Abstract
Background: Cluster analysis is a robust tool for exploring the underlining structures in data and grouping them with similar objects. In the researches of Functional Magnetic Resonance Imaging (fMRI), clustering approaches attempt to classify voxels depending on their time-course signals into a similar hemodynamic response over time.
Objective: In this work, a novel unsupervised learning approach is proposed that relies on using Enhanced Neural Gas (ENG) algorithm in fMRI data for comparison with Neural Gas (NG) method, which has yet to be utilized for that aim. The ENG algorithm depends on the network structure of the NG and concentrates on an efficacious prototype-based clustering approach.
Methods: The comparison outcomes on real auditory fMRI data show that ENG outperforms the NG and statistical parametric mapping (SPM) methods due to its insensitivity to the ordering of input data sequence, various initializations for selecting a set of neurons, and the existence of extreme values (outliers). The findings also prove its capability to discover the exact and real values of a cluster number effectively.
Results: Four validation indices are applied to evaluate the performance of the proposed ENG method with fMRI and compare it with a clustering approach (NG algorithm) and model-based data analysis (SPM). These validation indices include the Jaccard Coefficient (JC), Receiver Operating Characteristic (ROC), Minimum Description Length (MDL) value, and Minimum Square Error (MSE).
Conclusion: The ENG technique can tackle all shortcomings of NG application with fMRI data, identify the active area of the human brain effectively, and determine the locations of the cluster center based on the MDL value during the process of network learning.
Keywords: Enhanced Neural Gas (ENG), fMRI clustering technique, Minimum Description Length (MDL), Neural Gas (NG), Prototype-Based Clustering (PBC), Statistical Parametric Mapping (SPM).
Graphical Abstract
[1]
Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA 1990; 87(24): 9868-72.
[http://dx.doi.org/10.1073/pnas.87.24.9868] [PMID: 2124706]
[http://dx.doi.org/10.1073/pnas.87.24.9868] [PMID: 2124706]
[2]
Ogawa S, Tank DW, Menon R, et al. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci USA 1992; 89(13): 5951-5.
[http://dx.doi.org/10.1073/pnas.89.13.5951] [PMID: 1631079]
[http://dx.doi.org/10.1073/pnas.89.13.5951] [PMID: 1631079]
[3]
Kwong KK, Belliveau JW, Chesler DA, et al. Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc Natl Acad Sci USA 1992; 89(12): 5675-9.
[http://dx.doi.org/10.1073/pnas.89.12.5675] [PMID: 1608978]
[http://dx.doi.org/10.1073/pnas.89.12.5675] [PMID: 1608978]
[4]
Boxerman JL, Bandettini PA, Kwong KK, et al. The intravascular contribution to fMRI signal change: Monte Carlo modeling and diffusion-weighted studies in vivo. Magn Reson Med 1995; 34(1): 4-10.
[http://dx.doi.org/10.1002/mrm.1910340103] [PMID: 7674897]
[http://dx.doi.org/10.1002/mrm.1910340103] [PMID: 7674897]
[5]
Ogawa S, Lee TM, Barrere B. The sensitivity of magnetic resonance image signals of a rat brain to changes in the cerebral venous blood oxygenation. Magn Reson Med 1993; 29(2): 205-10.
[http://dx.doi.org/10.1002/mrm.1910290208] [PMID: 8429784]
[http://dx.doi.org/10.1002/mrm.1910290208] [PMID: 8429784]
[6]
Golay X, Kollias S, Stoll G, Meier D, Valavanis A, Boesiger P. A new correlation-based fuzzy logic clustering algorithm for fMRI. Magn Reson Med 1998; 40(2): 249-60.
[http://dx.doi.org/10.1002/mrm.1910400211] [PMID: 9702707]
[http://dx.doi.org/10.1002/mrm.1910400211] [PMID: 9702707]
[7]
Kahnt T, Chang LJ, Park SQ, Heinzle J, Haynes JD. Connectivity-based parcellation of the human orbitofrontal cortex. J Neurosci 2012; 32(18): 6240-50.
