Neural Network–Based Classification of Anesthesia/Awareness Using Granger Causality Features

This article investigates the signal processing part of a future system for monitoring awareness during surgery. The system uses features from the patients’ electrical brain activity (EEG) to discriminate between “anesthesia” and “awareness.” We investigate the use of a neural network classifier and Granger causality (GC) features for this purpose. GC captures anesthetic-induced changes in the causal relationships between pairs of signals from different brain areas. The differences in the pairwise causality estimated from the EEG activity are used as features for subsequent classification between “awake” and “anesthetized” states. EEG data from 31 subjects obtained during surgery and maintenance of anesthesia with propofol, sevoflurane, or desflurane, are classified using a neural network with one layer of hidden units. An average accuracy of 96% is obtained.

[1]  Max Hirshkowitz,et al.  Normal human sleep: an overview. , 2004, The Medical clinics of North America.

[2]  Nitish V. Thakor,et al.  Monotonicity of approximate entropy during transition from awareness to unresponsiveness due to propofol anesthetic induction , 2006, IEEE Transactions on Biomedical Engineering.

[3]  L. Angelini,et al.  Granger causality for circular variables , 2009, 0902.2341.

[4]  Xue Wang,et al.  Granger Causality between Multiple Interdependent Neurobiological Time Series: Blockwise versus Pairwise Methods , 2007, Int. J. Neural Syst..

[5]  James Theiler,et al.  Testing for nonlinearity in time series: the method of surrogate data , 1992 .

[6]  J. Sleigh,et al.  Monitoring consciousness: the current status of EEG-based depth of anaesthesia monitors. , 2007, Best practice & research. Clinical anaesthesiology.

[7]  J. Sleigh,et al.  Permutation entropy of the electroencephalogram: a measure of anaesthetic drug effect. , 2008, British journal of anaesthesia.

[8]  Karl J. Friston,et al.  Behavioral / Systems / Cognitive Connectivity Changes Underlying Spectral EEG Changes during Propofol-Induced Loss of Consciousness , 2012 .

[9]  Koujirou Yamamoto,et al.  Kidneys contribute to the extrahepatic clearance of propofol in humans, but not lungs and brain. , 2005, British journal of clinical pharmacology.

[10]  Katarzyna J. Blinowska,et al.  Determination of EEG activity propagation: pair-wise versus multichannel estimate , 2004, IEEE Transactions on Biomedical Engineering.

[11]  J. H. Fallon,et al.  Toward a Unified Theory of Narcosis: Brain Imaging Evidence for a Thalamocortical Switch as the Neurophysiologic Basis of Anesthetic-Induced Unconsciousness , 2000, Consciousness and Cognition.

[12]  Jamie Sleigh,et al.  The Bispectral Index , 1999 .

[13]  C. Greim,et al.  Prediction of depth of sedation and anaesthesia by the Narcotrend EEG monitor. , 2004, British journal of anaesthesia.

[14]  J. Sleigh,et al.  Practical Use of the Raw Electroencephalogram Waveform During General Anesthesia: The Art and Science , 2009, Anesthesia and analgesia.

[15]  E. John,et al.  Invariant Reversible QEEG Effects of Anesthetics , 2001, Consciousness and Cognition.

[16]  Joachim Gross,et al.  The effect of filtering on Granger causality based multivariate causality measures , 2010, NeuroImage.

[17]  Anthony G. Hudetz,et al.  Suppressing the mind , 2010 .

[18]  Anthony G. Hudetz,et al.  Are We Unconscious During General Anesthesia? , 2008, International anesthesiology clinics.

[19]  G N Kenny,et al.  Analysis of the EEG bispectrum, auditory evoked potentials and the EEG power spectrum during repeated transitions from consciousness to unconsciousness. , 1998, British journal of anaesthesia.

[20]  S. Dehaene,et al.  Converging Intracranial Markers of Conscious Access , 2009, PLoS biology.

[21]  Tarmo Lipping,et al.  Comparison of entropy and complexity measures for the assessment of depth of sedation , 2006, IEEE Transactions on Biomedical Engineering.

