Basis of the EEG Signal Juho ijl 1
Basis of the EEG Signal Juho Äijälä 1
Summary Intro to EEG Basics of the EEG-signal EEG frequency spectrum …brief intro to ERP’s 2
Intro to EEG § Electroencephalogram Electric Picture Brain § Electrodes on the scalp measure electrical activity generated by thousands of synchronised neurons § Direct non-invasive measure of neuronal activity! § Really good temporal resolution: sampling rates of 1024 hz – 4096 hz with modern systems https: //www. brightbraincentre. co. uk/electroencephalogram-eeg-brainwaves/ 3
Basis of the EEG-signal Some physics… • • • Electric potentials generated by neurons can be modeled with dipoles Dipole: a separation of electrical charges. Quantified by dipole moment (μ) Electric current flows from the negative pole to the positive Happens in neurons all the time: action potentials Primary vs. secondary current However, single current is too small to measure… https: //en. wikipedia. org/wiki/Dipole Becker (2014) 4
Basis of the EEG-signal Some neuroanatomy… • • • Neocortex consists of six distinct layers. Distinct (messy!) neuronal organisation and connections across layers. Luckily giant pyramidal cells projecting from layer 5 are lined perpendicular to the surface! https: //www. brightbraincentre. co. uk/electroe ncephalogram-eeg-brainwaves/ Kandell, Schwartz and Jessell (2000) 5
Basis of the EEG-signal Some physics + neuroanatomy… • • • How does the organisation of the pyramidal neurons help us? As noted, one dipole generated by one action potential is too small to measure… …the summation of tens of thousands is not Since pyramidal neurons point to the same direction, charges don’t cancel out We can measure the summed diapoles! Jackson and Bolger, (2014) 6
Basis of the EEG-signal Pre- vs. post-synaptic potentials • Pre and post-synaptic potentials differ in characteristics • Pre-synaptic: Short and biphasic • Post-synaptic: Longer and monophasic 7
Basis of the EEG-signal So we are measuring: • • • The summed dipoles generated (mostly) by the sychnorised post-synaptic potentials of tens of thousands of pyramidal neurons in Layer 5 … plus noise What does this correspond to? Can we localise the source? Adjamian (2014) 8 https: //www. brightbraincentre. co. uk/electroencephalogram-eeg-brainwaves/eeg-dipol
Basis of the EEG-signal How do we measure it? • • • Electrodes placed on scalp, standardised placement: 10 -20 system (More recent 10 -5) EEG uses differential amplifiers to produce each channel The way the electrodes are connected to the amplifiers are referred to as a montage Ueda, Sakai and Yanagisawa, (2019) Modir (2017) 9
Basis of the EEG-signal Standard recording derivations • • • Common reference derivation: a reference electrode is substracted from the scalp electrode. The same reference electrode is used for every amplifier Average reference derivation: Activity from all electrodes is summed, averaged and passed through a high value resistor. The resulting signal is used as the ’reference electrode’ Bipolar derivation: electrodes are sequentially linked together. E. g. from the back to the front. 10 https: //www. ebme. co. uk/articles/clinical-engineering/introduction-to-eeg
So what do we get from recording EEG? EEG frequency spectrum as a classification system § § Beta: Seen in a symmetrical distribution on both sides. Dominant when alert/anxious/eyes are open Alpha: Seen in posterior regions. Higher amplitude on the dominant hemisphere. Appears with relaxing/closed eyes Theta: ’slow activity’. Seen in sleep and children under 13 years old Delta: Lowest frequency/highest amplitude. Appears in stages 3 and 4 of sleep. https: //raphaelvallat. com/bandpower. html 11
Some applications § Sleeping disorders (Friedman, 1986) § Main tool for diagnosing epilepsy. Current research is looking at automated ways using machine learning. (Tiwari et al. , 2017) § Brain-Computer interfaces (Spüler, 2017) https: //www. britannica. com/science/electroencephalography https: //emedicine. medscape. com/article/1138154 -overview 12
ERP’s (briefly) § ERP = Event related potential § An EEG waveform associated with a certain action or mental event § Remember that EEG-data is noisy! § How can we examine small waveforms associated with specific events? § By a lot of repetition: random noise should cancel itself out, but systematic variance should remain! https: //medium. com/@mindpass 2050/the-stimulus-reaction-challenged 86 cd 57 e 22 fe 13
Basis of the MEG Signal Mercede Erfanian 14
Overview MEG basics EEG vs. MEG Advantages & Disadvantages Summary 15
