Impairments in emotional reactivity and behavioural flexibility are pervasive across disparate psychiatric conditions as traditionally defined. Here, I provide new evidence on how these processes are altered by neuromodulators in humans, with a primary focus on serotonin (5- HT; 5-hydroxytryptamine). Emotional reactions prepare the body for action. Some emotion is primitive, implicit, and critical for surviving threats, yet can inappropriately persist in times of safety. Other emotions are more complex, self-conscious and important in maintaining harmonious interpersonal relationships. At the same time, learned behaviours that are adaptive in the first instance, may become irrelevant or even disadvantageous as circumstances change. In Chapters 3 through 6, I report on experiments in healthy human volunteers that employed the dietary technique acute tryptophan depletion (ATD). ATD temporarily lowers serotonin synthesis and release by depleting its biosynthetic precursor tryptophan. Chapter 3 is a study of self-reported social emotion. ATD enhanced emotion in response to social injustice non-specifically; however, consideration of personality traits revealed that highly empathic participants reported more guilt under ATD, whereas individuals high in trait psychopathy demonstrated more annoyance. Chapter 4, in contrast, considers evolutionarily ancient automatic emotional reactions to threats. This was assayed instead by an objective measure, the skin conductance response (SCR). Here, ATD conversely attenuated the retention of Pavlovian conditioned emotional memory to threat. Traits again influenced this response: individuals more intolerant of uncertainty displayed the greatest attenuation of emotional reactions. Chapter 5 both extends the studies on emotion and bridges to the remaining empirical work by investigating reversal learning, an index of cognitive flexibility, in two experiments. Individuals again underwent Pavlovian (stimulus-outcome) threat conditioning, whereby one stimulus predicted threat, and another was safe. These contingencies then swapped (reversed). In a separate experiment, participants underwent instrumental (stimulus-response-outcome) conditioning on a deterministic schedule (the correct option was always correct), followed by reversal of the contingencies. ATD impaired both Pavlovian and instrumental reversal learning. Chapters 6 through 8 instead examine instrumental reversal learning that was probabilistic (the correct option was correct most but not all of the time), rather than deterministic. Chapter 6 expands on previous ATD studies of probabilistic reversal learning (PRL) in the literature, which had not found effects on choice behaviour. Despite nearly tripling the sample size, behaviour here assessed by conventional methods was unaffected, replicating previously published null results. Applying reinforcement learning (RL) models, however, revealed ATD elevated a basic perseverative tendency, referred to as “stimulus stickiness”; behaviour was more stimulus-bound and insensitive to the outcome of actions, consistent with the deterministic instrumental reversal impairment following ATD. Chapters 7 and 8 apply RL models as well, to existing datasets on PRL for comparison. Chapter 7 shows that healthy volunteers under lysergic acid diethylamide (LSD), which acts both at serotonin but also dopamine receptors, showed enhanced learning from positive feedback in particular, which was related to perseveration. Chapter 8 applies computational methods to PRL in clinical populations. RL modelling revealed a computational signature that dissociated PRL in stimulant use disorder (SUD) and obsessive-compulsive disorder (OCD): Individuals with SUD showed heightened stimulus stickiness, as occurred following ATD in healthy volunteers, whereas the OCD group (under serotonergic medication) demonstrated lower stimulus stickiness than healthy controls. Dopaminergic agents remediated a reward learning deficit in SUD, among other measures. The general discussion considers these various findings in terms of theories of central serotonin function, in relation to the animal literature, and its relevance to mental disorder. These results, collectively, advance knowledge of neurochemical and computational mechanisms underlying psychiatric conditions trans-diagnostically, with implications for revised psychiatric classifications in line with the Research Domain Criteria (RDoC).