Amphetamine modulates brain signal variability and working memory in younger and older adults

Abstract

Better-performing younger adults typically express greater brain signal variability relative to older, poorer performers. Mechanisms for age and performance-graded differences in brain dynamics have, however, not yet been uncovered. Given the age-related decline of the dopamine (DA) system in normal cognitive aging, DA neuromodulation is one plausible mechanism. Hence, agents that boost systemic DA [such as d-amphetamine (AMPH)] may help to restore deficient signal variability levels. Furthermore, despite the standard practice of counterbalancing drug session order (AMPH first vs. placebo first), it remains understudied how AMPH may interact with practice effects, possibly influencing whether DA up-regulation is functional. We examined the effects of AMPH on functional-MRI-based blood oxygen level-dependent (BOLD) signal variability (SD(BOLD)) in younger and older adults during a working memory task (letter n-back). Older adults expressed lower brain signal variability at placebo, but met or exceeded young adult SD(BOLD) levels in the presence of AMPH. Drug session order greatly moderated change-change relations between AMPH-driven SD(BOLD) and reaction time means (RT(mean)) and SDs (RT(SD)). Older adults who received AMPH in the first session tended to improve in RT(mean) and RT(SD) when SD(BOLD) was boosted on AMPH, whereas younger and older adults who received AMPH in the second session showed either a performance improvement when SD(BOLD) decreased (for RT(mean)) or no effect at all (for RT(SD)). The present findings support the hypothesis that age differences in brain signal variability reflect aging-induced changes in dopaminergic neuromodulation. The observed interactions among AMPH, age, and session order highlight the state- and practice-dependent neurochemical basis of human brain dynamics.

Keywords: aging; brain signal variability; dopamine; fMRI; working memory.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

PLS model of relation between SDBOLD, Age Group, AMPH, and Task Condition. Higher brain scores reflect higher BOLD signal variability. Error bars represent bootstrapped 95% confidence intervals (1,000× with replacement). Brain images are plotted in neurological orientation (left is Left). AMPH, amphetamine; BSR, bootstrap ratio.

Fig. 2.

Relation between AMPH-related changes in SDBOLD PLS brain scores (SDBOLD_within) and (i) AMPH-related changes in reaction time means (RTmean_within; Upper) and (ii) SDs (RTSD_within; Lower), moderated by age group and drug session order. Each subject in each group is represented by three points in the scatters, one for each n-back condition. Positive SDBOLD values indicate higher BOLD signal variability on AMPH relative to placebo. AMPH, amphetamine. Simple slope P values: (a) P = 0.039; (b) P = 0.672; (c) P = 0.005; (d) P = 0.412; (e) P = 0.001; (f) P = 0.0002; (g) P = 0.682; (h) P = 0.311.