Skip to main content

Machine learning identifies “rsfMRI epilepsy networks” in temporal lobe epilepsy



Experimental models have provided compelling evidence for the existence of neural networks in temporal lobe epilepsy (TLE). To identify and validate the possible existence of resting-state “epilepsy networks,” we used machine learning methods on resting-state functional magnetic resonance imaging (rsfMRI) data from 42 individuals with TLE.


Probabilistic independent component analysis (PICA) was applied to rsfMRI data from 132 subjects (42 TLE patients + 90 healthy controls) and 88 independent components (ICs) were obtained following standard procedures. Elastic net-selected features were used as inputs to support vector machine (SVM). The strengths of the top 10 networks were correlated with clinical features to obtain “rsfMRI epilepsy networks.”


SVM could classify individuals with epilepsy with 97.5% accuracy (sensitivity = 100%, specificity = 94.4%). Ten networks with the highest ranking were found in the frontal, perisylvian, cingulo-insular, posterior-quadrant, thalamic, cerebello-thalamic, and temporo-thalamic regions. The posterior-quadrant, cerebello-thalamic, thalamic, medial-visual, and perisylvian networks revealed significant correlation (r > 0.40) with age at onset of seizures, the frequency of seizures, duration of illness, and a number of anti-epileptic drugs.


IC-derived rsfMRI networks contain epilepsy-related networks and machine learning methods are useful in identifying these networks in vivo. Increased network strength with disease progression in these “rsfMRI epilepsy networks” could reflect epileptogenesis in TLE.

Key Points

• ICA of resting-state fMRI carries disease-specific information about epilepsy.

• Machine learning can classify these components with 97.5% accuracy.

• “Subject-specific epilepsy networks” could quantify “epileptogenesis” in vivo.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3



Cornus amonis


Fascia dentata


False discovery rate


Hippocampal gray matter volume


Independent component analysis


Independent components


Machine learning


Mesial temporal sclerosis


Probabilistic independent component analysis


Region of interest


Resting-state functional magnetic resonance imaging




Support vector machine


Temporal lobe epilepsy


  1. Chiang S, Haneef Z (2014) Graph theory findings in the pathophysiology of temporal lobe epilepsy. Clin Neurophysiol 125:1295–1305

    Article  PubMed  PubMed Central  Google Scholar 

  2. Liao W, Zhang Z, Pan Z et al (2010) Altered functional connectivity and small-world in mesial temporal lobe epilepsy. PLoS One 5:e8525

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Vlooswijk MC, Vaessen MJ, Jansen JF et al (2011) Loss of network efficiency associated with cognitive decline in chronic epilepsy. Neurology 77:938–944

    Article  CAS  PubMed  Google Scholar 

  4. Liao W, Zhang Z, Pan Z et al (2011) Default mode network abnormalities in mesial temporal lobe epilepsy: a study combining fMRI and DTI. Hum Brain Mapp 32:883–895

    Article  PubMed  Google Scholar 

  5. Widjaja E, Zamyadi M, Raybaud C, Snead OC, Smith ML (2013) Abnormal functional network connectivity among resting-state networks in children with frontal lobe epilepsy. AJNR Am J Neuroradiol 34:2386–2392

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Zhang Z, Lu G, Zhong Y et al (2009) Impaired perceptual networks in temporal lobe epilepsy revealed by resting fMRI. J Neurol 256:1705–1713

    Article  PubMed  Google Scholar 

  7. Waites AB, Briellmann RS, Saling MM, Abbott DF, Jackson GD (2006) Functional connectivity networks are disrupted in left temporal lobe epilepsy. Ann Neurol 59:335–343

    Article  PubMed  Google Scholar 

  8. Luo C, Li Q, Xia Y et al (2012) Resting state basal ganglia network in idiopathic generalized epilepsy. Hum Brain Mapp 33:1279–1294

    Article  PubMed  Google Scholar 

  9. Damoiseaux JS, Rombouts SA, Barkhof F et al (2006) Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci U S A 103:13848–13853

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Beckmann CF, Smith SM (2005) Tensorial extensions of independent component analysis for multisubject FMRI analysis. Neuroimage 25:294–311

    Article  CAS  PubMed  Google Scholar 

  11. Smith SM, Jenkinson M, Woolrich MW et al (2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23(Suppl 1):S208–S219

    Article  PubMed  Google Scholar 

  12. Cerliani L, Thomas RM, Aquino D, Contarino V, Bizzi A (2017) Disentangling subgroups of participants recruiting shared as well as different brain regions for the execution of the verb generation task: a data-driven fMRI study. Cortex 86:247–259

