Table 2.

Associations between schizophrenia-associated CNVs and CNV burden with cognition.

Predictor	Cognitive measure	Participants (CNV Carrier)	Regression coefficient	95% CI	Sig.
Schizophrenia-associated CNVs	Block design	2747 (21)	−10.1	−19.2, −0.9	0.031
	Digit span	–	–	–	–
	RAVLT immediate Recall	1906 (24)	−8.0	−13.3, −2.6	0.0036
	RAVLT delayed recall	1889 (24)	−3.3	−5.8, −0.7	0.0115
Genes affected by all CNVs (burden)	Block design	2747	−0.1	−0.3, 0.1	0.343
	Digit span	1270	0.03	−0.2, 0.2	0.775
	RAVLT immediate recall	1906	−0.03	−0.2, 0.1	0.634
	RAVLT delayed recall	1889	−0.01	−0.1, 0.1	0.854
Genes affected by deletions (burden)	Block design	2747	−0.4	−0.8, 0.01	0.056
	Digit span	1270	−0.1	−0.6, 0.3	0.544
	RAVLT immediate recall	1906	−0.2	−0.5, −0.1	0.0119
	RAVLT delayed recall	1889	−0.1	−0.2, −0.01	0.0661
Genes affected by duplications (burden)	Block design	2747	−0.001	−0.2, 0.2	0.992
	Digit span	1270	0.06	−0.1, 0.3	0.564
	RAVLT immediate recall	1906	0.02	−0.1, 0.2	0.774
	RAVLT delayed recall	1889	0.03	−0.04, 0.1	0.412

Figures in bold are below the multiple testing adjusted p-value threshold of 0.017.

Associations between known schizophrenia-associated CNVs and CNV burden with cognitive performance. For the schizophrenia-associated CNV analysis digit span was not examined as fewer than ten CNV carriers had available data. CNV burden was measured as number of genes affected by CNVs larger than >200 kb, with <1% frequency. All analyses are adjusted for the covariates age, sex, clinical group, centre and genetic relatedness (kinship matrix).