In Pavlovian fear conditioning, an initially neutral stimulus predicts the arrival of an aversive stimulus at a fixed time interval. Accumulating evidence indicates that in associative learning temporal relations between events are encoded. However the neural networks underlying the encoding and memory of interval durations are poorly understood. We addressed this question in odor fear conditioning which can be performed in young pups as well as in adult animals. In this task, an odor is presented to the animal and after a fixed interval (30 seconds) a mild footshock is applied. We first investigated the ontogenesis of interval timing at the behavioral level. We designed an experimental setup allowing the simultaneous recording of respiration, ultrasonic vocalizations (USVs) and freezing in rats. These parameters greatly improved the sensitivity of fear behavior assessment, thus increasing the probability of detecting transient anticipatory fear responses. Three ages were considered: 14-days-old (infants), 22-days-old (juveniles) and 75-days-old (adults). The data show that in adults, an anticipator y response develops after a few odor-shock pairings, characterized by a decrease in respiratory rhythm and an increase in USVs emission prior to shock delivery. Similar anticipatory responses might be present in juveniles, while infants might show no signs of interval timing, despite showing good learning of the odor-shock association. In parallel, current experiments are investigating learning-induced brain activation at these three ages, using 2 Deoxyglucose metabolic mapping. We are focusing on structures known to be involved in timing in adults: striatum, nucleus accumbens and prefrontal cortex.