Bayesian Brain in ADHD: potential catecholaminergic pathway of volatility estimation

The Bayesian Brain framework posits that the brain continuously updates its beliefs about a dynamic world by inferring hidden states, including multi-level volatility. The link between volatility and norepinephrine, a neurotransmitter which is imbalanced in attention-deficit/hyperactivity disorder (ADHD), warrants an investigation of the precise nature of volatility estimation in ADHD, which remains unexplored to date. Thus, the current study used a probabilistic associative learning (PAL) task to extract participant-specific belief trajectories and volatility estimates from adults with and without ADHD. We found only small evidence in favour of a difference in estimates for cue-outcome tonic volatility and even less evidence for differences in learning rate updates between adults with and without ADHD. Similarly, adults with and without ADHD showed comparable estimates for environmental volatility and response model parameters. Nonetheless, exploration revealed strong evidence for an increase in estimates for cue-outcome tonic volatility in adults with ADHD not currently taking ADHD medication. Furthermore, ADHD medication was associated with a decrease in estimates for cue-outcome tonic volatility, suggesting a causal relationship between the catecholaminergic imbalance in ADHD and cue-outcome tonic volatility processing. This provides crucial new evidence suggesting that the catecholaminergic imbalance in ADHD may specifically manifest as an altered processing of cue-outcome tonic volatility, offering a novel, computational target for understanding the disorder's neurobiology.