasticity cannot be fully understood and fruitfully studied without considering the temporal pattern of the injury inflicted on the brain[3].Slow-growing brain lesions do not provide the same level of clinical symptoms as acute brain lesions, even if they are in similar locations and are similar dimensions.Progression of low-grade glioma (LGG) is slow,usually 7-8 years.During this slow progression, a normal clinical outcome is observed in the majority of LGG patients.After a stroke, an acute brain injury, 70% of patients still exhibit mild to severe functional deficits after 11 years[4].Clinical data indicate that slow-growing lesions allow a much greater level of neural re-mapping then acute lesions, but mechanisms underlying this phenomenon remain unclear.In this study,we presented a 3-year follow-up language fMRI study of LGG patient during disease progression, after reoperation, after radiation therapy, and 1 year later.A preliminary account of this study already appeared[5].

CASE REPORT

A right-handed male (Edinburgh Inventory 78), aged 29 years, was subjected to neurosurgical intervention regarding LGG in the left frontal lobe, after incidental loss of consciousness.Histological analysis has proved oligodendroglioma grade II.The patient was not subjected to the proposed radiotherapy and/or chemotherapy treatment on his request.After tumour resection,the patient occasionally experienced a sudden speech arrest lasting approximately 20 seconds, but no loss of consciousness occurred.No other neurological deficits were present.Follow up of patient condition included magnetic resonance imaging (MRI)examination once a year.

The patient gave informed consent regarding undergoing scanning on a 3T Trio MR Unit (Siemens, Erlangen, Germany).In addition to conventional imaging, fMRI was performed for localization of eloquent language areas.

Experimental protocol

Beforescanning, the patient was instructed and rehearsed in the tasks and was asked to remain still in the scanner and perform specified language tasks silently.The functional images were obtained in axial planes at 3-second intervals while the patient alternatively rested and performed a specified language task for 30 seconds.For identification of language-related brain areas, two different block-designed language paradigms including verbal fluency and a language task with a semantic decision component were used.Both paradigms were performed in each follow-up fMRI examination.After scanning, the patient was asked whether he performed the specified task successfully.

Data acquisition and image processing

Technical parameters for the images were as follows:repetition time, 3 000 ms; echo time, 30 ms; matrix 64 × 64 field of view, 240 mm; and slice thickness, 3 mm.fMRI data were processed using FMRIB’s Software Library (Centre for Functional MRI of the Brain [FMRIB],Oxford, UK).For generation of the fMRI activation map,FMRI Expert Analysis Tool (version 5.63, part of the FMRIB’s Software Library) was used.Standard steps in pre-processing were applied: motion correction was done by the Motion Correction using FMRIB’s Linear Image Registration Tool[6], non-brain removal by the Brain Extraction Tool[7], spatial smoothing by a Gaussian kernel of full width at half maximum of 5 mm.

Mean-based intensity normalisation of all volumes was done by the same factor, high-pass temporal filtering(Gaussian-weighted least-squares straight line fitting,with sigma=30.0 seconds).Time-series statistical analysis was performed using the FMRIB’s Improved Linear Model with local autocorrelation correction[8].Z-statistic images were thresholded using clusters determined by Z ＞ 3 and a (corrected) cluster significance threshold ofP=0.05[9].

Three-year follow-up study of language adaptability(Figure 1)

Figure 1 Functional magnetic resonance imaging brain activation maps for verbal fluency and semantic decision tasks in a 29-year-old male with low-grade glioma in the left frontal lobe.(A, B) 7 years after first tumour operation; (C, D) 7 years and 11 months before reoperation; (E, F) 3 months after reoperation and before irradiation; (G, H) 1 year after reoperation and 7 months after irradiation; (I, J) 2 years after reoperation and 19 months after irradiation.

