Restoration of abnormal sleep EEG power in patients with insomnia disorder after 1Hz rTMS over left DLPFC

Xumeng Zhao, Jiayi Liu, Ziqiang Shao, Xiaoyang Liu, Zhen Wang, Kai Yuan, Bingqian Zhang, Yan Li, Xiaona Sheng, Yifei Zhu, Yansu GuoFrontiers in Psychiatry 2024 · DOI

TL;DR

This study investigates how repetitive transcranial magnetic stimulation (rTMS) affects abnormal sleep electroencephalogram (EEG) power in individuals with insomnia disorder (ID). The findings suggest that 1Hz rTMS over the left dorsolateral prefrontal cortex (DLPFC) can partially restore abnormal sleep EEG power, particularly in stage 2 non-rapid eye movement (NREM2) sleep, offering a potential therapeutic avenue for ID.

What Problem Does This Paper Address?

Insomnia disorder (ID) is a prevalent sleep disorder characterized by difficulties initiating and maintaining sleep, leading to significant daytime impairment. A leading theory for ID is "hyperarousal," suggesting excessive cortical activation. While electroencephalogram (EEG) power analysis has revealed abnormalities in ID patients, there's been limited focus on specific sleep stages like REM and NREM, and it remains unclear if these abnormalities can be reversed. This paper aims to address this gap by investigating the impact of 1Hz rTMS on sleep EEG power in ID patients, exploring whether this non-invasive brain stimulation can normalize these abnormal patterns.

Methods

The study employed an observational design comparing 26 patients with insomnia disorder (ID) to 26 healthy controls (HCs). For the ID group, overnight polysomnography (PSG) data and behavioral measures were collected before and after a 4-week course of 1Hz rTMS applied to the left dorsolateral prefrontal cortex (DLPFC).

Key methodological aspects include:

  • PSG Data Acquisition and Analysis: Standard PSG was used to record EEG, EOG, and EMG. EEG data was preprocessed (filtering, down-sampling) and sleep stages (Wake, N1, N2, N3, REM) were automatically scored using YASA and visually verified.
  • Spectral Analysis: The core of the analysis involved calculating the relative power of EEG in different frequency bands (delta, theta, alpha, sigma, beta, gamma) for each sleep stage and electrode. Relative power, as opposed to absolute power, was chosen for its sensitivity to individual differences and its ability to reflect normalized brain activity.
  • rTMS Procedure: A MagPro R30 TMS stimulator with an 8-shaped coil was used to deliver 1Hz rTMS to the left DLPFC. Treatment consisted of 20 sessions over 4 consecutive weeks, with each session lasting 30 minutes.
  • Statistical Analysis: Unpaired t-tests were used for baseline comparisons between ID patients and HCs. Paired t-tests were used to compare pre- and post-treatment data in ID patients. Pearson correlation analysis examined the relationship between behavioral measures and sleep EEG relative power. FDR correction was applied for multiple comparisons.

Datasets

The study utilized data from two groups:

  • 26 patients with Insomnia Disorder (ID): These participants met DSM-V criteria for insomnia, had a Pittsburgh Sleep Quality Index (PSQI) > 5, and an Insomnia Severity Index (ISI) > 8. They had not received other sleep-related treatments in the two weeks prior to and during the rTMS intervention.
  • 26 Healthy Controls (HCs): These participants were matched for age, sex, and education, with PSQI < 5 and ISI < 8, and no history of psychiatric or sleep disorders.

No new datasets were introduced; the study relied on collected PSG and behavioral data from these participants.

