Classification of Preschoolers with Low-Functioning Autism Spectrum Disorder Using Multimodal MRI Data

Kim, Johanna Inhyang; Bang, Sungkyu; Yang, Jin‐Ju; Kwon, Heejin; Jang, Soomin; Roh, Sungwon; Kim, Seok Hyeon; Kim, Mi Jung; Lee, Hyun Ju; Lee, Jongmin; Kim, Bung-Nyun

doi:10.1007/s10803-021-05368-z

Cited by 10 publications

(7 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Possible explanations are as follows: ML is good at processing big data, so complex rules may not be found in the case of limited data. At the same time, research on autism-related diagnostic models focuses on high-cost imaging materials such as MRI, while this study focuses on relatively easily available behavioral observations and socio-demographic data ( 37 , 53 , 54 ). In addition, selecting the best variable combination also has certain difficulties.…”

Section: Discussionmentioning

confidence: 99%

“…The existing predictive diagnosis of ASD or ASD comorbidities (e.g., attention deficit hyperactivity disorder) usually includes four aspects, disease prevention or risk factor identification, disease diagnosis, disease efficacy prediction, and disease prognosis prediction. Most of these diagnostic models use complex diagnostic-related data, such as expensive head MRI, EEG ( 37 – 39 ), and blood biochemical indicators. These data were used to build diagnostic models to diagnose ASD-related disorders and to determine their type or severity.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability

Song

Jiang

et al. 2022

Front. Psychiatry

View full text Add to dashboard Cite

BackgroundEarly detection of children with autism spectrum disorder (ASD) and comorbid intellectual disability (ID) can help in individualized intervention. Appropriate assessment and diagnostic tools are lacking in primary care. This study aims to explore the applicability of machine learning (ML) methods in diagnosing ASD comorbid ID compared with traditional regression models.MethodFrom January 2017 to December 2021, 241 children with ASD, with an average age of 6.41 ± 1.96, diagnosed in the Developmental Behavior Department of the Children’s Hospital Affiliated with the Medical College of Zhejiang University were included in the analysis. This study trained the traditional diagnostic models of Logistic regression (LR), Support Vector Machine (SVM), and two ensemble learning algorithms [Random Forest (RF) and XGBoost]. Socio-demographic and behavioral observation data were used to distinguish whether autistic children had combined ID. The hyperparameters adjustment uses grid search and 10-fold validation. The Boruta method is used to select variables. The model’s performance was evaluated using discrimination, calibration, and decision curve analysis (DCA).ResultAmong 241 autistic children, 98 (40.66%) were ASD comorbid ID. The four diagnostic models can better distinguish whether autistic children are complicated with ID, and the accuracy of SVM is the highest (0.836); SVM and XGBoost have better accuracy (0.800, 0.838); LR has the best sensitivity (0.939), followed by SVM (0.952). Regarding specificity, SVM, RF, and XGBoost performed significantly higher than LR (0.355). The AUC of ML (SVM, 0.835 [95% CI: 0.747–0.944]; RF, 0.829 [95% CI: 0.738–0.920]; XGBoost, 0.845 [95% CI: 0.734–0.937]) is not different from traditional LR (0.858 [95% CI: 0.770–0.944]). Only SVM observed a good calibration degree. Regarding DCA, LR, and SVM have higher benefits in a wider threshold range.ConclusionCompared to the traditional regression model, ML model based on socio-demographic and behavioral observation data, especially SVM, has a better ability to distinguish whether autistic children are combined with ID.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability

Song

Jiang

et al. 2022

Front. Psychiatry

View full text Add to dashboard Cite

show abstract

“…It is not uncommon for research designs to incorporate IQ-based functional classifications in their inclusion or exclusion criteria (e.g. Calhoun et al, 2020; Fung et al, 2021; Kim et al, 2023; Pereira et al, 2018; Zajic et al, 2018), and the phrases “low-” and “high-functioning autism” have been used widely in everyday communications as well. This approach is problematic given the consistently reported gap between cognitive and adaptive functioning in autistic individuals (Alvares et al, 2020; Klin et al, 2007; Tamm et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Using adaptive behavior scores to convey level of functioning in children with autism spectrum disorder: Evidence from the Study to Explore Early Development

