FV2. Cognition-related Parkinson’s disease pattern with functional MRI: Validation and clinical correlates

Abstract

Background Parkinson’s disease and related cognitive deficits have been associated with specific functional topographies termed functional PD-related pattern (fPDRP) and PD cognition-related pattern (fPDCP) ( Vo et al., 2017 ). Using multivariate analysis in resting-state functional MRI data this novel approach discriminated normal controls from PD patients based on their pattern expression and was associated with the severity of clinical measures. Aim We aim to investigate whether the functional PDCP topography is reproducible across different PD populations and whether fPDCP expression values can discriminate PD patients with and without mild cognitive impairment. Method Rs-fMRI scans from 22 PD patients in the on-state and 13 matched control subjects were analyzed using independent component analysis in conjunction with bootstrap resampling to identify a fPDCP topography in an independent cohort of PD patients from Berlin (Charite). Neuropsychological testing results were assessed and correlated with pattern expression. For further validation, subject scores for the fPDCP BERLIN pattern were computed in an independent group of 21 PD patients from North Shore (NS), USA. Subject scores and temporal dynamics were estimated for prospective cases with dual regression using the spatial maps generated from the Berlin derivation set and z-scored with regard to the normal control group. Results The fPDCP Berlin topography resembled the original fPDCP NS network topography ( Fig. 1 ). Subjects scores for the different fPDCP topographies from Berlin and North Shore were closely correlated in the Berlin sample (r = 0.67; p Fig. 2 ). fPDCP expression was significantly elevated in patients with MCI compared to their cognitively unimpaired counterparts in the NS cohort (p Fig. 2 ). Conclusions These findings suggest that fPDCP represents a stable disease related network topography across independent patient populations. This and the ability to discriminate PD patients with mild cognitive impairment indicate that fPDCP could be used as non-invasive imaging biomarker for cognitive dysfunction in PD patients.