A Multi-Level Fuzzy Explainable Prototype Network for Sex Classification Across Brain Atlases

Document Type : Research Paper

Author

Electrical and Computer engineering department, Tarbiat Modares University

10.22111/ijfs.2026.54363.9632

Abstract

Sex differences in resting-state functional brain organization have been widely reported, yet many deep learning models prioritize predictive performance over interpretability, limiting their ability to reveal underlying neurobiological mechanisms. We propose FXP-Net, a fuzzy explainable prototype-based neural network that combines temporal convolutional modeling with fuzzy prototype learning to enable intrinsically interpretable sex classification from resting-state fMRI. FXP-Net maps windowed ROI time series into a compact latent space in which class-specific prototypes are learned and activated through graded fuzzy membership functions, allowing each prediction to be explained via similarity to a small set of interpretable prototypes rather than opaque feature activations. To enforce interpretability, we introduce a prototype-regularized training objective that encourages intra-class clustering, inter-class separation, prototype diversity, and sparse activation. Across multiple brain atlases and repeated runs, FXP-Net achieves subject-level classification performance competitive with, and in several settings statistically superior to, a strong ConvNet baseline, with peak balanced accuracy exceeding 95 %. Importantly, a multi-level interpretability analysis, combining prototype-level and decision-level attributions reveals anatomically consistent patterns across Brainnetome, Glasser, and Gordon parcellations. Quantitative overlap analysis highlights regions within the dorsolateral prefrontal cortex (DLPFC), including lateral area 9/10 and area 46, as demonstrating robust cross-atlas agreement and consistently high importance. These findings are stable across runs and align with prior evidence linking DLPFC organization to sex-related differences in executive control and higher-order cognition. The Python implementation of the proposed FXP‑Net architecture is publicly available at: https://github.com/Mansooreh-Pakravan/FXPNet-Model.

