Group Transdiagnostic Cognitive Behavioural Therapy for Intolerance of Uncertainty Versus Applied Relaxation

The PHQ-9 (Kroenke et al., 2001) is a 9-item self-report measure of depressive symptom severity. Scores range from 0 to 27; higher scores indicate greater severity.

The WSAS (Mundt et al., 2002) is a 5-item self-report measure of functional impairment attributed to mental health problems across work, home management, social leisure, private leisure, and close relationships. Items are rated from 0 to 8. Higher scores indicate greater impairment.

The PSWQ (Meyer et al., 1990) is a 16-item self-report measure of pathological worry validated in GAD populations. Higher scores indicate more pervasive and uncontrollable worry.

The ARDUS (Wang et al., 2026) is a 37-item self-report measure of how often individuals experience negative affect (worry, anxiety, low mood, or concern) when faced with uncertainty across six everyday domains: academic/professional, climate, financial, health, political, and social. Items are rated on a 6-point Likert scale from 0 (Never) to 5 (Always), with an additional option indicating that the type of uncertainty has not been encountered in the past year. Higher scores reflect greater affective reactivity to daily uncertainty.

The Transport Volatility Learning Task is a probabilistic associative learning task assessing how participants track action-outcome contingencies in environments where outcome probabilities change over time. On each trial, participants encounter one of two cues (train lines) probabilistically associated with a binary outcome (delay vs no delay), with one cue carrying a higher probability of the aversive outcome at any given time. The underlying contingency reverses periodically, and volatility (the rate of contingency change) is manipulated across blocks to dissociate stable from volatile learning environments. Two task variants are administered to distinguish observational from instrumental learning: a prediction-only variant in which the cue is determined by the environment and participants predict the outcome, and a choice variant in which participants actively choose between cues and outcomes follow from their selection.

The SELT is an associative learning task assessing how individuals estimate and update beliefs under social-evaluative uncertainty (adapted from Lau et al., 2024). On each trial, participants are presented with a positive and a negative word and rate, on a continuous slider, the probability that an avatar (the "rater") will select that word to describe the target; they then receive feedback indicating which word the rater chose. The task crosses two referential conditions (self-referential, in which the participant is the target of evaluation, vs other-referential, in which a third character is the target) with rule conditions defining the underlying probability of positive feedback. After each block, participants make a global judgement of the rater's overall evaluative tendency, and at the end of the task participants make forced choices about preferred social partners for themselves and for the other-referential target.

The BART is a decision-making task in which participants accumulate points by inflating a virtual balloon and choose when to stop and bank their earnings (Lejuez et al., 2002). Each pump carries an unobserved probability of the balloon bursting and forfeiting that trial's points. Across blocks, the underlying probability distribution of the explosion point is manipulated to dissociate distinct components of uncertainty, including the mean and variance of the distribution.

The IST is a task in which participants actively sample binary information to reduce uncertainty before committing to a categorical decision (Hauser et al., 2017). On each trial, participants are presented with a set of 25 hidden samples drawn from an underlying binary distribution and choose, one at a time, how many to reveal before deciding which of two categories the source distribution belongs to. A log of revealed samples remains on screen throughout. Two conditions are administered: a free-sampling condition in which sampling is unconstrained, and a costly-sampling condition in which each sample reveal incurs an explicit monetary cost against a fixed per-trial budget, with correct and incorrect decisions yielding gains and losses scaled to the number of samples taken.

The Uncertain Face Selection Task is an emotion-categorisation task in which participants view facial expressions varying in perceptual ambiguity, achieved by morphing between happy and angry prototypes. The task comprises two blocks. In a forced-choice baseline block, participants categorise each briefly presented face as happy or angry under speeded response conditions, providing a measure of perceptual sensitivity to the morphed emotion continuum. In a subsequent free-response block, participants are given two additional options on each trial: re-viewing the same face before deciding (a checking response) or skipping the face entirely without categorising it (an avoidance response), with skipping carrying no time benefit so that it cannot be motivated by efficiency.

The ACT is a perceptual decision-making paradigm assessing how participants interpret ambiguous cues with uncertain outcomes (adapted from Aylward et al., 2019). On each trial, participants are presented with a visual cue that lands at one of several orientations and must categorise it as belonging to one of two classes (horizontal vs vertical) under time pressure. The two classes are associated with asymmetric reward magnitudes, such that one categorisation yields a higher payoff than the other. Cue orientations span a continuum from clearly belonging to one class through fully ambiguous (intermediate) angles, allowing participants' classification of the ambiguous cues to be modelled as a function of perceptual evidence and prior expectations.

The TPRLT is a probabilistic reward learning paradigm presented in a rodent-themed visual environment to advance cross-species translation of computational measures. On each trial, participants control an on-screen rat avatar and choose between two spatially-distinct response ports (left vs right). One port is associated with a higher probability of reward (a piece of cheese) at any given time, with the rewarded port determined probabilistically; the more rewarded option changes across the task, requiring participants to track and update reward contingencies through trial-and-error. Feedback (reward delivery vs no reward) is presented on each trial in a designated feedback zone.

