From our study, it is clear that DMN and SN connectivity decrease in PD patients regardless of their cognitive status. The CEN, however, exhibited similar connectivity across all groups. Regarding the decrease in DMN and SN connectivity in the PD patient groups, it should be noted that there were no significant differences between both groups of patients. These results add to previous reports showing a decrease in DMN connectivity in PD patients with normal cognition (Disbrow et al. 2014; Christopher et al. 2014; Yao et al. 2014; Tessitore et al. 2019), suggesting that the reduction in this network may not be exclusively linked to cognitive alterations. This could be partly explained by the accumulation of β-amyloid deposits in posterior brain regions, including parts of the DMN, observed as motor symptoms progress (Aracil-Bolaños et al. 2019). At the same time, the reduction in SN connectivity could also be associated with incipient basal ganglia alterations, as studies report parallel dopamine depletion in the striatum and insula (Monchi et al. 2007; Shine et al. 2013). Structurally, striatal neurons are highly interconnected with insular cortex neurons (Chikama et al. 1997; Fudge et al. 2005), a central SN node. The decrease in SN connectivity in patients with PD may thus reveal early alterations in both structures due to the loss of D2 signaling in the insula, which would interrupt the modulation of SN activity, impairing its function of coordinating interactions with other brain networks such as the DMN (Menon and Uddin 2010).
Although several studies have reported an association between cognitive decline and DMN alterations in PD patients (Rektorova et al. 2012), optimal MoCA performance may depend above all on adequate coordination of the DMN and SN. Whereas DMN interactions, lMPFC/rPL in particular, showed a negative correlation with MoCA total score (although this correlation was lost in the PD-MCI group), SN interactions such as lAI/rAI showed a positive correlation in PD patients with normal cognition. These findings are consistent with decreased inferior parietal to medial temporal FC within the DMN reported in a cohort of PD patients without cognitive impairment (Tessitore et al. 2012). The inverse relationship between MoCA global performance and DMN and SN connectivity is noteworthy, as it suggests that MoCA performance could serve as a cognitive indicator reflecting the adequate coordination of both networks. In healthy adult brains, SN connectivity increases and DMN connectivity is suppressed in the presence of external stimuli (Raichle et al. 2001; Greicius et al. 2003; Fox et al. 2005), indicating that inverse coupling between the SN and the DMN occurs in optimal cognition. This coupling, however, was not found in our PD-MCI cohort. The positive correlation between lAI/rSMG connectivity and MoCA total score, exclusive to this group, is nevertheless consistent with studies in PD patients with MCI and dementia (Christopher et al. 2014; Wolters et al. 2019; Aracil-Bolaños et al. 2019), suggesting that increased connectivity between these regions of the SN network may be a compensatory mechanism of the rSMG to counteract cognitive decline (Chen et al. 2015; Wolters et al. 2019).
Executive performance also involves posterior cortical FC in PD patients with MCI. Executive MoCA subdomain scores displayed a positive correlation with connectivity between the DMN’s bilateral parietal regions as well as between the lAI and the rSMG within the SN among PD-MCI. However, memory subdomain scores only correlated with SN (specifically lAI/ACC) connectivity in the PD-NC group, revealing FC pattern differences between both PD patient groups. Despite many executive disturbances being linked to prefrontal region alterations, the FC differences associated with MoCA subdomains exhibited by our groups could be explained by the dual syndrome hypothesis (Kehagia et al. 2013), which states that two different patterns of cognitive impairment may generally occur in patients with PD: an anterior, frontostriatal executive syndrome more common early in the disease and with a low rate of evolution to dementia, and a more posterior cortical syndrome, which has been related to worse cognitive prognoses over time (Williams-Gray et al. 2007; Tessitore et al. 2019). Likewise, our results suggest that executive decline involves significant changes in the SN at the level of the anterior insula and related areas such as the supramarginal gyrus and not only frontostriatal loop alterations. This has also been reported in recent studies comparing the FC of PD patients with and without cognitive impairment (Wolters et al. 2019; Aracil-Bolaños et al. 2019). Pathophysiological studies have also shown that PD patients with MCI differed from cognitively normal PD patients in the presence of greater striatal dopamine depletion related to greater D2 receptor loss in the insula (Christopher et al. 2014). Thus, our results for MoCA subdomains and RSNs add to the evidence pointing to cognitive decline in PD also being associated with disruptions in networks subserving frontoparietal cortical regions (Lewis et al. 2003; Lebedev et al. 2014; Caminiti et al. 2015).
It has been reported that dopaminergic therapy diminishes DMN integrity (Krajcovicova et al. 2012). However, in our study, no association was found between levodopa medications and their relationship to the explored networks, which limits our ability to elucidate the effects of dopamine on the interaction of cognitive RSNs. It must also be noted that the fMRI was performed in both the “ON” and “OFF” medication state among the recruited PD patients.