AbstractAnxiety disorders are among the most common mental illness in the western world with a major impact on disability. But their diagnosis has lacked objective biomarkers. We previously demonstrated a human anxiety process biomarker, goal-conflict-specific electroencephalography (EEG) rhythmicity (GCSR) in the stop-signal task (SST). Here we have developed and characterized an improved test appropriate for clinical group testing. We modified the SST to produce balanced numbers of trials in clearly separated stop-signal delay groups. As previously, right frontal (F8) GCSR was extracted as the difference in EEG log Fourier power between matching stop and go trials (that is, stop-signal-specific power) of a quadratic contrast of the three delay values (that is, power when stopping and going are in balanced conflict compared with the average of when stopping or going is greater). Separate experiments assessed drug sensitivity (n=34) and personality relations (n=59). GCSR in this new SST was reduced by three chemically distinct anxiolytic drugs (administered double-blind): buspirone (10 mg), triazolam (0.25 mg) and pregabalin (75 mg); had a frequency range (4–12 Hz) consistent with rodent model data; and positively correlated significantly with neuroticism and nonsignificantly with trait anxiety scores. GCSR, measured in our new form of the SST, should be suitable as a biomarker for one specific anxiety process in the testing of clinical groups and novel drugs and in the development of measures suitable for individual diagnosis.