How arousal shapes performance

How does arousal shape performance, and how flexible is this relationship across different contexts? In this thesis, I investigated the link between arousal and perceptual decision-making in humans, combining behavioral experiments, computational modeling, and pharmacological manipulations. I focused in particular on the idea, formalized by the Yerkes-Dodson law, that performance is optimal at intermediate arousal levels.
In Chapter 2, we found that performance peaked at intermediate levels of pupil-linked arousal across different tasks and modalities. A neurobiologically plausible model showed how this inverted U-shaped relationship could arise from arousal-driven disinhibition of excitatory populations. In Chapter 3, we examined the neurochemical basis of this relationship by manipulating catecholaminergic and cholinergic neuromodulation. Increased catecholaminergic activity raised arousal but preserved the inverted U-shape, suggesting that the brain flexibly adapts to fluctuations within arousal states. A refined model captured this effect via a catecholamine-sensitive population that shifted the arousal-performance curve. In Chapter 4, we addressed a potential confound: in the visual domain, pupil size itself can influence perception by modulating visual input. By showing that pupil size itself does not affect auditory decision-making, we demonstrated that auditory tasks provide a robust approach for studying pupil-linked arousal in isolation from such sensory confounds. Chapter 5 reports a survey among consciousness researchers, revealing a field marked by theoretical and methodological diversity and little consensus on foundational questions.
Together, these findings offer new insights into how arousal modulates performance and how the brain flexibly adapts to its internal state.