GrAMIA-Net: Gray Matter Adaptive Multi View Imbalance Aware Attention Network for Detection of Parkinson’s Disease

Document Type : Bioinformatics-Naghibzadeh

Authors

Elham Afarandeh ¹

Rahil Hosseini ²

samaneh yazdani ¹

¹ Department of Computer Engineering, NT.C., Islamic Azad University, Tehran, Iran.

² Department of Computer Engineering, ShQ.C., Islamic Azad University, Shahre Qods,Tehran, Iran.

10.22067/cke.2026.99262.1196

Abstract

Diagnosing Parkinson’s disease (PD) at its early stages from structural MRI remains a difficult task, primarily because of scanner-dependent variations, skewed class distributions, and the subtle anatomical alterations associated with early neurodegeneration. To improve MRI-based Parkinson’s disease classification under these conditions, GrAMIA-Net is proposed as a multi-view deep learning framework built upon gray matter representations derived from T1-weighted MRI. The framework learns complementary information from sagittal, coronal, and axial perspectives and employs a hierarchical attention mechanism to emphasize discriminative spatial cues within each view. In addition, an inter-view attention module is utilized to aggregate information across different anatomical planes effectively. To mitigate the adverse effects of class imbalance, an Adaptive Imbalance Optimization (AIBO) scheme is integrated into the training process, thereby improving learning for underrepresented samples. Validation on the multi-scanner and class-imbalanced PPMI cohort shows that the proposed framework attains an accuracy of 0.862, an F1-score of 0.919, and an AUC of 0.720, demonstrating reliable predictive capability across categories. These results suggest that GrAMIA-Net provides a practical and computationally efficient framework for PD detection using single-modality GM representations, with potential applicability in real-world neuroimaging settings.

Keywords

Adaptive Imbalance Optimization

Gray Matter Maps

Multi-View Learning

PD Detection

T1-weighted Structural Imaging

Subjects

• Bioinformatics

---

References

[1] M. Absalyamova, D. Traktirov, V. Burdinskaya, V. Artemova, Z. Muruzheva, and M. Karpenko. (2025). Molecular basis of the development of Parkinson's disease. Neuroscience. [Online]. pp. 292–300. Available: https://doi.org/10.1016/j.neuroscience.2024.12.009

[2] N. Vijiaratnam, T. Simuni, O. Bandmann, H. R. Morris, and T. Foltynie. (2021). Progress towards therapies for disease modification in Parkinson's disease. The Lancet Neurology. [Online]. 20(7), pp. 559–572. Available: https://doi.org/10.1016/s1474-4422(21)00061-2

[3] X. Zhang et al. (2025). Optimal channel selection based on ABC-SVM in Parkinson's disease detection using short-time resting state EEG. J Supercomput. [Online]. 81, p. 977. Available: https://doi.org/10.1007/s11227-025-07461-x

[4] S. J. Duncan, K. Marques, J. Fawkes, L. J. Smith, and D. T. Wilkinson. (2024). Galvanic vestibular stimulation modulates EEG markers of voluntary movement in Parkinson's disease. Neuroscience. [Online]. 555, pp. 178–183. Available: https://doi.org/10.1016/j.neuroscience.2024.07.048

[5] X. Liu et al. (2021). A dual-branch model for diagnosis of Parkinson's disease based on the independent and joint features of the left and right gait. Appl Intell. [Online]. 51(10), pp. 7221–7232. Available: https://doi.org/10.1007/s10489-020-02182-5

[6] P. Patil and W. R. Ford. (2024). Parkinson's disease recognition using decorrelated convolutional neural networks: Addressing imbalance and scanner bias in rs-fMRI data. Biosensors. [Online]. 14(5), p. 259. Available: https://doi.org/10.3390/bios14050259

[7] S. G. Ryman and K. L. Poston. (2020). MRI biomarkers of motor and non-motor symptoms in Parkinson's disease. Parkinsonism & Related Disorders. [Online]. 73, pp. 85–93. Available:  https://doi.org/10.1016/j.parkreldis.2019.10.002

[8] A. S. Talai, J. Sedlacik, K. Boelmans, and N. D. Forkert. (2021). Utility of multi-modal MRI for differentiating of Parkinson's disease and progressive supranuclear palsy using machine learning. Frontiers in Neurology. [Online]. 12, p. 648548. Available: https://doi.org/10.3389/fneur.2021.648548

[9] E. Porter, A. A. Roussakis, N. P. Lao-Kaim, and P. Piccini. (2020). Multimodal dopamine transporter (DAT) imaging and magnetic resonance imaging (MRI) to characterise early Parkinson's disease. Parkinsonism & Related Disorders. [Online]. 79, pp. 26–33. Available: https://doi.org/10.1016/j.parkreldis.2020.08.010

[10] B. Hazhirkarzar et al. (2024). Machine learning for prediction of Parkinson's disease progression using DAT-SPECT imaging and multimodal MRI. Journal of Nuclear Medicine. [Online]. 65(2), p. 242322.

