Heavy–Light Soft-Vote Fusion of EEG Heatmaps for Autism Spectrum Disorder Detection
Abstract
Autism spectrum disorder is a neurodevelopmental condition that affects social communication and behaviour, and diagnosis still relies on subjective behavioural assessment. Electroencephalography provides a noninvasive view of brain activity but is noisy and often analysed with handcrafted features or evaluation schemes that risk data leakage. This study proposes a deep learning pipeline that combines wavelet denoising, EEG-to-image encoding, and heavy-light decision fusion for autism detection from EEG. Sixteen-channel EEG from children and adolescents with autism and typically developing peers in the KAU dataset is denoised using discrete wavelet transform shrinkage, segmented into fixed 4 second windows, and rendered as pseudo colour heatmaps. These images are used to fine-tune five ImageNet pretrained architectures under a unified training protocol with 5-fold cross-validation. Heavy-light fusion combines one heavyweight backbone and one lightweight backbone through weighted soft voting on class posterior probabilities. The strongest single model, ConvNeXt Tiny, attains about 97.25 percent accuracy and 97.10 percent F1 score at the window level. The best heavy light pair, ConvNeXt plus ShuffleNet, reaches about 99.56 percent accuracy and 99.53 percent F1, with sensitivity and specificity in the 99 percent range. Fusion mainly reduces missed ASD windows without increasing false alarms, indicating complementary error patterns between heavy and light models. These findings show that the proposed denoise encode classify pipeline with heavy light fusion yields more robust autism EEG classification than individual backbones and can support EEG-based decision support in autism screening.
Downloads
References
S. Yun, “Advances, challenges, and prospects of electroencephalography-based biomarkers for psychiatric disorders: a narrative review,” Journal of Yeungnam Medical Science, vol. 41, no. 4, pp. 261–268, Oct. 2024, doi: 10.12701/jyms.2024.00668.
J. Shan et al., “A scoping review of physiological biomarkers in autism,” Front Neurosci, vol. 17, 2023, doi: 10.3389/fnins.2023.1269880.
J. Li et al., “Identification of autism spectrum disorder based on electroencephalography: A systematic review,” Comput Biol Med, vol. 170, p. 108075, 2024, doi: https://doi.org/10.1016/j.compbiomed.2024.108075.
Y. Xu, Z. Yu, Y. Li, Y. Liu, Y. Li, and Y. Wang, “Autism spectrum disorder diagnosis with EEG signals using time series maps of brain functional connectivity and a combined CNN–LSTM model,” Comput Methods Programs Biomed, vol. 250, p. 108196, 2024, doi: https://doi.org/10.1016/j.cmpb.2024.108196.
J. Rogala et al., “Enhancing autism spectrum disorder classification in children through the integration of traditional statistics and classical machine learning techniques in EEG analysis,” Sci Rep, vol. 13, no. 1, p. 21748, 2023, doi: 10.1038/s41598-023-49048-7.
S. Phadikar, N. Sinha, R. Ghosh, and E. Ghaderpour, “Automatic Muscle Artifacts Identification and Removal from Single-Channel EEG Using Wavelet Transform with Meta-Heuristically Optimized Non-Local Means Filter,” Sensors, vol. 22, no. 8, 2022, doi: 10.3390/s22082948.
M. Grobbelaar et al., “A Survey on Denoising Techniques of Electroencephalogram Signals Using Wavelet Transform,” Sep. 01, 2022, Multidisciplinary Digital Publishing Institute (MDPI). doi: 10.3390/signals3030035.
I. H. Elshekhidris, M. B. Mohamedamien, and A. Fragoon, “Wavelet Transforms for EEG Signal Denoising and Decomposition,” 2023.
A. Chaddad, Y. Wu, R. Kateb, and A. Bouridane, “Electroencephalography Signal Processing: A Comprehensive Review and Analysis of Methods and Techniques,” Jul. 01, 2023, Multidisciplinary Digital Publishing Institute (MDPI). doi: 10.3390/s23146434.
S. N. S. S. Daud and R. Sudirman, “Wavelet Based Filters for Artifact Elimination in Electroencephalography Signal: A Review,” Ann Biomed Eng, vol. 50, no. 10, pp. 1271–1291, 2022, doi: 10.1007/s10439-022-03053-5.
M. Melinda, M. Oktiana, Y. Yunidar, N. H. Nabila, and I. K. A. Enriko, “Classification of EEG Signal using Independent Component Analysis and Discrete Wavelet Transform based on Linear Discriminant Analysis,” International Journal on Informatics Visualization (JOIV), vol. 7, no. 3, pp. 830–838, Sep. 2023.
M. Melinda, F. H. Juwono, I. K. A. Enriko, M. Oktiana, S. Mulyani, and K. Saddami, “Application Of Continuous Wavelet Transform and Support Vector Machine for Autism Spectrum Disorder Electroencephalography Signal Classification,” Radioelectronic and Computer Systems, no. 3(107), pp. 73–90, 2023, doi: 10.32620/reks.2023.3.07.
