Ontogeny of place learning in children as measured in the Radial Arm Maze, Morris Search Task, and Open Field Task.

Abstract

Children and adults were tested on 3 place learning tasks. Children under the age of 7 were inferior to older subjects in solving the tasks by using spatial relational solutions, but subjects of all ages were equally proficient in solving the task by using simple stimulus-reward associations (cued solutions). Accurate performance on the cued versions suggests that neither the general response demands nor the large size of testing environments rendered the tasks differentially inappropriate for young children. Instead, the nature of the cognitive demands were responsible for different levels of performance across the age groups. Because, in animal studies, spatial relational solutions but not cued solutions of these tests require mature and undamaged medial temporal lobe structures, the results suggest that these systems are not fully developed in humans before approximately 7 years of age.

Development of allocentric spatial memory abilities in children from 18 months to 5 years of age

“…Even though we have previously shown that adults need no verbal instructions to understand the goal of our task and to exhibit successful performance ), a pilot study had shown that older children (48-60 months) were very uncomfortable not receiving any kind of verbal instruction or feedback, and were in fact performing worse than younger children (30-45 months) who seemed to have no problem understanding the goal of the task (i.e., find rewards at fixed spatial locations). Interestingly, Overman and colleagues found very similar results in a Morris search task (Overman et al, 1996). When not given information about the static nature of the goal location, 4-year-olds performed worse than 3-year-olds.…”

“…The majority of other studies have restricted their investigations of allocentric spatial capacities to children three years and older, and these studies have repeatedly shown that (1) the performance of 3-year-olds on allocentric spatial tasks is inferior to that of older children; (2) children show marked improvement in their allocentric capacities between four and six years of age; and (3) for most of the tasks employed, children's performance becomes adult-like sometime between seven and ten years of age (Bullens et al, 2010;Foreman et al, 1984Foreman et al, , 1990Lehnung et al, 1998;Leplow et al, 2003;Mandolesi, Petrosini, Menghini, Addona, & Vicari, 2009;Nardini, Burgess, Breckenridge, & Atkinson, 2006;Overman et al, 1996;Pentland, Anderson, Dye, & Wood, 2003;Smith et al, 2008). Thus, with the possible exception of Newcombe's study (Newcombe et al, 1998), no study has unequivocally demonstrated that children under three years of age are capable of solving a task which requires the use of an allocentric spatial representation, a fundamental component of episodic memory.…”

“…Aadland and colleagues also reported that the performance of a group of children under 30 months (n = 7; range 18 to 30 months) was slightly above chance (defined as 45% correct for choices 5-8) in a free-choice paradigm, but none of the children met a criterion of entering 7 different arms on the first 8 choices or completing 20 trials (Aadland, Beatty, & Maki, 1985). Finally, in a study by Overman and colleagues the performance of 2-year-olds (n = 6; range 20-30 months) varied between approximately 37% and 75% correct arm selection in the first eight choices of a free-choice paradigm (Overman, Pate, Moore, & Peuster, 1996). Although these data are suggestive that 2-year-olds possess a certain level of allocentric spatial competence, it is critical to consider the fact that age-dependent improvements in performance in the free-choice paradigm of the radial arm maze are often associated with a change in foraging strategy by children who begin visiting adjacent arms (Aadland et al, 1985).…”

“…Specifically, in order to solve the commonly-used radial arm maze (Aadland et al, 1985;Foreman et al, 1984Foreman et al, , 1990Mandolesi et al, 2009;Overman et al, 1996), participants must rely not only on their allocentric spatial memory to discriminate one arm from another, but also on their working memory to remember which arms they have already visited on any given trial (unless they solve the task by visiting immediately adjacent arms, see above). As working memory proficiency increases with age (Dempster, 1981;Gathercole, 1998;Hitch & Halliday, 1983;Siegel & Ryan, 1989), this may confound the results and interpretation of spatial memory experiments performed in the radial arm maze.…”

“…Perhaps more importantly, few studies implementing the radial arm maze in children have controlled for extramaze visual cues, as described above for the Balcomb et al (2011) study. For example, mazes were placed either in a room, where individual arms could be associated with visually distinctive objects that were located directly behind the arm (such as windows, doors, radiators, and cupboards; Aadland et al, 1985;Foreman et al, 1984Foreman et al, , 1990, or outdoors, where distinctive distal cues might be aligned with specific arms (such as buildings, trees, and parking areas; Overman et al, 1996), especially when participants were required to return to the center of the maze before choosing the next arm (thus allowing them a standardized view of each arm). When carried out under such conditions, the possibility that children can use a visual guidance strategy to identify specific maze arms cannot be ruled out (see Section 4.1 Task specificity for further discussion).…”