跳到主要內容

臺灣博碩士論文加值系統

(44.222.64.76) 您好!臺灣時間:2024/06/17 09:13
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:周懃
研究生(外文):Chin Chou
論文名稱:以演化型類神經網路結合回饋錯誤學習法則發展強健神經解碼進行前肢運動軌跡預測
論文名稱(外文):Robust Neural Decoding Forelimb Trajectory in Evolutionary Neural Networks with Feedback-error-learning Scheme
指導教授:楊世宏楊世宏引用關係陳右穎陳右穎引用關係
指導教授(外文):Shih-Hung YangYou-Yin Chen
學位類別:碩士
校院名稱:國立陽明大學
系所名稱:生物醫學工程學系
學門:工程學門
學類:生醫工程學類
論文種類:學術論文
論文出版年:2018
畢業學年度:106
語文別:英文
論文頁數:54
中文關鍵詞:腦機介面神經解碼錯誤迴授演化式演算法遞歸類神經網路
外文關鍵詞:Brain Machine InterfacesNeural DecodingError FeedbackEvolutionary AlgorithmRecurrent Neural Network
相關次數:
  • 被引用被引用:0
  • 點閱點閱:141
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
在腦機介面中,神經訊號與動作指令間的關係式會隨著時間改變,這對神經解碼器長期的使用造成很大的挑戰,因為解碼器的解碼效果會不穩定並且必須更新,因而中斷腦機介面的使用。之前研究提出一個用很多天數的訓練資料來建立的解碼器,可以提升解碼器對於神經訊號與動作指令間的關係式隨時間改變的穩定性。但在少天數的訓練資料下,目前並沒有解碼器能有長期穩定的表現。因此本研究想開發出一個用少天數的訓練資料就能有長期穩定表現的解碼器。本研究提出使用具有錯誤迴授之演化型類神經網路用於神經解碼。此解碼器將前一刻的誤差輸入解碼器中調整當下的輸入。我們設計一個大鼠壓桿的實驗,並用提出的解碼器透過大鼠動作皮質區的神經訊號來解碼其手部軌跡,以驗證提出的解碼器的表現。本研究用兩天的訓練資料來訓練解碼器並用五天的測試資料來測試解碼器的表現。實驗結果顯示,與一般腦機介面常用的不具錯誤迴授之遞歸神經網路相較,本研究提出的解碼器在解碼的準確度與穩定性上皆有顯著提升。此結果證明本研究提出的解碼器用少天數的訓練資料就能有長期穩定表現。
Changes in the functional mapping between neural activities and kinematic parameters over time poses a challenge to current neural decoder, which requires the decoder to be retrained and therefore interrupts the use of brain machine interfaces (BMIs). Previous work proposed to a decoder robust to changes in functional mappings, but the robustness of that decoder comes from training with many days of training data. If trained with few days of training data, that decoder won’t be robust to changes in functional mappings. Therefore, we want to propose a decoder robust to changes in functional mappings, but only trained with few days of training data. In this study, we proposed an evolutionary neural network with error feedback, ECPNN-EF, as a neural decoder. ECPNN-EF took previous error as an input to the decoder to modify the output of the decoder. The decoder was validated in a lever-pressing task for rat to decode rat’s forelimb movement from its neural activities. Two days of data were used to train the decoder and five days of data were used to test the decoder. The results comparing the performance across five days showed that the performance of ECPNN-EF was significantly higher than that of the standard recurrent neural network without error feedback, which is commonly used in BMI. This suggested that when trained with few days of training data, ECPNN-EF can be a decoder robust to changes in functional mappings.