[11] Y. Zhou, S. Tinaz, and H. D. Tagare. (2021). Robust Bayesian analysis of early-stage Parkinson's disease progression using DaTscan images. IEEE Transactions on Medical Imaging. [Online]. 40(2), pp. 549–561. Available: https://doi.org/10.1109/tmi.2020.3031478

[12] E. Adeli et al. (2016). Joint feature-sample selection and robust diagnosis of Parkinson's disease from MRI data. NeuroImage. [Online]. 141, pp. 206–219. Available: https://doi.org/10.1016/j.neuroimage.2016.05.054

[13] A. A. Vijayakumari, H. H. Fernandez, and B. L. Walter. (2023). MRI-based multivariate gray matter volumetric distance for predicting motor symptom progression in Parkinson's disease. Scientific Reports. [Online]. 13, p. 19554. Available: https://doi.org/10.1038/s41598-023-46992-2

[14] M. Mohsen et al. (2024). Advanced MRI-based evaluation of gray and white matter changes in Parkinson's disease. Egyptian Journal of Radiology and Nuclear Medicine. [Online]. 55, p. 165. Available: https://doi.org/10.1186/s43055-024-01340-w

[15] D. Payares-Garcia, J. Mateu, and W. Schick. (2023). NeuroNorm: An R package to standardize multiple structural MRI. Neurocomputing. [Online]. 550, p. 126493. Available:  https://doi.org/10.2139/ssrn.4374278

[16] F. J. Martinez-Murcia, J. M. Górriz, J. Ramírez, and A. Ortiz. (2018). Convolutional neural networks for neuroimaging in Parkinson's disease: Is preprocessing needed? International Journal of Neural Systems. [Online]. 28(10), p. 1850035. Available: https://doi.org/10.1142/s0129065718500351 

[17] S. Soltaninejad, I. Cheng, and A. Basu. (2018). Towards the identification of Parkinson's disease using only T1 MR images. In Smart Multimedia. [Online]. Cham, Switzerland: Springer, pp. 145–156. Available: https://doi.org/10.48550/arXiv.1806.07489 

[18] R. Kumari, J. Singh, and A. Gosain. (2024). Diagnosis of Parkinson disease using ensemble methods for class imbalance problem. Presented at ICDAM 2023. [Online]. Available:  https://doi.org/10.1016/j.compbiomed.2023.107031 

[19] S. J. MacEachern and N. D. Forkert. (2021). Machine learning for precision medicine. Genome. [Online]. 64(4), pp. 416–425. Available: https://doi.org/10.1139/gen-2020-0131

[20] S. K. Sahu and A. S. Chowdhury. (2023). Multi-modal multi-class Parkinson disease classification using CNN and decision level fusion. arXiv. [Online]. Available: https://doi.org/10.48550/arXiv.2307.02978

[21] D. Jin et al. (2019). Attention-based 3D convolutional network for Alzheimer's disease diagnosis and biomarkers exploration. Presented at 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019). [Online]. Available: https://doi.org/10.1109/ISBI.2019.8759455

[22] M. Camacho et al. (2023). Explainable classification of Parkinson's disease using deep learning trained on a large multi-center database of T1-weighted MRI datasets. NeuroImage: Clinical. [Online]. 38, p. 103405. Available: https://doi.org/10.1016/j.nicl.2023.103405

[23] S. Soltaninejad. (2020). "Intelligent Parkinson's disease classification and progress monitoring using non-invasive techniques." Ph.D. dissertation, Dept. Comput. Sci., Univ. Alberta, Edmonton, AB, Canada. Available: https://doi.org/10.7939/r3-41sk-s535

[24] G. Solana-Lavalle and R. Rosas-Romero. (2021). Classification of PPMI MRI scans with voxel-based morphometry and machine learning to assist in the diagnosis of Parkinson's disease. Computer Methods and Programs in Biomedicine. [Online]. 198, p. 105793. Available: https://doi.org/10.1016/j.cmpb.2020.105793 

[25] H. W. Loh et al. (2021). Application of deep learning models for automated identification of Parkinson's disease: A review (2011–2021). Sensors. [Online]. 21(21), p. 7034. Available: https://doi.org/10.3390/s21217034