[http://dx.doi.org/10.1523/JNEUROSCI.0257-12.2012] [PMID: 22553030]
[http://dx.doi.org/10.1523/JNEUROSCI.0257-12.2012] [PMID: 22553030]
[8]
Goutte C, Toft P, Rostrup E, Nielsen F, Hansen LK. On clustering fMRI time series. Neuroimage 1999; 9(3): 298-310.
[http://dx.doi.org/10.1006/nimg.1998.0391] [PMID: 10075900]
[http://dx.doi.org/10.1006/nimg.1998.0391] [PMID: 10075900]
[9]
Ngan SC, Hu X. Analysis of functional magnetic resonance imaging data using self-organizing mapping with spatial connectivity. Magn Reson Med 1999; 41(5): 939-46.
[http://dx.doi.org/10.1002/(SICI)1522-2594(199905)41:5<939::AID-MRM13>3.0.CO;2-Q] [PMID: 10332877]
[http://dx.doi.org/10.1002/(SICI)1522-2594(199905)41:5<939::AID-MRM13>3.0.CO;2-Q] [PMID: 10332877]
[10]
Meyer-Bäse A, Lange O, Wismüller A, Ritter H. Model-free functional MRI analysis using topographic independent component analysis. Int J Neural Syst 2004; 14(4): 217-28.
[http://dx.doi.org/10.1142/S0129065704002017] [PMID: 15372699]
[http://dx.doi.org/10.1142/S0129065704002017] [PMID: 15372699]
[11]
Ngan SC, Yacoub ES, Auffermann WF, Hu X. Node merging in Kohonen’s self-organizing mapping of fMRI data. Artif Intell Med 2002; 25(1): 19-33.
[http://dx.doi.org/10.1016/S0933-3657(02)00006-4] [PMID: 12009261]
[http://dx.doi.org/10.1016/S0933-3657(02)00006-4] [PMID: 12009261]
[12]
Peltier SJ, Polk TA, Noll DC. Detecting low-frequency functional connectivity in fMRI using a self-organizing map (SOM) algorithm. Hum Brain Mapp 2003; 20(4): 220-6.
[http://dx.doi.org/10.1002/hbm.10144] [PMID: 14673805]
[http://dx.doi.org/10.1002/hbm.10144] [PMID: 14673805]
[13]
Dimitriadou E, Barth M, Windischberger C, Hornik K, Moser E. A quantitative comparison of functional MRI cluster analysis. Artif Intell Med 2004; 31(1): 57-71.
[http://dx.doi.org/10.1016/j.artmed.2004.01.010] [PMID: 15182847]
[http://dx.doi.org/10.1016/j.artmed.2004.01.010] [PMID: 15182847]
[14]
Meyer-Baese A, Wismueller A, Lange O. Comparison of two exploratory data analysis methods for fMRI: unsupervised clustering versus independent component analysis IEEE transactions on information technology in biomedicine: a publication of the IEEE Engineering in Medicine and Biology Society 2004; 8(3): 387-98.
[http://dx.doi.org/10.1109/TITB.2004.834406]
[http://dx.doi.org/10.1109/TITB.2004.834406]
[15]
Wismüller A, Meyer-Bäse A, Lange O, Auer D, Reiser MF, Sumners D. Model-free functional MRI analysis based on unsupervised clustering. J Biomed Inform 2004; 37(1): 10-8.
[http://dx.doi.org/10.1016/j.jbi.2003.12.002] [PMID: 15016382]
[http://dx.doi.org/10.1016/j.jbi.2003.12.002] [PMID: 15016382]
[16]
Gonzalez AI, D’Anjou A, Garcia-Sebastian MT, et al. SOM and neural gas as graduated nonconvexity algorithms.Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Springer Verlag 2006; pp. 1143-52.
[http://dx.doi.org/10.1007/11751595_120]
[http://dx.doi.org/10.1007/11751595_120]
[17]
Lachiche N, Kommet J, Korczak J, et al. Neuronal clustering of brain fMRI images. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 2005; pp. 300-5.
