student id 08069470
play

Student ID: 08069470 Supervised By: Prof. Wen J. LI Date: April 30 - PowerPoint PPT Presentation

M.Sc. Project for Biomedical Engineering By: SHAN Qing Student ID: 08069470 Supervised By: Prof. Wen J. LI Date: April 30 th , 2010 Outline Basic Knowledge on EEG Brain Waves EEG Measurement EEG Applications Existing EEG-Based


  1. M.Sc. Project for Biomedical Engineering By: SHAN Qing Student ID: 08069470 Supervised By: Prof. Wen J. LI Date: April 30 th , 2010

  2. Outline Basic Knowledge on EEG • Brain Waves • EEG Measurement • EEG Applications Existing EEG-Based Motion Control Technology • Overview • Case I: Using Mu and Beta Rhythm • Case II: Using Event-Related Evoked Potential • Conclusion How Can MEMS Technology Help • Problems of existing commercial electrodes • Case I: MEMS Based Silicon Spiked Electrode Array • Case II: Dry Electrode Using CNT Arrays • Summary on the Advantages of MEMS EEG Electrodes Future Potentials Conclusion Q & A April 30th, 2010 M.Sc. Project Presentation 2

  3. April 30th, 2010 M.Sc. Project Presentation 3

  4. Brain Waves Basics* • Amplitude ranging from 0.5 ~ 100µV (peak to peak) • Cerebral EEG falls in range of 1 ~ 20Hz • Categorized based on frequency • Beta : > 13Hz; Alpha: 8 ~ 13Hz; Theta: 4~8 Hz; Delta: 0.5 ~ 4Hz Beta Wave* • Dominant during normal state of wakefulness with open eyes • Closely linked to motor behavior Alpha Wave* • Induced by closing eyes and by relaxation Mu Rhythm* • Alpha-range Activity that is seen over the sensorimotor cortex • Characteristics relates to the movement of the contralateral arm or mental imagery of movement of the contralateral arm * Reference [1-3] April 30th, 2010 M.Sc. Project Presentation 4

  5. EEG Measurement The electroencephalogram (EEG) is defined as electrical activity of an alternating type recorded from the scalp surface after being picked up by metal electrodes and conductive media.* Types of Electrodes* • Disposable (pre-gelled, and gel-less) • Reusable Disc Electrodes (require conductive gel; usually made of Ag/Ag-Cl) • Saline Electrode • Headbands and Electrode Caps • Invasive Needle Electrodes Figure 1. EEG Measurement Setup* * Reference [1-3] April 30th, 2010 M.Sc. Project Presentation 5

  6. EEG Applications Clinical Applications* • Monitor alertness, coma, and brain death • Locate areas of damage after head injury, stroke, tumor, etc. • Investigate epilepsy and locate seizure origin • Etc., Video 1. MindFlex* Neuromarketing* ITEM PRICE EPOC USD 500 Force USD 60.54 Trainer MindFlex USD 79.99 * Reference [1][7][8] April 30th, 2010 M.Sc. Project Presentation 6

  7. April 30th, 2010 M.Sc. Project Presentation 7

  8. Overview Degraded signal Complexity of signal Extremely variable signal Motion control through recognizing the brain’s intentions from EEG is probably NOT possible in the foreseeable future*  Alternatives  Motion Control based on Mu and Beta Rhythm  Motion Control based on Event-Related Potentials * Reference [6][9] April 30th, 2010 M.Sc. Project Presentation 8

  9. Case I: Using Mu and Beta Rhythm This case demonstrated a noninvasive Brain Computer Interface (BCI) that can provide 2D motion control using beta and mu rhythms recorded from the scalp. The project is carried out in the Wadsworth Center at State University of New York.* Over 60 Years of Study • People can learn to control certain features of the EEG • Alpha Wave can be induced by relaxation (Working principle of Force Trainer and MindFlex) • Mu and Beta Rhythm wax and wane in association with actual movement or imagination of movement * Reference [4][10] April 30th, 2010 M.Sc. Project Presentation 9

  10. Case I: Using Mu and Beta Rhythm Figure 2. Study Protocol* R H R V : Right Side Amplitude L H L HV : Left Side Amplitude w RH w LH : Weight a H a V b H b V : Tunable Coefficient Figure 3. Topographical and Spectral Properties of User’s EEG Control* * Reference [10] April 30th, 2010 M.Sc. Project Presentation 10

  11. Case I: Using Mu and Beta Rhythm  4 Subjects Participated  Performance gradually improved over training sessions Video 2. Cursor Controlled by User* Figure 4. Cursor Trajectories of Users* * Reference [10] April 30th, 2010 M.Sc. Project Presentation 11

  12. Case II: Using Event-Related Potentials Project done by Universidad de Zaragoza, Spain. The team developed a non-invasive brain-actuated wheelchair that relies on P300 neurophysiological protocol.* What is Event-Related Potentials (ERP)? • Measured brain response that is directly the result of a thought or perception. * • Any stereotyped electrophysiological response to an internal or external stimulus.* • No need to provide motor or verbal response • Focusing on the change in electrophysiological signal that occurs immediately following the stimulus event * Reference [5][11] April 30th, 2010 M.Sc. Project Presentation 12

