阿茲海默症(Alzheimer's disease, AD)是一種非正常老化的大腦疾病，患有此疾病的患者是無法被根治的。輕度認知障礙(Mild Cognitive Impairment, MCI)常被喻為失智症與健康者之間的過渡期，提早接受治療是有機會完全痊癒的，然而MCI是不容易被診斷的，為了讓醫師更容易診斷，本研究目的為應用機器學習方法以分辨AD、MCI與健康對照組(Health Control, HC)的二元分類器。首先，本研究透過留一參與者交叉驗證選取超參數，比較自我回歸(Autoregressive, AR)模型、線性與非線性的拉蓋多爾-沃爾泰拉自我回歸(Laguerre-Volterra Autoregressive, LVAR)模型三種特徵擷取方法在AD-HC、MCI-HC以及AD-MCI三種二元分類器的性能表現。實驗結果顯示，以非線性的LVAR模型的核為特徵，在AD-HC以及MCI-HC這兩組分類器中可以得到最高的準確率分別為0.74以及0.57，在AD-MCI分類器中，以AR模型的係數為特徵，可得出最高準確率為0.66；其次，本研究以主要動態模式(Principal Dynamic Modes, PDMs)分析，線性化平均之相關的非線性函數(Associated Nonlinear Functions, ANFs)斜率作為分類器分類的指標。實驗結果顯示，在這三種二元分類器的最高準確率落在0.55至0.6之間。本研究結論為非線性的LVAR核作為分類器特徵可提升分類器性能，此外透過PDMs分析利用線性化平均的ANFs斜率作為分類器的指標，透過delta、low alpha和low beta頻段的global PDMs所對應的ANFs作為分類器分類的指標是有利於分類器的分類。

Three classifiers are established in this study, i.e. Alzheimer's Disease (AD) vs. Healthy Controls (HC), Mild Cognitive Impairment (MCI) vs. Healthy Controls (HC), and AD vs. MCI, by using Lasso logistic regression. Moreover, several different features such as coefficients from the autoregressive (AR) model, kernels of linear Laguerre-Volterra AR (LVAR) model, and kernels of nonlinear LVAR model are applied for comparing the classifiers' performance. Additionally, the use of slopes from associated nonlinear functions (ANFs) in principal dynamic modes (PDMs) analysis as indicators for classification in all three classifiers is also explored. The results of this study indicate that using nonlinear LVAR model kernels as features in the classifiers significantly increases the classification performance for AD-HC and MCI-HC categories. Furthermore, classifiers built on the slopes of three specific global PDMs, corresponding to the delta, alpha, and low beta frequency bands, achieve accuracy ranging from 0.55 to 0.62 across all three classifiers.

[1] J. J. Manly, M.-X. Tang, N. Schupf, Y. Stern, J.-P. G. Vonsattel, and R. Mayeux, “Frequency and course of mild cognitive impairment in a multiethnic community,” Ann Neurol., vol. 63, no. 4, pp. 494–506, Apr. 2008, doi: 10.1002/ana.21326.
[2] R. C. Petersen, “Mild Cognitive Impairment,” 2016.
[3] T.-T. Trinh, C.-F. Tsai, Y.-T. Hsiao, C.-Y. Lee, C.-T. Wu, and Y.-H. Liu, “Identifying Individuals With Mild Cognitive Impairment Using Working Memory-Induced Intra-Subject Variability of Resting-State EEGs,” Front. Comput. Neurosci., vol. 15, p. 700467, Aug. 2021, doi: 10.3389/fncom.2021.700467.
[4] P.-N. Yu, C. Y. Liu, C. N. Heck, T. W. Berger, and D. Song, “A sparse multiscale nonlinear autoregressive model for seizure prediction,” J. Neural Eng., vol. 18, no. 2, p. 026012, Apr. 2021, doi: 10.1088/1741-2552/abdd43.
[5] Y. Kang, J. Escudero, D. Shin, E. Ifeachor, and V. Marmarelis, “Principal Dynamic Mode Analysis of EEG Data for Assisting the Diagnosis of Alzheimer’s Disease,” IEEE J. Transl. Eng. Health Med., vol. 3, pp. 1–10, 2015, doi: 10.1109/JTEHM.2015.2401005.
[6] V. Z. Marmarelis, “Identification of nonlinear biological systems using laguerre expansions of kernels,” Ann Biomed Eng, vol. 21, no. 6, pp. 573–589, Nov. 1993, doi: 10.1007/BF02368639.
[7] G. Valenza, L. Citi, E. P. Scilingo, and R. Barbieri, “Point-Process Nonlinear Models With Laguerre and Volterra Expansions: Instantaneous Assessment of Heartbeat Dynamics,” IEEE Trans. Signal Process., vol. 61, no. 11, pp. 2914–2926, Jun. 2013, doi: 10.1109/TSP.2013.2253775.
[8] D. Song, R. H. M. Chan, V. Z. Marmarelis, R. E. Hampson, S. A. Deadwyler, and T. W. Berger, “Nonlinear Dynamic Modeling of Spike Train Transformations for Hippocampal-Cortical Prostheses,” IEEE Trans. Biomed. Eng., vol. 54, no. 6, pp. 1053–1066, Jun. 2007, doi: 10.1109/TBME.2007.891948.