Acknowledgements i
摘要 ii
Abstract iii
Table of Contents iv
List of Tables vii
List of Figures viii
Symbols x
Chapter 1. Introduction 1
1.1 Overview 1
1.2 Literature Review 3
1.3 Motivations and Objectives 6
1.4 Thesis Organization 8
Chapter 2. Materials and Methods 9
2.1 Architecture of Neural Decoding of Forelimb Movement 9
2.2 Electrode Implantation and Animal Training 10
2.2.1 Animal Preparation 10
2.2.2 Animal Training and Behavioral Task 10
2.2.3 Surgical Procedures 11
2.3 Data Recording and Preprocessing 12
2.4 ECPNN-EF as a Neural Decoder 16
2.4.1 Error Feedback 17
2.4.2 Time-Lags and Temporal Order of the Neural Signal 18
2.4.3 ECPNN-EF Training 18
2.4.3.1 Encoding Scheme 20
2.4.3.2 Population Initialization 21
2.4.3.3 Network Crossover 22
2.4.3.4 Network Mutation 23
2.4.3.5 CBP 24
2.4.3.6 ABSS 25
2.4.3.7 Termination Criterion 25
2.5 Performance Evaluation 26
Chapter 3. Results 29
3.1 Training ECPNN-EF with Different Numbers of Days of Data 29
3.2 Performance Optimization for ECPNN-EF 30
3.2.1 Effects of Different Time-Lags, and Temporal Orders of the Instantaneous Firing Rate 30
3.2.2 Effects of Different Types and Tapped Delay Lines of Error Feedback 31
3.2.3 ECPNN-EF Trained with Different Probability of Crossover and Mutation 32
3.2.4 Evolution Progress of ECPNN-EF 33
3.2.5 Effects of CBP and ABSS 35
3.3 Effects of Interconnected Hidden Units 36
3.4 Effects of ECPA 37
3.5 Effects of Error Feedback 38
3.6 Comparison with Other Work 39
Chapter 4. Discussion 41
4.1 Training ECPNN-EF with Different Numbers of Days of Data 41
4.2 Performance Optimization for ECPNN-EF 42
4.2.1 Effects of Different Time-Lags, and Temporal Orders of the Instantaneous Firing Rate 42
4.2.2 Effects of Different Types and Tapped Delay Lines of Error Feedback 42
4.2.3 Effects of Different Probability of Crossover and Mutation 43
4.2.4 Evolution Progress of ECPNN-EF 44
4.2.5 Effects of CBP and ABSS 45
4.3 Effects of Interconnected Hidden Units 46
4.4 Effects of ECPA 46
4.5 Effects of Error Feedback 47
4.6 Comparison with Other Work 47
References 49


List of Tables
Table 1. Training parameters used for ECPNN-EF 20
Table 2. Experimental data characteristics 27
Table 3. Performance of ECPNN-EF, ECPNN-EFWC, and ECPNN-EFWA 35


List of Figures
Figure 1. Examples of successful BMI applications 1
Figure 2. System architecture of Brain Machine Interfaces (BMIs) 3
Figure 3. Illustrations of decoder retraining and adaptive decoders 5
Figure 4. The architecture of neural decoding for forelimb movement in a water-reward lever-pressing task 9
Figure 5. Specification of Plexiglas behavioral box 11
Figure 6. Microwire electrode array and the location for electrode array implantation 12
Figure 7. Spike sorting and trajectory tracking 14
Figure 8. An example of rat’s neural activities and forelimb movement over time 15
Figure 9. ECPNN-EF as a neural decoder 17
Figure 10. Flow chart of training ECPNN-EF 19
Figure 11. An example of NN coding 21
Figure 12. An example of network crossover 23
Figure 13. An example of network mutation 24
Figure 14. Performance of ECPNN-EF trained with different numbers of days of training data 29
Figure 15. Performance of ECPNN-EF decoded with different time-lags (33 ms/lag) and temporal orders of the instantaneous firing rate 31
Figure 16. Performance of ECPNN-EF with different types and tapped delay lines (TDLs) of error feedback 32
Figure 17. Performance of ECPNN-EF trained with different probability of crossover and mutation 33
Figure 18. The training performance and validation performance with respect to generation during training the ECPNN-EF 34
Figure 19. Performance comparison of ECPNN-EF and ECPNN-EFWI 36
Figure 20. Performance comparison of ECPNN-EF and RNN-EF 37
Figure 21. Performance comparison of ECPNN-EF and ECPNN 39
Figure 22. Performance comparison of ECPNN-EF and RNN 40
1. Kao, Jonathan C., et al. "Information systems opportunities in brain–machine interface decoders." Proceedings of the IEEE 102.5 (2014): 666-682.
2. Chapin, John K., et al. "Real-time control of a robot arm using simultaneously recorded neurons in the motor cortex." Nature neuroscience 2.7 (1999): 664-670.
3. Collinger, Jennifer L., et al. "High-performance neuroprosthetic control by an individual with tetraplegia." The lancet 381.9866 (2013): 557-564.
4. Bouton, Chad E., et al. "Restoring cortical control of functional movement in a human with quadriplegia." Nature 533.7602 (2016): 247-250.
5. Leuthardt, Eric C., et al. "A brain–computer interface using electrocorticographic signals in humans." Journal of neural engineering 1.2 (2004): 63.
6. Leuthardt, Eric C., et al. "Evolution of brain-computer interfaces: going beyond classic motor physiology." Neurosurgical focus 27.1 (2009): E4
7. Wang, Zuoguan, et al. "Decoding onset and direction of movements using electrocorticographic (ECoG) signals in humans." Frontiers in neuroengineering 5 (2012): 15-27.
8. Gilja, Vikash, et al. "Clinical translation of a high-performance neural prosthesis." Nature medicine 21.10 (2015): 1142-1145.
9. Pandarinath, Chethan, et al. "High performance communication by people with paralysis using an intracortical brain-computer interface." eLife 6 (2017): 1-27.
