Subject C T6
Figure 4.8: AEP comparison between beamforming and dipole-fitting results. We show subject C with experiments for AEP. A positive peak is at 94 ms, where stimulus on-set time is 0 ms. The dipole-fitting results is roughly drawn on the MRI figures; where we can see the difference between beamforming and dipole-fitting results.
Chapter 5
Conclusions
We have developed an efficient and accurate EEG source estimation technique. Ac-cording to our experiments using phantom data, the localization error using our method is very close to that using the most realistic BEM forward model and the MUSIC inverse estimation. But our method is more efficient because there is no need to calculate the time-consuming BEM model. In Chapter 2, we mentioned that the OS method proposed by Mosher et al. uses the BEM forward model and requires a set of significant dipole sources to estimate the OS. In the proposed method we eliminated the primary current term because it is irrelevant to the shell geometry. Also, only the inner skull surface mesh is required to estimate the OS, such that estimation can be simplified very much. Furthermore, the in-ner skull surface mesh can be approximated by using the EEG sensor location. Thus the proposed method is convenient, particularly when the MRI of the subject is not available to be used for surface extraction.
Using Equation 2.15 to calculate inner skull surface point normal nre(i) will depend on the triangle superficial measures surrounded with the target mesh point. In fact, inner skull surface normals is assumed to be equal to scalp surface normals which can be estimated from EEG sensor locations. When EEG sensor distribution is unbalanced, the calculated normals will not be precise. We can adopting “Gaussian curvature” to estimate the surface curvature including locations and normals of surface points. By using Gaussian curvature, accuracy of nre(i) will be improve, but it costs more time than just using Equation 2.15. A trade-off between using Equation 2.15 and Gaussian curvature will be judged in following works.
In Section 4.1.2, we have compared source localization error with Berg method and Sun method to determine which method is better. Localization result shows that Sun method has smaller error than Berg method. And, Sun method uses a closed-form solution rather than a non-linear search formula used in Berg method to generate the pre-processing parameters [25, 29]. Thus Sun method is more efficient and stable than Berg method obviously. In short, whether we mention to accuracy or efficiency, Sun method performs better than Berg method does.
Bibliography
[1] P. Berg and M. Scherg. A fast method for forward computation of multiple-shell spherical head models. Electroencephalography and Clinical Neurophysiology, 90, 1994.
[2] D. A. Brody, F. H. Terry, and R. E. Ideker. Eccentric dipole in a spherical medium:
Generalized expression for surface potentials. IEEE Transactions on Biomedical En-gineering, 20:141–143, 1973.
[3] D. K. Cheng. Field and Wave Electromagnetics. Addison Wesley, 2 edition, 1989.
[4] Edwin K. P. Chong and Stanislaw H. Zak. An Introduction to Optimization. John Wiley & Sons, 2 edition, 2001.
[5] B. N. Cuffin and D. Cohen. Comparison of the magnetoencephalogram and electroen-cephalogram. Electroencephalography and Clinical Neurophysiology, 47:132–146, 1979.
[6] A. Delorme and S. Makeig. EEGLAB: An open source toolbox for analysis of signal-trial EEG dynamics. Journal of Neuroscience Methods, 134:9–21, 2004.
[7] J. J. Ermer, J. C. Mosher, S. Baillet, and R. M. Leahy. Rapidly recomputable EEG forward models for realistic head shapes. Physics in Medicine and Biology, 46:1265–
1281, 2001.
[8] D. B. Geselowitz. On bioelectric potentials in an inhomogeneous volume conductor.
Biophysical Journal, 7:1–11, 1967.
[9] S. Ghahramani. Fundamentals of Probability. Prentice Hall, 3 edition, 2000.
[10] M. Hamalainen and R. Hari. Brain Mapping. Brain Research Unit, Low Temperature Laboratory, Helsinki University of Technology, 2 edition, 2002.
[11] M. Hamalainen, R. Hari, J. Ilmoniemi, J. Knuutila, and O. V. Lounasmaa.
