Addiction is considered a "bio-psycho-socialspiritual disorder"  for which complete recovery cannot be assured. Although addiction was computationally characterized as a non-reversible process [2,3], behavioral evidences support the possibility of recovery [4, 5, 6,7]. We thus propose to consider addiction as a non-monotonic dynamical disease. Previously, we presented a multiscale computational model that combines three scales of observations: behavior, cognition, and neuropsychology. This model evaluates the drug-seeking behavior in virtual subjects. We use our model to analyze dynamical properties of two typical virtual subjects. One has become an addict at the age of 17 and has since expressed a severe multi relapse pattern. The other one was exposed to addictive behavior at age 36 but managed to cease drug seeking. The results are encouraging for further explorations into the individual dynamics in addicts.