Table 4.

Summary of the reviewed CAP cycle classification methodologies, presenting the article’s best-performing results. The table sorts the articles by classification approach (first thresholds, then FSM, followed by conventional machine learning) and performance. At the of the table, the mean and standard deviation of the metrics (reported by at least two studies) are presented. Abbreviations: bruxism (B), derivation (D), female (F), healthy (H), male (M), subjects (S), threshold-based (TB), with ages between (WAB), years old (YO)

Article	Population			Classification	Performance metrics (%)
	Subjects	Characteristics			Acc	Sen		Spe		AUC		Other
[38]	16~	H S using C3-A2 or the C4-A1 D		TB	-	-		-		-		CAP rate error: 16.9
[31]	4	H S		TB	88.5	-		-		-		-
[33]	4	H S (2 M and 2 F) using C4-A1 D		TB	89.8	89.8		95.0		-		-
[56]	15 ~ WAB 23 and 42 YO (mean 32.2)	H S (6 M and 9 F) using C3-A2 or the C4-A1 D		FSM	67.9	50.1		89.5		70.3		-
[44]	14 ~ WAB 23 and 78 YO	Using C3-A2 or the C4-A1 D		FSM	75.0	-		-		-		-
[54]	19 ~ WAB 23 and 78 YO (mean 40.6)	S (15 H and 4 with SDB, 10 M and 9 F) using C3-A2 or the C4-A1 D		FSM	76.3	71.4		84.2		77.8		Positive predictive value: 67.8, negative predictive value: 77.7, diagnostic odds ratio: 13.3
[65]	19~	S (15 H and 4 with SDB, 10 M and 9 F) using C3-A2 or the C4-A1 D		FSM	77.7	72.5		80.5		-		CAP rate percentage error: 17.2
[45]	14 ~ WAB 23 and 78 YO	S (9 H, 4 with SDB, and 1 with B) using C3-A2 or the C4-A1 D		FSM	79.0	-		-		-		-
[53]	14~	S (9 H, 4 with SDB, and 1 with B) using C3-A2 or the C4-A1 D		FFNN	62.0	67.0		59.0		-		-
[52]	8~	S (4 H and 4 with SDB) using C4-A1 D		SVM	81.0	-		-		-		-
Mean ± standard deviation	4.0 ± 0.0 S			TB	89.2 ± 0.6		-		-		-		-
	12.0 ± 2.8 S			FSM	77.0 ± 1.4		72.0 ± 0.5		82.4 ± 1.9		-		-
	16.8 ± 2.3 S			Conventional machine learning	71.5 ± 9.5		-		-		-		-

~Reccordings from the CAP sleep database [6, 30]