[http://dx.doi.org/10.1007/11590316_43]
[http://dx.doi.org/10.1007/11590316_43]
[18]
Ichihashi H, Honda K, Notsu A, et al. Exploratory approach to fMRI study with fuzzy clustering and general linear model. IEEE International Conference on Fuzzy Systems. 1167-74.
[http://dx.doi.org/10.1109/FUZZY.2006.1681857]
[http://dx.doi.org/10.1109/FUZZY.2006.1681857]
[19]
Korczak J. Interactive mining of functional MRI data. Proceedings - International Conference on Signal Image Technologies and Internet Based Systems, SITIS. 912-7.
[http://dx.doi.org/10.1109/SITIS.2007.151]
[http://dx.doi.org/10.1109/SITIS.2007.151]
[20]
Davoudi H, Taalimi A, Fatemizadeh E. Extracting activated regions of fMRI data using unsupervised learning. Proceedings of the International Joint Conference on Neural Networks. 641-5.
[http://dx.doi.org/10.1109/IJCNN.2009.5178805]
[http://dx.doi.org/10.1109/IJCNN.2009.5178805]
[21]
Aljobouri HK, Jaber HA, Koçak OM, Algin O, Çankaya I. Clustering fMRI data with a robust unsupervised learning algorithm for neuroscience data mining. J Neurosci Methods 2018; 299: 45-54.
[http://dx.doi.org/10.1016/j.jneumeth.2018.02.007] [PMID: 29471065]
[http://dx.doi.org/10.1016/j.jneumeth.2018.02.007] [PMID: 29471065]
[22]
Sun X, Xu W. Fast implementation of DeLong’s algorithm for comparing the areas under correlated receiver operating characteristic curves. IEEE Signal Process Lett 2014; 21(11): 1389-93.
[http://dx.doi.org/10.1109/LSP.2014.2337313]
[http://dx.doi.org/10.1109/LSP.2014.2337313]
[23]
Baumgartner R, Windischberger C, Moser E. Quantification in functional magnetic resonance imaging: fuzzy clustering vs. correlation analysis. Magn Reson Imaging 1998; 16(2): 115-25.
[http://dx.doi.org/10.1016/S0730-725X(97)00277-4] [PMID: 9508268]
[http://dx.doi.org/10.1016/S0730-725X(97)00277-4] [PMID: 9508268]
[24]
Bock H-H, Diday E. Analysis of Symbolic Data 2000; 153-65.
[http://dx.doi.org/10.1007/978-3-642-57155-8]
[http://dx.doi.org/10.1007/978-3-642-57155-8]
[25]
Lindquist MA. The statistical analysis of fMRI Data. Stat Sci 2008; 23(4): 439-64.
[http://dx.doi.org/10.1214/09-STS282]
[http://dx.doi.org/10.1214/09-STS282]
[26]
Liao W, Chen H, Yang Q, Lei X. Analysis of fMRI data using improved self-organizing mapping and spatio-temporal metric hierarchical clustering. IEEE Trans Med Imaging 2008; 27(10): 1472-83.
[http://dx.doi.org/10.1109/TMI.2008.923987] [PMID: 18815099]
[http://dx.doi.org/10.1109/TMI.2008.923987] [PMID: 18815099]
[27]
Pereira F, Mitchell T, Botvinick M. Machine learning classifiers and fMRI: a tutorial overview. Neuroimage 2009; 45(Suppl 1.): S199-209.
[http://dx.doi.org/10.1016/j.neuroimage.2008.11.007] [PMID: 19070668]
[http://dx.doi.org/10.1016/j.neuroimage.2008.11.007] [PMID: 19070668]
[28]
Katwal SB, Gore JC, Rogers BP. Analyzing fMRI data with graph-based visualizations of self-organizing maps. Proceedings - International Symposium on Biomedical Imaging. 1577-80.
[http://dx.doi.org/10.1109/ISBI.2011.5872703]
[http://dx.doi.org/10.1109/ISBI.2011.5872703]
[29]
Qin AK, Suganthan PN. Enhanced neural gas network for prototype-based clustering. Pattern Recognit 2005; 38(8): 1275-88.