In the Aversive Learning Task participants choose between two visually-distinct options (street images) on each trial. One option carries a higher probability of an aversive event (a pickpocketing incident resulting in a monetary loss), and participants must learn which option is currently safer through trial-and-error feedback. The safer option changes periodically across the task, so participants must continuously update their beliefs. Participants begin with an initial endowment, and accumulated losses are deducted from this balance. Two referential conditions are administered: a self-referential condition in which choices determine the participant's own outcome, and an other-referential condition in which the participant chooses on behalf of another participant whose monetary bonus is determined by the participant's performance, and reciprocally receives choices made on their behalf. This distinguishes between aversive learning for self from aversive learning for other.

The BFNE (Leary, 1983) is a 12-item self-report measure of apprehension about being evaluated unfavourably by others. Items are rated from 1 (not at all characteristic of me) to 5 (extremely characteristic of me). Higher scores indicate greater fear of negative evaluation.

The TEC (Nijenhuis et al., 2002) is an inventory of 29 potentially traumatic events across the lifespan, spanning emotional neglect, emotional abuse, physical abuse, sexual harassment, sexual abuse, bodily threat, loss, and war-related experiences. Each event is rated for occurrence, age at occurrence, and subjective impact from 1 to 5, yielding indices of trauma frequency and severity.

The DERS (Gratz \& Roemer, 2004) is a 36-item self-report measure of problems in emotion regulation. Items are rated from 1 to 5 across six subscales covering nonacceptance of emotional responses, difficulty engaging in goal-directed behaviour, impulse control difficulties, lack of emotional awareness, limited access to regulation strategies, and lack of emotional clarity. Higher scores indicate greater difficulties.

The ERS (Nock et al., 2008) is a 21-item self-report measure of the sensitivity, intensity, and persistence of emotional responses. Items are rated from 0 to 4. Higher scores indicate greater emotional reactivity.

The RRS-10 (Treynor et al., 2003) is a 10-item self-report measure of ruminative responses to low mood. Items are rated from 1 to 4 and yield brooding and reflection subscale scores. Higher scores indicate greater rumination.

The BALCI (Radomsky \& Gagné, 2020) is a 21-item self-report measure of negative beliefs about losing control over one's thoughts, behaviour, emotions, and bodily functions, and about the importance of staying in control. Items are rated from 0 to 4. Higher scores indicate stronger beliefs about losing control.

The WBI (Mahoney et al., 2016) is a self-report measure of behaviours used to prevent, control, or avoid worry, including checking, reassurance seeking, avoidance, and overpreparation. Items are rated from 0 to 4. Higher scores indicate greater use of worry-related behaviours.

The SCCS (Campbell et al., 1996) is a 12-item self-report measure of the degree to which self-beliefs are clearly defined, internally consistent, and temporally stable. Items are rated from 1 to 5. Higher scores indicate greater self-concept clarity.

The ETMCQ (Campbell et al., 2021) is a 15-item self-report measure of an individual's stance towards accepting knowledge communicated by others. Items are rated from 1 to 7 and yield three subscale scores for trust, mistrust, and credulity.

The ECR-RS (Fraley et al., 2011) is a 9-item self-report measure of adult attachment. Items are rated from 1 to 7 and yield scores on two dimensions, attachment-related avoidance and attachment-related anxiety. Higher scores indicate greater attachment insecurity on each dimension. Here it is administered with reference to close relationships in general.

The BCSS (Fowler et al., 2006) is a 24-item self-report measure of core evaluative beliefs about the self and others. Respondents endorse each belief (no or yes) and, if endorsed, rate conviction from 1 to 4, yielding four subscale scores for negative-self, positive-self, negative-other, and positive-other schemas.

MAIA-2 (Mehling et al., 2018) is a self-report measure of interoceptive body awareness. In the 24-item form used here, items are rated from 0 to 5, spanning facets including noticing, attention regulation, emotional awareness, self-regulation, body listening, and trusting. Higher scores indicate greater interoceptive awareness.

The GRACES-Q (Shalev et al., unpublished) is a 25-item self-report measure of reaction and grief-like responses to changes to the self following the onset of a mental health condition. After identifying the earliest condition, its time of onset, and two open-ended reflections on the self before and after onset, respondents rate statements from 0 to 5 across five subscales covering identity discontinuity and self-alienation, longing and grieving the past self, grieving the future self, acceptance and moving forward, and dual-process oscillation.

The PSST (Steiner et al., 2003) is a 19-item self-report screening measure of premenstrual symptoms, comprising 14 items covering emotional, behavioural, and physical symptoms and 5 items covering interference with daily functioning. Each item is rated from 1 (not at all) to 4 (severe).