Machine learning-based ability to classify psychosis and early stages of disease through parenting and attachment-related variables is associated with social cognition

Table 2 Features entering the machine learning algorithms and their respective characterization in all study samples

	All subjects (mean ± SD)	HC (mean ± SD)	PSY (mean ± SD)	HC vs. PSY p value
A. Discovery sample
PBI maternal care	25.14 ± 6.39	25.77 ± 5.81	23.82 ± 7.39	0.144
PBI maternal overprot	16.30 ± 7.54	15.21 ± 7.09	18.56 ± 8.03	0.033*
PBI paternal care	20.57 ± 7.79	20.28 ± 7.75	21.18 ± 7.94	0.583
PBI paternal overprot	14.97 ± 7.22	13.80 ± 7.06	17.35 ± 7.07	0.018*
ECR avoidance	46.59 ± 16.28	44.06 ± 16.64	51.88 ± 14.32	0.020*
ECR anxiety	65.61 ± 21.22	61.30 ± 20.11	74.62 ± 20.92	0.002*

	All subjects (mean ± SD)	HC (mean ± SD)	ESD (mean ± SD)	HC vs. ESD p value
B. Validation clinical sample
PBI maternal care	27.18 ± 6.49	28.67 ± 5.30	24.20 ± 7.63	0.002*
PBI maternal overprot	16.18 ± 6.65	15.32 ± 6.61	17.90 ± 6.49	0.082
PBI paternal care	22.76 ± 8.06	23.42 ± 7.68	21.43 ± 8.74	0.273
PBI paternal overprot	13.30 ± 7.09	13.35 ± 7.14	13.20 ± 7.11	0.925
ECR avoidance	44.37 ± 18.26	41.40 ± 17.59	50.30 ± 18.43	0.028*
ECR anxiety	66.17 ± 19.85	63.43 ± 19.58	71.63 ± 19.56	0.064

	All subjects (mean ± SD)	HC (mean ± SD)	FHR (mean ± SD)	HC vs. FHR p value
C. Validation familial risk sample
PBI maternal care	26.33 ± 6.71	25.65 ± 6.88	27.69 ± 6.41	0.368
PBI maternal overprot	15.44 ± 7.62	17.27 ± 7.01	11.77 ± 7.71	0.032*
PBI paternal care	23.69 ± 6.53	24.50 ± 6.17	22.08 ± 7.18	0.280
PBI paternal overprot	14.28 ± 8.17	15.35 ± 7.46	12.15 ± 9.39	0.255
ECR avoidance	45.21 ± 16.42	43.88 ± 14.12	47.85 ± 20.67	0.542
ECR anxiety	61.51 ± 17.81	61.96 ± 16.63	60.62 ± 20.65	0.827

Significant between-groups differences (p < 0.05) are marked with (*)
PBI = Parental Bonding Instrument; ECR = Experiences in Close Relationships; overprot = overprotection; HC = Healthy Controls; PSY = Patients with Psychosis; ESD = Early Stages of Disease; FHR = Familial High Risk; SD = standard deviation

ISSN: 2050-7283