Five-institution study of automated classification of pathological slowing from adult scalp electroencephalograms

Pathological slowing in the electroencephalogram (EEG) is widely investigated for the diagnosis of neurological disorders. Currently, the gold standard for slowing detection is the visual inspection of the EEG by experts. However, visual inspection is time-consuming and subjective. Moreover, there is shortage of EEG experts worldwide. To address those issues, we propose three automated approaches to detect slowing in EEG: unsupervised learning system (ULS), supervised shallow learning system (SSLS), and supervised deep learning system (SDLS). These systems are evaluated on single-channel segments (channel-level), multi-channel segments (segment-level), and entire EEGs (EEG-level). The ULS performs prediction via spectrum features. By contrast, the SSLS and SDLS detect slowing at individual channels via a channel-level slowing detector, and leverage the channel-level detections for detections on the level of segments and full EEGs. We evaluate the systems through Leave-One-Subject-Out (LOSO) cross-validation (CV) and Leave-One-Institution-Out (LOIO) CV on four datasets from the US, Singapore, and India. The SDLS achieved the best overall results: LOIO CV mean balanced accuracy (BAC) of 71.9%, 75.5%, and 82.0% at channel-, segment- and EEG-level, and LOSO CV mean BAC of 73.6%, 77.2%, and 81.8% at channel-, segment-, and EEG-level. The channel- and segment-level performance is comparable to the intra-rater agreement (IRA) of an expert of 72.4% and 82%. The SDLS can process a 30-minutes EEG in 4 seconds, and may be deployed to assist clinicians in interpreting EEGs.

[1]  Katherine Noe.  Most Experts Agree … But What About Other EEG Readers? , 2020, Epilepsy currents.

[2]  V. Mathews,et al.  Magnetic Resonance Imaging (MRI)and Electroencephalographic (EEG) Findings in a Cohort of Normal Children With Newly Diagnosed Seizures , 2006, Journal of child neurology.

[3]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[4]  Chao Wu,et al.  DeepSleepNet: A Model for Automatic Sleep Stage Scoring Based on Raw Single-Channel EEG , 2017, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[5]  Bao-Liang Lu,et al.  Emotional state classification from EEG data using machine learning approach , 2014, Neurocomputing.

[6]  Justin Dauwels,et al.  Automated Detection of Interictal Epileptiform Discharges from Scalp Electroencephalograms by Convolutional Neural Networks , 2020, Int. J. Neural Syst..

[7]  Bernhard Schölkopf,et al.  A tutorial on support vector regression , 2004, Stat. Comput..

[8]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[9]  Zhenghao Chen,et al.  On Random Weights and Unsupervised Feature Learning , 2011, ICML.

[10]  LuBao-Liang,et al.  Emotional state classification from EEG data using machine learning approach , 2014 .

[11]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[12]  E. Bahar,et al.  EEG changes in patients with intracranial tumors and seizures symptom at Mohammad Hoesin Hospital Palembang , 2019, Journal of Physics: Conference Series.

[13]  Shouqian Sun,et al.  Single-trial EEG classification of motor imagery using deep convolutional neural networks , 2017 .

[14]  Cuntai Guan,et al.  Quantitative EEG as Biomarkers for the Monitoring of Post-Stroke Motor Recovery in BCI and tDCS Rehabilitation , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[15]  A. Kaszniak,et al.  Cerebral atrophy, EEG slowing, age, education, and cognitive functioning in suspected dementia , 1979, Neurology.

[16]  Ettore Beghi,et al.  Inter-rater reliability of the EEG reading in patients with childhood idiopathic epilepsy , 2005, Epilepsy Research.

[17]  Kim Dremstrup,et al.  EMD-Based Temporal and Spectral Features for the Classification of EEG Signals Using Supervised Learning , 2016, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[18]  Simon Finnigan,et al.  Defining abnormal slow EEG activity in acute ischaemic stroke: Delta/alpha ratio as an optimal QEEG index , 2016, Clinical Neurophysiology.

[19]  Nitesh V. Chawla,et al.  SMOTE: Synthetic Minority Over-sampling Technique , 2002, J. Artif. Intell. Res..

[20]  Laurent Vercueil,et al.  A convolutional neural network for sleep stage scoring from raw single-channel EEG , 2018, Biomed. Signal Process. Control..

[21]  K. Iyer.  Effective assessments of electroencephalography during stroke recovery: contemporary approaches and considerations. , 2017, Journal of neurophysiology.

[22]  Tonio Ball,et al.  Machine-learning-based diagnostics of EEG pathology , 2020, NeuroImage.

[23]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[24]  Yike Guo,et al.  Automatic Sleep Stage Scoring with Single-Channel EEG Using Convolutional Neural Networks , 2016, ArXiv.

[25]  J. Friedman.  Stochastic gradient boosting , 2002 .

[26]  Joseph Picone,et al.  Improved EEG event classification using differential energy , 2015, 2015 IEEE Signal Processing in Medicine and Biology Symposium (SPMB).

[27]  Hiroshi Shibasaki,et al.  A revised glossary of terms most commonly used by clinical electroencephalographers and updated proposal for the report format of the EEG findings. Revision 2017 , 2017, Clinical neurophysiology practice.

[28]  Nigel P Dowdall.  “Most experts agree” , 2008, BMJ : British Medical Journal.

[29]  Sebastian Raschka,et al.  Python Machine Learning , 2015 .

[30]  Andrew S. Blum,et al.  Focal and Generalized Slowing, Coma, and Brain Death , 2007 .

[31]  John Thomas,et al.  EEG CLassification Via Convolutional Neural Network-Based Interictal Epileptiform Event Detection , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[32]  Ana Catarina Fonseca,et al.  Quantitative EEG and functional outcome following acute ischemic stroke , 2018, Clinical Neurophysiology.

[33]  Arthur C. Grant,et al.  EEG interpretation reliability and interpreter confidence: A large single-center study , 2014, Epilepsy & Behavior.

[34]  R. Kakigi,et al.  Source localization of posterior slow waves of youth using dipole modeling , 2012, Psychiatry and Clinical Neurosciences.

[35]  H. Soininen,et al.  Slowing of EEG in Parkinson's disease. , 1991, Electroencephalography and clinical neurophysiology.

[36]  Sandy Middleton,et al.  Acute single channel EEG predictors of cognitive function after stroke , 2017, PloS one.

[37]  Guy Nagels,et al.  Quantitative EEG in ischemic stroke: Correlation with infarct volume and functional status in posterior circulation and lacunar syndromes , 2011, Clinical Neurophysiology.

[38]  Joseph Picone,et al.  The Temple University Hospital EEG Data Corpus , 2016, Front. Neurosci..

[39]  Y. Xing,et al.  Transcranial Doppler combined with quantitative EEG brain function monitoring and outcome prediction in patients with severe acute intracerebral hemorrhage , 2018, Critical Care.

[40]  Maarten De Vos,et al.  DNN Filter Bank Improves 1-Max Pooling CNN for Single-Channel EEG Automatic Sleep Stage Classification , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[41]  Lauren C. Frey,et al.  Electroencephalography (EEG): An Introductory Text and Atlas of Normal and Abnormal Findings in Adults, Children, and Infants , 2016 .

[42]  Eamonn J. Keogh,et al.  Rapid annotation of interictal epileptiform discharges via template matching under Dynamic Time Warping , 2016, Journal of Neuroscience Methods.

[43]  Surya Ganguli,et al.  Exact solutions to the nonlinear dynamics of learning in deep linear neural networks , 2013, ICLR.