Key Insights and Ideas

  • Hyperarousal in Sleep: The study reinforces the hyperarousal theory of insomnia by demonstrating widespread abnormalities in sleep EEG relative power in ID patients compared to HCs, particularly in the beta and gamma bands, indicative of increased cortical activity.
  • Stage-Specific Abnormalities: The paper highlights that these EEG abnormalities are not uniform across all sleep stages, with distinct patterns observed in NREM2, NREM3, and REM sleep.
  • rTMS as a Restorative Tool: The core idea is that rTMS, a non-invasive brain stimulation technique, might be able to "rewire" or normalize these disrupted EEG patterns, offering a potential therapeutic benefit beyond subjective symptom relief.
  • NREM2 Sleep as a Target: The findings suggest that NREM2 sleep is a particularly responsive stage to 1Hz rTMS, showing improvements in delta and beta band relative power. This points to a specific window for intervention.
  • Limited Effect on Deeper Sleep: The study notes that deeper sleep stages like N3 might be less susceptible to the effects of 1Hz rTMS, potentially requiring more intensive or prolonged interventions.

Key Results

  • Baseline Differences: ID patients showed significantly higher PSQI, ISI, BDI, and BAI scores compared to HCs. Objectively, they exhibited higher WASO, N2%, and awakenings, and lower N3% at baseline.
  • EEG Abnormalities: At baseline, ID patients had significantly higher relative power in beta and gamma bands across multiple electrodes and sleep stages (N2, N3, REM) compared to HCs. They also showed lower delta band power in REM and N3 sleep.
  • Correlations: Sleep efficiency (SE) was negatively correlated with gamma band power in N2 and N3 sleep. Alpha band power in REM sleep was positively correlated with sleep onset latency (SOL).
  • rTMS Effects: After 1Hz rTMS treatment, ID patients showed significant improvements in subjective sleep quality (PSQI, ISI, BDI, BAI) and some objective sleep measures (SE, SOL). Crucially, rTMS led to a significant increase in delta band power and a decrease in beta band power in N2 sleep. However, these changes did not pass FDR correction.
  • Hemispheric Differences: Preliminary, uncorrected findings suggested potential improvements in the right frontal and occipital regions (F4, O2) after rTMS, but not in the directly stimulated left frontal region (F3).

Limitations and Future Work

The authors acknowledge several limitations:

  • Limited Electrode Coverage: The study used a limited number of EEG electrodes, which restricts the spatial resolution of the findings. Future studies could benefit from a denser electrode montage.
  • Lack of Sham Group: The observational design did not include a sham rTMS group or a control group receiving intervention, making it difficult to definitively attribute all observed changes solely to the rTMS treatment.
  • Single PSG Measurement: PSG was performed only once before and after treatment, which might not capture the full variability of sleep patterns. Future work could incorporate multiple PSG sessions or sleep diaries.
  • Causality: The study cannot definitively establish whether sleep quality improvements preceded or followed the EEG changes, or if they occurred simultaneously.

Future work could involve:

  • Larger, Randomized Controlled Trials: Including sham-controlled groups to establish causality.
  • Denser Electrode Arrays: To obtain more comprehensive spatial information.
  • Longer-Term Follow-up: To assess the durability of rTMS effects.
  • TMS-EEG Integration: To directly investigate the immediate effects of TMS on EEG.
  • Targeted Interventions: Exploring different rTMS parameters or targets for deeper sleep stages.

Summary updated

Abstract

The relative power of sleep EEG exhibited a significant correlation with sleep measures in ID patients, and demonstrated notable differences from HCs across the delta, beta, and gamma frequency bands. Furthermore, our findings suggest that rTMS treatment may partially ameliorate relative power abnormalities in patients with ID.
BibTeX
@inproceedings{zhao2024restoration,
  author = {Xumeng Zhao and Jiayi Liu and Ziqiang Shao and Xiaoyang Liu and Zhen Wang and Kai Yuan and Bingqian Zhang and Yan Li and Xiaona Sheng and Yifei Zhu and Yansu Guo},
  title = {{Restoration of abnormal sleep EEG power in patients with insomnia disorder after 1Hz rTMS over left DLPFC}},
  booktitle = {Frontiers in Psychiatry},
  year = {2024},
  doi = {10.3389/fpsyt.2024.1431837},
}
Figure