Table 4. Sensitivity analyses of the main outcomes [SMD (95%CI)].

Outcomes	Trials number	Effects		Heterogeneity		Egger’s est
		SMD (95%CI)	P	I2(%)	P	a	P
Music therapy
Using fixed effect model	41	-0.50 (-0.58, -0.43)	<0.001	83	<0.001	-2.82(-4.71, -0.93)	0.005
Using trim and fill analysis	41	-0.66 (-0.86, -0.46)	<0.001	-	<0.001	-	-
Excluding the paper without information regarding music therapy (Chirico et al., 2020; Koelsch et al., 2010; Toccafondi et al., 2017; Porter et al., 2017)	37	-0.66 (-0.88, -0.43)	<0.001	82.2	<0.001	-3.03(-5.26, -0.81)	0.009
Excluding the papers with high bias (Toccafondi et al., 2017 and Fancourt et al., 2019)	39	-0.69 (-0.91, -0.47)	<0.001	83.6	<0.001	-2.95(-5.04, -0.86)	0.007
Excluding the papers with small sample size (20< n<30)	35	-0.57 (-0.77, -0.38)	<0.001	81.3	<0.001	2.22(-4.53, 0.08)	0.058
Excluding the studies focused on the people with a severe mental disease (Choi et al., 2008; Cheung et al. 2019)	32	-0.64(-0.86, -0.42)	<0.001	82.1	<0.001	‘-2.54(-4.67, -0.40)	0.022
Excluding the papers using an infrequently used scale (Erkkilä et al., 2011; Chen et al., 2015; Cheung et al., 2019; Chirico et al., 2020; Park et al., 2015; Sigurdardóttir et al., 2019; Wu et al., 2019; Low et al., 2020)	34	-0.62 (-0.84, -0.39)	<0.001	83.2	<0.001	-2.63(-4.67, -0.60)	0.013
Music medicine
Using fixed effect model	16	-0.86(-0.98, -0.73)	<0.001	95.4	<0.001	-5.78(-11.65, 0.10)	0.053
Using trim and fill analysis	16	-1.33(-1.96, -0.70)	<0.001	-	<0.001	-	-
Excluding the papers with small sample size (20< n<30) [49]	15	-1.32(-1.98, -0.66)	<0.001	95.7	<0.001	-6.09(-12.53, 0.36)	0.062
Excluding the papers using an infrequently used scale (Chen et al., 2015)	14	-1.25(-1.92, -0.57)	<0.001	95.7	<0.001	-5.71(-12.38, 0.98)	0.98