Chronic escitalopram in healthy volunteers has specific effects on reinforcement sensitivity: a double-blind, placebo-controlled semi-randomised study

Abstract

Several studies of the effects on cognition of selective serotonin reuptake inhibitors (SSRI), administered either acutely or sub-chronically in healthy volunteers, have found changes in learning and reinforcement outcomes. In contrast, to our knowledge, there have been no studies of chronic effects of escitalopram on cognition in healthy volunteers. This is important in view of its clinical use in major depressive disorder (MDD) and obsessive-compulsive disorder (OCD). Consequently, we aimed to investigate the chronic effect of the SSRI, escitalopram, on measures of ‘cold’ cognition (including inhibition, cognitive flexibility, memory) and ‘hot cognition’ including decision-making and particularly reinforcement learning. The study, conducted at the University of Copenhagen between May 2020 and October 2021, used a double-blind placebo-controlled design with 66 healthy volunteers, semi-randomised to receive either 20 mg of escitalopram (n = 32) or placebo (n = 34), balanced for age, sex and intelligence quotient (IQ) for at least 21 days. Questionnaires, neuropsychological tests and serum escitalopram measures were taken. We analysed group differences on the cognitive measures using linear regression models as well as innovative hierarchical Bayesian modelling of the Probabilistic Reversal Learning (PRL) task. The novel and important finding was that escitalopram reduced reinforcement sensitivity compared to placebo on both the Sequential Model-Based/Model-Free task and the PRL task. We found no other significant group differences on ‘cold’ or ‘hot’ cognition. These findings demonstrate that serotonin reuptake inhibition is involved in reinforcement learning in healthy individuals. Lower reinforcement sensitivity in response to chronic SSRI administration may reflect the ‘blunting’ effect often reported by patients with MDD treated with SSRIs. Trial Registration: NCT04239339.

Common roles for serotonin in rats and humans for computations underlying flexible decision-making

“…The punishment learning rate was increased for both repeated 5 mg/kg citalopram and sub-chronic citalopram in rats only. Reinforcement sensitivity was increased following 10 mg/kg of citalopram and decreased during sub-chronic treatment in rats, agreeing with our own recent analysis of chronic escitalopram treatment in humans 57 , although this parameter was also shown to be decreased in the present analysis following acute 20mg escitalopram in humans.…”

“…Langley et al . 57 have recently shown diminished reinforcement sensitivity in healthy humans following chronic – at least 21 days – of 20 mg escitalopram performing the same PRL task and modelled in an identical fashion – this reduction is hence the same direction as for the acute dose in humans and sub-chronic dosing in rats. Although this parallel between single and chronic dosing in humans was unexpected, it is notable that reinforcement sensitivity in rats following sub-chronic dosing was also decreased.…”

“…The participants were semi-randomized into the treatment group ( n = 32) receiving 20 mg escitalopram or the control group receiving the placebo for 3 to 5 weeks. The conventional analysis identified no significant group differences 57 . After computational modelling, we found that the chronic use of SSRI reduced reinforcement sensitivity compared to placebo ( n = 34) in healthy volunteers (MD = -2.7673 [90% HDI, −5.2846 to −0.3959]), but had no effect on reward/punishment learning rates or stimulus stickiness (0 ∈ 75% HDI) 57 .…”

“…Whilst the effects of serotonin on reinforcement sensitivity revealed by the present analyses were ostensibly more difficult to interpret -underscoring that stickiness is a distinct mechanism -there is an intriguing parallel with a recent study. Langley et al 57 have recently shown diminished reinforcement sensitivity in healthy humans following chronic -at least 21 days -of 20 mg escitalopram performing the same PRL task and modelled in an identical fashion -this reduction is hence the same direction as for the acute dose in humans and subchronic dosing in rats. Although this parallel between single and chronic dosing in humans was unexpected, it is notable that reinforcement sensitivity in rats following sub-chronic dosing was also decreased.…”

“…As reported in our recent publication for the effect of chronic use of SSRI on behavioral flexibility by a double-blind, placebo-control, semi-randomized study 57 , the computational modelling approach was applied to the behavioral data of the same probabilistic reversal learning task in healthy volunteers. The participants were semi-randomized into the treatment group ( n = 32) receiving 20 mg escitalopram or the control group receiving the placebo for 3 to 5 weeks.…”