[26] T. N. Kinyanjui, K. Mugoye, and R. Kibuku. (2025). Multi-head self-attention fusion network for enhanced multi-class crop disease classification. Journal of AI and Data Mining. [Online]. 13(2), pp. 227–240. Available: https://doi.org/10.22044/jadm.2025.15689.2687 

[27] M. Golovanevsky, C. Eickhoff, and R. Singh. (2022). Multimodal attention-based deep learning for Alzheimer's disease diagnosis. Journal of the American Medical Informatics Association. [Online]. 29(12), pp. 2014–2022. Available: https://doi.org/10.1093/jamia/ocac168 

[28] V. Chauhan, A. Tiwari, and A. Kumar. (2025). An attention mechanism-based hybrid TimeAttentionBiLSTM architecture for long-term traffic forecasting. J Supercomput. [Online]. 81, p. 1257. Available: https://doi.org/10.1007/s11227-025-07747-0

[29] Y. Dai, Y. Gao, and F. Liu. (2021). TransMed: Transformers advance multi-modal medical image classification. Diagnostics. [Online]. 11(8), p. 1384. Available: https://doi.org/10.3390/diagnostics11081384 

[30] Z. Liu, H. Lu, X. Pan, M. Xu, R. Lan, and X. Luo. (2022). Diagnosis of Alzheimer's disease via an attention-based multi-scale convolutional neural network. Knowledge-Based Systems. [Online]. 238, p. 107942. Available:  https://doi.org/10.1016/j.knosys.2021.107942

[31] Y. Niu, N. Chen, H. Zhu, G. Li, and Y. Chen. (2024). Brain connectivity and time-frequency fusion-based auditory spatial attention detection. Neuroscience. [Online]. 560, pp. 397–405. Available: https://doi.org/10.1016/j.neuroscience.2024.09.017

[32] Z. Liu, H. Lu, X. Pan, M. Xu, R. Lan, and X. Luo. (2022). Diagnosis of Alzheimer's disease via an attention-based multi-scale convolutional neural network. Knowledge-Based Systems. [Online]. 238, p. 107942. Available: https://doi.org/10.1016/j.knosys.2021.107942 

[33] S. Datta, M. K. Acharaya, B. Chakrabarty, and N. Ghosh. (2024). MRI voxel-based morphometry to classify brain structural alterations in Parkinson's disease. Presented at IEEE 21st India Council International Conference (INDICON). [Online]. Available: https://doi.org/10.1109/INDICON63790.2024.10958550 

[34] K. Marek et al. (2011). The Parkinson Progression Marker Initiative (PPMI). Progress in Neurobiology. [Online]. 95(4), pp. 629–635. Available:  https://doi.org/10.1016/j.pneurobio.2011.09.005

[35] L. Dular, F. Pernuš, and Ž. Špiclin. (2024). Extensive T1-weighted MRI preprocessing improves generalizability of deep brain age prediction models. Computers in Biology and Medicine. [Online]. 173, p. 108320. Available:  https://doi.org/10.1101/2023.05.10.540134 

[36] J. Liu, H. Du, Q. Bi, H. Liao, and Y. Pan. (2022). MEST: Multi-plane embedding and spatial-temporal transformer for Parkinson's disease diagnosis. Presented at IEEE International Conference on Bioinformatics and Biomedicine (BIBM). [Online]. Available: https://doi.org/10.1109/BIBM55620.2022.9995093

[37] K. Aghajani. (2024). Deep learning approach for robust voice activity detection: Integrating CNN and self-attention with multi-resolution MFCC. Journal of AI and Data Mining. [Online]. 12(3), pp. 337–347. Available: https://doi.org/10.22044/jadm.2024.14839.2582

[38] M. S. Hossain and A. P. Paplinski. (2021). Dual focal loss to address class imbalance in semantic segmentation. Neurocomputing. [Online]. 462, pp. 69–87. Available: https://doi.org/10.1016/j.neucom.2021.07.055

[39] X. Cui et al. (2020). Gray matter atrophy in Parkinson's disease and the parkinsonian variant of multiple system atrophy: A combined ROI- and voxel-based morphometric study. Clinics. [Online]. 75, p. e1505. Available: https://doi.org/10.6061/clinics/2020/e1505

[40] N. U. Islam and R. Khanam. (2024). Machine learning-based Parkinson's disease detection and progression evaluation using gray and white matter segmentation from 3D MRI. Journal of Engineering and Applied Sciences. [Online]. Available:  https://doi.org/10.1186/s44147-024-00522-7