Pas de résumé / The thesis investigated the development of children’s global/local processing hierarchical patterns introduced by Navon (1977). The objectives were to understand more comprehensively the developmental characteristics of children’s perception through their global and local processing of hierarchical patterns, by considering the effects of age, stimuli properties, duration of exposure to the stimuli and gender in a perceptual task and a drawing task. These effects were tested in 3 different populations: typically developing children, children with mental retardation and early blind children. The results revealed that typically developing children attended to both the local and global level of processing but these modes of spatial information processing operated independently. In a first step, children before 4 years of age showed dominance of local processing and then a more global processing developed at 4 years of age, and at 5 years of age integrated responses began to emerge. Early blind children showed similar developmental characteristics, although there was a protracted period of local processing dominance. Indeed, these children mainly produced local responses at ages of between 6 and 10 years, and then developed more global responses at 11-12 years and continued to integrate the two levels of analysis at later ages. On the other hand, global dominance was shown in children with mental retardation and their development was affected more by mental age than by chronological age. Moreover, their responses were shown to be sensitive to the fact that meaningful object could be located at the local level, enhancing local processing in this case. These results need further confirmations as the studies of global/local processing in atypical children are not numerous. In particular, the effect of duration of exposure to the stimuli should be further analyzed, because this factor did not seem to have a great effect in our experiments while it seemed more powerful in other studies carried out with adults. Replication of the study with children with mental retardation appears also important to plan for future work, because we can have some doubt relatively the absence of modification through ages of the way these children perceive hierarchical patterns. Finally, defining more precisely what may underlie the gender differences seems also worth to explore since gender did not show a major effect in our results.

The thesis investigated the development of children's global/local processing hierarchical patterns introduced by Navon (1977). The objectives were to understand more comprehensively the developmental characteristics of children's perception through their global and local processing of hierarchical patterns, by considering the effects of age, stimuli properties, duration of exposure to the stimuli and gender in a perceptual task and a drawing task. These effects were tested in 3 different populations: typically developing children, children with mental retardation and early blind children. The results revealed that typically developing children attended to both the local and global level of processing but these modes of spatial information processing operated independently. In a first step, children before 4 years of age showed dominance of local processing and then a more global processing developed at 4 years of age, and at 5 years of age integrated responses began to emerge. Early blind children showed similar developmental characteristics, although there was a protracted period of local processing dominance. Indeed, these children mainly produced local responses at ages of between 6 and 10 years, and then developed more global responses at 11-12 years and continued to integrate the two levels of analysis at later ages. On the other hand, global dominance was shown in children with mental retardation and their development was affected more by mental age than by chronological age. Moreover, their responses were shown to be sensitive to the fact that meaningful object could be located at the local level, enhancing local processing in this case. These results need further confirmations as the studies of global/local processing in atypical children are not numerous. In particular, the effect of duration of exposure to the stimuli should be further analyzed, because this factor did not seem to have a great effect in our experiments while it seemed more powerful in other studies carried out with adults. Replication of the study with children with mental retardation appears also important to plan for future work, because we can have some doubt relatively the absence of modification through ages of the way these children perceive hierarchical patterns. Finally, defining more precisely what may underlie the gender differences seems also worth to explore since gender did not show a major effect in our results.