Cortical networks of parkinsonian gait: a metabolic and functional connectivity study

Franziska Pellegrini, Nicoló Gabriele Pozzi, Chiara Palmisano, Giorgio Marotta, Andreas K. Buck, Stefan Haufe, Ioannis U. IsaiasmedRxiv 2023 · DOI

TL;DR

Researchers have mapped the "roadblocks" and "rerouting" of brain signals that occur when people with Parkinson’s Disease (PD) walk. By combining metabolic imaging with high-density EEG, the study reveals that parkinsonian gait is characterized by a breakdown in communication between motor control centers (like the SMA) and sensory areas, forcing the brain to rely on visual processing as a compensatory "detour."

What Problem Does This Paper Address?

While we often think of walking as a simple, automated task, it actually requires a sophisticated "symphony" of neural synchronization across the cerebral cortex. In Parkinson’s Disease, this symphony falls out of tune, leading to hypokinetic gait (slow, small steps) and freezing.

Most previous research focused on subcortical structures like the basal ganglia. However, we still don't fully understand how the cortex—the brain's higher-level processing center—adapts or fails during actual overground walking. The challenge is that imaging a moving brain is difficult; you can't put a walking person in an MRI machine. This paper addresses this gap by using a multimodal approach to capture brain activity in action.

Methods

The authors employed a clever "tri-modal" experimental setup with 10 PD patients in a "meds-off" state (after pausing dopaminergic medication):

  1. Metabolic Imaging (PET): Patients were injected with [18F]FDG, a radioactive glucose tracer, while walking or resting. Since the tracer "locks" into the brain cells active during the task, the researchers could scan the patients after they finished walking to see which regions were most active.
  2. High-Density EEG: Patients wore a 64-channel EEG cap to record electrical activity during walking. The researchers used Source Localization (LCMV) to map scalp signals back to specific brain regions and Time-Reversed spectral Granger Causality (TRGC) to determine the direction of information flow between those regions.
  3. Kinematics: Inertial measurement units (IMUs) and EMG sensors tracked stride duration and muscle activity to correlate brain patterns with actual physical performance.

Datasets

This study used a primary clinical dataset consisting of 10 individuals with idiopathic Parkinson’s Disease (5 female, 5 male).

  • Clinical Scores: Motor symptoms were quantified using the Unified Parkinson’s Disease Rating Scale (UPDRS-III).
  • Comparison: The study used an intra-subject design, comparing the same patients in two conditions: "Resting" (sitting) vs. "Walking" (unperturbed overground walking on a 60m path).

Key Insights and Ideas

  • The SMA Roadblock: The Supplementary Motor Area (SMA) is the brain's "conductor" for automated movement. The study found a significant reduction in information flow from the SMA to sensory and motor areas during walking. This suggests that the primary deficit isn't just "less activity," but a failure to integrate motor plans with sensory feedback.
  • Visual Rerouting: As a compensation for the SMA failure, the brain appears to "reroute" information through visuomotor areas. Specifically, the researchers observed increased connectivity from the SMA to the cuneus and superior occipital gyrus.
  • The Lingual Gyrus "Detour": The study suggests that the lingual gyrus acts as a critical hub in this compensatory network. When the standard motor pathways are blocked, the brain uses visual information (processed via the inferior fronto-occipital fasciculus) to "push" locomotor information forward to the motor cortex.
  • Connectivity vs. Performance: The slower the patient's gait (longer stride duration), the worse the information flow was from the SMA to the precuneus. This directly links network "roadblocks" to clinical symptoms.

Key Results

  • Metabolic Activation: Walking significantly increased glucose metabolism in the SMA, paracentral lobule, precuneus, and several occipital regions compared to rest.
  • Kinematic Correlation: There was a strong positive correlation ($r = 0.80, p = 0.017$) between the severity of motor symptoms (UPDRS-III) and stride duration.
  • Directed Connectivity: In the beta frequency band (associated with motor control), walking was characterized by:
  • Decreased flow: SMA $\rightarrow$ Paracentral/Postcentral/Parietal areas.
  • Increased flow: SMA $\rightarrow$ Visuomotor areas (Cuneus/Occipital).
  • Hypokinetic Markers: Poor information flow from the calcarine to the lingual gyrus was specifically associated with slower, parkinsonian gait.

Limitations and Future Work

The authors acknowledge several limitations:

  • Sample Size: With only 10 patients, the findings are a "proof of concept" that requires validation in larger cohorts.
  • Lack of Control Group: The study did not include healthy age-matched controls, making it hard to distinguish between "normal walking networks" and "Parkinson's-specific adaptations."
  • Cortical Focus: EEG is limited to the cortex; the study could not directly measure the real-time influence of deep-brain structures like the basal ganglia or thalamus during the walk.
  • Future Directions: The researchers suggest that future work should investigate how dopaminergic medication or deep brain stimulation (DBS) "unblocks" these pathways or if visual-cueing therapies (like laser shoes) work by specifically targeting the lingual gyrus "detour" identified here.

Summary updated

Abstract

Abstract Objective Locomotion is an automated voluntary movement sustained by coordinated neural synchronization across a distributed brain network. The cerebral cortex is central for adapting the locomotion pattern to the environment and alterations of cortical network dynamics can lead to gait impairments. Gait problems are a common symptom with a still unclear pathophysiology and represent an unmet therapeutical need in Parkinson’s disease. Little is known about the cortical network dynamics of locomotor control in these patients. Methods We studied the cortical basis of parkinsonian gait by combining metabolic brain imaging with high-density EEG recordings and kinematic measurements performed at rest and during unperturbed overground walking. Results We found significant changes in functional connectivity between frontal, sensorimotor and visuomotor cortical areas during walking as compared to resting. Specifically, hypokinetic gait was associated with poor information flow from the supplementary motor area (SMA) to precuneus and from calcarine to lingual gyrus as well as high information flow from calcarine to cuneus. Interpretation Our findings support a role for visuomotor integration processes in PD-related hypokinetic gait and suggest that reinforcing visual information may act as a compensatory strategy to allow SMA-related feedforward locomotor control in PD.
BibTeX
@misc{pellegrini2023cortical,
  author = {Franziska Pellegrini and Nicoló Gabriele Pozzi and Chiara Palmisano and Giorgio Marotta and Andreas K. Buck and Stefan Haufe and Ioannis U. Isaias},
  title = {{Cortical networks of parkinsonian gait: a metabolic and functional connectivity study}},
  year = {2023},
  doi = {10.1101/2023.10.09.23296653},
  howpublished = {medRxiv},
}
Figure