Autism Spectrum Disorder Detection by Hybrid Convolutional Recurrent Neural Networks from Structural and Resting State Functional MRI Images

Table 4

Recent studies in the literature with ASD.

Studies	Method	Year	Modality	Accuracy (%)
Zhao et al. [12]	SVM	2020	f-MRI	83
Mostafa et al. [72]	LDA	2019	f-MRI	77.70
Liu et al. [73]	DFC + MTFS	2020	rs-f-MRI	76.80
Wang et al. [74]	MLP + ensemble learning	2020	f-MRI	74.52
Sun et al. [75]	No superparameter FCN	2021	rs-f-MRI	71.74
Shi et al. [76]	Three-way decision model	2021	f-MRI	71.35
Grana and Silva [77]	SVC	2021	rs-f-MRI	71.10
Spera et al. [78]	L-SVM	2019	rs-f-MRI	71
Reiter et al. [79]	RF	2021	rs-f-MRI	67.81
Chaitra et al. [80]	RCE-SVM	2020	f-MRI	67.30
Brahim and Farrugia [81]	RBF-SVC	2020	rs-f-MRI	66.70
Sun et al. [82]	RBF + SVM	2021	f-MRI	59.10
Kazeminejad and Sotero [83]	NEG + MLP	2020	rs-f-MRI	58.70
Dekhil et al. [15]	Correlation analysis	2021	rs-f-MRI	95–100
Dekhil et al. [18]	k-NN, random forest	2019	s-MRI, f-MRI	81
Jamwal et al. [84]	DBN	2022	s-MRI, f-MRI	—
Traut et al. [85]	Logistic regression, SVC	2022	s-MRI, f-MRI	79
Dadi et al. [86]	K-means, wards algorithm, CanICA, DictLearn	2019	rs-f-MRI	86
Abraham et al. [87]	K-means, wards algorithm, ICA	2017	rs-f-MRI	67
Gogoi et al. [88]	VGG16, inception v3, ResNet50	2023	MRI	94
Han et al. [89]	Cross, supervised, LOSO, Fed_DA,	2023	rs-f-MRI, s-MRI	69.37
Manikantan and Jaganathan [90]	Graphical neural networks	2023	rs-f-MRI, s-MRI	69.45
Deng et al. [91]	GAN	2023	f-MRI	71
Dhinagar et al. [92]	Metalearning	2023	MRI	85.70
Artiles et al. [93]	Multiple linear regression	2023	rs-f-MRI	76.40
Quiang et al. [94]	Hierarchical FBN	2023	f-MRI	82.10
Jönemo et al. [95]	3D CNN	2023	rs-f-MRI	80
Kunda et al. [96]	MIDA, ridge classifier, logistic regression, SVM	2023	rs-f-MRI	73

SVM: support vector machine; LDA: linear discriminant analysis; DFC: dynamic functionally connected; MTFS: multitask feature selection; MLP: multilayer perceptron; FCN: functionally connected network; L-SVM: linear kernel support vector machine; RF: random forest; RCE: recursive cluster removal; RBF: radial basis function; NEG: negative correlation matrix; k-NN: k-nearest neighbors; DBN: deep belief network; CanICA: canonical independent component analysis; DictLearn: dictionary learning; ICA: independent component analysis; VGG: very deep convolutional networks; LOSO: leave one site out; Fed_DA: federated domain adaptation; GAN: adversarial generation network; FBN: functional brain networks; CNN: convolutional neural network; MIDA: maximum independence domain adaptation.