    Article  PubMed  Google Scholar 

  13. Li S, Tian J, Li M et al (2018) Altered resting state connectivity in right side frontoparietal network in primary insomnia patients. Eur Radiol 28:664–672

    Article  PubMed  Google Scholar 

  14. Panda R, Bharath RD, Upadhyay N, Mangalore S, Chennu S, Rao SL (2016) Temporal dynamics of the default mode network characterize meditation-induced alterations in consciousness. Front Hum Neurosci 10:372

    Article  PubMed  PubMed Central  Google Scholar 

  15. Schölvinck ML, Maier A, Ye FQ, Duyn JH, Leopold DA (2010) Neural basis of global resting-state fMRI activity. Proc Natl Acad Sci U S A 107:10238–10243

    Article  PubMed  PubMed Central  Google Scholar 

  16. Rodionov R, De Martino F, Laufs H et al (2007) Independent component analysis of interictal fMRI in focal epilepsy: comparison with general linear model-based EEG-correlated fMRI. Neuroimage 38:488–500

    Article  CAS  PubMed  Google Scholar 

  17. Simon P (2013) Too big to ignore: the business case for big data. John Wiley & Sons, Inc. New Jersey

  18. Arbabshirani MR, Plis S, Sui J, Calhoun VD (2017) Single subject prediction of brain disorders in neuroimaging: promises and pitfalls. Neuroimage 145:137–165

    Article  PubMed  Google Scholar 

  19. Tognin S, Pettersson-Yeo W, Valli I et al (2013) Using structural neuroimaging to make quantitative predictions of symptom progression in individuals at ultra-high risk for psychosis. Front Psychiatry 4:187

    PubMed  Google Scholar 

  20. van der Burgh HK, Schmidt R, Westeneng HJ, de Reus MA, van den Berg LH, van den Heuvel MP (2017) Deep learning predictions of survival based on MRI in amyotrophic lateral sclerosis. Neuroimage Clin 13:361–369

    Article  PubMed  Google Scholar 

  21. Chen CP, Keown CL, Jahedi A et al (2015) Diagnostic classification of intrinsic functional connectivity highlights somatosensory, default mode, and visual regions in autism. Neuroimage Clin 8:238–245

    Article  PubMed  PubMed Central  Google Scholar 

  22. Kaufmann T, Skåtun KC, Alnaes D et al (2015) Disintegration of sensorimotor brain networks in schizophrenia. Schizophr Bull 41:1326–1335

    Article  PubMed  PubMed Central  Google Scholar 

  23. Ryali S, Chen T, Supekar K, Menon V (2012) Estimation of functional connectivity in fMRI data using stability selection-based sparse partial correlation with elastic net penalty. Neuroimage 59:3852–3861

    Article  PubMed  Google Scholar 

  24. Ng B, Vahdat A, Hamarneh G, Abugharbieh R (2010) Generalized sparse classifiers for decoding cognitive states in fMRI. In: Wang F, Yan P, Suzuki K, Shen D (eds) Machine Learning in Medical Imaging. Lecture Notes in Computer Science, vol 6357. Springer, Berlin

  25. Sochat V, Supekar K, Bustillo J, Calhoun V, Turner JA, Rubin DL (2014) A robust classifier to distinguish noise from fMRI independent components. PLoS One 9:e95493

    Article  PubMed  PubMed Central  Google Scholar 

  26. Beckmann CF, DeLuca M, Devlin JT, Smith SM (2005) Investigations into resting-state connectivity using independent component analysis. Philos Trans R Soc Lond Ser B Biol Sci 360:1001–1013

    Article  Google Scholar 

  27. Beckmann CF, Mackay CE, Filippini N, Smith SM (2009) Group comparison of resting-state FMRI data using multi-subject ICA and dual regression. Neuroimage 47:S148

    Article  Google Scholar 

  28. Chollet F (2015) Keras: deep learning library for theano and tensorflow. Available via

  29. Zou H, Hastie T (2005) Regularization and variable selection via the elastic net. J R Statist Soc B 67:301–320

    Article  Google Scholar 

  30. Barrat A, Barthélemy M, Pastor-Satorras R, Vespignani A (2004) The architecture of complex weighted networks. Proc Natl Acad Sci U S A 101:3747–3752

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Suppa P, Anker U, Spies L et al (2015) Fully automated atlas-based hippocampal volumetry for detection of Alzheimer’s disease in a memory clinic setting. J Alzheimers Dis 44:183–193

