Home | People | Papers | Projects | Methods | Links | Graduate Programs | Subjects

The neural development and organization of abstract word recognition

Behavioral study. ALIs are known to affect the speed with which subjects match strings of letters: Subjects are slower at distinguishing strings in which an uppercase letter is replaced by the lowercase form of the same letter ("HYLE", "hyle"), compared with strings in which the uppercase form is replaced with the lowercase form of a different letter ("HYLE", "hile", Besner et al., 1984). If ALIs arise from exposure to common contexts, then one should be able to produce artificial ALIs in a completely novel alphabet with common contexts training. We tested this prediction by giving naive subjects extensive exposure to strings of Japanese characters in which arbitrary pairs of characters appeared in common contexts (unbeknownst to the subjects).

Subjects first performed a computerized pretest in which they indicated whether pairs of Japanese character strings were identical or not to provide a pre-training baseline. They then received six weeks of paper-and-pencil training in which they performed the same matching task. Unbeknownst to the subjects, two pairs of characters appeared in common visual contexts in their training packets. Half the subjects received one pairing and the other half received an orthogonal pairing. After training subjects again performed the computerized matching task.

In the pretest the common contexts pairs were on average distinguished slightly faster than the different contexts pairs (this effect was not significant). In the post-test however, the opposite pattern was observed. This training type by test interaction was significant.

This result is analogous to the standard ALI effects on string matching in which subjects are slower to distinguish strings that share the same ALIs. In this case, training with common contexts in novel alphabet produced the effect. These results support the hypothesis that common contexts training leads to the development on an abstract representation (which in turn impairs matching performance).

Neuroimaging | Theory | Computational Model | Behavioral Studies | Discussion | Papers