Sensibilidad y Especificidad del test cognitivo “Ice Cream” en adultos con autismo leve

Abstract

Introduction.

The literature has pointed out the increase and efficacy of virtual reality techniques in the assessment and intervention of people with Autism Spectrum Disorder (ASD), given their ability to better capture and interpret information through the visual pathway. On the other hand, the absence of studies in ASD with the Ice Cream (IC) cognitive test is notable.

Objectives.

To analyse the sensibility and specificity of the IC virtual reality test to assess Executive Functions (EEFF) in adults diagnosed with ASD with mild impairment (grade 1). Additionally, a comparison with a traditional paper and pencil test, the Five Digit Test (FDT), was included.

Methods.

The sample consisted of 32 participants (ASD grade 1: n = 16; controls:
n = 16), 40.6% were female and 59.4% were male. The mean age of the control group (CG) was: M = 26.375 (SD = 10.776) and that of the experimental group (EG) was: M = 27.750 (SD = 9.747). The assessment instruments used were: 1) clinical and sociodemographic questionnaire, 2) IC virtual reality test and, 3) a paper and pencil test, FDT.

The assessments were carried out under the supervision of a trained psychologist and
under controlled conditions at the Hogar Virgen de los Reyes and at the Loyola Andalucía University (ULA). Mean comparisons were carried out to analyse the differences between
groups. In addition, the UMAP technique was used to ablate and reduce variables, as well
as a Machine Learning mechanism, and the Support Vector Machine (SVM) algorithm was used to predict IC’s ability to classify subjects appropriately into CG and EG.

Results.

The performance of the EG was generally lower that the CG, especially in Processing Speed. High heterogeneity was observed within the EG in Flexibility, Planning and Working memory, ranging from very high to very low performers. To generate the IC classification model, the initial variables were reduced to two UMAP 6 variables. IC provided good discrimination between the two groups, however, it sometimes generated false positives, it diagnosed people with ASD grade 1 who did not actually have a diagnosis, while FDT generated false negatives.

Conclusions.

The results demonstrated the efficacy of the IC virtual reality tool in the assessment of EEFF in adults with ASD grade 1, providing promising indications. However, it is essential to continue with research in this direction, generating increasingly valid and reliable tools applicable to the therapeutic context.