Seven years after tumour resection (at the patient age of 36), MRI revealed the existence of recurrent tumour formation.In addition to conventional MRI, functional blood oxygen level-dependent (BOLD) MRI was performed for localization of language-related brain areas.Mapping of eloquent language cortex included both verbal fluency and semantic decision tasks.Tasks were administered in three periods of 30 seconds each.Both tasks revealed left hemisphere dominance for language(Figures 1A, B).

A second MRI was performed 11 months later in preparation for neurosurgical intervention.fMRI assessment of language-related brain areas was performed according to the same protocol as previously reported.Conventional MRI showed a further increase of lesion diameter.Functional evaluation of language indicated a significant difference compared with the first examination.Specifically, the second fMRI examination revealed right hemi-sphere dominance for the semantic decision task and mixed dominance for the verbal fluency task (Figures 1C,D).At 10 days after this examination, subtotal tumour resection was performed.Histological examination showed that tumour type and grade remain unchanged.At 3 months after tumour resection, conventional MRI and fMRI assessments of language were repeated according to the established protocol.Conventional MRI showed resection of more than 90% of tumour mass,while eloquent language areas in the left hemisphere identified by fMRI were preserved.BOLD response to specified tasks revealed dominance of the left hemisphere for both language tasks (Figures 1E, F).After reoperation, the patient had no aphasic episodes nor any other neurological deficit.The patient agreed to undergo radiation therapy.At 1 year after reoperation and 7 months after irradiation of 55 Gy, a control examination with both conventional and functional segments was performed.Conventional MRI showed good response to radiation therapy.Assessment of language-related areas revealed left hemisphere dominance for the verbal fluency task and mixed dominance for the semantic decision task (Figures 1G, H).

The next control MRI and fMRI performed 1 year later showed no tumour progression and again re-established left hemisphere dominance for both language tasks(Figures 1I, J).

DISCUSSION

Cortical changes occur after brain injuries to compensate for loss of brain function and to optimise performance of certain functions to reduce neurological deficits.The phenomenon of brain recovery after injuries has been studied from different points of view and using brain injuries of different etiologies (stroke, LGG, multiple sclerosis)[10-12].The most frequent cortical change that occurs after brain injury is increased activity in the region participating in the realisation of the specific tasks, which can be explained by the engagement of parallel cortical tracts[13].It is also possible that cortical areas that previously were not included in the realisation of certain functions become active during the execution of the task that involves the injured cortical area[14].The idea that even a mature brain has a number of redundant subsystems that are involved in recovery after brain injury is described in the concept of neural equivalence[15].

In this reported case study, a high level of adaptation of the language-related brain network to a progressive slow-growing lesion is evident.Although there was no evidence of adaptive changes in the first examination(Figures 1A, B) and in the second examination 11 months later, the language-related brain network was extended to language areas of the contralateral hemisphere for the verbal fluency task and was almost completely shifted to the contralateral hemisphere for the language task with a semantic component (Figures 1C,D).Mapping of language-related cortex 3 months after reoperation and before irradiation showed re-established dominance of the left hemisphere for both language tasks (Figures 1E, F).On the other hand, at 7 months after irradiation, the language-related brain network was again extended to the non-dominant hemisphere in a similar way as during the period of disease progression(Figures 1G, H).This could be explained by a disturbed white matter tract after irradiation.Actually, Harriset al[16]performed a study of white matter changes following radiotherapy in adult LGG patients and found that fractional anisotropy (FA), as a measure of impairment of white matter tracts, declined progressively at 3 and 8 months after radiotherapy and recovered at 14 months to values identical to those at 3 months, for dose bins ＞ 55 Gy and 50-55 Gy.According to Harriset al[16], we performed fMRI examinations at 7 months after 55 Gy irradiation in a period of progressive decrease of FA.fMRI exam performed 19 months after irradiation again showed left hemisphere dominance for both language tasks, which is in accordance with Harris’s findings of recovered white matter at 14 months after radiotherapy.fMRI has been used to study recovery from acute lesions(most often stroke).There is a large amount of data showing evidence for functional reorganisation of brain areas close to or distant from a lesion.In patients with acute lesions, recovery is often incomplete, and task-related brain activation studies such as fMRI should be carefully interpreted due to the impaired performance and lack of baseline measurements.In patients with slow-growing lesions, such as LGG, functional compensation can be impressive, and deficits are generally less severe than in patients with acute onset lesions, but there are only a few reports that compared fMRI maps before and after surgery or during disease progression and during therapy[17-18].

