Hybrid implementation of the fastICA algorithm for high-density EEG using the capabilities of the intel architecture and CUDA programming

• Autorzy: Gajos-Balińska Anna , Wójcik Grzegorz M. , Stpiczyński Przemysław

Identyfikatory

DOI

10.7494/csci.2023.24.4.5539

• Języki publikacji: EN

Abstrakty

EN

High-density electroencephalographic (EEG) systems are utilized in the study of the human brain and its underlying behaviors. However, working with EEG data requires a well-cleaned signal, which is often achieved through the use of independent component analysis (ICA) methods. The calculation time for these types of algorithms is the longer the more data we have. This article presents a hybrid implementation of the fastICA algorithm that uses parallel programming techniques (libraries and extensions of the Intel processors and CUDA programming), which results in a significant acceleration of execution time on selected architectures.

Słowa kluczowe

EN

ICA; EEG; BLAS; MKL; OpenMP; Intel Cilk Plus; CUDA

• Wydawca: Wydawnictwa AGH
• Czasopismo: Computer Science
• Rocznik: 2023
• Tom: T. 24 (4)
• Strony: 455--472
• Opis fizyczny: Bibliogr. 28 poz., rys., tab., wykr.

Twórcy

autor

Gajos-Balińska Anna

• Maria Curie-Sklodowska University, Department of Neuroinformatics and BiomedicalEngineering, Institute of Computer Science, ul. Akademicka 9, 20-033 Lublin, Poland

autor

Wójcik Grzegorz M.

• Maria Curie-Sklodowska University, Department of Neuroinformatics and BiomedicalEngineering, Institute of Computer Science, ul. Akademicka 9, 20-033 Lublin, Poland

autor

Stpiczyński Przemysław

• Maria Curie-Sklodowska University, Department of Information Systems Software,Institute of Computer Science, ul. Akademicka 9, 20-033 Lublin, Poland

