Wavelet-Based Bracketing, Time–Frequency Beta Burst Detection: New Insights in Parkinson's Disease

Parkinson’s Disease Research Report: Beta Burst Behavior Analysis Using a Wavelet-Based Time-Frequency Detection Framework

Background

Parkinson’s disease (PD) is a prevalent neurodegenerative disorder characterized by motor dysfunction, including tremors, rigidity, and bradykinesia. Recent studies have linked motor impairments in PD to excessive synchronization of neuronal activity within the beta frequency band (13–35 Hz). While traditionally viewed as sustained elevation, beta activity in PD is now understood to manifest as transient, exaggerated bursts of variable magnitude and duration. Current detection methods primarily focus on single frequency peaks within the beta band, potentially overlooking other critical information. This study introduces a novel wavelet-based method for identifying and analyzing beta bursts across a broad frequency range and explores their correlation with motor impairments.

Paper Source

This research was conducted by Tanmoy Sil, Ibrahem Hanafi, and colleagues, affiliated with institutions including the University Hospital Würzburg, Germany, and Imperial College London, UK. It was published in Neurotherapeutics (Volume 20, 2023) and made available online on October 11, 2023.

Methodology

Subjects

The study recruited seven patients (one female) with advanced PD, averaging 57.57 years of age and a disease duration of 9.29 years. All participants underwent Medtronic SenSight™ deep brain stimulation (DBS) electrode implantation, and chronic local field potential (LFP) recordings were collected at least three months post-surgery.

Data Acquisition

LFP data were recorded in a resting state at a sampling rate of 250 Hz for 20.9 seconds. Patients discontinued antiparkinsonian medication at least 12 hours prior to recordings, with DBS stimulation halted 30 minutes beforehand. Data were processed using wavelet transformations to create time-frequency spectrograms, with bursts identified through thresholding at the 80th percentile.

Wavelet Decomposition and Burst Identification

Morlet wavelets (10 cycles, 10–40 Hz range, 1 Hz resolution) were used for time-frequency decomposition. Bursts were defined as time-frequency regions exceeding the power threshold. Key metrics included: - Burst duration (Δt): Temporal extent of a burst. - Frequency range (Δf): Frequency span of a burst. - Burst power: Maximum amplitude within a burst region.

Statistical Analysis

Correlations between burst metrics and clinical motor scores (MDS-UPDRS III) were analyzed using Pearson’s correlation coefficient and repeated-measures ANOVA (RM-ANOVA).

Results

Beta Burst Characteristics Differ by Frequency Range

Significant differences were observed between low beta (13–20 Hz) and high beta (21–35 Hz) bursts: 1. Low beta bursts showed longer durations (Δt), broader frequency ranges (Δf), and positive correlations with motor scores. 2. High beta bursts exhibited negative correlations between burst metrics and motor scores.

Additionally, low beta bursts had a higher likelihood of occurrence and greater bilateral symmetry than high beta bursts.

Clinical Correlations

Long-duration low beta bursts positively correlated with motor symptom severity, suggesting a pathological role. Conversely, longer-duration high beta bursts were negatively correlated, indicating a possible compensatory mechanism.

Methodological Advantages

The wavelet-based method effectively detects beta bursts across a wide frequency range, providing broader insights than single-frequency methods. The novel frequency range metric (Δf) offers an additional dimension for burst characterization.

Implications

This study provides critical insights into the pathological mechanisms and therapeutic targets for PD: 1. Pathological Mechanisms: Low beta activity may reduce information coding capacity through excessive synchrony, contributing to motor impairments. 2. DBS Optimization: Identifying bursts with long durations or wide frequency ranges in the low beta band could guide parameter adjustments in DBS programming. 3. Novel Metrics: The introduction of Δf enriches beta burst analysis and enhances the understanding of their role in PD.

Future Directions

Further research could utilize long-term LFP recordings and high-resolution EEG to explore information flow in low beta activity and its causal links to motor symptoms. Additionally, investigating the effects of pharmacological interventions on Δf and burst characteristics may provide deeper insights into PD pathology.

Conclusion

This study presents a novel wavelet-based framework for detecting beta bursts and analyzing their properties. It highlights the contrasting behaviors of low and high beta bursts and their distinct correlations with motor impairments. The findings offer valuable perspectives for advancing PD research and optimizing DBS interventions.