[http://dx.doi.org/10.1016/j.patcog.2004.12.007]
[http://dx.doi.org/10.1016/j.patcog.2004.12.007]
[30]
Zacks JM, Braver TS, Sheridan MA, et al. Human brain activity time-locked to perceptual event boundaries. Nat Neurosci 2001; 4(6): 651-5.
[http://dx.doi.org/10.1038/88486] [PMID: 11369948]
[http://dx.doi.org/10.1038/88486] [PMID: 11369948]
[31]
Hasson U, Nir Y, Levy I, Fuhrmann G, Malach R. Intersubject synchronization of cortical activity during natural vision. Science 2004; 303(5664): 1634-40.
[http://dx.doi.org/10.1126/science.1089506] [PMID: 15016991]
[http://dx.doi.org/10.1126/science.1089506] [PMID: 15016991]
[32]
Bartels A, Zeki S. Functional brain mapping during free viewing of natural scenes. Hum Brain Mapp 2004; 21(2): 75-85.
[http://dx.doi.org/10.1002/hbm.10153] [PMID: 14755595]
[http://dx.doi.org/10.1002/hbm.10153] [PMID: 14755595]
[33]
Malinen S, Hlushchuk Y, Hari R. Towards natural stimulation in fMRI--issues of data analysis. Neuroimage 2007; 35(1): 131-9.
[http://dx.doi.org/10.1016/j.neuroimage.2006.11.015] [PMID: 17208459]
[http://dx.doi.org/10.1016/j.neuroimage.2006.11.015] [PMID: 17208459]
[34]
Lee MH, Hacker CD, Snyder AZ, et al. Clustering of resting state networks. PLoS One 2012; 7(7): e40370.
[http://dx.doi.org/10.1371/journal.pone.0040370] [PMID: 22792291]
[http://dx.doi.org/10.1371/journal.pone.0040370] [PMID: 22792291]
[35]
Friston KJ, Frith CD, Liddle PF, Frackowiak RS. Functional connectivity: the principal-component analysis of large (PET) data sets. J Cereb Blood Flow Metab 1993; 13(1): 5-14.
[http://dx.doi.org/10.1038/jcbfm.1993.4] [PMID: 8417010]
[http://dx.doi.org/10.1038/jcbfm.1993.4] [PMID: 8417010]
[36]
Friston KJ, Poline JB, Holmes AP, Frith CD, Frackowiak RS. A multivariate analysis of PET activation studies. Hum Brain Mapp 1996; 4(2): 140-51.
[http://dx.doi.org/10.1002/(SICI)1097-0193(1996)4:2<140::AID-HBM5>3.0.CO;2-3] [PMID: 20408193]
[http://dx.doi.org/10.1002/(SICI)1097-0193(1996)4:2<140::AID-HBM5>3.0.CO;2-3] [PMID: 20408193]
[37]
Sychra JJ, Bandettini PA, Bhattacharya N, Lin Q. Synthetic images by subspace transforms. I. Principal components images and related filters. Med Phys 1994; 21(2): 193-201.
[http://dx.doi.org/10.1118/1.597374] [PMID: 8177152]
[http://dx.doi.org/10.1118/1.597374] [PMID: 8177152]
[38]
Backfrieder W, Baumgartner R, Sámal M, Moser E, Bergmann H. Quantification of intensity variations in functional MR images using rotated principal components. Phys Med Biol 1996; 41(8): 1425-38.
[http://dx.doi.org/10.1088/0031-9155/41/8/011] [PMID: 8858728]
[http://dx.doi.org/10.1088/0031-9155/41/8/011] [PMID: 8858728]
[39]
McKeown MJ, Makeig S, Brown GG, et al. Analysis of fMRI data by blind separation into independent spatial components. Hum Brain Mapp 1998; 6(3): 160-88.