Magnetoencephalography-theory, instrumentation, and applications to noninvasive studies of the working human brain. Review of Modern Physics, 65:413–497, 1993.
[12] M. S. Hamalainen and J. Sarvas. Realistic conductivity geometry model of the human head for interpretation of neuromagnetic data. IEEE Transactions on Biomedical Engineering, 36:165–171, 1989.
[13] M. X. Huang, J. C. Mosher, and R. M. Leahy. A sensor-weighted overlapping-sphere head model and exhaustive head model comparison for MEG. Physics in Medicine and Biology, 44:423–440, 1999.
[14] C. Humphries, S. Enghoff, Tzyy-Ping Jung, Te-Won Lee, and T. Bell. EEGLAB, http://sccn.ucsd.edu/eeglab/. Institute for Neural Computation at the University of California San Diego, 2002.
[15] A. Hyvarinen, J. Karhunen, and E. Oja. Independent Component Analysis. John Wiley & Sons, 2001.
[16] D. C. Lay. Linear Algebra and Its Applications. Addison Wesley, 2 edition, 1998.
[17] R. M. Leahy, J. C. Mosher, M. E. Spencer, M. X. Huang, and J. D. Lewine. A study of dipole localization accuracy for MEG and EEG using a human skull phantom.
Electroencephalography and Clinical Neurophysiology, 107, 1998.
[18] H. Y. Liu, Y. S. Chen, L. F. Chen, and J. C. Hsieh. Statistical mapping of cortical activities using minimum-variance maximum-discrimination spatial filtering. In The Joint Meeting of the 5th International Conference on Bioelectromagnetism and the 5th International Symposium on Noninvasive Functional Source Imaging within the Human Brain and Heart, Minneapolis, Minnesota, USA, May 2005.
BIBLIOGRAPHY 49
[19] N. Lynnerup. Cranial thickness in relation to age, sex and general body build in a danish forensic sample. Forensic Science International, 177:45–51, 2001.
[20] J. C. Mosher, R. M. Leahy, and P. S. Lewis. EEG and MEG: Forward solutions for inverse methods. IEEE Transactions on Biomedical Engineering, 46:245–259, 1999.
[21] R. D. Pascual-Marqui, C. M. Michel, and D. Lehmann. Low resolution electromag-netic tomography: A new method for localizing electrical activity in the brain. Inter-national Journal of Psychophysiology, 18:49–65, 1994.
[22] W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery. Numerical Recipes in C. Cambridge University Press, 1989.
[23] S. E. Robison and J. Vrba. Functional neuroimaging by synthetic aperture magne-tometry (SAM). Technical Report, CTF Systems Inc., Port Coquitlam, BC, Canada, 1998.
[24] Y. Salu, L. G. Cohen, D. Rose, S. Sato, C. Kufta, and M. Hallett. An improved method for localizing electric brain dipoles. IEEE Transactions on Biomedical Engineering, 37:699–705, 1994.
[25] Mingui Sun. An efficient algorithm for computing multishell spherical volume con-ductor models in EEG dipole source localization. IEEE Transactions on Biomedical Engineering, 44:1243–1252, 1997.
[26] P. C. Teo, G. Sapiro, and B. A. Wandell. Creating connected representations of cor-tical gray matter for functional MRI visualization. IEEE Transactions on Medical Imaging, 16:852–863, 1997.
[27] B. D. Van Veen, W. V. Drongelen, M. Yuchtman, and Akifumi Suzuki. Localization of brain electrical actvity via linearly constrained minimum variance spatial filtering.
IEEE Transactions on Biomedical Engineering, 44:867–880, 1997.
[28] J. Vrba and S. E. Robison. Differences between synthetic aperture magnetometry (SAM) and linear Beamformers. In Biomag 2000, 12th International Conference on Biomagnetism, Helsinki University of Technology, Espoo, Finland, 2000.
[29] Zhi Zhang. A fast method to compute surface potentials generated by dipoles within multilayer anisotropic spheres. Physics in Medicine and Biology, 40, 1994.