Identified changes in functional topology of the language-related brain network in this patient could be interpreted according to an actual neurocognitive model of brain functions as dynamic large-scale networks that are organized around cortical epicentres.In this theory, continuous modification of neuronal networks is an essential property of the brain to store and update information, to acquire new skills to optimise and automate information processing, and to adapt to structural changes (for example aging or brain tumour)[2,19].Network analysis of the functional connectivity in Alzheimer’s disease and schizophrenia patients revealed a significant difference in patients compared with a control group[20-21].These observed differences in network topology suggest that monitoring changes in the network can be a useful tool in understanding different pathologies of the brain.Bosmaet al[22]correlated neurocognitive function to theoretical parameters of functional brain networks of LGG patients using resting-state magnetoencephalography.Their findings suggest widespread changes in the strength and spatial organization of brain network in LGG patients.fMRI can provide information about spatial and temporal aspects of neural activity.The spatial extent of activation ranges from millimetres (firing patterns of group of neurons) to centimetres (interactions between cortical regions).In a similar manner, temporal processes can be represented on a scale from milliseconds (firing patterns of group of neurons, synchronization, and cognitive processes) to weeks (for example recovery from loss of brain function due to stroke or surgery) and to months or years (functional reshaping due to growth of LGG as in our example).An important factor of time and concept of plastic functional topography usually was not included in actual clinical models[2].

This finding, as well as other similar clinical examples[18,23], suggests that the language network of the non-dominant hemisphere is anatomically and functionally connected to the language network of the dominant hemisphere.In normal conditions, it is passively involved,and its integrity is not essential to the language function.Moreover, it seems that inclusion of the non-dominant hemisphere in language generation is a process rather than an immediate action.Successful compensation of language function could be conditioned by preservation of essential network nodes and essential cortical tracts during the critical period.According to Jordan and Rumelhart[24], compensatory mechanisms in preserving vital brain functions are implemented through a process of supervised learning.In this essential network, nodes or the eloquent brain could play a role of “distal teacher”.Recent advances in noninvasive evaluation of brain activity, as well as new models of structural and functional brain connectivity, could provide new possibilities in the evaluation and treatment of patients with brain injuries of different etiologies.

It should be emphasized that, without detailed clarification of the oxygen metabolism during neuronal activity and vascular reorganization in the pathologic brain, fMRI could overlook activated areas in patients with brain disorders, leading to incorrect conclusions concerning functional reorganization.A multimodal approach that includes other techniques (such as magnetoencephalography and near-infrared spectroscopy)[25]may be useful in further evaluation of this phenomenon.

Author contributions:Olivera Sveljo was responsible for the research design, data analysis, literature search, and paper writing.Katarina Koprivsek participated in the research design and data analysis.Milos Lucic was in charge of data analysis and manuscript supervision.

Conflicts of interest:None declared.

[1]Zhu X, Lee J, Wong J, et al.Poststroke DNA immunization against neurite growth inhibitors is beneficial to the recovery from local cerebral ischemia in rats.Neural Regen Res.2007;2:65-69.

[2]Rutten GJ, Ramsey NF.The Role of functional magnetic resonance imaging in brain surgery.Neurosurg Focus.2010;28: E4.

[3]Dismurget M, Bonnetblanc F, Duffau H.Contrasting acute and slow growing lesions a new door to brain plasticity.Brain.2007;130:898-914.

[4]Varona JF, Bermejo F, Guerra JM, et al.Long-term prognosis of ischemic stroke in young adults.Study of 272 cases.J Neurol.2004;251:1507-1514.