[http://dx.doi.org/10.1002/(SICI)1097-0193(1998)6:3<160::AID-HBM5>3.0.CO;2-1] [PMID: 9673671]
[http://dx.doi.org/10.1002/(SICI)1097-0193(1998)6:3<160::AID-HBM5>3.0.CO;2-1] [PMID: 9673671]
[40]
McKeown MJ, Jung TP, Makeig S, et al. Spatially independent activity patterns in functional MRI data during the stroop color-naming task. Proc Natl Acad Sci USA 1998; 95(3): 803-10.
[http://dx.doi.org/10.1073/pnas.95.3.803] [PMID: 9448244]
[http://dx.doi.org/10.1073/pnas.95.3.803] [PMID: 9448244]
[41]
McKeown MJ. Detection of consistently task-related activations in fMRI data with hybrid independent component analysis. Neuroimage 2000; 11(1): 24-35.
[http://dx.doi.org/10.1006/nimg.1999.0518] [PMID: 10686114]
[http://dx.doi.org/10.1006/nimg.1999.0518] [PMID: 10686114]
[42]
Calhoun VD, Adali T, Pearlson GD, Pekar JJ. Spatial and temporal independent component analysis of functional MRI data containing a pair of task-related waveforms. Hum Brain Mapp 2001; 13(1): 43-53.
[http://dx.doi.org/10.1002/hbm.1024] [PMID: 11284046]
[http://dx.doi.org/10.1002/hbm.1024] [PMID: 11284046]
[43]
Hyvarinen A, Karhunen J, Oja E. Independent component analysis J. Wiley 2001.
[http://dx.doi.org/10.1002/0471221317]
[http://dx.doi.org/10.1002/0471221317]
[44]
Marrelec G, Benali H, Ciuciu P, Pélégrini-Issac M, Poline JB. Robust Bayesian estimation of the hemodynamic response function in event-related BOLD fMRI using basic physiological information. Hum Brain Mapp 2003; 19(1): 1-17.
[http://dx.doi.org/10.1002/hbm.10100] [PMID: 12731100]
[http://dx.doi.org/10.1002/hbm.10100] [PMID: 12731100]
[45]
Jain AK, Murty MN, Flynn PJ. Data clustering: A review.ACM Computing Surveys. 1999; pp. 264-323.
[http://dx.doi.org/10.1145/331499.331504]
[http://dx.doi.org/10.1145/331499.331504]
[47]
Kohonen T. Self-organized formation of topologically correct feature maps. Biol Cybern 1982; 43(1): 59-69.
[http://dx.doi.org/10.1007/BF00337288]
[http://dx.doi.org/10.1007/BF00337288]
[48]
Martinetz TM, Berkovich SG, Schulten KJ. ; Neural-gas’ network for vector quantization and its application to time-series prediction. IEEE Trans Neural Netw 1993; 4(4): 558-69.
[http://dx.doi.org/10.1109/72.238311] [PMID: 18267757]
[http://dx.doi.org/10.1109/72.238311] [PMID: 18267757]
[51]
Ashburner J, Barnes G, Chen C-C, et al. SPM8 Manual The FIL Methods Group (and honorary members). 2013 .http://www.fil.ion.ucl.ac.uk/spm/
[52]
Anderberg MR. Cluster Analysis for Applications. 1st edition. Academic Press. Academic Press 1973.
[53]
Skudlarski P, Constable RT, Gore JC. ROC analysis of statistical methods used in functional MRI: individual subjects. Neuroimage 1999; 9(3): 311-29.
[http://dx.doi.org/10.1006/nimg.1999.0402] [PMID: 10075901]
[http://dx.doi.org/10.1006/nimg.1999.0402] [PMID: 10075901]
[54]
Frey BJ, Dueck D. Clustering by passing messages between data points. Science 2007; 315(5814): 972-6.
[http://dx.doi.org/10.1126/science.1136800] [PMID: 17218491]
[http://dx.doi.org/10.1126/science.1136800] [PMID: 17218491]
[55]
Rissanen J. A universal prior for integers and estimation by minimum description length. Ann Stat 1983; 11(2): 416-31.
[http://dx.doi.org/10.1214/aos/1176346150]
[http://dx.doi.org/10.1214/aos/1176346150]
Article Metrics
9