Table 3.

Predicting 6-month symptoms/behavior controlling for baseline scores.

	UPPS-P total	MASQ-AD	YMRS	HAM-D	HAM-A	Negative urgency	Lack of premeditation	Lack of perseverance	Sensation seeking	Positive urgency
Gender	*	*	*	*	*
Age	*	*	−0.62 (0.25)	*	*
Education	*	*	*	*	*
Medication load change	*	*	*	*	*
Baseline symptoms/behaviors	0.98 (0.0000)	0.73 (0.0002)	0.37 (0.18)	0.59 (0.001)	0.53 (0.002)	0.83 (0.0000)	0.70 (0.0000)	0.98 (0.0000)	1.11 (0.0000)	0.91 (0.0000)
R Amygdala anger	*	*	*	*	−9.7 (0.057)
L Amygdala anger	*	0.75 (0.18)	*	*	*
L vlPFC anger	*	0.80 (0.16)	*	*	*
L vmPFC anger	*	0.47 (0.12)	*	*	*
R Amygdala fear	*	*	−4.7 (0.055)	*	−9.5 (0.081)
L Amygdala fear	*	*	*	*	*
L vlPFC fear	−0.34 (0.27)	*	*	*	*
L vmPFC fear	*	*	*	*	*
R Amygdala sad	*	*	*	*	*
L Amygdala sad	0.50 (0.017)	0.55 (0.23)	*	*	*	0.41 (0.17)	0.049 (0.86)	0.51 (0.098)	0.34 (0.27)	0.38 (0.29)
L vmPFC sad	*	0.50 (0.078)	*	*	*
R Amygdala happy	*	*	*	*	*
L Amygdala happy	0.30 (0.025)	*	*	*	*

Beta and FDR adjusted p values in linear regression models are reported for nonzero coefficients from the elastic net regression models. Asterisks indicate zero coefficients from the elastic net regression models. Each of the five impulsivity/other symptom models was run separately. FDR correction for multiple tests was performed across self-report (impulsivity and anhedonia) and clinician-rated symptoms (mania/hypomania, depression, and anxiety) (indicated by asterisks and values given in italics). Each neural measure was tested separately controlling for demographic variables. For significant relationships between neural activity and UPPS-P total score, relationships with UPPS-P subscales were examined, controlling for those demographic variables that were related to UPPS-P total score.