Embedded Files
Jutika Borah, Abdur R. Fayjie, Rajkumar Saini, Debarun Chakraborty, Bhabesh Deka, Foteini Liwicki
jutikaborah13@gmail.com
Introduction 🟢
Parkinson’s disease is the second most common neurodegenerative disorder, causing progressive motor dysfunction along with associated non-motor symptoms such as cognitive impairment. While slow and insidious cognitive decline can occur at any stage of the disease, rapid decline or mental changes are critical. Early cognitive impairments affecting attention, visuospatial processing, memory, and executive function are typical [1] and can increase the risk of early progression to dementia. The underlying neurophysiological mechanisms of cognitive decline in Parkinson’s disease remain unclear. While slowing and frequency variations on EEG have been repeatedly described, network-level directional interactions across brain regions remain unclear.
Challenges 🔴
The identification of disease progression, as well as the discovery of biomarkers to better predict cognitive decline, and identify patients at risk for early and rapid cognitive impairment.
Background ⚫
While EEG slowing is a recognized risk factor, we move beyond simple power analysis by leveraging Granger Causality (GC) on resting-state EEG to map directional functional network flow. We applied GC across all frequency bands and functionally relevant regions, specifically modelling neural influence between the motor, cognitive, temporal, and parietal-occipital networks. This cross-spectral causal investigation reveals how Parkinson's disease reconfigures brain network hierarchy: for instance, weakened cognitive to parietal-occipital drive indicates declining visuospatial integration. In contrast, altered motor ←→ cognitive influence reflects difficulty integrating movement planning with executive control.
🟣 Proposed Method 🟣
🟣 Proposed Method 🟣
Our framework utilizes GC on multi-band, region-averaged EEG signals to quantify the directional strength of neural interactions and statistical significance. We hypothesize that abnormal causal drive originating from the cognitive network towards the temporal, motor, and visuospatial regions serves as an objective biomarker for emerging cognitive impairments in Parkinson’s disease.
The methodology involves a targeted preprocessing pipeline (Independent Component Analysis (ICA), referencing, epoching) followed by Granger causality estimation across the cognitive, temporal, motor, and parietal-occipital networks, providing a scalable signature for pathophysiological change.
Findings 🟠
This study presents a region-wise EEG analysis integrating spectral band power with Granger Causality to investigate directional network interactions related to cognitive decline in Parkinson’s disease. Moving beyond simple power analysis, the causality framework quantifies directional influence between functionally relevant cortical networks, including those involved in mental, motor, temporal, and parietal-occipital functions.
Preliminary analyses reveal pathological signatures, including increased slowing in Delta and Theta bands, reduced inter-regional coherence, and significantly altered directional influence in cortical networks. Specific to causality, findings indicate weakened cognitive → parietal-occipital drive, consistent with declining visuospatial integration, and altered motor ←→ cognitive influence, reflecting a drawback in executive control over movement planning.
Our core findings involve the identification of pathological slow-wave hyperconnectivity, characterized by significant alterations in the strength and direction of delta/theta oscillations. The work addresses the urgent clinical need to map the underlying network hierarchy reconfiguration in Parkinson’s disease as revealed by slowing EEG, providing an objective means of identifying patients at risk of cognitive impairment or progressive cognitive changes.
Limitations 🔴
Assumptions of linearity and stationarity constraints GC; integrating it with frequency-resolved and anatomically informed analysis enhances interpretability and clinical relevance. Differentiating cognitive decline or memory changes in normal aging from those of mild cognitive impairment or cognitive impairment in a patient with Parkinson’s disease.
Conclusion ⚫
Multi-band, region-specific causal connectivity reveals early disruption of cognitive influence in Parkinson’s disease, particularly toward visuospatial and motor systems. Our GC-based framework provides an interpretable, objective measure of pathological network breakdown, supporting EEG-derived biomarkers for tracking cognitive decline.
References:
[1] Dag Aarsland, Lucia Batzu, Glenda M Halliday, Gert J Geurtsen, Clive Ballard, K Ray Chaudhuri, and Daniel Weintraub. Parkinson disease-associated cognitive impairment. Nature reviews Disease primers, 7(1), 2021.
Page updated
Report abuse