10. Georgopoulos, Apostolos P., et al. "Primate motor cortex and free arm movements to visual targets in three-dimensional space. II. Coding of the direction of movement by a neuronal population." Journal of neuroscience 8.8 (1988): 2928-2937.
11. Salinas, Emilio, et al. "Vector reconstruction from firing rates." Journal of computational neuroscience 1.1 (1994): 89-107.
12. Koyama, Shinsuke, et al. "Bayesian decoding of neural spike trains." Annals of the institute of statistical mathematics 62.1 (2010): 37-59.
13. Bishop, Gary, and Greg Welch. "An introduction to the kalman filter." Proc of SIGGRAPH, Course 8.27599-23175 (2001): 41-87.
14. Wu, Wei, et al. "Neural decoding of cursor motion using a Kalman filter." Advances in neural information processing systems (2003): 133-140.
15. Homer, Mark L., et al. "Mixing decoded cursor velocity and position from an offline Kalman filter improves cursor control in people with tetraplegia." Neural engineering (NER), 6th international IEEE/EMBS conference on. IEEE (2013): 715-718.
16. Golub, Matthew D., et al. "Motor cortical control of movement speed with implications for brain-machine interface control." Journal of neurophysiology 112.2 (2014): 411-429.
17. Sanchez, Justin C., et al. "Interpreting spatial and temporal neural activity through a recurrent neural network brain-machine interface." IEEE transactions on neural systems and rehabilitation engineering 13.2 (2005): 213-219.
18. Sanchez, J. C., et al. "Simultaneus prediction of four kinematic variables for a brain-machine interface using a single recurrent neural network." Engineering in medicine and biology society 2 (2004): 5321-5324.
19. Sussillo, David, et al. "A recurrent neural network for closed-loop intracortical brain–machine interface decoders." Journal of neural engineering 9.2 (2012): 026027.
20. Kifouche, Abdessalam, et al. "Decoding Hand Trajectory from Primary Motor Cortex ECoG Using Time Delay Neural Network." International conference on engineering applications of neural networks. springer, cham (2014): 237-247.
21. Kim, Sung-Phil, et al. "Statistical analysis of the non-stationarity of neural population codes." Biomedical robotics and biomechatronics (2006): 811-816.
22. Simeral, J. D., et al. "Neural control of cursor trajectory and click by a human with tetraplegia 1000 days after implant of an intracortical microelectrode array." Journal of neural engineering 8.2 (2011): 025027.
23. Perge, János A., et al. "Intra-day signal instabilities affect decoding performance in an intracortical neural interface system." Journal of neural engineering 10.3 (2013): 036004.
24. Perge, János A., et al. "Reliability of directional information in unsorted spikes and local field potentials recorded in human motor cortex." Journal of neural engineering 11.4 (2014): 046007.
25. Wodlinger, B., et al. "Ten-dimensional anthropomorphic arm control in a human brain− machine interface: difficulties, solutions, and limitations." Journal of neural engineering 12.1 (2014): 16011.
26. Barrese, James C., et al. "Failure mode analysis of silicon-based intracortical microelectrode arrays in non-human primates." Journal of neural engineering 10.6 (2013): 066014.
27. Fernández, Eduardo, et al. "Acute human brain responses to intracortical microelectrode arrays: challenges and future prospects." Frontiers in neuroengineering 7 (2014): 24-29.
28. Jackson, Andrew, et al. "Long-term motor cortex plasticity induced by an electronic neural implant." Nature 444.7115 (2006): 56-60.
29. Salatino, Joseph W., et al. "Glial responses to implanted electrodes in the brain." Nature biomedical engineering 1.11 (2017): 862–877.
30. Jarosiewicz, Beata, et al. "Virtual typing by people with tetraplegia using a self-calibrating intracortical brain-computer interface." Science translational medicine 7.313 (2015): 313ra179-313ra179.
31. Jarosiewicz, Beata, et al. "Advantages of closed-loop calibration in intracortical brain–computer interfaces for people with tetraplegia." Journal of neural engineering 10.4 (2013): 046012.
32. Orsborn, Amy L, et al. "Closed-loop decoder adaptation shapes neural plasticity for skillful neuroprosthetic control." Neuron 82.6 (2014): 1380-1393.
33. Dangi, Siddharth, et al. "Design and analysis of closed-loop decoder adaptation algorithms for brain-machine interfaces." Neural computation 25.7 (2013): 1693-1731.
34. Li, Zheng, et al. "Adaptive decoding for brain-machine interfaces through Bayesian parameter updates." Neural computation 23.12 (2011): 3162-3204.
35. Nuyujukian, Paul, et al. "Performance sustaining intracortical neural prostheses." Journal of neural engineering 11.6 (2014): 066003.