Keywords

Main Subjects

References

[1] A. Abrol, Z. Fu, M. Salman, R. Silva, V. D. Calhoun, Deep learning encodes robust discriminative neuroimaging
representations to outperform standard machine learning, Nature Communications, 12 (2021), 353. https://doi.
org/10.1038/s41467-020-20655-6
[2] E. J. Bacon, et al., Neuroimage analysis using artificial intelligence approaches: A systematic review,
Medical and Biological Engineering and Computing, 62(9) (2024), 2599-2627. https://doi.org/10.1007/
s11517-024-03097-w
[3] D. Badre, A. D.Wagner, Left ventrolateral prefrontal cortex and the cognitive control of memory, Neuropsychologia,
45(13) (2007), 2883-2901. https://doi.org/10.1016/j.neuropsychologia.2007.06.015
[4] N. Bibi, J. Courtney, K. McGuinness, Enhancing brain disease diagnosis with XAI: A review of recent studies,
ACM Transactions on Computing for Healthcare, 6(2) (2025), 1-35. https://doi.org/10.1145/3709152
[5] I. D. Borlea, R. E. Precup, F. Dragan, A. B. Borlea, Centroid update approach to K-means clustering, Advances
in Electrical and Computer Engineering, 17(4) (2017). https://doi.org/10.4316/AECE.2017.04001
[6] R. Botvinik-Nezer, et al., Variability in the analysis of a single neuroimaging dataset by many teams, Nature, 582
(2020), 84-88. https://doi.org/10.1038/s41586-020-2314-9
[7] R. L. Buckner, J. R. Andrews-Hanna, D. L. Schacter, The brain’s default network: Anatomy, function, and relevance
to disease, Annals of the New York Academy of Sciences, 1124(1) (2008), 1-38. https://doi.org/10.1196/ annals.1440.011
[8] P. W. Burgess, I. Dumontheil, S. J. Gilbert, The gateway hypothesis of rostral prefrontal cortex (area 10) function,
Trends in Cognitive Sciences, 11(7) (2007), 290-298. https://doi.org/10.1016/j.tics.2007.05.004
[9] N. C¸ a˘gman, S. Engino˘glu, Fuzzy soft matrix theory and its application in decision making, Iranian Journal of Fuzzy
Systems, 9(1) (2012), 109-119. https://doi.org/10.22111/ijfs.2012.229
[10] N. C¸ a˘gman, S. Engino˘glu, F. C¸ ıtak, Fuzzy soft set theory and its applications, Iranian Journal of Fuzzy Systems,
8(3) (2011), 137-147. https://doi.org/10.22111/ijfs.2011.292
[11] H. S. Chiang, D. H. Shih, B. Lin, M. H. Shih, An APN model for arrhythmic beat classification, Bioinformatics,
30(12) (2014), 1739-1746. https://doi.org/10.1093/bioinformatics/btu101
[12] M. Corbetta, G. L. Shulman, Control of goal-directed and stimulus-driven attention in the brain, Nature Reviews
Neuroscience, 3(3) (2002), 201-215. https://doi.org/10.1038/nrn755
[13] E. Dhamala, K. W. Jamison, M. R. Sabuncu, A. Kuceyeski, Sex classification using long-range temporal dependence
of resting-state functional MRI time series, Human Brain Mapping, 41(13) (2020), 3567-3579. https://doi.org/ 10.1002/hbm.25030
[14] A. Etkin, T. Egner, R. Kalisch, Emotional processing in anterior cingulate and medial prefrontal cortex, Trends in
Cognitive Sciences, 15(2) (2011), 85-93. https://doi.org/10.1016/j.tics.2010.11.004
[15] L. Fan, et al., The human brainnetome Atlas: A new brain atlas based on connectional architecture, Cerebral
Cortex, 26(8) (2016), 3508-3526. https://doi.org/10.1093/cercor/bhw157
[16] E. S. Finn, et al., Functional connectome fingerprinting: Identifying individuals using patterns of brain connectivity,
Nature Neuroscience, 18 (2015), 1664-1671. https://doi.org/10.1038/nn.4135
[17] S. Gadgil, Q. Zhao, A. Pfefferbaum, et al., Spatio-temporal graph convolution for resting-state fMRI analysis,
Medical Image Computing and Computer Assisted Intervention – MICCAI 2020, 12267 (2020), 528-538. https:
//doi.org/10.1007/978-3-030-59728-3_52
[18] M. Gholami, M. Faramarzi, N. Alipour, M. Pakravan, Node embedding extraction for causal brain graphs in fMRI
data, In Proceedings of the 29th International Computer Conference, Computer Society of Iran (CSICC), (2025),
1-6. https://doi.org/10.1109/CSICC65765.2025.10967434
[19] M. F. Glasser, et al., The minimal preprocessing pipelines for the human connectome project, NeuroImage, 80
(2013), 105-124. https://doi.org/10.1016/j.neuroimage.2013.04.127
[20] M. F. Glasser, et al., A multi-modal parcellation of human cerebral cortex, Nature, 536 (2016), 171-178. https:
//doi.org/10.1038/nature18933
[21] E. M. Gordon, et al., Generation and evaluation of a cortical area parcellation from resting-state correlations,
Cerebral Cortex, 26(1) (2016), 288-303. https://doi.org/10.1093/cercor/bhu239
[22] R. M. Hutchison, et al., Dynamic functional connectivity: Promise, issues, and interpretations, NeuroImage, 80
(2013), 360-378. https://doi.org/10.1016/j.neuroimage.2013.05.079
[23] M. Khosla, K. Jamison, G. H. Ngo, A. Kuceyeski, M. R. Sabuncu, Machine learning in resting-state fMRI analysis,
Magnetic Resonance Imaging, 64 (2019), 101-121. https://doi.org/10.1016/j.mri.2019.05.031
[24] P. Y. Kim, J. Kwon, et al., SwiFT: Swin 4D fMRI Transformer, Advances in Neural Information Processing
Systems (NeurIPS 2023), 36 (2023), 42015-42037. https://doi.org/10.48550/arXiv.2307.05916
[25] B. H. Kim, J. C. Ye, J. J. Kim, Learning dynamic graph representation of brain connectome with spatio-temporal
attention, Advances in Neural Information Processing Systems (NeurIPS 2021), 34 (2021), 4314-4327. https:
//proceedings.neurips.cc/paper/2021/hash/229754a0ca4d715694200427845774a3-Abstract.html
