Poster #2 - Children's and Adults’ Performance on a 4-Item Flexible Item Selection Task

• In Event: Poster Session 14

Abstract

Executive functions (EFs) are top-down cognitive processes responsible for the voluntary control of thought and action (Diamond, 2013). Working memory (hold information in mind), cognitive flexibility (hold different perspectives and switch between them) and inhibitory control (inhibit prepotent responses) are believed to constitute core related, yet separable, EFs (Miyake et al., 2000). EFs develop throughout life, characterized by periods of sudden and gradual improvements, and declines. Unfortunately, direct comparisons across ages are hampered by the lack of measures useable with a wide age range, because of ceiling or floor effects.
The current study examined Time 1 performance in longitudinal sample of 4-to-12-year-olds (N=1909) and 35 undergraduates on a modified version of the Flexible Item Selection Task (FIST; Jacques & Zelazo, 2001), a measure of cognitive flexibility initially developed for 3- to 5-year-olds. On each trial of the original FIST, children are shown three items, two that match on one dimension (e.g., color), two that match on a different dimension (e.g., shape), and a third dimension (e.g., size) is constant across items. Children are told to “pick 2 pictures that go together in one way” (Selection 1), and once selected, “pick 2 pictures that go together, but in another way” (Selection 2). Accordingly, children must select one pivot item on two dimensions. Children reach ceiling performance on this version by 6.
Hence, a 4-item computerized FIST was developed. Participants were presented with 2 demos, 2 practice (with feedback), and 18 test trials in which they saw four pictures and had to find two matching pairs (as in the original task). However, on this 4-item version, half the trials were flexible trials and half were nonflexible trials. Flexible trials are like the original such that a pivot item match one item on one dimension and another on a different dimension. For example, in Fig.1 (Left), Items 1-3 (left-to-right) match on color and Items 1-4 match on shape (Item 1 is the pivot). The fourth item matches the others on the irrelevant dimension (size). In nonflexible trials, however, one pair matches on one dimension and the remaining pair matches on a different dimension. In Fig.1 (Right), Items 1-4 match on shape, and Items 2-3 match on color, with no overlap and therefore, no need to consider an item from multiple perspectives.
A 10(age) x 2(gender) x 2(trial type) x 2(selection) mixed ANOVA, using mean performance across trials revealed main effects of age, F(9,1923)=112.51, p<.001, MSe=8.27, Partial η2=.35, and gender F(1,1923)=4.37, p=.04, MSe=8.27, Partial η2=.002 (girls better than boys), although gender did not interact with other variables. Selection and Trial Type were also significant (ps<.001), and subsumed under a significant 3-way interaction with age, F(9,1923)=4.62, p<.001, MSe=2.19, Partial η2=.02. As predicted, difficulties were more prevalent over development for Selection 2 of flexible trials (Fig.2).
In short, switch accuracy on flexible trials on the 4-item FIST follows a protracted developmental course. Moreover, the task demonstrates sufficient variability to be used across development without changing scoring methods (accuracy vs. reaction time), tasks or versions.

Author

• Sophie Jacques, Dalhousie University
  Presenting Author