A Non-Parametric Approach for the Activation Detection of Block Design fMRI Simulated Data Using Self-Organizing Maps and Support Vector Machine

Abstract

Functional magnetic resonance imaging (fMRI) is a popular method to probe the functional organization of the brain using hemodynamic responses. In this method, volume images of the entire brain are obtained with a very good spatial resolution and low temporal resolution. However, they always suffer from high dimensionality in the face of classification algorithms. In this work, we combine a support vector machine (SVM) with a self-organizing map (SOM) for having a feature-based classification by using SVM. Then, a linear kernel SVM is used for detecting the active areas. Here, we use SOM for feature extracting and labeling the datasets. SOM has two major advances: (i) it reduces dimension of data sets for having less computational complexity and (ii) it is useful for identifying brain regions with small onset differences in hemodynamic responses. Our non-parametric model is compared with parametric and non-parametric methods. We use simulated fMRI data sets and block design inputs in this paper and consider the contrast to noise ratio (CNR) value equal to 0.6 for simulated datasets. fMRI simulated dataset has contrast 1–4% in active areas. The accuracy of our proposed method is 93.63% and the error rate is 6.37%.

Transformer-based dimensionality reduction

“…In this information age, a large amount of data has been generated in various fields, including education, medical care, Internet, social media and business Transformer-based dimensionality reduction [1]. These data are usually high-dimensional, heterogeneous, complex and massive [2], and they have different forms, such as text, digital image, voice signal and video.…”

Transformer-based dimensionality reduction

“…In this information age, a large amount of data has been generated in various fields, including education, medical care, Internet, social media and business [1]. These data are usually highdimensional, heterogeneous, complex and massive [2], and they have different forms, such as text, digital image, voice signal and video.…”

The use of machine learning and deep learning algorithms in functional magnetic resonance imaging—A systematic review

“…I G U R E 1 0 Number of fMRI studies for various mental diseases T A B L E 5 Classification of fMRI studies based on Machine Learning Algorithms . (2016),Arbabshirani et al (2014),Bahrami and Shamsi (2017),Bi et al (2018),Brodersen et al (2012),Chanel et al (2016),Chen et al (2017),Cox and Savoy (2003),De Martino et al (2007, 2008, Deng et al(2018), Dosenbach et al (2010) Gao et al (2017), Garner et al (2019), Guo, Xie, Cheng, and Zhao (2009), Ji, Liu, Wang, and Tang (2004), Johnston, Steele, et al (2015), Johnston, Tolomeo, et al (2015), Khazaee A (2017), LaConte, Strother, Cherkassky, Anderson, and Hu (2005), Li et al (2017), Linn et al (2016), Lorbert and Ramadge (2013), Lu et al (2016), Månsson et al (2015), Matsubara et al (2019), Meier et al (2012), Misaki, Luh, and Bandettini (2013), Mourao-Miranda, Bokde, Born, Hampel, and Stetter (2005), Peltier, Lisinski, Noll, and LaConte (2009), Qureshi et al (2017), Rubin-Falcone et al (2018), Sacchet et al (2015), Sankar et al (2016), Schrouff, Kussé, Wehenkel, Maquet, and Phillips (2012), Tan et al (2015), Tang et al (2013), Vergun et al (2013), Wang et al (2019), Wang, Ren, and Zhang (2017), Wang et al (2007), Wang et al (2009), Wetherill et al (2019), Wu et al (2015), and Zafar, Malik, Shuaibu, ur Rehman, and Dass…”