[9] D. Song, R. H. M. Chan, V. Z. Marmarelis, R. E. Hampson, S. A. Deadwyler, and T. W. Berger, “Nonlinear modeling of neural population dynamics for hippocampal prostheses,” Neural Networks, vol. 22, no. 9, pp. 1340–1351, Nov. 2009, doi: 10.1016/j.neunet.2009.05.004.
[10] D. Song et al., “Identification of sparse neural functional connectivity using penalized likelihood estimation and basis functions,” Journal of computational neuroscience, vol. 35, pp. 335–357, 2013.
[11] I. A. Langmoen and P. Andersen, “The hippocampal slice in vitro. A description of the technique and some examples of the opportunities it offers,” Electrophysiology of isolated mammalian CNS preparations, pp. 51–105, 1981.
[12] Z. Sankari, H. Adeli, and A. Adeli, “Intrahemispheric, interhemispheric, and distal EEG coherence in Alzheimer’s disease,” Clinical Neurophysiology, vol. 122, no. 5, pp. 897–906, May 2011, doi: 10.1016/j.clinph.2010.09.008.
[13] N. V. Ponomareva, G. I. Korovaitseva, and E. I. Rogaev, “EEG alterations in non-demented individuals related to apolipoprotein E genotype and to risk of Alzheimer disease,” Neurobiology of Aging, vol. 29, no. 6, pp. 819–827, Jun. 2008, doi: 10.1016/j.neurobiolaging.2006.12.019.
[14] G. Adler, S. Brassen, and A. Jajcevic, “EEG coherence in Alzheimer’s dementia,” J Neural Transm, vol. 110, no. 9, pp. 1051–1058, Sep. 2003, doi: 10.1007/s00702-003-0024-8.
[15] B. Jelles, Ph. Scheltens, W. M. Van Der Flier, E. J. Jonkman, F. H. L. Da Silva, and C. J. Stam, “Global dynamical analysis of the EEG in Alzheimer’s disease: Frequency-specific changes of functional interactions,” Clinical Neurophysiology, vol. 119, no. 4, pp. 837–841, Apr. 2008, doi: 10.1016/j.clinph.2007.12.002.
[16] “關於失智症-認識失智症.” http://www.tada2002.org.tw/About/IsntDementia (accessed Jul. 15, 2023).
[17] “| MoCA 臨床數據.” https://mocacognition.com/moca-clinic-data/ (accessed Jul. 15, 2023).
[18] https://www.facebook.com/verywell, “Montreal Cognitive Assessment (MoCA) Test for Dementia,” Verywell Health. https://www.verywellhealth.com/alzheimers-and-montreal-cognitive-assessment-moca-98617 (accessed Jul. 15, 2023).
[19] C. Lehmann et al., “Application and comparison of classification algorithms for recognition of Alzheimer’s disease in electrical brain activity (EEG),” Journal of Neuroscience Methods, vol. 161, no. 2, pp. 342–350, Apr. 2007, doi: 10.1016/j.jneumeth.2006.10.023.
[20] B. Oltu, M. F. Akşahin, and S. Kibaroğlu, “A novel electroencephalography based approach for Alzheimer’s disease and mild cognitive impairment detection,” Biomedical Signal Processing and Control, vol. 63, p. 102223, Jan. 2021, doi: 10.1016/j.bspc.2020.102223.
[21] E. Parvinnia, M. Sabeti, M. Zolghadri Jahromi, and R. Boostani, “Classification of EEG Signals using adaptive weighted distance nearest neighbor algorithm,” Journal of King Saud University - Computer and Information Sciences, vol. 26, no. 1, pp. 1–6, Jan. 2014, doi: 10.1016/j.jksuci.2013.01.001.
[22] D. T. Westwick and R. E. Kearney, Identification of nonlinear physiological systems, vol. 7. John Wiley & Sons, 2003.
[23] D. Song and T. W. Berger, “Identification of Nonlinear Dynamics in Neural Population Activity,” in Statistical Signal Processing for Neuroscience and Neurotechnology, Elsevier, 2010, pp. 103–128. doi: 10.1016/B978-0-12-375027-3.00004-1.
[24] D. Song et al., “Hippocampal microcircuits, functional connectivity, and prostheses,” Recent Advances On the Modular Organization of the Cortex, pp. 385–405, 2015.
[25] P.-N. Yu, S. A. Naiini, C. N. Heck, C. Y. Liu, D. Song, and T. W. Berger, “A sparse Laguerre-Volterra autoregressive model for seizure prediction in temporal lobe epilepsy,” in 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Orlando, FL, USA: IEEE, Aug. 2016, pp. 1664–1667. doi: 10.1109/EMBC.2016.7591034.
[26] V. Z. Marmarelis, Nonlinear dynamic modeling of physiological systems, vol. 10. John Wiley & Sons, 2004.
[27] “LYSIS – BMSR Biomedical Simulations Resource.” https://bmsr.usc.edu/software/lysis/ (accessed Jul. 15, 2023).
[28] D. Moretti, “Individual analysis of EEG frequency and band power in mild Alzheimer’s disease,” Clinical Neurophysiology, vol. 115, no. 2, pp. 299–308, Feb. 2004, doi: 10.1016/S1388-2457(03)00345-6.