MEG: introduction § Electroencephalogram (EEG) electrodes § Scalp recording of electrical activity of cortex => waveform signals § Microvolts (µV) – small! § Role of EEG in neuroimaging: § Identify neural correlates § Diagnose epilepsy, sleep disorders, anaesthesia, coma, brain death § Magnetoencephalography § Direct external recordings of magnetic fields created by electrical currents in cortex § Measured in f. T to p. T § Role of MEG in neuroimaging: § Neural correlates of cognitive/perceptual processes § Localise affected regions before surgery(? ), determine regional and network functionality http: //www. admin. ox. ac. uk/estates/capitalprojects/previouscapitalprojects/megsca nner/ 16
MEG: basis of the signal § EEG and MEG both measure the neuronal activities but EEG detects synchronised electrical activity of large groups of neurons, whereas MEG detects the tiny changes in magnetic fields § Recall: large pyramidal neurons in layer V of cortex, arranged in parallel, similarlyoriented, perpendicular to surface, fire synchronously Tiege & Zlobinski, 2006 § Dipolar current flow generates a magnetic field. TRY IT: ‘Right hand grip’! § 10, 000 to 50, 000 active neurons required for detectable signal http: //www. youtube. com/wat ch? v=CPj 4 j. JACe. Is Ochi et al. 2011 § Scalp topography: - Influx maxima ‘source’ - Efflux maxima ‘sink’ 17
MEG: tangential vs. radial § MEG magnetic field not distorted by conductive properties of scalp/head radial § MEG coil not sensitive to perfectly radial sources tangential MEG pick-up coils § But in practice, only a small proportion (<1%) of cell populations are perfectly radial – i. e. on top of gyri Tiege and Zlobinski, 2006 18
MEG: scale of magnetic field § MEG signal is tiny! Interference from heartbeat! § Interference from electrical equipment, traffic, the earth, participant’s heartbeat etc. § Requires magnetically shield rooms and supersensitive magnetometers 19
MEG: magnetically shielded room (MSR) Brock & Sowman (2014) § 3, 5 or 6 layers with different magnetic properties to protect from different frequencies of magnetic interference 20
MEG is super-cool § SQUID § Superconducting QUantum Interference Device, immersed in super-cool liquid helium § Sensitive to field changes in order of femto-Tesla (10 -15) § Superconductive ring with two Josephson junctions § Flux transformers (coils) - Magnetometers Gradiometers (planar/axial) 21 http: //www. csiro. au/~/media/CSIROau/Images/Maps%20%20 Graphs/SQUID_CESRE_ind/High_Resolution. gif
MEG: flux transformers scalp Axial magnetometer Axial/planar gradiometers (1 st order) Single superconducting coil – highly sensitive but affected by environmental noise Two oppositely-wound coils – environmental noise affects both electrodes : no net noise. Sources from cortex affect coils differentially http: //www. youtube. com/watch? v=CPj 4 j. JACe. Is 22
MEG: applications § Excellent spatial resolution good for functional mapping of specific cortex (M 1, V 1) during behavioural, cognitive, perceptive tasks § Surgical planning (? ) in patients with brain tumours or intractable epilepsy § Research into whole-brain network connectivity Millisecond temporal resolution 23 de Pasquale et al (2010)
EEG vs. MEG EEGMEG EEG Signal magnitude 10 m. V (easily detectable) EEG Measurement Secondary currents Signal purity Distortion by skull/scalp 10 f. T (magnetic shielding required) Primary currents Little effect by skull/scalp Temporal resolution ~1 ms Spatial resolution ~1 cm <1 cm Experimental flexibility Moves with subject Dipole orientation Tangential and radial Subject must remain stationary Tangential better 24
EEG/MEG advantages Non-invasive Direct measurements of neuronal function (unlike f. MRI) High temporal resolution (1 ms or less, 1000 x better than f. MRI) Easy to use clinically (adults, children) Quiet! (can study auditory processing) Affordable, EEG is portable Subjects can perform tasks sitting up (more natural than MRI scanner) https: //www. colbertnewshub. com/2013/04/05/april-4 -2013 -dr-francis-collins/ https: //medicalxpress. com/news/2015 -02 -brain-imaging-links-language-chromosome. html 25
EEG/MEG disadvantages Not as good spatial localisation as f. MRI, CT Sensitivity depth only ~4 cm (c. f. whole brain sensitivity of f. MRI) - Sensitivity loss proportional to square of distance from sensor 3 D Source reconstruction is ill-posed? forward and inverse problems https: //ngp. usc. edu/files/2013/06/Syed__EEG_MEG. pdf 26
Forward & inverse problems Forward modelling: easy! Neuronal activity/ Current density EEG/MEG Sensor data https: //www. youtube. com/watc h? v=Aog. BOXt. Xk 1 s SOLUTION: Use forward models for inverse problem. Source localisation models and algorithms; iterative source reconstruction 27