36. Bishop, William, et al. "Self-recalibrating classifiers for intracortical brain–computer interfaces." Journal of neural engineering 11.2 (2014): 026001.
37. Chestek, Cynthia A., et al. "Long-term stability of neural prosthetic control signals from silicon cortical arrays in rhesus macaque motor cortex." Journal of neural engineering 8.4 (2011): 045005.
38. Sussillo, David, et al. "Making brain–machine interfaces robust to future neural variability." Nature communications (2016): 3162-3204.
39. Connor, Jerome T., et al. "Recurrent neural networks and robust time series prediction." IEEE transactions on neural networks 5.2 (1994): 240-254.
40. Mahmud, Mohammad Sultan, and Meesad, Phayung. "An innovative recurrent error-based neuro-fuzzy system with momentum for stock price prediction." Soft computing 20.10 (2016): 4173-4191.
41. Waheeb, Waddah, et al. "Ridge Polynomial Neural Network with Error Feedback for Time Series Forecasting." PloS one 11.12 (2016): 458-462.
42. Hornik, Kurt, et al. "Multilayer feedforward networks are universal approximators." Neural networks 2.5 (1989): 359-366.
43. Haykin, Simon. "Neural networks: a comprehensive foundation." Prentice hall PTR (1994)
44. Jaeger, Herbert, and Harald Haas. "Harnessing nonlinearity: Predicting chaotic systems and saving energy in wireless communication." Science 304.5667 (2004): 78-80.
45. Waibel, Alex, et al. "Phoneme recognition using time-delay neural networks." IEEE transactions on acoustics, speech, and signal processing 37.3 (1989): 328-339.
46. Kwok, Tin-Yau, and Yeung, Dit-Yan. "Constructive algorithms for structure learning in feedforward neural networks for regression problems." IEEE transactions on neural networks 8.3 (1997): 630-645.
47. Reed, Russell. "Pruning algorithms-a survey." IEEE transactions on neural networks 4.5 (1993): 740-747.
48. Islam, Md Monirul, et al. "A new adaptive merging and growing algorithm for designing artificial neural networks." IEEE transactions on systems, man, and cybernetics, part B (cybernetics) 39.3 (2009): 705-722.
49. Yang, Shih-Hung, and Chen, Yon-Ping. "An evolutionary constructive and pruning algorithm for artificial neural networks and its prediction applications." Neurocomputing 86 (2012): 140-149.
50. Huang, De-Shuang, and Du, Ji-Xiang. "A constructive hybrid structure optimization methodology for radial basis probabilistic neural networks." IEEE transactions on neural networks 19.12 (2008): 2099-2115.
51. Masutti, Thiago AS, and Leandro N. de Castro. "Neuro-immune approach to solve routing problems." Neurocomputing 72.10 (2009): 2189-2197.
52. Kaylani, Assem, et al. "AG-ART: An adaptive approach to evolving ART architectures." Neurocomputing 72.10 (2009): 2079-2092.
53. Angeline, Peter J., et al. "An evolutionary algorithm that constructs recurrent neural networks." IEEE transactions on neural networks 5.1 (1994): 54-65.
54. Lin, Hui-Ching, et al. "Central thalamic deep-brain stimulation alters striatal-thalamic connectivity in cognitive neural behavior." Frontiers in neural circuits 9 (2016): 87-98.
55. Wu, Wei, et al. "Modeling and decoding motor cortical activity using a switching Kalman filter." IEEE transactions on biomedical engineering 51.6 (2004): 933-942.
56. Rumelhart, David E., et al. "Learning representations by back-propagating errors." Nature 323.6088 (1986): 533-536.
57. Prechelt, Lutz. "Automatic early stopping using cross validation: quantifying the criteria." Neural networks 11.4 (1998): 761-767.
58. Churchland, Mark M., et al. "Preparatory activity in premotor and motor cortex reflects the speed of the upcoming reach." Journal of neurophysiology 96.6 (2006): 3130-3146.
59. Wu, Wei, et al. "Bayesian population decoding of motor cortical activity using a Kalman filter." Neural computation 18.1 (2006): 80-118.
60. Yang, Shih-Hung, et al. "A sliced inverse regression (SIR) decoding the forelimb movement from neuronal spikes in the rat motor cortex." Frontiers in neuroscience 10 (2016): 556-565.
61. Li, Yun, and Alexander Häuβler. "Artificial evolution of neural networks and its application to feedback control." Artificial intelligence in engineering 10.2 (1996): 143-152.
62. Sheng, Weiguo, et al. "An adaptive memetic algorithm with rank-based mutation for artificial neural network architecture optimization." IEEE access 5 (2017): 18895-18908.
63. Elizondo, D., et al. "A survey of partially connected neural networks." International journal of neural systems 8.05n06 (1997): 535-558.
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top