Table 2.

Association between region of interest connectivity strength and clinical effectiveness

Region of interest	Discriminant power	By chance probability	P-value	OR (95% CI)
Superior frontal gyrus (n = 18)	0.89	0.64	0.014*	6.4 (1.3–30.6)
Thalamus (n = 20)	0.80	0.64	0.085	2.9 (0.9–9.6)
Substantia nigra (n = 19)	0.58	0.64	0.813	0.69 (0.2–2.0)
Brainstem (n = 20)	0.45	0.64	0.979	0.35 (0.1–1.0)

Discriminant power was defined as a proportion of trials that successfully identified a contact with clinical efficacy, by selecting the contact with the highest connectivity to each relevant region of interest, within each lead. n represents the number of leads included in the analysis; leads without connectivity to the region of interest were excluded. Significance defined as *P < 0.05. P-value was obtained by testing the null hypothesis that contacts identified by the discriminant analysis were selected by chance. The null distribution was Poisson binomial with success probabilities of either 0.75, 0.5 or 0.25, depending on the proportion of effective contacts in a lead. OR of association using electrode connectivity strength as a measure of exposure and clinical effectiveness or ineffectiveness as the outcome.