Increased Dynamic Amplitude of Low Frequency Fluctuation in Primary Insomnia

Xianyun Meng, Jianjun Zheng, Yingpeng Liu, Yi Yin, Kelei Hua, Shishun Fu, Yunfan Wu, Guihua JiangFrontiers in Neurology 2020 · DOI

TL;DR

This study reveals that individuals with primary insomnia exhibit increased temporal variability in their brain's intrinsic activity, particularly in regions associated with memory and emotion. By analyzing dynamic functional magnetic resonance imaging (fMRI) data, the researchers found that this "instability" in brain function is more pronounced in insomniacs and correlates with anxiety levels. This work highlights the importance of considering the dynamic nature of brain activity, not just its average state, to understand neurological conditions like insomnia.

What Problem Does This Paper Address?

Primary insomnia (PI) is a widespread sleep disorder affecting a significant portion of the population, characterized by persistent difficulty falling asleep, staying asleep, or experiencing non-restorative sleep. Despite its prevalence, the underlying physiological mechanisms of PI remain poorly understood. While resting-state functional magnetic resonance imaging (fMRI) has been a valuable tool in exploring PI, most previous studies have focused on static measures of brain activity. This approach overlooks the fact that brain activity is inherently dynamic and fluctuates over time, even during rest. The authors aim to address this gap by investigating the temporal variability of intrinsic brain activity in individuals with PI, hypothesizing that this dynamic aspect might hold crucial clues to the disorder's pathophysiology.

Methods

The researchers employed a novel approach using resting-state fMRI to capture the dynamic nature of brain activity. The core technique is dynamic Amplitude of Low Frequency Fluctuation (dALFF). Unlike traditional ALFF, which assumes brain activity is stationary, dALFF quantifies the temporal variability of intrinsic brain activity.

Here's a breakdown of their method:

  • Data Acquisition: Resting-state fMRI data was collected from 59 patients with PI and 47 matched healthy controls (HCs).
  • dALFF Calculation: The fMRI time series was segmented into sliding windows. For each window, the ALFF was calculated. This process was repeated across all windows, generating a series of dALFF maps for each subject.
  • Temporal Variability Measurement: The variance of these dALFF maps across time was then computed for each voxel. This variance serves as a measure of the temporal variability or "instability" of intrinsic brain activity.
  • Group Comparison: The variance of dALFF was compared between the PI group and the HC group using statistical analyses, controlling for age, sex, and education.
  • Correlation Analysis: The researchers also investigated the relationship between the altered dALFF and clinical symptom severity, specifically anxiety levels, using linear regression.
  • Validation: To ensure the robustness of their findings, they performed validation steps, including assessing the correlation between dALFF and head motion (mean frame-wise displacement) and re-analyzing data with different window lengths for dALFF calculation.

Datasets

The study utilized a dataset comprising resting-state fMRI scans from:

  • 59 patients diagnosed with Primary Insomnia (PI) according to DSM-IV criteria.
  • 47 matched healthy controls (HCs).

The paper does not introduce any new benchmark datasets.

Key Insights and Ideas

  • Beyond Static Measures: The study powerfully demonstrates that focusing solely on average brain activity (static measures) can miss crucial information about neurological conditions. The dynamics of brain activity are equally, if not more, important for understanding disorders like insomnia.
  • Brain Instability in Insomnia: The core finding of increased temporal variability (variance of dALFF) in key brain regions suggests that the brains of individuals with PI are less stable and more prone to fluctuations in their intrinsic activity. This could reflect a state of chronic hyperarousal or a disrupted regulatory mechanism.
  • Hippocampus and Insula Involvement: The specific regions showing increased dALFF variance – the bilateral hippocampus, parahippocampus, right putamen, and right anterior insula – are critical for memory, emotion processing, and self-awareness. This points to how insomnia might impact these cognitive and emotional functions.
  • Anxiety Link: The correlation between increased dALFF variance in the right putamen and higher anxiety scores (SAS) suggests a potential neurobiological link between anxiety and the dynamic instability of brain activity in PI. This could explain why anxiety is a common comorbidity and symptom of insomnia.

Key Results

  • Patients with PI showed significantly increased variance of dALFF in the bilateral hippocampus extending to the parahippocampus, the right putamen, and the right anterior insula cortex compared to healthy controls.
  • The variance of dALFF in the right putamen was positively correlated with Self-rating Anxiety Scale (SAS) scores in patients with PI (R² = 0.079, p = 0.032 uncorrected).
  • There were no significant differences in mean head motion (FD) between the PI and HC groups, and no significant correlation between dALFF variance and head motion in the PI group, suggesting that head motion did not confound the results.
  • Validation with different parameter selections confirmed the reliability of the increased dALFF findings.

Limitations and Future Work

The authors acknowledge several limitations:

  • Single Dataset: The study was conducted on a single dataset, and they recommend future research using independent sample sets to validate their findings.
  • Lack of Polysomnography (PSG): The study did not include PSG for patients with PI. PSG is the gold standard for objectively measuring sleep architecture and could provide a more comprehensive understanding of the sleep disturbances in the participants.
  • Emotional Function: While they found a correlation with anxiety, they suggest that more work is needed to fully elaborate the role of altered dALFF in the neuromechanism of PI, particularly concerning abnormal emotional function.

Future work could focus on:

  • Replicating these findings in larger, independent cohorts.
  • Integrating dALFF analysis with PSG data to link brain activity dynamics directly to sleep stages and quality.
  • Exploring the specific contributions of different brain regions to the observed dynamic instability and their impact on cognitive and emotional symptoms.
  • Investigating potential interventions that might modulate this dynamic brain activity in PI.

Summary updated

Abstract

The physiological mechanism underlying primary insomnia (PI) is poorly understood. Resting-state functional magnetic resonance imaging (fMRI) has emerged as a powerful tool to explore PI. However, previous studies ignore the dynamics of the brain activity. In the current study, we aimed to explore altered dynamic intrinsic brain activity in PI. Fifty-nine patients with PI and 47 matched healthy controls (HCs) were recruited and underwent resting-state fMRI. The variance of dynamic amplitude of low frequency fluctuation (dALFF) maps across time was calculated to measure the temporal variability of intrinsic brain activity and then compared between patients with PI and HCs. As a result, patients with PI presented increased variance of dALFF in the bilateral hippocampus extending to the parahippocampus, the right putamen and the right anterior insula cortex. In addition, the variance of dALFF in the right putamen was positively correlated with Self-rating Anxiety Scale (SAS) score in PI. Our results revealed increased instability of intrinsic activity in PI.
BibTeX
@inproceedings{meng2020increased,
  author = {Xianyun Meng and Jianjun Zheng and Yingpeng Liu and Yi Yin and Kelei Hua and Shishun Fu and Yunfan Wu and Guihua Jiang},
  title = {{Increased Dynamic Amplitude of Low Frequency Fluctuation in Primary Insomnia}},
  booktitle = {Frontiers in Neurology},
  year = {2020},
  doi = {10.3389/fneur.2020.00609},
}
Figure