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

Table 3

Recent studies in the literature non-ASD.

Studies	Method	Year	Modality	Data set	Accuracy (%)
Hosseini-Asl et al. [47]	3D CNN	2018	3D s-MRI	Alzheimer	100
Wang et al. [48]	CNN	2018	s-MRI	Alzheimer	97.65
Duc et al. [49]	3D CNN	2020	f-MRI	Alzheimer	85.27
Spasov et al. [50]	3D CNN	2018	s-MRI	Alzheimer	99
Liu et al. [51]	ResNet/3D DenseNet	2020	s-MRI	Alzheimer	88.9
Farooq et al. [52]	GoogleNet/ResNet	2017	s-MRI	Alzheimer	98.88
Korolev et al. [53]	3D CNN	2017	3D s-MRI	Alzheimer	80
Senanayake et al. [54]	ResNet/DenseNet/GoogleNet	2018	3D MRI	Alzheimer	79
Zou et al. [55]	3D CNN	2017	rs-f-MRI	Hyperactivity	65.67
Zou et al. [56]	3D CNN	2017	f-MRI—s-MRI	Hyperactivity	69.15
Chen et al. [26]	3D CNN/2D CNN	2019	EEG	Hyperactivity	90.29
Campese et al. [57]	SVM, 2D CNN	2019	2D/3D s-MRI	Bipolar	86.30
Choi et al. [58]	3D CNN	2017	Computerized tomography	Parkinson	98.8
Wang et al. [21]	LSTM	2018	Array collection data	Alzheimer	99
Dakka et al. [59]	LSTM	2017	4D f-MRI	Schizophrenia	66.4
Kumar et al. [60]	RNN	2019	CT—MRI—PET	Alzheimer/Autism	91.9
Talathi [61]	GRU	2017	EEG	Epilepsy	99.6
Che et al. [62]	GRU	2017	Parkinson dataset	Parkinson	95.7
Yao et al. [63]	IndRNN	2019	EEG	Epilepsy	87
Zhao et al. [64]	GAN	2020	f-MRI	Mental	82.1
Shi et al. [65]	CNN/RNN	2019	Resting state EEG	Parkinson	82.89
Kim et al. [66]	SVM, logistic, k-NN, Naïve Bayes, random forest, AdaBoost, GBM, XGBoost	2023	PET validated EEG	Alzheimer	90.9
De Nedai et al. [67]	Unsupervised machine learning	2023	f-MRI	Obsessive-compulsive disorder	65.9
Hámori et al. [68]	SVM	2023	Event related potential (ERP)	Hyperactivity	78
Cha et al. [8]	Light GBM, XGBoost GBM	2023	Diffusion tensor imaging (DTI)	Obsessive-compulsive disorder	76.72
Belić et al. [69]	Sensor-based image	2023	k-NN	Parkinson	85.18
Escobar-Ipuz et al. [70]	EEG	2023	XGB, k-NN, decision tree, Naïve Bayes	Epilepsy	98.13

CNN: convolutional neural network; 3D: three-dimensional; 2D: two-dimensional; ResNet: residual networks; DenseNet: densely connected networks; MRI: magnetic resonance imaging; f-MRI: functional MRI; s-MRI: structural MRI; EEG: electroencephalography; SVM/SVC: support vector machine; CT: computed tomography; PET: positron emission tomography; RNN: recurrent neural network; LSTM: long-short term memory; GRU: gated recurrent unit; IndRNN: independent RNN; GAN: generative adversarial network; k-NN: k-Nearest Neighbors; GBM: gradient boosting machine; XGB: extreme gradient boosting.