Measuring efficacy and behavioural actions of new antidepressants (ADs) is greatly enhanced by having videotaped records of assessment interviews. This study describes a revised version of the Video Interview Behaviour Evaluation Scales (VIBES), shortened to make it more applicable to clinical trials. The method focuses on physically expressive, motor and social behavioural aspects of the depressive disorder. The Brief version permits juxtaposing of baseline and outcome interviews during the same viewing session thereby reducing the role of memory in the rater's observations. The method provides measures of behavioural components and four new severity dimensions, Social withdrawal-motor retardation, Anxiety-agitation, Hostility, Depressive mood-cognitive impairment. Viewing a patient's series of baseline, during treatment, and outcome interviews can be conducted in about 1 h. This study compared the VIBES with the Hamilton Depression Rating Scale (HAMD) in (1) assessing the efficacy of the selectively targeted ADs, desipramine and paroxetine, and placebo in depressed in-patients and (2) determining onset and nature of the drugs' early behavioural actions. The findings showed (1) components of the method to be reliable; (2) the VIBES to be more sensitive than the HAMD in measuring efficacy; (3) the methods to be equally sensitive in detecting early clinical actions of the two drugs; (4) the VIBES more informative in identifying discrete behavioural aspects of the disorder that are impacted by the drugs; and (5) that in differentiating the drugs' behavioural actions, desipramine was indicated to initially 'stimulate', i.e. effect motor activity and depressed mood, and paroxetine to reduce global severity and anxiety. The study shows the VIBES to be capable of uncovering behavioural mechanisms underlying AD's capacity to resolve depressive disorder.