The psychobiological model of personality developed by Cloninger, including four dimensions of temperament and three dimensions of character, allows to explore personality factors associated with depressive disorders. The three main dimensions of temperament are Novelty Seeking (NS), ie the tendency towards excitement in response to novel or rewarding stimuli, Harm Avoidance (HA) hypothesized to represent the tendency to respond intensely to signals of adverse stimuli, and Reward Dependence reflecting the tendency to respond intensely to signals of reward and to maintain behavior previously associated with reward. These personality traits are hypothetically related to underlying neurotransmetter systems (especially NS to dopaminergic function and HA to serotonergic function). The two main dimensions of character are Self-Directedness (SD) and Cooperativeness (C), measuring maturity traits respectively concerning individual and social adaptation; thus they are negatively correlated with the risk of personality disorder for a given patient. Many studies have been carried out with the Temperament and Character Inventory (TCI), or with the previous Tridimensional Personality Questionnaire (TPQ), in depressive disorders with cross-sectional but also with short-term and long-term longitudinal designs. They show primarily that patients with history of depressive disorders, even in euthymic phase, have very high Harm Avoidance scores. In prospective studies conducted in depressive patients, even after remission of the depressive episodes, the Harm Avoidance scores are lower than before treatment, but still elevated compared to the general population. The patients who fail to respond to antidepressant treatments have generally higher Harm Avoidance scores before treatment than the others. Overall, various results support four types of potent relationships between Harm Avoidance and depression: an influence of state on trait measure, a pathoplastic effect of Harm Avoidance on depressive expression, a vulnerability model (Harm Avoidance representing a susceptibility factor for depression), and a scar model with elevated Harm Avoidance scores even after remission of acute depressive symptoms. Other temperament dimensions, Novelty Seeking and Reward Dependence, are not consistently associated with depression characteristics nor affected by state effects, but Self-Directedness character dimension is often low when compared to subjects without depressive disorders, reflecting maladaptive personalities frequently associated to depressive disorders. Some studies suggest that low Self-Directedness scores can be predictive of suicidal behaviors. In cross-sectional studies, Harm Avoidance scores are highly positively correlated with depression intensity (r=0.23 to 0.67) and Self-Directedness scores are highly negatively correlated with depression (r=- 0.37 to - 0.60). Some studies suggest that the temperamental dimensions assessed by the TCI could have a predictive value for the response to antidepressants treatments, but this result is controversial and needs further research. For example, a study conducted in 84 patients with major depressive disorder showed that temperament type based on TPQ traits explained 25% of the response to treatment variance: patients with high scores on both Harm Avoidance and Reward Dependence had a favourable response to either clomipramine or desipramine. Studies of the relationship of temperament dimensions to biological markers of depression have also been published. It has been shown for example that Reward Dependence and Harm Avoidance scores are significantly predictive of morning hypercortisolemia in depressed subjects. More specifically, correlations have been obtained between platelet serotonergic markers (5-HT2a receptors) and Harm Avoidance scores also in depressed patients. In conclusion, Harm Avoidance seems to be a vulnerability factor or at least an associated factor to depressive disorders. This temperament dimension is supposed to be highly heritable, and to be linked to the serotonergic system variations.