A computational psychiatry approach identifies how alpha-2A noradrenergic agonist Guanfacine affects feature-based reinforcement learning in the macaque

Abstract

Noradrenaline is believed to support cognitive flexibility through the alpha 2A noradrenergic receptor (a2A-NAR) acting in prefrontal cortex. Enhanced flexibility has been inferred from improved working memory with the a2A-NA agonist Guanfacine. But it has been unclear whether Guanfacine improves specific attention and learning mechanisms beyond working memory, and whether the drug effects can be formalized computationally to allow single subject predictions. We tested and confirmed these suggestions in a case study with a healthy nonhuman primate performing a feature-based reversal learning task evaluating performance using Bayesian and Reinforcement learning models. In an initial dose-testing phase we found a Guanfacine dose that increased performance accuracy, decreased distractibility and improved learning. In a second experimental phase using only that dose we examined the faster feature-based reversal learning with Guanfacine with single-subject computational modeling. Parameter estimation suggested that improved learning is not accounted for by varying a single reinforcement learning mechanism, but by changing the set of parameter values to higher learning rates and stronger suppression of non-chosen over chosen feature information. These findings provide an important starting point for developing nonhuman primate models to discern the synaptic mechanisms of attention and learning functions within the context of a computational neuropsychiatry framework.

Double Dissociation of Nicotinic α7 and α4/β2 Sub-receptor Agonists for Enhancing Learning and Attentional Filtering of Distraction

“…norepinephrine receptors could be one route α7 could improve learning. Strong support for this scenario comes from a previous study in which we found a similar faster reversal learning performance at optimal doses of the α-2A noradrenergic agonist Guanfacine (Hassani SA et al 2017). In that study, an increased reversal learning speed was the most prominent behavioral effect during prolonged testing at effective doses.…”

“…we computed the learning curves for each reversal block using an ideal observer statistics that calculated how consistently the monkeys choices were rewarded across sequences of trials (see Methods, and (Hassani SA et al 2017)). We then fit the resulting learning curves with a hyperbolic ratio function to extract the trial number at which the learning reached 50% of the maximum asymptotic performance value, which could be early or late, corresponding to fast versus slow learning ( Fig.…”

Feature Specific Prediction Errors and Surprise across Macaque Fronto-Striatal Circuits during Attention and Learning

“…Figure S1A, S3A, B). To discern trial-by-trial encoding of prediction errors we fitted an attention weighting reinforcement learning model to the choice data of the monkeys (Figure 1D, see Methods) (Hassani et al, 2017;Leong et al, 2017;Wilson and Niv, 2011). We then correlated the model derived reward prediction errors (RPEs) with the neural firing rates during the 0-1.5 sec.…”

“…Biasing attention to those features causing outcomes that have been most unexpected can optimize attentional sampling in the long run to those stimuli with the most reward-predictive features (Daddaoua et al, 2016;Dayan et al, 2000;Ghazizadeh et al, 2016). The mechanisms underlying this attentional optimization through reinforcement learning have been explored in recent studies suggesting that attentional guidance by prediction errors is facilitated when value predictions are already biased towards those feature dimensions that are most likely reward predictive (Hassani et al, 2017;Leong et al, 2017;Niv et al, 2015;Wilson and Niv, 2011). Instead of attending all possible feature dimensions of a stimulus equally, prioritizing those dimensions that most prominently are reward predictive dramatically enhances the learning speed (Farashahi et al, 2017;Kruschke and Hullinger, 2010).…”

Multi-Neuromodulator Measurements across Fronto-Striatal Network Areas of the Behaving Macaque using Solid-Phase Microextraction

“…The animals were engaged in an over-trained 130 attention tasks in which they would use saccadic eye movements to acquire juice reward (Fig S1). 131 The specifics of the task are described elsewhere (Hassani et al 2017). Both animals showed stable 132 performance and acquired similar reward volumes on all recording days.…”