[22]  Eiji Shimizu,et al.  Approximate Entropy in the Electroencephalogram during Wake and Sleep , 2005, Clinical EEG and neuroscience.

[23]  Aneta Stefanovska,et al.  Neuronal synchrony during anesthesia: a thalamocortical model. , 2008, Biophysical journal.

[24]  R. Llinás,et al.  Bursting of thalamic neurons and states of vigilance. , 2006, Journal of neurophysiology.

[25]  J. Sleigh,et al.  The bispectral index: a measure of depth of sleep? , 1999, Anesthesia and analgesia.

[26]  C. Granger.  Testing for causality: a personal viewpoint , 1980 .

[27]  Hualou Liang,et al.  Short-window spectral analysis of cortical event-related potentials by adaptive multivariate autoregressive modeling: data preprocessing, model validation, and variability assessment , 2000, Biological Cybernetics.

[28]  A Matsuki,et al.  Hypnotic endpoints vs. the bispectral index, 95% spectral edge frequency and median frequency during propofol infusion with or without fentanyl. , 1999, European journal of anaesthesiology.

[29]  Mingzhou Ding,et al.  Evaluating causal relations in neural systems: Granger causality, directed transfer function and statistical assessment of significance , 2001, Biological Cybernetics.

[30]  Hans-Ulrich Dodt,et al.  Anaesthesia Monitoring by Recurrence Quantification Analysis of EEG Data , 2010, PloS one.

[31]  M. Roizen,et al.  Use of the BIS Monitor to Detect Onset of Naturally Occurring Sleep , 2004, Journal of Clinical Monitoring and Computing.

[32]  J. Geweke,et al.  Measures of Conditional Linear Dependence and Feedback between Time Series , 1984 .

[33]  C Prys-Roberts,et al.  Anaesthesia: a practical or impractical construct? , 1987, British Journal of Anaesthesia.

[34]  Jesse Gilbert,et al.  Analysis: Theory and Practice , 2013 .

[35]  A. Seth,et al.  Behaviour of Granger causality under filtering: Theoretical invariance and practical application , 2011, Journal of Neuroscience Methods.

[36]  Steven L. Bressler,et al.  Wiener–Granger Causality: A well established methodology , 2011, NeuroImage.

[37]  Gordon Pipa,et al.  EEG under anesthesia - Feature extraction with TESPAR , 2009, Comput. Methods Programs Biomed..

[38]  A. Yli-Hankala,et al.  Description of the Entropy™ algorithm as applied in the Datex‐Ohmeda S/5™ Entropy Module , 2004, Acta anaesthesiologica Scandinavica.

[39]  G. Tononi,et al.  Breakdown in cortical effective connectivity during midazolam-induced loss of consciousness , 2010, Proceedings of the National Academy of Sciences.

[40]  E. Kochs,et al.  Detection of awareness in surgical patients with EEG-based indices--bispectral index and patient state index. , 2003, British journal of anaesthesia.

[41]  Stanislas Dehaene,et al.  The Global Neuronal Workspace Model of Conscious Access: From Neuronal Architectures to Clinical Applications , 2011 .

[42]  Seunghwan Kim,et al.  Characterization of Phase Transition in the Thalamocortical System during Anesthesia-Induced Loss of Consciousness , 2012, PloS one.

[43]  G. Tononi.  Consciousness as Integrated Information: a Provisional Manifesto , 2008, The Biological Bulletin.

[44]  L. Baccalá,et al.  Overcoming the limitations of correlation analysis for many simultaneously processed neural structures. , 2001, Progress in brain research.

[45]  F. L. D. Silva,et al.  EEG analysis: Theory and Practice , 1998 .

[46]  Anthony G. Hudetz,et al.  Cortical Disintegration Mechanism of Anesthetic-Induced Unconsciousness , 2009 .

[47]  Sean L. Hill,et al.  The Sleep Slow Oscillation as a Traveling Wave , 2004, The Journal of Neuroscience.

[48]  S. Bressler,et al.  Granger Causality: Basic Theory and Application to Neuroscience , 2006, q-bio/0608035.

