Table 4.

Performance evaluation of image-modality based ML models for dementia

S.No	Authors	Year	D_ID$_{}^{\ast }$	Feature	Model	Metrics	Acc.%	Spec.%	Sens.%
1	Dashtipour et al. [70]	2021	01	CNN	BiLSTM	k-fold	93.19	93.00	93.00
2	Helaly et al. [71]	2021	02	CNN	BMIC	MCC	97.00	98.00	94.00
3	Vandenberghe et al. [72]	2012	03	Voxels	SVM	MCC	100	92.00	85.00
4	Odusami et al. [73]	2021	04	ReLU	CNN	MCC	98.74	100	97.24
5	Zheng et al. [74]	2019	05	LASSO	SVM	ROC	87.17	87.17	82.65
6	Basheer et al. [75]	2019	06	KPCA	CNN	MCC	92.30	NA	NA
7	Leong et al. [76]	2019	07	Boruta	RF	MCC	94.39	100	84.24
8	Battineni et al. [77]	2019	08	RBF	SVM	MCC	68.75	NA	NA
9	Mathotaarachchi et al. [78]	2017	9	${\chi }^2$	RUSRF	F1-score	90.60	86.50	70.80
10	Aruna and Chitra [79]	2015	10	ICA	SVM	MCC	90.24	NA	NA
11	Herzog et al. [80]	2021	11	std, RMS	CNN	MCC	90.50	90.00	93.00
12	Battineni et al. [81]	2020	12	ANN	NB	ROC	94.20	89.56	89.92
13	Ma et al. [82]	2021	13	FPDS	CNN	MCC	89.30	67.00	84.50
14	Bidani et al. [83]	2019	14	K-means	DNN	MCC	81.94	NA	NA
15	Moscoso et al. [84]	2019	15	NA	ML	AUC	84.00	71.00	78.00
16	De Bruijne [90]	2016	16	DNN	SVM	ROC	65.00	88.00	80.00
17	Mohammed et al. [86]	2020	16	IFS	ML	ROC	90.00	93.00	98.00
18	Salvatore et al. [87]	2015	13	PCA	SVM	k-fold	76.00	NA	NA
19	Katako et al. [88]	2018	09	ISDA	SVM	AUC	94.50	95.00	84.00
20	Moscoso et al. [89]	2019	16	FSL FAST	RF	MCC	76.00	89.80	88.90
21	Tong et al. [91]	2014	11	mi-Graph	SVM	k-Fold	89.00	92.60	84.90
22	Akhila et al. [92]	2017	10	SFTA	ANN	MCC	97.50	97.50	97.50
23	Chen and Pham [93]	2013	16	mi-Graph	Markov	k-Fold	78.30	76.00	79.00
24	Patil and Yardi [94]	2013	NA	DCT	ANN	MCC	100	NA	NA
25	Gulhare et al. [95]	2017	06	niblack	DNN	MCC	96.60	NA	NA

D_ID$_{}^{\ast }$: is a reference number of Dataset ID