Factors Contributing to Success and Failure in MRI Assessment in Individuals with Low-Functioning Autism

Background: Although accumulating evidence indicates altered brain phenotypes in autism spectrum disorder (ASD), the majority of neuroimaging research has focused on individuals with relatively intact cognitive and language function (i.e., ‘high-functioning’). Those having more severe phenotypes (i.e., ‘low-functioning’, including minimally verbal and intellectually disabled individuals, are generally understudied and overlooked. This dearth of the knowledge hinders unraveling the etiologies underlying the heterogeneity in phenotypic expressions.

Objectives: We aimed to identify the factors and individual features which contribute to the success and failure in MRI assessment.

Methods: To answer this specific question and further investigate the expression of brain phenotypes in this understudied population, minimally verbal individuals and those with intellectual disabilities were recruited. These individuals were intentionally grouped in the ‘low-functioning’ group considering the sample size and dimensional features of verbal language expression levels, and were contrasted with the ‘high-functioning’ individuals with ASD (full-IQ>70), and neurotypical group. All participants received comprehensive assessment of clinical presentations (including the ADI-R, ADOS, and a myriad of parent-rated questionnaires measuring autistic features, adaptive functioning, and co-occurring behavioral/emotional problems), cognitive abilities (Wechsler and Leiter scales), and MRI (T1w, resting-state and task MRI, and diffusion imaging). Failure to complete MRI assessment was defined as inadequate quality (notable motion artifacts after several runs) of 6-minute T1w image, participants' active cancellation of scans, or inability to be compliant.

Results: The sample (age 7-30 years) consisted of 33 high-functioning, 27 low-functioning participants with ASD (17 fulfilled the stringent criteria of minimally verbal status), and 27 neurotypical controls. Among them, 32 (97%) high-functioning, 20 (74%) low-functioning individuals (12 minimally verbal), and all neurotypical controls successfully completed MRI assessments. All low-functioning individuals had practiced to lie still before scans, and were accompanied by the parents or therapists during scans. 20 (all low-functioning, including all unsuccessful cases) had received the mock scan before the official assessment; 3 of them failed to comply with mock scans. All participants who failed to complete scans, alongside 3 individuals in the low-functioning group having tolerable quality of image data, had several runs of T1w scans. Overall, low-functioning, relative to high-functioning individuals, had higher autistic severity based on the ADI-R and ADOS. These two patient groups did not differ in ratio of sex, or psychiatrist-diagnosed comorbidity, or most of parents-rated questionnaires, except higher internalizing and externalizing problems in high-functioning individuals. Benchmarking against the low-functioning individuals successfully completing MRI assessment (N=20), those who failed (N=7) had younger age (10±1.5 vs. 19±4.9 years). These two subgroups on the low-functioning spectrum did not differ in the severity of autistic or ADHD symptoms (regardless of the raters), or intellectual function.

Conclusions: Our preliminary findings suggest that younger age lead to higher failures in completing MRI assessment. Company by the family members or therapists, and several scan attempts may facilitate higher successful rates for scans. The mock practice could help the low-functioning individuals better accustom to the scan environment, and also helps the researchers to decide which participants might have higher chances to finish MRI scans.