Introduction: Short-term psychiatric effects have been described forcannabis, but rarely under clinical setting.Aims: We report three cases of healthy male young subjects who wereoccasional cannabis users without known psychiatric history who developedanxiety attacks, panic crisis, and transient perceptual disturbances (visualhallucinations, depersonalization, paranoid feelings and derealization)following oral or smoking administration of cannabis. In contrast to mostother case reports where subjects characteristics and history of drug use,circumstances, doses and blood concentrations are unknown, the three casesreported here happened under experimental conditions.Methods: Among the three cases, two happened among a series of 8 healthymale volunteers included in a study testing the psychomotor effects of oralcannabis (one case received 20 mg dronabinol and the other a decoction of16.6 mg THC). The third critical case happened in a series of 16 subjectsinvolved in a study to assess the effects of cannabis smoking on a trackingtask carried out during a functional magnetic resonance imaging (fMRI)experiment. The two studies were approved by the ethical commissions.Results: no opiates, amphetamines, cocaine, and benzodiazepines werefound in urine and no breath alcohol was detected before each session. Theingested (16.6 mg THC or 20 mg dronabinol) or inhaled dose (joint=0.8 g,11% THC, 10 puffs of 2 s, no tobacco added), the time-events of effects onbehavior, willingness to drive, and performance as well as the cannabinoidwhole blood levels were documented. While the oral route of administrationachieved only limited blood concentrations (less than 4.7 ng/mL (dronabinol)and 3.9 ng/mL (THC)), significant psychotic reactions occurred. In contrast,following inhalation, much larger THC blood levels (peak concentration:143 ng/mL) were found. The psychotic symptoms started at the end ofthe inhalation phase and continued during the distribution phase when theTHC levels rapidly decreased from 143 to 17 ng/mL in 16 min. Two tabletsof Temesta 1 mg (lorazepam) were then successively administered to thisvolunteer to ease anxiety and hallucinations. In this latter case, MRI andfMRI brain imaging performed before cannabis smoking did not discloseany obvious anatomical or brain functioning anomalies. All 3 cases werewithdrawn from the studies because of these unwanted side effects.Conclusions: The THC and dronabinol maximal blood levels were found tobe poor predictors of possible psychiatric side-effects. Furthermore, theseunwanted symptoms seem to be more frequent than expected, despite therelative medium dose of cannabinoids administered and the careful selection of the subjects.