[26] J. Kwon, J. Seo, H. Wang, T. Moon, S. Yoo, J. Cha, Predicting task-related brain activity from resting-state brain
dynamics with fMRI Transformer, Imaging Neuroscience, 3 (2025). https://doi.org/10.1162/imag_a_00440
[27] R. Leech, D. J. Sharp, The role of the posterior cingulate cortex in cognition and disease, Brain, 137(1) (2014),
12-32. https://doi.org/10.1093/brain/awt162
[28] M. Leeming, J. Suckling, Deep learning for sex classification in resting-state and task functional brain networks
from the UK Biobank, NeuroImage, 241 (2021), 118409. https://doi.org/10.1016/j.neuroimage.2021.118409
[29] N. Leonardi, D. Van De Ville, On spurious and real fluctuations of dynamic functional connectivity during rest,
NeuroImage, 104 (2015), 430-436. https://doi.org/10.1016/j.neuroimage.2014.09.007
[30] X. Li, X. Zhou, N. Dvornek, M. Zhang, S. Gao, J. Zhuang, et al., BrainGNN: Interpretable brain graph neural
network for fMRI analysis, Medical Image Analysis, 74 (2021), 102233. https://doi.org/10.1016/j.media.
2021.102233
[31] V. Menon, M. D’Esposito, The role of PFC networks in cognitive control and executive function, Neuropsychopharmacology, 47 (2022), 90-103. https://doi.org/10.1038/s41386-021-01152-w
[32] R. Nazari, M. Salehi, A. Shoeibi, An explainable connectome convolutional transformer for multimodal autism
spectrum disorder classification, International Journal of Neural Systems, 35(8) (2025), 2550043. https://doi.
org/10.1142/S0129065725500431s
[33] J. Ning, Neural network-based pattern recognition in the framework of edge computing, Romanian Journal of Information Science and Technology, 2024(1) (2024), 106-119. https://doi.org/10.59277/ROMJIST.2024.1.08
[34] M. Pakravan, FuzzyCAL: A fuzzy logic enhanced causal attention GNN for robust cocaine use disorder classification,
Iranian Journal of Fuzzy Systems, 22(6) (2025), 167-182. https://doi.org/10.22111/ijfs.2025.52840.9343
[35] M. Pakravan, Uncertainty aware type II fuzzy graph modeling of resting state fMRI uncovers robust sex differences,
Journal of Neuroscience Methods, 431 (2026). https://doi.org/10.1016/j.jneumeth.2026.110745
[36] M. Pakravan, Explainable graph attention network on resting state fMRI dynamic functional connectivity reveals
cerebellar hub breakdown in chronic knee osteoarthritis pain, Iranian Journal of Science and Technology, Transactions
of Electrical Engineering, 50 (2026), 289-300. https://doi.org/10.1007/s40998-025-01012-z
[37] U. Pervaiz, R. Vidaurre, M. W. Woolich, M. E. Smith, Optimising network modelling methods for fMRI, NeuroImage,
211 (2020), 116604. https://doi.org/10.1016/j.neuroimage.2020.116604
[38] J. D. Power, et al., Methods to detect, characterize, and remove motion artifact in resting-state fMRI, NeuroImage,
84 (2014), 320-341. https://doi.org/10.1016/j.neuroimage.2013.08.048
[39] S. Ramezanzadeh, A. Memariani, S. Saati, Data envelopment analysis with fuzzy random inputs and outputs: A
chance-constrained programming approach, Iranian Journal of Fuzzy Systems, 2(2) (2005), 21-35. https://doi.
org/10.22111/IJFS.2005.479
[40] S. J. Ritchie, et al., Sex differences in the adult human brain: Evidence from 5,216 UK Biobank participants,
Cerebral Cortex, 28(8) (2018), 2959-2975. https://doi.org/10.1093/cercor/bhy109
[41] S. Ryali, Y. Zhang, et al., Deep learning models reveal replicable and behaviorally relevant sex differences in brain
organization, Proceedings of the National Academy of Sciences, 121(9) (2024), e2310012121. https://doi.org/
10.1073/pnas.2310012121
[42] M. A. Saket, M. Pakravan, Cross-atlas identification of narrative hubs via multi-embedding graph models in fMRI
data, Neuroinformatics, 24 (2026), 29. https://doi.org/10.1007/s12021-026-09787-0
[43] S. S. Salehi, et al., There is no single functional atlas even for a single individual: Functional parcel definitions
change with task, NeuroImage, 208 (2020), 116366. https://doi.org/10.1016/j.neuroimage.2019.116366
[44] E. Tjoa, C. Guan, A survey on explainable artificial intelligence (XAI): Toward medical XAI, IEEE Transactions
on Neural Networks and Learning Systems, 32(11) (2021), 4793-4813. https://doi.org/10.1109/TNNLS.2020.
3027314
[45] D. C. Van Essen, et al., The WU-Minn human connectome project: An overview, NeuroImage, 80 (2013), 62-79.
https://doi.org/10.1016/j.neuroimage.2013.05.041
[46] C. Wang, A. U. Yaari, et al., BrainBERT: Self-supervised representation learning for intracranial recordings, arXiv
preprint arXiv:2302.14367, (2023). https://doi.org/10.48550/arXiv.2302.14367
[47] S. Weis, K. R. Patil, et al., Sex classification by resting-state brain connectivity, Cerebral Cortex, 30(2) (2020),
824-835. https://doi.org/10.1093/cercor/bhz129
[48] L. A. Zadeh, Fuzzy sets, Information and Control, 8(3) (1965), 338-353. https://doi.org/10.1016/
S0019-9958(65)90241-X
[49] A. Zalesky, A. Fornito, L. Cocchi, L. L. Gollo, M. Breakspear, Time-resolved resting-state brain networks, Proceedings
of the National Academy of Sciences, 111(28) (2014), 10341-10346. https://doi.org/10.1073/pnas.
1400181111
[50] L. Zhao, Embedding human brain function via transformer, Medical Image Computing and Computer-Assisted
Intervention (MICCAI), 13431 (2022), 366-375. https://doi.org/10.1007/978-3-031-16431-6_35
Iranian Journal of Fuzzy Systems
Volume 23, Issue 3
May and June 2026
Pages 165-191

Files

Share

How to cite

Statistics