INTRODUCTION

Graph signal processing (GSP) has emerged as a powerful framework for linking time-varying functional brain activity from resting-state fMRI (rs-fMRI) to the underlying neural architecture derived from structural connectivity assessed with DWI [1]. We applied GSP to examine structural-functional coupling in patients with Parkinson’s disease (PwPD).

METHODS

MRI data from 17 healthy controls (HC) and 41 PwPD were employed. rs-fMRI preprocessing followed standardized pipelines (fMRIPrep [2], XCP-D [3]). Structural connectivity matrices were constructed using Schaefer’s parcellation with 200 regions and seven networks [4].

Within the GSP framework, functional activity was treated as graph signals on SC-defined nodes, and the Graph Fourier Transform was used to decompose these signals into structural harmonics. Graph spectral filtering isolated signal components with stronger or weaker structural coupling. We derived the individual Structural-Decoupling Index (SDI) for each node to quantify this coupling. SDI values near 0 indicate balance, while negative and positive values reflect strong coupling and decoupling, respectively.

A Wilcoxon rank-sum test (FDR-corrected, p<0.05) was applied to asses’ differences in SDI between PwPD and HC. Finally, a Random Forest classifier with 5-fold cross-validation was used to classify individuals based on SDI.

RESULTS

No significant differences were found between groups across the seven functional networks. However, the classifier achieved a balanced accuracy of 81%, an area under the receiver operating characteristic curve of 0.88, a sensitivity of 0.90, and a specificity of 0.73 (all averaged across folds). The five best features selected in the model represented SDI in the sensorimotor network.

CONCLUSION

SDI did not capture cross-sectional differences between HC and PwPD. However, it enabled remarkable classification performance, particularly in the sensorimotor areas, which are most affected by the disease.

REFERENCES

[1] M.G. Preti et al., “Decoupling of brain function from structure reveals regional behavioral specialization in humans”, Nat Commun, 18, Oct. 2019, doi: https://doi.org/10.1038/s41467-019-12765-7.

[2] O. Esteban et al., “FMRIPrep: a robust preprocessing pipeline for functional MRI,” Nat. Methods, vol. 16, no. 1, pp. 111–116, Jan. 2019, doi: 10.1038/s41592-018-0235-4.

[3] K. Mehta et al., “XCP-D: A robust pipeline for the post-processing of fMRI data,” Imaging Neurosci., vol. 2, pp. 1–26, Aug. 2024, doi: 10.1162/imag_a_00257.

[4] BT. Yeo et al., “The organization of the human cerebral cortex estimated by intrinsic functional connectivity”, J Neurophysiol, 8, Jun. 2011, doi: 10.1152/jn.00338.2011.