Bipolar disorder, mood stabilizers and cognitive flexibility: Translationally dissecting illness from drug effects.

Bipolar disorder (BD) effects on cognition are confounded by the putative cognitive impact of its major pharmacological treatments, given the neurotrophic potential of mood stabilizers, particularly lithium. We examined the area of cognitive flexibility (CF), aiming to disentangle BD from medication effects, using translational methodology. CF was assessed by CANTAB-IED (intra- extra-dimensional shift; Study 1, euthymic BD participants) and its animal analog (Study 2, rats). Both studies included groups (1) control, (2) lithium, chronic, current treatment (LI-CHRON-C, A: > 2 years, N = 32; B: 2 months, N = 11); (3) valproate, chronic, current treatment (VPA-CHRON-C, A: > 2 years, N = 30; B: 2 months, N = 12). Study 2 included 2 additional groups; Group 4: LI-CHRON-PAST (2 months, stopped 1 month pretest, N = 13); Group 5: LI-ACUTE (LI on test days only, N = 13). In Study 1, neither total nor stage (discrimination: D; reversal R; intra- extra-dimensional shifts: IED) IED errors differed between groups [Kruskal-Wallis: H(2, N = 94) 0.95 > p > 0.65]. Similarly in Study 2, errors did not differentiate the 5 pharmacological groups. Differences emerged only between LI-ACUTE and Controls in response latencies (D, R, IED ANOVAS: 0.002 > p > 0.0003; contrasts D, R: p = 0.002, 0.0001). In conclusion, LI and VPA BD patients were indistinguishable from Controls in IED errors, as were animals treated with LI-CHRON, current or past, or VPA-CHRON-C vs Controls. LI-ACUTE treatment produced significant latency deficits vs Controls. Within the limitations of translational comparisons, our results suggest that the normal CF noted in euthymic BDs is not attributable to mood stabilizer effects.