Counteracting uncertainty: exploring the impact of anxiety on updating predictions about environmental states

Biological Cybernetics (2025) 119:8 https://doi.org/10.1007/s00422-025-01006-4 ORIGINAL ARTICLE Counteracting uncertainty: exploring the impact of anxiety on updating predictions about environmental states David Harris1 · Tom Arthur1 · Mark Wilson1 · Ben Le Gallais1 · Thomas Parsons1 · Ally Dill1 · Sam Vine1 Received: 28 June 2024 / Accepted: 28 January 2025 © The Author(s) 2025 Abstract Anxious emotional states disrupt decision-making and control of dexterous motor actions. Computational work has shown that anxiety-induced uncertainty alters the rate at which we learn about the environment, but the subsequent impact on the predictive beliefs that drive action control remains to be understood. In the present work we tested whether anxiety alters predictive (oculo)motor control mechanisms. Thirty participants completed an experimental task that consisted of manual interception of a projectile performed in virtual reality. Participants were subjected to conditions designed to induce states of high or low anxiety using performance incentives and social-evaluative pressure. We measured subsequent effects on physiological arousal, self-reported state anxiety, and eye movements. Under high pressure conditions we observed visual sampling of the task environment characterised by higher variability and entropy of position prior to release of the projectile, consistent with an active attempt to reduce uncertainty. Computational modelling of predictive beliefs, using gaze data as inputs to a partially observable Markov decision process model, indicated that trial-to-trial updating of predictive beliefs was reduced during anxiety, suggesting that updates to priors were constrained. Additionally, state anxiety was related to a less deterministic mapping of beliefs to actions. These results support the idea that organisms may attempt to counter anxiety-related uncertainty by moving towards more familiar and certain sensorimotor patterns. Keywords Gaze · Stress · Eye tracking · Bayesian · Predictive Processing 1 Introduction Both dextrous motor actions and control of our visual system are thought to depend on predictions about future states of the world and our own body (Wolpert and Flanagan 2001; Shadmehr et al. 2010; Adams et al. 2013). Skilled movement also depends on the ability to flexibly adapt and update those predictions according to new contexts or new sensory evidence. As anxious emotional states can bias the way in which we make and update predictions (Cornwell et al. 2017; Hein et al. 2021; Hein and Herrojo Ruiz 2022), they can disrupt motor actions with potentially damaging consequences (Harris et al. 2023b). It is well known that during highly pressurised situations, such as a job interview or the final of a sporting event, Communicated by Benjamin Lindner. B 1 David Harris School of Public Health and Sport Sciences, Medical School, University of Exeter, St Luke’s Campus, Exeter EX1 2LU, UK people can experience drastic breakdowns in task performance (Beilock and Carr 2001; Nieuwenhuys and Oudejans 2012; Payne et al. 2018). In the present work we sought to understand the impact of anxiety on predictive (oculo)motor control mechanisms. We first outline theoretical approaches that have described anxiety in terms of uncertainty or entropy and how sensorimotor control might reflect an active attempt to resolve that uncertainty. Anxiety is a negative emotional response to a perceived threat (Eysenck 2013; Grupe and Nitschke 2013). It is often characterised along cognitive (worry) and somatic (physiological arousal) dimensions. Anxiety can lead to distinct difficulties learning about the world and making decisions (Bishop 2007; Carleton 2016), and can particularly disrupt the ability to adapt to changing task conditions, especially in individuals with an intolerance of uncertainty (Browning et al. 2015; Huang et al. 2017; Hein et al. 2021). One potential explanation for these effects is that anxiety may alter the way in which we make predictions about the world around us. For instance, several theoretical accounts have characterised anxiety as closely related to situational estimates of uncertainty 0123456789().: V,-vol 123 8 Page 2 of 16 and a (perceived) inability to reliably predict the absence of threats (Grupe and Nitschke 2013; de Berker et al. 2016; Cornwell et al. 2017; Seriès 2019; Lawson et al. 2021). For instance, Hirsh et al.’s. (2012) Entropy Model of Uncertainty (EMU) describes uncontrollable or unpredictable situations as creating an aversive high-entropy state, in which an organism experiences a reduced ability to predict successive states (e.g., sensory outcomes) based on the current state. According to the EMU, organisms find such uncertainty aversive and experience it subjectively as anxiety. Research in this vein therefore conceptualises anxiety as an epistemic emotion, that is arising from our engagement with knowledge, learning, or the need to reduce uncertainty (Miceli and Castelfranchi 2005). The EMU characterisation of anxiety aligns closely with other dynamical systems approaches like the Free Energy Principle (FEP) and Active Inference, which similarly conceptualise the goal of a cognitive-behavioural system as the minimization of internal entropy (or ‘free energy’) and an increase in prediction success (Friston 2009; Friston et al. 2010; Clark 2013). From this perspective, to cope with instability (e.g., during anxiety) organisms should seek to return to familiar low-entropy states. Under the FEP (Friston 2010), this tendency to return to low entropy states is, by definition, true of a self-organising system as they must resist dissipative forces (i.e., states of high entropy) to maintain their own integrity and existence. One way to achieve this goal in the face of anxiety-induced uncertainty, is to adopt behaviours like withdrawal or avoidance to evade the aversive stimulus. Alternatively, the organism may seek out new perceptual information that will disambiguate any uncertainty (e.g., looking under the bed to check for monsters). These uncertainty-reducing behaviours can sometimes be detrimental, such as when a sportsperson’s movements become rigid and constrained (Harris et al. 2023b), but they may also be beneficial, enabling faster belief updating (e.g., Behrens et al. 2007)), or neutral, as in cases of task-irrelevant uncertainty reduction (e.g., Miceli and Castelfranchi 2005). It has previously been suggested that an adaptive response in times of uncertainty is to modify the rate with which you update your beliefs about the world; learning faster allows one’s model of the world to better reflect the current ‘reality’ of an environment (Behrens et al. 2007). Neurocomputational work has illustrated that this is indeed the case for perceptual learning (Lawson et al. 2021) where learning rates (i.e., the speed at which we revise prior expectations) appear to be increased by noradrenergic up-regulation of prediction error signal (...truncated)