Randomness as a behavioral marker of attentional control during the attention training technique and its neural...

The Attention Training Technique (ATT) is a core intervention of metacognitive therapy (MCT) effective in reducing repetitive negative thinking by engaging the lateral prefrontal cortex (PFC). The current study tests whether a parallel random number generation (RNG) task can serve as a behavioral marker of attentional engagement.

We conducted two studies using the fMRI-adapted ATT paradigm combined with a parallel RNG task. In the behavioral study (N = 36), we compared two task versions with short (21 presses) versus long (45 presses) trials. In the fMRI study (N = 42), the ATT paradigm was performed with either parallel RNG or a motor control task. Markers of randomness (i.e., Redundancy, Repetition Gap, and Coupon scores) were compared between ATT conditions. Trait AC was measured using the Attentional Control Scale (ACS).

Randomness was reduced during ATT compared to control trials in both studies (p <.021). Additional analyses showed that 1) across conditions, two of the RNG scores correlated with indices from a standard RNG task (p <.029); 2) ATT-control differences in RNG performance were detected in short and long trials of the ATT paradigm; and 3) ATT-control differences in Redundancy were associated with self-reported trait AC (p = .039). Further, RNG selectively activated regions within the dorsal attention network (p_FWE  <.05) and dual-task interference attenuated ATT-related task-network activation.

These findings establish RNG as a feasible and sensitive behavioral marker of ATT engagement and offer the opportunity to link behavioral performance during ATT to individual differences in AC. However, dual-task demands limit the use of RNG as a behavioral marker, particularly in fMRI which calls for alternative, behavior-independent measures of attentional effort.