prof. olga sourina - imiimi.ntu.edu.sg/.../documents/wang_qiang_2011.pdf · 2017-08-11 ·...
TRANSCRIPT
Wang QiangElectrical & Electronic Engineering, NTUInstitute for Media Innovation, NTU
Prof. Olga Sourina
Conclusion
Proposed Research Method
Background & Motivation
EEG
Brain Maps Frontal LobeArousalreaction to self and environmentExecutive functioning and judgmentsEmotional response and stabilityLanguage usage
Parietal LobeVirtual and touch perceptionObject Manipulation
Occipital LobeVirtual perception
Temporal LobeAuditory perception
BCI application in Game
•P300
•Motor Image
•Frequency bands power
•SSVEP
Neurofeedback
Desktop
ComputerPlayerEmotiv
EEG Device
Wireless
Communication
Psychological Disorders Attention Deficit Hyperactivity Disorder (ADHD) [1]
Patients have problem in attention and focus.
Autistic spectrum Disorders (ASD) [2]Patients have problem in social interaction and communication
Substance Use Disorders (SUD)[3]Abuse of alcohol or drug
[1] T. Fuchs, N. Birbaumer, W. Lutzenberger, J. H. Gruzelier and J. Kaiser, "Neurofeedback treatment for attention-deficit/hyperactivity disorder in children: A comparison with methylphenidate," Applied Psychophysiology Biofeedback, vol. 28, pp. 1-12, 2003.
[2] R. Coben, M. Linden and T. E. Myers, "Neurofeedback for autistic spectrum disorder: A review of the literature," Applied Psychophysiology Biofeedback, vol. 35, pp. 83-105, 2010.
[3] T. M. Sokhadze, R. L. Cannon and D. L. Trudeau, "EEG biofeedback as a treatment for substance use disorders: Review, rating of efficacy, and recommendations for further research," Applied Psychophysiology Biofeedback, vol. 33, pp. 1-28, 2008.
One Channel EEG
Algorithms
Application Fields
Attention Level Recognition Stage Researching nonlinear EEG features for attention level recognition to
replace the linear EEG feature.
Neurofeedback Implementation Stage Implement 2D and 3D neurofeedback games based on previous findings
in EEG feature.
Real-world Application Stage Expanding the application field from psychological disorder treatment to
normal personal applications.
Conclusion
Proposed Research Method
Background & Motivation
Fractal Dimension Model The fractal dimension is a measurement of complexity. The changes of pattern
embedded in signal can be noticed by changes in FD value. The higher fractal dimension is, the more the structure observed proves “irregular”,
conversely a low dimension will characterize a more regular structure. In 2-D space the FD should be limited in this region 1 <= FD <= 2.
Box-counting Method: In box-counting method[1], the fractal dimension value is evaluated from the time-
amplitude space directly by counting the normalized boxes occupied by the signals. N(d) is proportional to d-FD and the best line fitting method was also used to calculate the fractal dimension value after counting the box.
[1] A. Block, W. Von Bloh and H. J. Schellnhuber, "Efficient box-counting determination of generalized fractal dimensions," Physical Review A, vol. 42, pp. 1869-1874, 1990.
Higuchi Method: In Higuchi Method[1], m-dimension phase space is reconstructed by embedding
time-delay information, i.e. construct poly-phase subsequences from the recording signals . The capacity number occupying the space is calculated according the length over the subsequences. L(k) is proportional to k-FD. Fractal dimension value can be calculated by the least-square linear best fitting line procedure.
[1] T. Higuchi, "Approach to an irregular time series on the basis of the fractal theory," Physica D: Nonlinear Phenomena, vol. 31, pp. 277-283, 1988.
Comparison: Computational Cost:
Box-counting method is quite faster than Higuchi method.
Accuracy:Higuchi method is more accurate than Box-counting method.
(b)
Experiment Setting up: Hardware: Emotiv EEG device. (128 Hz sampling frequency, 16-bit
resolution)
Electrodes: O1 position according to the 10-20 international system. (Occipital Lobe)
Subjects: 5 subjects aged from 22 – 30.
Brain state: Relax(Distraction State) vs. concentration (working math problem).
.
ROC[1] curve illustration:.
[1] T. Fawcett, "An introduction to ROC analysis," Pattern Recognition Letters, vol. 27, pp. 861-874, 2006.
Roc Curves for 5 subjects in Concentration Experiments
Comparison for 4 methods in brain state recognition
[1] Lutsyuk,N.V., Eismont,E.V., Pavlenko,V.B, " Modulation of attention in healthy children using a course of eeg-feedback sessions," Neurophysiology,vol. 38(5-6), pp. 389-395, 2006.
[2] Pop-Jordanov,J., Pop-Jordanova,N, " Neurophysical substrates of arousal and attention," .CognitiveProcessing. 10(1) , pp. 71-79, 2009.
Higuchi Box-counting
Brain Rate [1]
θ/β ratio [2]
Accuracy mean 0.8811 0.8654 0.8493 0.8026
var 0.0077 0.0140 0.0078 0.0190
Best Threshold mean 1.9331 1.6048 15.8071 1.6701
var 0.0025 0.0022 35.05 0.2557
Time Consuming mean 0.0172 0.0050 0.0035 0.0043
Game Structure
Objective:
•Immersive 2D or 3D environment
for feedback and therapy
•Recognition of the concentration
state from EEG signals
•Generate visual feedback
according to the brain state.
Neurofeedback Game
Conclusion and future work
Proposed Research Method
Background & Motivation
In this research, we studied fractal dimension model and implemented the neurofeedback algorithm based on Higuchi method and box-counting method.
Fractal dimension model was applied in brain states (concentration and relax) recognition and been proven better than other EEG linear features.
We proposed and implemented the original neurofeedback 2D game “Brain Chi”, “Pipe” and 3D game “Dance Robot”, “Escape” based on fractal dimension model that could be used for entertainment.
Implement more nonlinear EEG features like dynamic analysis and compare the results with fractal dimension model.
Extent the fractal dimension model to Renyi Entropy based fractal dimension model. Proposed our own method to calculate the FD value.
Combine the fractal dimension model with the classical Neurofeedbackalgorithm. Employ the fractal dimension model in each EEG frequency band individually.
Apply the neurofeedback game based on fractal dimension model in pain management.
Construct database based for reference on fractal dimension.
[1] Q. Wang, O. Sourina and M. K. Nguyen, "Fractal Dimension Based Algorithm for Neurofeedback Games," in Proc. CGI 2010. SP25, Singapore, 2010.
[2] Q. Wang, O. Sourina and M. K. Nguyen, "EEG-based "Serious" Games Design for Medical Applications," in Proc. 2010 Int. Conf. on Cyberworlds, Singapore, 2010, pp. 270-276.
[3] O. Sourina, Q. Wang and M. K. Nguyen, "EEG-based "Serious" Games and Monitoring Tools for Pain Management," in Proc. Medicine Meets Virtual Reality 18, Newport Beach, California, USA, In Press, 2011.
[4] O. Sourina, Q. Wang, Y. Liu and M. K. Nguyen, "A Real-time Fractal-based Brain State Recognition from EEG and its Applications," in Proc. Biosignals, Rome, Italy, In Press, 2011.
[5] Q. Wang, O. Sourina and M. K. Nguyen, "Fractal Dimension Based Neurofeedback," Visual Computer, Accepted, 2011.
[6] K. Kith, O. Sourina and Q. Wang, "Fractal-based Algorithm for Ocular Artifacts Detection in the EEG without an EOG Channel," Computers in Biology and Medicine, Under Review, 2011.