F. Fahmi, M. Melinda, P. D. Purnamasari, E. Elizar, and A. Rafiki, “Recognition of EEG Features in Autism Disorder Using SWT and Fisher Linear Discriminant Analysis,” Diagnostics, vol. 15, no. 18, p. 2291, Sep. 2025, doi: 10.3390/diagnostics15182291.
H. T. Lee, H. R. Cheon, S. H. Lee, M. Shim, and H. J. Hwang, “Risk of data leakage in estimating the diagnostic performance of a deep-learning-based computer-aided system for psychiatric disorders,” Sci Rep, vol. 13, no. 1, Dec. 2023, doi: 10.1038/s41598-023-43542-8.
G. Brookshire et al., “Data leakage in deep learning studies of translational EEG,” Front Neurosci, vol. Volume 18-2024, 2024, doi: 10.3389/fnins.2024.1373515.
R. Kessler, A. Enge, and M. A. Skeide, “How EEG preprocessing shapes decoding performance,” Commun Biol, vol. 8, no. 1, Dec. 2025, doi: 10.1038/s42003-025-08464-3.
N. S. Amer and S. B. Belhaouari, “Exploring new horizons in neuroscience disease detection through innovative visual signal analysis,” Sci Rep, vol. 14, no. 1, p. 4217, 2024, doi: 10.1038/s41598-024-54416-y.
M. A. Bravo-Ortiz et al., “SpectroCVT-Net: A convolutional vision transformer architecture and channel attention for classifying Alzheimer’s disease using spectrograms,” Comput Biol Med, vol. 181, p. 109022, 2024, doi: https://doi.org/10.1016/j.compbiomed.2024.109022.
E. Vafaei, F. Nowshiravan Rahatabad, S. K. Setarehdan, and P. Azadfallah, “Extracting a Novel Emotional EEG Topographic Map Based on a Stacked Autoencoder Network,” J Healthc Eng, vol. 2023, 2023, doi: 10.1155/2023/9223599.
N. Bajaj and J. Requena Carrión, “Deep Representation of EEG Signals Using Spatio-Spectral Feature Images,” Applied Sciences (Switzerland), vol. 13, no. 17, Sep. 2023, doi: 10.3390/app13179825.
A. Rafiki et al., “Implementation of Vision Transformer for Early Detection of Autism Based on EEG Signal Heatmap Visualization,” Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics, vol. 7, no. 1, pp. 102–112, 2025.
J. Li, J. Chen, Y. Tang, C. Wang, B. A. Landman, and S. K. Zhou, “Transforming medical imaging with Transformers? A comparative review of key properties, current progresses, and future perspectives,” Med Image Anal, vol. 85, p. 102762, 2023, doi: https://doi.org/10.1016/j.media.2023.102762.
Z. Li, R. Zhang, Y. Zeng, L. Tong, R. Lu, and B. Yan, “MST-net: A multi-scale swin transformer network for EEG-based cognitive load assessment,” Brain Res Bull, vol. 206, p. 110834, 2024, doi: https://doi.org/10.1016/j.brainresbull.2023.110834.
Y. Mehmood and U. I. Bajwa, “Brain tumor grade classification using the ConvNext architecture,” Digit Health, vol. 10, p. 20552076241284920, 2024, doi: 10.1177/20552076241284920.
Y. A. Saadoon, M. Khalil, and D. Battikh, “Predicting Epileptic Seizures Using EfficientNet-B0 and SVMs: A Deep Learning Methodology for EEG Analysis,” Bioengineering, vol. 12, no. 2, 2025, doi: 10.3390/bioengineering12020109.
S. K. Prabhakar, J. J. Lee, and D.-O. Won, “Ensemble Fusion Models Using Various Strategies and Machine Learning for EEG Classification,” Bioengineering, vol. 11, no. 10, 2024, doi: 10.3390/bioengineering11100986.
A. Karim, S. Ryu, and I. cheol Jeong, “Ensemble learning for biomedical signal classification: a high-accuracy framework using spectrograms from percussion and palpation,” Sci Rep, vol. 15, no. 1, p. 21592, 2025, doi: 10.1038/s41598-025-05027-8.
M. Salvi et al., “Multi-modality approaches for medical support systems: A systematic review of the last decade,” Information Fusion, vol. 103, Mar. 2024, doi: 10.1016/j.inffus.2023.102134.
T. Wu, X. Kong, Y. Zhong, and L. Chen, “Automatic detection of abnormal EEG signals using multiscale features with ensemble learning,” Front Hum Neurosci, vol. Volume 16-2022, 2022, doi: 10.3389/fnhum.2022.943258.
K. Munadi et al., “A Deep Learning Method for Early Detection of Diabetic Foot Using Decision Fusion and Thermal Images,” Applied Sciences (Switzerland), vol. 12, no. 15, Aug. 2022, doi: 10.3390/app12157524.
