Abstract
Anxiety disorders are not only a public health concern but also a major economic burden. Globally, anxiety disorders affect around 301 million people, covering 4.05% of the total population. The estimated annual direct cost of managing these healthcare burdens is $6.5 trillion (2.08% of all healthcare-related expenditures), with an additional $5.4 trillion for indirect costs. Despite the prevalence of anxiety disorders, the current diagnostic system is largely subjective and lacks objective biomarkers. McNaughton and his group previously developed an anxiolytic-sensitive biomarker; Goal-Conflict-Specific Rhythmicity (GCSR) derived from EEG recording in the Stop Signal Task (SST). However, GCSR in the SST lacks test-retest stability and does not have effective control of approach-avoidance tendencies. Therefore, it cannot be used in the clinic as a reliable diagnostic tool. Our aim was to overcome reliability issues raised in the SST by introducing control of approach-avoidance tendencies into newly modified one-way and two-way virtual predator escape tasks. In the escape tasks, behavioural conflict was calculated by contrasting escape responses time differences, and GCSR was extracted as the difference in EEG power between early and late virtual predators. Behavioural conflict and GCSR (low; 3-7Hz, high; 9-15Hz) were examined and their drug sensitivities were assessed for both low and high reward conditions. Three specific anxiolytics (buspirone, pregabalin and triazolam) and anxiolytic panicolytics (fluoxetine/citalopram) were administered in two double-blind experiments. Drug sensitivity was consistently found at the usual right frontal site for the high (9-15Hz) frequency range in both the low and high reward conflict measures. However, drugs lacked effect on behavioural measures. Importantly, the anxiolytic panicolytic fluoxetine shared its sensitivity pattern with other specific anxiolytics, whereas citalopram differed left-frontally (F3). Furthermore, the combined analysis of the two experiments failed to replicate EEG baseline conflict measures. Separately, a test-retest reliability experiment showed conflict variability across weekly sessions, respectively. Therefore, a two-way escape task – motivated by the “omission of punishment” rather than presentation of monetary reward – was tested as a more balanced test for the GCSR theory. The two-way task elicited predictable behavioural and electrophysiological results, and it largely found that trait anxiety significantly influences both the behavioural expression of conflict and its neural underpinnings. Future validation of the two-way escape task including reliability check and pharmacological challenges, may establish it as an effective behavioural and neural biomarker for clinical translation and treatment evaluation of anxiety disorders. However, our results and those of others suggest that further paradigm development may be required. Particularly, modifications to key task components (e.g., predator features, reward cues or threats) could be targeted to improve testing reliability across multiple sessions, enhancing pharmacological sensitivity and better integration between behavioural and conflict-related neural activity.