    Article  PubMed  Google Scholar 

  32. Eickhoff SB, Stephan KE, Mohlberg H et al (2005) A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. Neuroimage 25:1325–1335

    Article  PubMed  Google Scholar 

  33. Smith SM, Fox PT, Miller KL et al (2009) Correspondence of the brain's functional architecture during activation and rest. Proc Natl Acad Sci U S A 106:13040–13045

    Article  PubMed  PubMed Central  Google Scholar 

  34. Barba C, Rheims S, Minotti L et al (2016) Reply: temporal plus epilepsy is a major determinant of temporal lobe surgery failures. Brain 139:e36

    Article  PubMed  Google Scholar 

  35. Kelly RE Jr, Alexopoulos GS, Wang Z et al (2010) Visual inspection of independent components: defining a procedure for artifact removal from fMRI data. J Neurosci Methods 189:233–245

    Article  PubMed  PubMed Central  Google Scholar 

  36. Griffanti L, Douaud G, Bijsterbosch J et al (2017) Hand classification of fMRI ICA noise components. Neuroimage 154:188–205

    Article  PubMed  Google Scholar 

  37. Feigin A, Kaplitt MG, Tang C et al (2007) Modulation of metabolic brain networks after subthalamic gene therapy for Parkinson’s disease. Proc Natl Acad Sci U S A 104:19559–19564

    Article  PubMed  PubMed Central  Google Scholar 

  38. Hillary FG, Rajtmajer SM, Roman CA et al (2014) The rich get richer: brain injury elicits hyperconnectivity in core subnetworks. PLoS One 9:e104021

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Pitkänen A, Sutula TP (2002) Is epilepsy a progressive disorder? Prospects for new therapeutic approaches in temporal-lobe epilepsy. Lancet Neurol 1:173–181

    Article  PubMed  Google Scholar 

  40. Scharfman HE (2007) The neurobiology of epilepsy. Curr Neurol Neurosci Rep 7:348–354

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Dyhrfjeld-Johnsen J, Santhakumar V, Morgan RJ, Huerta R, Tsimring L, Soltesz I (2007) Topological determinants of epileptogenesis in large-scale structural and functional models of the dentate gyrus derived from experimental data. J Neurophysiol 97:1566–1587

    Article  PubMed  Google Scholar 

  42. Salinsky M, Kanter R, Dasheiff RM (1987) Effectiveness of multiple EEGs in supporting the diagnosis of epilepsy: an operational curve. Epilepsia 28:331–334

    Article  CAS  PubMed  Google Scholar 

  43. Javidan M (2012) Electroencephalography in mesial temporal lobe epilepsy: a review. Epilepsy Res Treat 2012:637430

    PubMed  PubMed Central  Google Scholar 

  44. Fergus P, Hussain A, Hignett D, Al-Jumeily D, Abdel-Aziz K, Hamdan H (2016) A machine learning system for automated whole-brain seizure detection. Appl Comput Inf 12:70–89

    Google Scholar 

  45. Focke NK, Yogarajah M, Symms MR, Gruber O, Paulus W, Duncan JS (2012) Automated MR image classification in temporal lobe epilepsy. Neuroimage 59:356–362

    Article  PubMed  Google Scholar 

  46. Chiang S, Levin HS, Haneef Z (2015) Computer-automated focus lateralization of temporal lobe epilepsy using fMRI. J Magn Reson Imaging 41:1689–1694

    Article  PubMed  Google Scholar 

Download references


We acknowledge the Department of Science and Technology, Government of India for providing the 3T MRI scanner for research. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. We also acknowledge research fellows Mr. Aditya Jayashankar and Mr. Sunil K. Khokhar for their help in analysis.


The authors state that this work has not received any funding.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Tapan K. Gandhi.

Ethics declarations


The scientific guarantor of this publication is Dr. Rose Dawn Bharath, Additional Professor, Neuroimaging and Interventional Radiology, NIMHANS, Bengaluru-29, India.

Conflict of interest

The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article.

Statistics and biometry

One of the authors has significant statistical expertise.

Informed consent

Written informed consent was obtained from all subjects (patients) in this study.

Ethical approval

Institutional Review Board approval was obtained.


• Prospective

• Case-control study

• Performed at one institution

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material


(DOCX 1024 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bharath, R.D., Panda, R., Raj, J. et al. Machine learning identifies “rsfMRI epilepsy networks” in temporal lobe epilepsy. Eur Radiol 29, 3496–3505 (2019).

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Temporal lobe epilepsy
  • Magnetic resonance imaging
  • Support vector machine
  • Seizures