Deep Learning Approach for Classification and Interpretation of Autism Spectrum Disorder

Prasad, P. K.; Khare, Yash; Dadi, Kamalaker; Vinod, P.; Surampudi, Bapi Raju

doi:10.1109/ijcnn55064.2022.9892350

Cited by 11 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequently, interpretability could aid in detecting patterns in fMRI data that indicate the presence of autism ( 13 ), leading to more accurate diagnoses and personalized treatment plans. Recently, the integrated Gradients (IG) and Deep LIFT techniques were utilized to identify the correlations between brain regions that contribute most to the classification task ( 14 ). Nowadays, a hybrid deep learning framework was proposed to improve classification accuracy and interpretability simultaneously ( 15 ).…”

Section: Introductionmentioning

confidence: 99%

Interpretation of deep non-linear factorization for autism

Chen

Yin

2023

Front. Psychiatry

View full text Add to dashboard Cite

Autism, a neurodevelopmental disorder, presents significant challenges for diagnosis and classification. Despite the widespread use of neural networks in autism classification, the interpretability of their models remains a crucial issue. This study aims to address this concern by investigating the interpretability of neural networks in autism classification using the deep symbolic regression and brain network interpretative methods. Specifically, we analyze publicly available autism fMRI data using our previously developed Deep Factor Learning model on a Hibert Basis tensor (HB-DFL) method and extend the interpretative Deep Symbolic Regression method to identify dynamic features from factor matrices, construct brain networks from generated reference tensors, and facilitate the accurate diagnosis of abnormal brain network activity in autism patients by clinicians. Our experimental results show that our interpretative method effectively enhances the interpretability of neural networks and identifies crucial features for autism classification.

show abstract

Section: Introductionmentioning

confidence: 99%

Interpretation of deep non-linear factorization for autism

Chen

Yin

2023

Front. Psychiatry

View full text Add to dashboard Cite

show abstract

“…A multilayer perceptron (MLP) based classification model with auto encoder pretraining was utilised to distinguish ASD from Typically Developing (TD) using MRI images from the ABIDE-1 dataset (15) . Pre-trained deep convolutional neural networks (CNNs) including GoogleNet, AlexNet, MobileNet, and SqueezeNet achieved validation accuracy of 75%, 75.84%, 79.45%, and 82.98% in detecting the scalograms generated by EEG signals (16) .…”

Section: Introductionmentioning

confidence: 99%

Enco – Standardization Data Pre-Processing Technique in Autism Spectrum Disorder Detection

Kavitha,

Kasthuri

2023

IJST

View full text Add to dashboard Cite

Background/Objectives: The goal of this study was to create an Enco-Standardization technique that would produce accurate data and improve the diagnosis of Autism Spectrum Disorder (ASD).This method uses mean values to replace missing values in a dataset and improves them by combining label encoding and conventional scaling techniques. Methods: The ASD dataset, which has 704 instances and 21 attributes, is used in this study. Training and testing are divided by the dataset (80%-20%). As an imputation strategy in this dataset, missing values are located and replaced with the mean value. Attributes are encoded using the Enco-Standardization methodology using a label encoding technique that changes non-numeric variables into numeric ones. After that, the data were scaled into a machine-readable format to standardise it. Different machine learning classifier models are compared to the hybrid strategy of encoding and scaling techniques. Based on the accuracy found using machine learning classifier models, the dataset acquired using the Enco-Standardization technique is assessed. Findings: The dataset needs to be accurate and relevant in order to increase accuracy and decrease computing time. The findings of the Enco-Standardization methodology showed a good pre-processing method with accuracy values of 98% for Naive Bayes (NB), 71% for K Nearest Neighbour (KNN), 74% for Support Vector Machine (SVM), 97% for Linear Regression (LR), 100% for Decision Tree (DT), and 100% for Random Forest (RF). The deletion of missing values improves performance in KNN (94%), SVM (95.9%), LR, DT, and RF (100%) but decreases the number of instances in the dataset, rendering the model ineffective. Novelty: The data in a dataset are transformed and encoded using the proposed Enco-Standardization pre-processing technique, which increases the precision of the data analysis process in ASD prediction. Data discrepancies are avoided by using this eco-standardization technique.

show abstract

“…All the aforementioned CNN-based models, which were intensively trained on the ImageNet dataset, which contains 14 million images divided into 1000 categories, are used to extract attributes from the photographs in the Kaggle autistic image dataset (https://www.kaggle.com/general/123978). Using fMRI scans from the ABIDE-1 dataset, Prased et al [26] classified Using a multilayer perceptron (MLP) based classification model with autoencoder pretraining, ASD is distinguished from Typically Developing (TD). The suggested method identified the correlations between brain regions that contribute most to the categorization problem…”

mentioning

confidence: 99%

Efficient Net-based Transfer Learning Technique for Facial Autism Detection

Saeed Mian

2023

SCPE

View full text Add to dashboard Cite

Autism Spectrum Disorder is a neurological disorder in which an individual faces life-long effects in communication and interaction with others. Nowadays, the Autism Spectrum disorder ratio is increasing drastically more than ever before. Autism can be identified at all developmental levels as a ”behavioural condition,” and its symptoms often arise between the ages of two and four. The ASD issue starts during puberty and persists through adolescence and adulthood. Children with ASD use both nonverbal and verbal behaviour to communicate, and they struggle with joint attention and social reciprocity. Children with autism are frequently socially isolated as a result of these problems. Through very expensive and time-consuming screening exams, autism spectrum features can be identified. As one of the possible mirrors of the brain, children’s faces can be utilised as a biomarker and as a quick and convenient technique for the early identification of ASD. An effective, genuine, and automatic method of face-based spectrum disorder identification is required. In this study we compare the transfer learning approach used for autism identification with the convolutional neural network (CNN)-based efficient-net strategy to identify autistic children using facial images. We used an open-source Kaggle dataset and evaluated the model performance in terms of accuracy, confusion matrix, precision, recall, and F1 measure. Efficient shows an accuracy of 97% on the benchmark dataset and beats the baseline technique of transfer learning-based approaches. This study can be used to help medical professionals validate their initial screening procedures for finding youngsters with ASD disease.

show abstract

Deep Learning Approach for Classification and Interpretation of Autism Spectrum Disorder

Cited by 11 publications

References 25 publications

Interpretation of deep non-linear factorization for autism

Interpretation of deep non-linear factorization for autism

Enco – Standardization Data Pre-Processing Technique in Autism Spectrum Disorder Detection

Efficient Net-based Transfer Learning Technique for Facial Autism Detection

Contact Info

Product

Resources

About