Furnier,

Ellis Weismer,

Rubenstein

et al. 2023

Autism

View full text Add to dashboard Cite

We examined relationships between measures of adaptive behavior, cognitive ability, and autism symptom severity in 1458 preschool-aged children with autism from the Study to Explore Early Development. While publications commonly describe autistic children as “low-” or “high-functioning” based on cognitive ability, relying solely on cognitive scores may obscure meaningful variation in functioning. We found significant heterogeneity in adaptive behavior scores of children with cognitive scores both above and below the threshold of two or more standard deviations below the population mean specified in the diagnostic criteria for intellectual disability (ID). Although cognitive and adaptive behavior scores were strongly associated in our sample, considerable variation in overall adaptive behavior and more than half in socialization and motor skills was unaccounted for by cognitive ability, autism symptom severity, and other covariates. Among children who could be designated “low-functioning” based on cognitive scores, 39.7% had composite adaptive behavior scores indicating no significant delays, while among those who might be designated “high-functioning,” 9.0% had significant delays in overall adaptive behavior and 22.2% in socialization. These results suggest adaptive behavior scores capture variations in the autism phenotype not accounted for by other measures we considered. Lay Abstract Autistic people are often described as “low-” or “high-functioning” based on their scores on cognitive tests. These terms are common in publications and in everyday communication. However, recent research and feedback from the autistic community suggests that relying on cognitive ability alone to describe functioning may miss meaningful differences in the abilities of autistic children and adults and in the kinds of support they may need. Additional methods are needed to describe “functioning” in autistic children. We examined whether scores from a test measuring adaptive behaviors would provide information on the functional abilities of children with autism that is different from cognitive ability and autism symptom severity. Adaptive behaviors include age-appropriate skills that allow people to function in their everyday lives and social interactions. We found that a large amount of the variation in adaptive behavior scores was not explained by cognitive development, autism symptom severity, and behavioral and emotional problems. In addition, there was a wide range of adaptive ability levels in children with autism in our study, including in those with low, average, or high cognitive scores. Our results suggest that adaptive behavior scores could provide useful information about the strengths and support needs of autistic children above and beyond measures of cognitive ability and autism symptom severity. Adaptive behavior scores provide important information on the needs of autistic people.

show abstract

“…Machine learning-based analysis of MRI data was useful in classification of ASD patients from typically developing ones. Combination of two types of data had improved classification accuracy about 10% [24].…”

Section: Methodsmentioning

confidence: 99%

Autism Classification using Visual and Behavioral Data

Hriti

Shaer

Momin

et al. 2022

Preprint

View full text Add to dashboard Cite

1AbstractCurrently Autism Spectrum Disorder (ASD) is diagnosed via the combination of multiple medical tools and screening tests that require extensive amounts of time and money. Autism diagnosis can be formulated as a typical machine learning classification problem between ASD patients and a control group consisting of neurotypical individuals. In order for this to yield accurate results, large datasets with different modalities are required. However, the unavailability of such robust datasets stands as a threat to this automated diagnosis. To resolve this, we propose a method of Autism Classification using Visual and Behavioral Data. The proposed technique relates datasets of two modalities (visual and behavioral) collected from similar participants by generating common attributes among the records and distributing these records into sub classes. Then records within these subclasses are combined to form an integrated dataset. Finally, decision level fusion is performed on the multimodal data. The main contribution of our work can be outlined as follows: an accuracy of 97.57% in autism classification has been obtained from the integrated data, which is higher than detection from only visual data, we have shown that combining data within sub classes based on common attributes is more accurate than combining them arbitrarily, and finally, we have introduced a novel, integrated multimodal dataset in the ASD domain.

show abstract

Classification of Preschoolers with Low-Functioning Autism Spectrum Disorder Using Multimodal MRI Data

Cited by 10 publications

References 79 publications

A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability

A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability

Using adaptive behavior scores to convey level of functioning in children with autism spectrum disorder: Evidence from the Study to Explore Early Development

Autism Classification using Visual and Behavioral Data

Contact Info

Product

Resources

About