Md. H. R. Rabbani and S. Md. R. Islam, “Deep learning networks based decision fusion model of EEG and fNIRS for classification of cognitive tasks,” Cogn Neurodyn, vol. 18, no. 4, pp. 1489–1506, 2024, doi: 10.1007/s11571-023-09986-4.
M. Zakir Ullah and D. Yu, “Grid-tuned ensemble models for 2D spectrogram-based autism classification,” Biomed Signal Process Control, vol. 93, p. 106151, 2024, doi: https://doi.org/10.1016/j.bspc.2024.106151.
M. J. Alhaddad et al., “Diagnosis Autism by Fisher Linear Discriminant Analysis FLDA via EEG,” 2012.
M. Murias, S. J. Webb, J. Greenson, and G. Dawson, “Resting State Cortical Connectivity Reflected in EEG Coherence in Individuals With Autism,” Biol Psychiatry, vol. 62, no. 3, pp. 270–273, 2007, doi: https://doi.org/10.1016/j.biopsych.2006.11.012.
E. V. Orekhova et al., “Excess of High Frequency Electroencephalogram Oscillations in Boys with Autism,” Biol Psychiatry, vol. 62, no. 9, pp. 1022–1029, Nov. 2007, doi: 10.1016/j.biopsych.2006.12.029.
R. Coben, A. R. Clarke, W. Hudspeth, and R. J. Barry, “EEG power and coherence in autistic spectrum disorder,” Clinical Neurophysiology, vol. 119, no. 5, pp. 1002–1009, May 2008, doi: 10.1016/j.clinph.2008.01.013.
Y. Xia, K. Li, D. Li, J. Nan, and R. Lu, “An Improved VMD and Wavelet Hybrid Denoising Model for Wearable SSVEP-BCI,” 2024. [Online]. Available: www.ijacsa.thesai.org
D. L. Donoho, “De-Noising by Soft-Thresholding,” 1995.
S. G. Chang, B. Yu, and M. Vetterli, “Adaptive wavelet thresholding for image denoising and compression,” IEEE Transactions on Image Processing, vol. 9, no. 9, pp. 1532–1546, 2000, doi: 10.1109/83.862633.
Y. Huang, P. Wen, B. Song, and Y. Li, “Real-Time Depth of Anaesthesia Assessment Based on Hybrid Statistical Features of EEG,” Sensors, vol. 22, no. 16, 2022, doi: 10.3390/s22166099.
W. Liu, K. Jia, and Z. Wang, “Graph-based EEG approach for depression prediction: integrating time-frequency complexity and spatial topology,” Front Neurosci, vol. Volume 18-2024, 2024, doi: 10.3389/fnins.2024.1367212.
L. Cao et al., “A Novel Deep Learning Method Based on an Overlapping Time Window Strategy for Brain–Computer Interface-Based Stroke Rehabilitation,” Brain Sci, vol. 12, no. 11, Nov. 2022, doi: 10.3390/brainsci12111502.
S. Y. Ke et al., “Classification of autism spectrum disorder using electroencephalography in Chinese children: a cross-sectional retrospective study.,” Front Neurosci, vol. 18, p. 1330556, 2024, doi: 10.3389/fnins.2024.1330556.
T. Xu, Y. Zhou, Z. Hou, and W. Zhang, “Decode Brain System: A Dynamic Adaptive Convolutional Quorum Voting Approach for Variable-Length EEG Data,” Complexity, vol. 2020, 2020, doi: 10.1155/2020/6929546.
J. Duan, J. Xiong, Y. Li, and W. Ding, “Deep learning based multimodal biomedical data fusion: An overview and comparative review,” Information Fusion, vol. 112, p. 102536, 2024, doi: https://doi.org/10.1016/j.inffus.2024.102536.
Y. Dong et al., “Subject-Independent EEG Classification of Motor Imagery Based on Dual-Branch Feature Fusion,” Brain Sci, vol. 13, no. 7, 2023, doi: 10.3390/brainsci13071109.
I. Jemal, L. Abou-Abbas, K. Henni, A. Mitiche, and N. Mezghani, “Domain adaptation for EEG-based, cross-subject epileptic seizure prediction,” Front Neuroinform, vol. Volume 18-2024, 2024, doi: 10.3389/fninf.2024.1303380.
A. M. Alghamdi, M. U. Ashraf, A. A. Bahaddad, K. A. Almarhabi, W. A. Al Shehri, and A. Daraz, “Cross-subject EEG signals-based emotion recognition using contrastive learning,” Sci Rep, vol. 15, no. 1, p. 28295, 2025, doi: 10.1038/s41598-025-13289-5.
Copyright (c) 2026 Melinda Melinda, Syahrul Gazali, Yuwaldi Away, Aufa Rafiki, W.K Wong, Muliyadi Muliyadi, Siti Rusdiana
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlikel 4.0 International (CC BY-SA 4.0) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
.png)
.png)
.png)
.png)
.png)
