A 10.13µJ/Classification 2-Channel Deep Neural Network Based SoC for Negative Emotion Outburst Detection of Autistic Children

Aslam, Abdul Rehman; Altaf, Muhammad Awais Bin

doi:10.1109/tbcas.2021.3113613

Cited by 24 publications

(7 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Promising medical therapies, including behavioral interventions and antipsychotic medications, have been implicated in the treatment of ASD over the past decades. However, drug resistance against ASD may occur frequently if traditional classifications based on gender and comorbidities are used to guide anti-ASD treatment ( Anderson et al, 2007 ; Ospina et al, 2008 ; Woolfenden et al, 2012 ; Aslam and Altaf, 2019 ; 2021 ). Therefore, accurate differentiation of ASD clusters at the molecular level will improve the understanding of the heterogeneity in ASD and is vital to guide the individualized treatment of ASD.…”

Section: Discussionmentioning

confidence: 99%

Identification of Ferroptosis-Related Molecular Clusters and Immune Characterization in Autism Spectrum Disorder

et al. 2022

View full text Add to dashboard Cite

Introduction: Autism spectrum disorder (ASD) is a neurodevelopmental disorder with clinical presentation and prognostic heterogeneity. Ferroptosis is a regulated non-apoptotic cell death program implicated in the occurrence and progression of various diseases. Therefore, we aimed to explore ferroptosis-related molecular subtypes in ASD and further illustrate the potential mechanism.Methods: A total of 201 normal samples and 293 ASD samples were obtained from the Gene Expression Omnibus (GEO) database. We used the unsupervised clustering analysis to identify the molecular subtypes based on ferroptosis-related genes (FRGs) and evaluate the immune characteristics between ferroptosis subtypes. Ferroptosis signatures were identified using the least absolute shrinkage and selection operator regression (LASSO) and recursive feature elimination for support vector machines (SVM-RFE) machine learning algorithms. The ferroptosis scores based on seven selected genes were constructed to evaluate the ferroptosis characteristics of ASD.Results: We identified 16 differentially expressed FRGs in ASD children compared with controls. Two distinct molecular clusters associated with ferroptosis were identified in ASD. Analysis of immune infiltration revealed immune heterogeneity between the two clusters. Cluster2, characterized by a higher immune score and a larger number of infiltrated immune cells, exhibited a stronger immune response and was markedly enriched in immune response-related signaling pathways. Additionally, the ferroptosis scores model was capable of predicting ASD subtypes and immunity. Higher levels of ferroptosis scores were associated with immune activation, as seen in Cluster2. Lower ferroptosis scores were accompanied by relative immune downregulation, as seen in Cluster1.Conclusion: Our study systematically elucidated the intricate correlation between ferroptosis and ASD and provided a promising ferroptosis score model to predict the molecular clusters and immune infiltration cell profiles of children with ASD.

show abstract

Section: Discussionmentioning

confidence: 99%

Identification of Ferroptosis-Related Molecular Clusters and Immune Characterization in Autism Spectrum Disorder

et al. 2022

View full text Add to dashboard Cite

show abstract

“…They can be broadly categorized into software or hardwarebased solutions. The hardware-based solutions include hardware accelerators and on-chip applications for emotion recognition, which have gained a lot of attention in recent years Aslam and Altaf, 2021a). These hardware applications require some special considerations when selecting ML or deep learning (DL) models:…”

Section: Emotions Classificationmentioning

confidence: 99%

Channels and Features Identification: A Review and a Machine-Learning Based Model With Large Scale Feature Extraction for Emotions and ASD Classification

et al. 2022

Self Cite

View full text Add to dashboard Cite

Autism Spectrum Disorder (ASD) is characterized by impairments in social and cognitive skills, emotional disorders, anxiety, and depression. The prolonged conventional ASD diagnosis raises the sheer need for early meaningful intervention. Recently different works have proposed potential for ASD diagnosis and intervention through emotions prediction using deep neural networks (DNN) and machine learning algorithms. However, these systems lack an extensive large-scale feature extraction (LSFE) analysis through multiple benchmark data sets. LSFE analysis is required to identify and utilize the most relevant features and channels for emotion recognition and ASD prediction. Considering these challenges, for the first time, we have analyzed and evaluated an extensive feature set to select the optimal features using LSFE and feature selection algorithms (FSA). A set of up to eight most suitable channels was identified using different best-case FSA. The subject-wise importance of channels and features is also identified. The proposed method provides the best-case accuracies, precision, and recall of 95, 92, and 90%, respectively, for emotions prediction using a linear support vector machine (LSVM) classifier. It also provides the best-case accuracy, precision, and recall of 100% for ASD classification. This work utilized the largest number of benchmark data sets (5) and subjects (99) for validation reported till now in the literature. The LSVM classification algorithm proposed and utilized in this work has significantly lower complexity than the DNN, convolutional neural network (CNN), Naïve Bayes, and dynamic graph CNN used in recent ASD and emotion prediction systems.

show abstract

“…Moreover, DT ensembles have shown superior performance in other neural tasks such as Parkinsonian tremor detection using local field potentials (LFP) [9], [17] and migraine state classification from somatosensory evoked potentials (SSEP) [10]. ML models have also been explored for neural signal classification in applications such as emotion detection [11], [12], sleep stage classification [13], and for predicting memory dysfunction [18] and mental fatigue [19] (to potentially trigger a neurostimulation therapy). In [11], a convolutional neural network (CNN) SoC with online training capability was implemented for emotion recognition, Fig.…”

Section: A Symptom Detectionmentioning

confidence: 99%

“…Combined with an external feature extraction processor, the CNN classifier achieved an accuracy of 83.36% in a binary emotion detection task. A 4-layer neural network classifier was recently reported for emotion detection in autistic children [12]. This 2-channel EEG processor achieved a classification accuracy of 85.2% while consuming 10.1µJ/class.…”

Section: A Symptom Detectionmentioning

confidence: 99%

Closed-Loop Neural Interfaces with Embedded Machine Learning

Zhu

Shin

Shoaran

2020

Preprint

View full text Add to dashboard Cite

Neural interfaces capable of multi-site electrical recording, on-site signal classification, and closed-loop therapy are critical for the diagnosis and treatment of neurological disorders. However, deploying machine learning algorithms on lowpower neural devices is challenging, given the tight constraints on computational and memory resources for such devices. In this paper, we review the recent developments in embedding machine learning in neural interfaces, with a focus on design trade-offs and hardware efficiency. We also present our optimized treebased model for low-power and memory-efficient classification of neural signal in brain implants. Using energy-aware learning and model compression, we show that the proposed oblique trees can outperform conventional machine learning models in applications such as seizure or tremor detection and motor decoding.

show abstract

A 10.13µJ/Classification 2-Channel Deep Neural Network Based SoC for Negative Emotion Outburst Detection of Autistic Children

Cited by 24 publications

References 42 publications

Identification of Ferroptosis-Related Molecular Clusters and Immune Characterization in Autism Spectrum Disorder

Identification of Ferroptosis-Related Molecular Clusters and Immune Characterization in Autism Spectrum Disorder

Channels and Features Identification: A Review and a Machine-Learning Based Model With Large Scale Feature Extraction for Emotions and ASD Classification

Closed-Loop Neural Interfaces with Embedded Machine Learning

Contact Info

Product

Resources

About