Summary § Direct, non-invasive measures of cortical electrical activity EEG: secondary currents, MEG: magnetic fields § Good spatial & temporal resolution § Depth sensitivity? Add thalamus, hippocampus, amygdala to MEG source reconstruction models (!) § Spontaneous or evoked neural activity; § Applications in epilepsy, sleep, Alzheimer’s disease biomarkers(? ), schizophrenia(? ), autism(? ), whole-brain functional networks 28
Thank you! 29
Sources: EEG • Adjamian, P. (2014). The application of electro-and magneto-encephalography in tinnitus research– methods and interpretations. Frontiers in neurology, 5, 228. • Freedman, R. R. (1986). EEG power spectra in sleep-onset insomnia. Electroencephalography and clinical neurophysiology, 63(5), 408 -413. • Jackson, A. F. , & Bolger, D. J. (2014). The neurophysiological bases of EEG and EEG measurement: A review for the rest of us. Psychophysiology, 51(11), 1061 -1071. • Jurcak, V. , Tsuzuki, D. , & Dan, I. (2007). 10/20, 10/10, and 10/5 systems revisited: their validity as relative head-surface-based positioning systems. Neuroimage, 34(4), 1600 -1611. • Modir, Aslan. (2017). NIRS signal evaluation in order to epileptic seizure detection. 10. 13140/RG. 2. 2. 23784. 32007. • Spüler, M. (2017). A high-speed brain-computer interface (BCI) using dry EEG electrodes. Plo. S one, 12(2). • Tiwari, A. K. , Pachori, R. B. , Kanhangad, V. , & Panigrahi, B. K. (2016). Automated diagnosis of epilepsy using key-point-based local binary pattern of EEG signals. IEEE journal of biomedical and health informatics, 21(4), 888 -896. • Thomas, Jessell, Siegelbaum, S. , & Hudspeth, A. J. (2000). Principles of neural science (Vol. 4, pp. 1227 -1246). E. R. Kandel, J. H. Schwartz, & T. M. Jessell (Eds. ). New York: Mc. Graw-hill. • Ueda, K. , Sakai, Y. , & Yanagisawa, H. (2019). Quantitative evaluation of sense of discrepancy to operation response using event-related potential. ar. Xiv preprint ar. Xiv: 1907. 01827. 30
Sources: MEG Brock J and Sowman P (2014) Meg for Kids: Listening to Your Brain with Super-Cool SQUIDs. Frontiers for Young Minds. 2(10) de Pasquale, F. , Della Penna, S. , Snyder, A. Z. , Lewis, C. , Mantini, D. , Marzetti, L. , … Corbetta, M. (2010). Temporal dynamics of spontaneous MEG activity in brain networks. Proceedings of the National Academy of Sciences , 107(13), 6040– 6045. da Silva, F. L. , (2013). EEG and MEG: Relevance to Neuroscience. Neuron 80(1), 1112– 1128. de Tiege, X. , and Zlobinski, I. (2006). What do we measure with EEG and MEG? . Unpublished manuscript, Institute of Neurology, University College London, United Kingdom. Retrieved from: http: //slideplayer. com/slide/6086213/ Kallara (2012) Biomedical Engineering Module-1 Unpublished teaching slides from: https: //www. slideshare. net/subkal/biomedical-engineering-mod 1 Kandel, E. R. , Schwartz, J. H. , & Jessell, T. M. (1991). Principles of Neural Science. Neurology Malmivuo, Jaakko & Plonsey, Robert. (1995). Bioelectromagnetism - Principles and Applications of Bioelectric and Biomagnetic Fields. Oxford University Press, NY Ochi, A. , Go, C. Y. , and Otsubo, H. , (2011). Clinical MEG Analyses for Children with Intractable Epilepsy, Magnetoencephalography, Dr. Elizabeth Pang (Ed. ), Smith, S. J. M. (2005). EEG in the diagnosis, classification, and management of patients with epilepsy. Journal of Neurology, Neurosurgery, and Psychiatry, 76 Suppl 2(suppl 2), ii 2 -7. (and Dr. Sofie Meyer) 31
Sources (cont) Images from: https: //en. wikipedia. org/wiki/Dipole https: //www. brightbraincentre. co. uk/electroencephalogram-eeg-brainwaves/eeg-dipoles https: //www. ebme. co. uk/articles/clinical-engineering/introduction-to-eeg https: //raphaelvallat. com/bandpower. html https: //www. britannica. com/science/electroencephalography https: //emedicine. medscape. com/article/1138154 -overview https: //medium. com/@mindpass 2050/the-stimulus-reaction-challenge-d 86 cd 57 e 22 fe http: //opencc. co. uk/blog/out-of-touch-manual-keypads-and-controllers-face-competition-from-new-hands-freecomputer-interfaces/ http: //www. gensat. org/imagenavigator. jsp? image. ID=29099 http: //www. lucid. ac. uk/news-and-events/blogs/how-to-study-language-why-do-we-put-electrodes-on-people-sheads/ http: //www. admin. ox. ac. uk/estates/capitalprojects/previouscapitalprojects/megscanner/ http: //www. youtube. com/watch? v=CPj 4 j. JACe. Is http: //www. csiro. au/~/media/CSIROau/Images/Maps%20%20 Graphs/SQUID_CESRE_ind/High_Resolution. gif https: //www. colbertnewshub. com/2013/04/05/april-4 -2013 -dr-francis-collins/ https: //medicalxpress. com/news/2015 -02 -brain-imaging-links-language-chromosome. html https: //www. youtube. com/watch? v=Aog. BOXt. Xk 1 s 32
- Slides: 32