No abstract

No abstract

The relations between recall performance and specific network and spatial representations of memory were investigated in serial and free recall paradigms. The structural representations were derived from relatedness ratings by using the alternating least squares scaling (ALSCAL), Kruskal-Tfoung-Shepard-Torgerson (KYST; Kruskal & Wish, 1978), and Kruskal multidimensional scaling algorithms and the Pathfinder network scaling algorithm. In the serial recall task, list organization defined by the network yielded faster learning than organization defined by multidimensional space, although both types of organizations yielded increases in the number of items recalled. In free recall, the network representations were predictive of recall order even with the original ratings partialed out. The spatial representations, however were not independently predictive of recall. The results suggest that Pathfinder networks better capture the relations important for recall than do spatial representations. Also, Pathfinder networks provide information about memory organization that is not directly available in pairwise ratings.

No abstract

No abstract