[5]Sveljo O, Koprivsek K, Lucić MA, et al.Functional magnetic resonance imaging of cortical changes in a low-grade glioma patient.Vojnosanit Pregl.2010;67:941-944.

[6]Jenkinson M, Bannister P, Brady M, et al.Improved optimization for the robust and accurate linear registration and motion correction of brain images.Neuroimage.2002;17:825-841.

[7]Smith S.Fast robust automated brain extraction.Hum Brain Mapp.2002;17:143-155.

[8]Woolrich MW, Ripley BD, Brady JM, et al.Temporal autocorrelation in univariate linear modelling of fMRI Data.Neuroimage.2001;14:1370-1386

[9]Worsley KJ, Evans AC, Marrett S, et al.A three-dimensional statistical analysis for CBF activation studies in human brain.J Cereb Blood Flow Metab.1992;12:900-918.

[10]Murphy TH, Corbett D.Plasticity during stroke recovery: from synapse to behaviour.Nat Rev Neurosci.2009;10:861-872.

[11]Duffau H.Lessons from brain mapping in surgery for low-grade glioma: insights into associations between tumour and brain plasticity.Lancet Neurol.2005;4:476-486.

[12]Filippi M, Rocca MA.Disturbed function and plasticity in multiple sclerosis as gleaned from functional magnetic resonance imaging.Curr Opin Neurol.2003;16:275-282.

[13]Feydy A, Carlier R, Roby-Brami A, et al.Longitudinal study of motor recovery after stroke, recruitment and focusing of brain activation.Stroke.2002;33:1610-1617.

[14]Pascual-Leone A, Amedi A, Fregini F, et al.The plastic human brain cortex.Annu Rev Neurosci.2005;28:377-401.

[15]Slavin S, Laurence S, Stein DG.Another look at vicariation.In:Finger S, LeVere TE, Almli CR, Stein DG, eds.Brain Injury and Recovery: Theoretical and Controversial Issues.New York: Plenum Press.1988:165-178.

[16]Harris M, Kumar S, Raj MK, et al.Serial diffusion tensor imaging to characterize radiation-induced changes in normal-appearing white matter following radiotherapy in patients with adult low-grade gliomas.Radiat Med.2008;26:140-150.

[17]Krainik A, Duffau H, Capelle L, et al.Role of the healthy hemisphere in recovery after resection of the supplementary motor area.Neurology.2004;62:1323-1332.

[18]Pillai JJ.Insights into adult postlesional language cortical plasticity provided by cerebral blood oxigen level-dependent functional MR Imaging.Am J Neuroradiol.2010;31:990-996.

[19]Honey CJ, Sporns O.Dynamical consequences of lesions in cortical networks.Hum Brain Mapp.2008l;29:802-809.

[20]Supekar K, Menon V, Rubin D, et al.Network analysis of intrinsic functional brain connectivity in Alzheimer’s disease.PLoS Comput Biol.2008;4:e1000100.

[21]Basset DS, Bullmore E, Verchinski BA, et al.Hierarchical organisation of human cortical networks in health and schizophrenia.J Neurosci.2008;28:9239-9248.

[22]Bosma I, Reijneveld JC, Klein M, et al.Disturbed functional brain networks and neurocognitive function in low-grade glioma patients: a graph theoretical analysis of resting-state MEG.Nonlinear Biomed Phys.2009;3:9.

[23]Bhatnagar S, Andy OJ.Language in the nondominant right hemisphere.Arch Neurol.1983;40:728-731.

[24]Jordan MI, Rumelhart DE.Forward models: supervised learning with a distal teacher.Cogn Sci.1992;16:307-354.

[25]Murata Y, Sakatani K, Katayama Y, et al.Decreases of blood oxygenation level- dependent signal in the activated motor cortex during functional recovery after resection of a glioma.Am J Neuroradiol.2004;25:1242-1246.