  13. Case II: Using Event-Related Potentials Training Process • Present to all possible stimulus • EEG signal after the stimulus will be recorded • Feature Vector will be extracted for each stimulus Real-time Recognition Process • Options flash one by one • EEG signal after each stimulus will be taken to compare with the feature vector of that stimulus type • Stimulus of the greatest probability will be regard as the choice Figure 5. Module Diagram of the Brain-Actuated Wheelchair* * Reference [11] April 30th, 2010 M.Sc. Project Presentation 13

  14. Case II: Using Event-Related Potentials Figure 6. Subject Navigating along Hallway by Concentrating on Graphic Interface on Screen* Experimental Result • The system was used and validated by five healthy subjects in three consecutive steps: screening, virtual environment driving and wheelchair driving sessions • All subjects accomplished two different tasks with relative easiness • Recognition accuracy of higher than 90% within one hour * Reference [11] April 30th, 2010 M.Sc. Project Presentation 14

  15. Conclusion Frequency Domain Analysis Amplitude of the EEG in a particular frequency band Commands (Rhythm) Time Domain Analysis The form or magnitude of the voltage changed evoked by a stereotyped Commands stimulus (Evoked Potential) April 30th, 2010 M.Sc. Project Presentation 15

  16. April 30th, 2010 M.Sc. Project Presentation 16

  17. Problems of Existing Commercial Electrodes Require conductive gel for adhesion and lowering electrode-skin interface impedance • Conductivity of gel gradually decrease due to the hardened of gel • Uncomfortable and inconvenient • Cause skin red and itchy feeling Require careful skin preparation before experiments April 30th, 2010 M.Sc. Project Presentation 17

  18. Case I: MEMS Based Silicon Spiked Electrode Array A novel MEMS EEG sensor for drowsiness detection application. Joint project between National Chiao-Tung University, Taiwan & University of California, San Diego.  Designed to pierce the stratum corneum (SC) into the electrically conducting tissue layer stratum germinativum (SG), but not reach the dermis layer so as to avoid pain or bleeding.* * Reference [13] April 30th, 2010 M.Sc. Project Presentation 18

  19. Case I: MEMS Based Silicon Spiked Electrode Array Figure 7. Fabrication Process* Figure 8. SEM of the Fabricated Dry Electrodes* Fabrication Process* • Fabricated on a silicon wafer with high aspect ratio • Thick PR film patterned with circular dots to provide etching hard mask • Isotropic etching to obtain probe tip • Anisotropic etching to obtain probe shaft • Wet etching to release hard mask at the probe tip • Coat Ti/Pt using DC sputtering • Mounted on flexible PCB using silver glue * Reference [13] April 30th, 2010 M.Sc. Project Presentation 19

  20. Case I: MEMS Based Silicon Spiked Electrode Array Figure 9. Positions of Wet and Dry Electrodes & Raw EEG Data Recorded* Results • The recorded signals by dry electrodes are extremely comparable to those obtained by corresponding wet electrodes Figure 10. The EEG power spectrum* * Reference [13] April 30th, 2010 M.Sc. Project Presentation 20

  21. Case II: Dry Electrode Using CNT Arrays A new dry electrode sensor for surface biopotential applications. The project is done by University of Barcelona, Spain and University of Surrey, UK.  This project team aimed at designing a dry electrophysiology sensor using CNT to eliminate skin preparation and gel application requirements in order to reduce noise while improving wearability.* * Reference [14] April 30th, 2010 M.Sc. Project Presentation 21

  22. Case II: Dry Electrode Using CNT Arrays Prototype • The MWCNT arrays were mounted on commercial active electrodes with onsite amplifiers and connect to commercial off-the-shelf Figure 11. Electrode Prototype Design* recording equipment • MWCNT arrays were grown on highly doped Si substrate using plasma-enhanced chemical vapor deposition (PECVD) of acetylene over an iron catalyst Figure 12. Electron microscope image of the MWCNT array* * Reference [14] April 30th, 2010 M.Sc. Project Presentation 22

  23. Case II: Dry Electrode Using CNT Arrays Test Results • Noise measured by the new electrodes is low and rather similar to that of the commercial electrodes • Human tests carried out for both spontaneous EEG recording and ERPs Figure 13. Conventional Wet Electrode (BIOSEMI) & Prototype Electrode (ENOBIO)* * Reference [14] April 30th, 2010 M.Sc. Project Presentation 23

  24. Advantages of MEMS EEG Electrodes No Longer needs Conductive Gel Allow smaller electrode size  Higher selectivity Decrease electrode-skin interface impedance April 30th, 2010 M.Sc. Project Presentation 24

  25. April 30th, 2010 M.Sc. Project Presentation 25

Recommend


More recommend