Children with Autism Spectrum Disorder (ASD) Demonstrate Cognitive Flexibility in Categorizing Foods

Background: Cross-classification, i.e., categorizing the same object in multiple ways (Nguyen & Girgis, 2014), is a basic aspect of cognitive flexibility that has not been studied in children with Autism Spectrum Disorder (ASD). Deficits in cognitive flexibility have been found in ASD, but the nature and magnitude of these deficits is unclear, due to the varied ability levels of participants, complexity of tasks, and stimulus domains used (Landry & Al-Taie, 2016; Memari et al., 2013). Given the impairments in executive functioning, rigidity, and perseverative thought and behavior observed in individuals with ASD (Hill, 2004), impairments in cross-classification may be expected.

Objectives:

To investigate basic categorization and cognitive flexibility among children with ASD employing a highly familiar stimulus domain – food.

• To investigate the accuracy of children with ASD when categorizing food items into taxonomic, script, and evaluative categories.
• To investigate the flexibility of children with ASD when cross-classifying the same food item into more than one of the above categories.

To improve on previous methodologies by using restricted age and IQ ranges.

Methods:

Participants

19 boys, 2 girls diagnosed with ASD

Age: M=9.62 years; SD=1.66

FSIQ M-composite-score=99.29; SD=13.42

Categorization Experiment (Following Nguyen & Murphy (N&M), 2003)

Children viewed 28 triads of color photographs of foods comprising 12 taxonomic (e.g., fruit, meats), 12 script (based on setting/time of eating), and 4 evaluative (healthy/unhealthy) groupings. Each triad consisted of a unique target food and two choices: a choice sharing a taxonomic, script, or evaluative relationship with the target; and an unrelated choice. Children were asked to identify the choice that was the same food type as the target.

Cross-Classification Experiment (Following N&M)

Children performed a categorization task like above, but this time each target food was presented first in either a taxonomic or script triad and later presented again in the opposite triad. Identifying the correct choice in both types of triads for the same target was deemed correct cross-classification.

Results:

Categorization Experiment

Mean Percentages of Correct Categorization:

			Grp 1	Grp 2
	All(N=21)	t(20)	7-9yo(n=11)	10-12yo(n=10)	Grp1Vs2
	M(SD)	compared to 50%	M(SD)	M(SD)	t(19)
Taxonomic:	93(08)	24.46*	92(07)	93(09)	-00.25
Script:	87(16)	10.61*	89(11)	85(20)	00.51
Evaluative:	83(24)	06.33*	89(17)	78(30)	01.06
Overall:	89(12)	14.81*	90(07)	88(16)	00.51

*p<.000

Comparing our 7-9yo sample (M=90) with the N&M (M=85) neurotypical sample of 7yo’s yielded a significant difference (t(10)=2.64, p=.025).

A 2(age group)X3(category type)ANOVA produced no main effect for age (F(1)=.54, p=.473) or interaction. Category type (F(2)=3.61, p=.037) was significant.

Cross-Classification Experiment

Mean Percentages of Correct Cross-Classification:

		t(20)
	M(SD)	compared to 25%
All(N=21):	70(20)	10.38*
7-9yo(n=11):	73(14)	11.32*
10-12yo(n=10):	67(25)	05.24**

*p<.000 **p<.001

There was no difference in overall performance on cross-classification between 7-9yo’s and 10-12yo’s (t(19)=0.69, p=.498).

Conclusions: Our children, with ASD and average IQs, scored above chance (50%) when categorizing by each category type and our 7-9yo’s outperformed an age-matched neurotypical sample. Our children performed above chance (25%) when cross-classifying the same food into more than one category, thus demonstrating categorization ability as well as cognitive flexibility when tested using highly familiar stimuli.