Functional reorganization of the locomotor network in Parkinson patients with freezing of gait

Abstract

Freezing of gait (FoG) is a transient inability to initiate or maintain stepping that often accompanies advanced Parkinson's disease (PD) and significantly impairs mobility. The current study uses a multimodal neuroimaging approach to assess differences in the functional and structural locomotor neural network in PD patients with and without FoG and relates these findings to measures of FoG severity. Twenty-six PD patients and fifteen age-matched controls underwent resting-state functional magnetic resonance imaging and diffusion tensor imaging along with self-reported and clinical assessments of FoG. After stringent movement correction, fifteen PD patients and fourteen control participants were available for analysis. We assessed functional connectivity strength between the supplementary motor area (SMA) and the following locomotor hubs: 1) subthalamic nucleus (STN), 2) mesencephalic and 3) cerebellar locomotor region (MLR and CLR, respectively) within each hemisphere. Additionally, we quantified structural connectivity strength between locomotor hubs and assessed relationships with metrics of FoG. FoG+ patients showed greater functional connectivity between the SMA and bilateral MLR and between the SMA and left CLR compared to both FoG- and controls. Importantly, greater functional connectivity between the SMA and MLR was positively correlated with i) clinical, ii) self-reported and iii) objective ratings of freezing severity in FoG+, potentially reflecting a maladaptive neural compensation. The current findings demonstrate a re-organization of functional communication within the locomotor network in FoG+ patients whereby the higher-order motor cortex (SMA) responsible for gait initiation communicates with the MLR and CLR to a greater extent than in FoG- patients and controls. The observed pattern of altered connectivity in FoG+ may indicate a failed attempt by the CNS to compensate for the loss of connectivity between the STN and SMA and may reflect a loss of lower-order, automatic control of gait by the basal ganglia.

Figure 1. Locomotor hubs used as regions of interest for functional connectivity analysis.

X = 7 is chosen for display as it is the only slice on which all ROIs have some overlap. Please refer to the Methods section for coordinates describing the centroid of each sphere.

Figure 2. Structural connectivity ROIs in the A) sagittal, B) coronal, and C) axial plane.

The SMA, as defined by Mayka et(2006) is shown in blue. The STN, as defined by Prodoehl et al (2008) is shown in yellow. The MLR, as defined by Fling et al. (2013), is shown in purple. All ROIs are overlaid on the MNI 1 mm template and are displayed in neurologic convention (right = right). Similar ROIs for the SMA and MLR were used in the left hemisphere (not displayed here).

Figure 3. Functional connectivity strength between the SMA and the other locomotor hubs.

FoG+ patients show greater connectivity to the MLR and CLR, but reduced connectivity to the STN. **P<0.001, *P<0.05.

Figure 4. Structural connectivity between the SMA and right STN, as assessed by diffusion tensor imaging.

A) Normalized z-scores for structural connectivity strength between the rSTN – SMA. HC have significantly greater connectivity than either FoG− or FoG+. B) Rendered fiber tracts (red) from one healthy control demonstrating the structural connectivity within the hyperdirect loop between the SMA and STN (both displayed in green). *P<0.05. HC = healthy controls.

Figure 5. Associations between locomotor network functional connectivity and freezing severity.

A) Functional connectivity strength of the lMLR – SMA loop was positively correlated with clinical rating of FoG severity during the single task turning condition (r = 0.71). B) Functional connectivity strength of the rMLR – SMA loop was also positively correlated with objective sensor measurement of FoG severity during dual-task turning (0.76).