A modified version of the dimensional change card sort task tests cognitive flexibility in children (Homo sapiens) and cotton-top tamarins (Saguinus oedipus)

Abstract

A modified Dimensional Change Card Sort (DCCS) task was used to test cognitive flexibility in adult cotton-top tamarins and children aged 19 months to 60 months. Subjects had to infer a rule from the experience of selecting between two cards to earn a reward, and the pairs of stimuli defined the rule (e.g., pick blue ones, not red ones, or pick trucks, not boats). Two different tests measured subjects' ability to shift to a reversal of the rule (intradimensional shift) and to shift to a new rule defined by a dimension previously irrelevant (interdimensional shift). Both adult tamarins and children aged 49-60 months were able to learn the initial rule and switch to a reversal and to a rule based on a different dimension. In contrast, the two younger groups of children, aged 19-36 months and aged 37-48 months, could switch when a reversal was imposed but took significantly longer to learn a new rule on a former irrelevant dimension. Experiment 2 presented a wider set of novel stimuli which shared some features with the original set to further explore the basis of rule learning. The result was that tamarins and 52- to 60-month-old children both chose novel stimuli that fit the rule and had no a priori associative strength, suggesting a rule application not solely based on associative strength. Importantly, novel items introduced some risk for choice, and children showed themselves to be risk-averse, whereas tamarins were risk-prone within a novel context. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Figure 1:. An Example of Training and the Two Tests Involving Rule-Shifting (Intradimensional/Reversal, and Interdimensional Shift to the Formerly Irrelevant Dimension)

Note. Hands represent choice responses that are correct in each phase. The intradimensional and interdimensional shifts displayed trials in which only one relevant dimension was varied, and trials in which both relevant and irrelevant dimensions were varied.

Figure 2:. Mean Trials to Criterion for Children in the Three Age Groups in the Three Conditions (Training, Intradimensional Shift, Interdimensional Shift)

Note. Error bars show standard errors. * Significance at p <0.05. ** Significance at p < 0.01. See the online article for the color version of this figure.

Figure 3:. Mean Trials to Criterion for Monkeys Across the Three Conditions (Training, Intradimensional Shift, Interdimensional Shift), Separated by the Dimension Trained Initially (Color or Shape)

Note. Error bars show standard errors. *Significance at p < 0.05. See the online article for the color version of this figure.

Figure 4:. Example Pairs From Experiment 2

Note. Training trials were rewarded for responses to the blue boat or blue truck. Tests used a mixture of training trials rewarded accordingly, with test trials that were rewarded pseudorandomly on a 50% schedule independent of type of response.

Figure 5:. Mean Accuracies by Children (Blue) and Monkeys (Orange) in Training and Test Blocks

Note. Error bars show standard errors and are drawn only downward to avoid needing to exceed 100% correct on the y axis. Dashed line indicates chance level responding. ** Significantly above chance at p < 0.01. * Significantly above chance at p < 0.05, with the exception of RED S−:Novel Novel for children, which is significantly below chance. See the online article for the color version of this figure.