[49]  Andreas Ziehe,et al.  Constructing Surrogate Data to Control for Artifacts of Volume Conduction for Functional Connectivity Measures , 2010 .

[50]  Q. C. Meng,et al.  A Conserved Behavioral State Barrier Impedes Transitions between Anesthetic-Induced Unconsciousness and Wakefulness: Evidence for Neural Inertia , 2010, PloS one.

[51]  J. Bruhn,et al.  Approximate Entropy as an Electroencephalographic Measure of Anesthetic Drug Effect during Desflurane Anesthesia , 2000, Anesthesiology.

[52]  Julius Georgiou,et al.  EEG-Based Automatic Classification of ‘Awake’ versus ‘Anesthetized’ State in General Anesthesia Using Granger Causality , 2012, PloS one.

[53]  Gudrun Stockmanns,et al.  Electroencephalographic Order Pattern Analysis for the Separation of Consciousness and Unconsciousness: An Analysis of Approximate Entropy, Permutation Entropy, Recurrence Rate, and Phase Coupling of Order Recurrence Plots , 2008, Anesthesiology.

[54]  J. Bruhn,et al.  Depth of anaesthesia monitoring: what's available, what's validated and what's next? , 2006, British journal of anaesthesia.

[55]  Thomas Standl,et al.  Comparative Evaluation of the Datex-Ohmeda S/5 Entropy Module and the Bispectral Index® Monitor during Propofol–Remifentanil Anesthesia , 2004, Anesthesiology.

[56]  L. Voss,et al.  Using Permutation Entropy to Measure the Electroencephalographic Effects of Sevoflurane , 2008, Anesthesiology.

[57]  I. Rampil.  A Primer for EEG Signal Processing in Anesthesia , 1998, Anesthesiology.

[58]  M. Carskadon,et al.  Chapter 2 - Normal Human Sleep : An Overview , 2005 .

[59]  UnCheol Lee,et al.  The directionality and functional organization of frontoparietal connectivity during consciousness and anesthesia in humans , 2009, Consciousness and Cognition.

[60]  M. Boly,et al.  Granger Causality Analysis of Steady-State Electroencephalographic Signals during Propofol-Induced Anaesthesia , 2012, PloS one.

[61]  Nassib G. Chamoun,et al.  An introduction to bispectral analysis for the electroencephalogram , 1994, Journal of Clinical Monitoring.

[62]  T. Seppänen,et al.  Time–frequency properties of electroencephalogram during induction of anesthesia , 2008, Neuroscience Letters.

[63]  C. Granger.  Investigating causal relations by econometric models and cross-spectral methods , 1969 .

[64]  S. Sengupta.  Introduction to Applied Statistical Signal Analysis , 1991 .

[65]  Julius Georgiou,et al.  Using Granger Causality to Characterise Bidirectional Interactions in the Human Brain during Induction of Anaesthesia , 2011, BIOSIGNALS.

[66]  J. Sleigh,et al.  Modelling general anaesthesia as a first-order phase transition in the cortex. , 2004, Progress in biophysics and molecular biology.

[67]  Erik W. Jensen,et al.  EEG complexity as a measure of depth of anesthesia for patients , 2001, IEEE Trans. Biomed. Eng..

[68]  J. Jakobsson,et al.  Nitrous oxide does not alter bispectral index: study with nitrous oxide as sole agent and as an adjunct to i.v. anaesthesia. , 1999, British journal of anaesthesia.

[69]  Olga A Imas,et al.  Volatile Anesthetics Enhance Flash-induced &ggr; Oscillations in Rat Visual Cortex , 2005, Anesthesiology.

[70]  UnCheol Lee,et al.  Preferential Inhibition of Frontal-to-Parietal Feedback Connectivity Is a Neurophysiologic Correlate of General Anesthesia in Surgical Patients , 2011, PloS one.

[71]  J. Sleigh,et al.  Analysis of depth of anesthesia with Hilbert–Huang spectral entropy , 2008, Clinical Neurophysiology.