Bibliografia

• [1] Albajes-Eizagirre A., Dubreuil Vall L., David I.S., Riera A., Soria-Frisch A.,Dunne S., Ruffini G.:EEG/ERP analysis: methods and applications, CRC Press,Boca Raton, 2014.
• [2] Brown G.D., Yamada S., Sejnowski T.J.: Independent component analysis atthe neural cocktail party,Trends in Neurosciences, vol. 24(1), pp. 54–63, 2001.doi: 10.1016/S0166-2236(00)01683-0.
• [3] Delorme A., Sejnowski T., Makeig S.: Enhanced detection of artifacts in EEGdata using higher-order statistics and independent component analysis,Neuroim-age, vol. 34(4), pp. 1443–1449, 2007. doi: 10.1016/j.neuroimage.2006.11.004.
• [4] Dickter C.L., Kieffaber P.D.:EEG methods for the psychological sciences, SAGEKnowledge, Los Angeles, 2014.
• [5] Duch W., Nowak W., Meller J., Osi ́nski G., Dobosz K., Miko lajewski D.,W ́ojcik G.M.:Computional approach to understanding autism spectrumdisorders,Computer Science, vol. 13(2), pp. 47–61, 2012.doi: 10.7494 /csci.2012.13.2.47.
• [6] Gajos A., W ́ojcik G.M.: Independent component analysis of EEG data for EGIsystem,Bio-Algorithms and Med-Systems, vol. 12(2), pp. 67–72, 2016.
• [7] Gajos-Bali ́nska A., W ́ojcik G.M., Stpiczy ́nski P.: Concept of independentcomponent analysis algorithm parallelisation. In: M. Bubak, M. Turala, K. Wiatr(eds.),Proceedings of Cracow Grid Workshop – CGW’15, pp. 55–56, 2015.
• [8] Gajos-Bali ́nska A., W ́ojcik G.M., Stpiczy ́nski P.: Parallel independent componentanalysis algorithm – performance comparison for EEG signal. In: M. Bubak,M. Turala, K. Wiatr (eds.),CGW Workshop ’17, Krak ́ow, Poland, October 23–25,2017: proceedings, pp. 33–34, 2017.
• [9] Gajos-Bali ́nska A., W ́ojcik G.M., Stpiczy ́nski P.: High performance optimizationof independent component analysis algorithm for EEG data,Lecture Notes inComputer Science, vol. 10777, pp. 495–504, 2018.
• [10] Gajos-Bali ́nska A., W ́ojcik G.M., Stpiczy ́nski P.: Cooperation of CUDA andIntel multi-core architecture in the independent component analysis algorithmfor EEG data,Bio-Algorithms and Med-Systems, vol. 16(3), 2020. doi: 10.1515/bams-2020-0044.
• [11] Hyv ̈arinen A.: Fast and robust fixed-point algorithms for independent componentanalysis,IEEE Transactions on Neural Networks, vol. 10(3), pp. 626–634, 1999.
• [12] Hyv ̈arinen A., Oja E.: Independent component analysis: algorithms and appli-cations,Neural Networks, vol. 13(4), pp. 411–430, 2000. doi: 10.1016/s0893-6080(00)00026-5.
• [13] Kawala-Janik A., Bauer W., Al-Bakri A., Haddix C., Yuvaraj R., Cichon K.,Podraza W.: Implementation of low-pass fractional filtering for the purpose ofanalysis of electroencephalographic signals. In:Conference on Non-integer OrderCalculus and Its Applications, pp. 63–73, Springer, 2017.
• [14] Lastovetsky A., Szustak L., Wyrzykowski R.: Model-based optimization ofEULAG kernel on Intel Xeon Phi through load imbalancing,IEEE Trans-actions on Parallel and Distributed Systems, vol. 28(3), pp. 787–797, 2016.doi: 10.1109/TPDS.2016.2599527.
• [15] Miko lajewska E., Miko lajewski D.: Integrated IT environment for people withdisabilities: a new concept,Open Medicine, vol. 9(1), pp. 177–182, 2014.doi: 10.2478/s11536-013-0254-6.
• [16] Miko lajewska E., Miko lajewski D.: The prospects of brain – computer interfaceapplications in children,Open Medicine, vol. 9(1), pp. 74–79, 2014. doi: 10.2478/s11536-013-0249-3.
• [17] NetStation acquisition. Technical manual. Documentation, EGI, 2011.
• [18] Rahman R.:Intel Xeon Phi coprocessor architecture and tools: the guide forapplication developers, Apress, Berkely, CA, USA, 2013.
• [19] Rojek I., Macko M., Mikolajewski D., S ́aga M., Burczynski T.: Modern methodsin the field of machine modelling and simulation as a research and practical issuerelated to Industry 4.0,The Bulletin of the Polish Academy of Sciences TechnicalSciences, vol. 69(2), 2021. doi: 10.24425/bpasts.2021.136717.
• [20] Szustak L.: Strategy for data-flow synchronizations in stencil parallel computa-tions on multi-/manycore systems,The Journal of Supercomputing, vol. 74(4),pp. 1534–1546, 2018.
• [21] Szustak L., Bratek P.: Performance portable parallel programming of heteroge-neous stencils across shared-memory platforms with modern Intel processors,TheInternational Journal of High Performance Computing Applications, vol. 33(3),pp. 534–553, 2019. doi: 10.1177/1094342019828153.
• [22] Tadeusiewicz R. (ed.):Neurocybernetyka teoretyczna, Wydawnictwa Uniwer-sytetu Warszawskiego, 2009.
• [23] Ungureanu M., Bigan C., Strungaru R., Lazarescu V.: Independent componentanalysis applied in biomedical signal processing,Measurement Science Review,vol. 4(2), p. 18, 2004.
• [24] Wojcik G.M., Masiak J., Kawiak A., Kwasniewicz L., Schneider P., Polak N.,Gajos-Balinska A.: Mapping the Human Brain in Frequency Band Analysis ofBrain Cortex Electroencephalographic Activity for Selected Psychiatric Disor-ders,Frontiers in Neuroinformatics, vol. 12, 2018. doi: 10.3389/fninf.2018.00073.
• [25] Wojcik G.M., Masiak J., Kawiak A., Kwasniewicz L., Schneider P., Postepski F.,Gajos-Balinska A.: Analysis of Decision-Making Process Using Methods of Quan-titative Electroencephalography and Machine Learning Tools,Frontiers in Neu-roinformatics, vol. 13, 2019. doi: 10.3389/fninf.2019.00073.
• [26] Wojcik G.M., Masiak J., Kawiak A., Schneider P., Kwasniewicz L., Polak N.,Gajos-Balinska A.: New Protocol for Quantitative Analysis of Brain Cortex Elec-troencephalographic Activity in Patients With Psychiatric Disorders,Frontiersin Neuroinformatics, vol. 12, 2018. doi: 10.3389/fninf.2018.00027.
• [27] W ́ojcik G.M.: Selected methods of quantitative analysis in electroencephalog-raphy. In: I. Roterman-Konieczna (ed.),Simulations in Medicine, pp. 35–54,De Gruyter, 2020. doi: 10.1515/9783110667219-003.
• [28] Wyrzykowski R., Szustak L., Rojek K.: Parallelization of 2D MPDATA EULAGalgorithm on hybrid architectures with GPU accelerators,Parallel Computing,vol. 40(8), pp. 425–447, 2014. doi: 10.1016/j.parco.2014.04.009.