Literature Sharing | Treating Insomnia: Research on the Neural Mechanism and Efficacy Prediction of taVNS
Original Article Link: Predicting Treatment Response to Transcutaneous Auricular Vagus Nerve Stimulation in Insomnia Patients: A Multivoxel Pattern Analysis Based on Resting‑State Functional Connectivity.pdf
Cannot fall asleep, light‑sleeping, frequent awakenings, frequent dreams… Insomnia is no longer a trivial issue that "goes away when you get tired", but a real signal of imbalanced brain networks.
Among numerous non‑invasive physical therapies, transcutaneous auricular vagus nerve stimulation (taVNS) has become a popular option for insomnia treatment owing to its advantages of safety, convenience and freedom from drug dependence. However, a key question has long remained unresolved: given the same taVNS treatment, why do some patients achieve remarkable effects while others show hardly any improvement? Can we identify suitable candidates before treatment?
Recently, a research team from Guangzhou University of Chinese Medicine published an important study in BMC Psychiatry. Using resting‑state functional magnetic resonance imaging (fMRI), the team provided clear answers: the core mechanism of taVNS for insomnia treatment lies in remodeling insula‑related networks; baseline functional connectivity of the left insula‑visual/auditory circuit can directly predict therapeutic response, serving as an important biomarker for precise screening of responsive patients.
This means that physical therapy for insomnia has truly entered a new era featuring predictability, subject screening and precise intervention, instead of blind trials.
Research Background: Critical Unmet Need in Insomnia Treatment
Primary insomnia is essentially characterized by long‑term abnormal hyperactivity of the brain salience network, with the insula acting as its core hub. The insula modulates vigilance as well as the balance between visual, auditory and default‑mode networks. In insomnia patients, the insula often stays in a state of "hypersensitivity and hyper‑connectivity", preventing the brain from relaxing and falling asleep.
Although taVNS has been proven to improve insomnia, unclear neural mechanisms and the inability to predict treatment outcomes in advance constitute major bottlenecks for its clinical promotion.
This study addresses these core issues: to reveal whether taVNS exerts therapeutic effects via insula‑network remodeling, and to identify neuroimaging biomarkers capable of predicting treatment response beforehand for precise screening of potential responders.
Study Design: Scientific and Rigorous, Double‑Blind Controlled Design for Higher Credibility
Study Subjects
A total of 67 patients with primary insomnia completed the trial: 34 in the real taVNS group and 33 in the sham stimulation (staVNS) group. All participants met DSM‑5 diagnostic criteria with PSQI scores > 7. Patients with secondary insomnia or craniocerebral diseases were excluded.
Intervention Protocol
Figure 1: Auricular stimulation sites for taVNS. Orange dots: cymba conchae and cavum conchae (effective stimulation sites); blue dot: cauda helicis (sham‑stimulation site); taVNS: transcutaneous auricular vagus nerve stimulation
Stimulation sites: real‑stimulation group received stimulation on left cavum conchae (vagus‑nerve‑innervated area); sham‑stimulation group received stimulation on cauda helicis (without vagus‑nerve innervation);
Treatment duration: 30 minutes per session, twice daily, 5 days per week, for 4 consecutive weeks;
Full double‑blind design: participants, operators and evaluators were all blinded to group assignment
Assessment and Analysis
Figure 2: Flowchart of prediction analysis. taVNS: transcutaneous auricular vagus nerve stimulation; FC: functional connectivity; PSQI: Pittsburgh Sleep Quality Index; CV: cross‑validation; SVR: support vector regression; CAL: calcarine sulcus; SOG: superior occipital gyrus; MOG: middle occipital gyrus; MTG: middle temporal gyrus
Clinical scales: PSQI, ISI, SAS, SDS
Brain imaging: 3.0T resting‑state fMRI; seed‑based functional‑connectivity analysis with bilateral insula as seeds
Prediction model: support vector regression (SVR) multivoxel pattern analysis, 10‑fold cross‑validation plus 5000‑permutation test
Key Results: Dual Breakthroughs in Clinical Efficacy and Neural Mechanisms
After 4‑week intervention, sleep improvements in the taVNS group were significantly greater than those in the sham‑stimulation group:
The total PSQI score decreased by 6.94 points, nearly twice the reduction observed in the sham‑stimulation group
Significant improvements were observed in sleep duration, sleep efficiency, sleep fragmentation and anxiety status
Table 1: Baseline clinical variables for taVNS and StaVNS groups
Clinical outcome indicators of the two groups (pre‑treatment vs. post‑treatment)
Neural Mechanism Validated: taVNS Remodels Insula Functional Networks
For the first time, this study clearly demonstrates that taVNS achieves brain‑network remodeling by reducing abnormal functional connectivity between the left insula and multiple brain regions. Brain regions exhibiting significantly decreased functional connectivity include:
Visual network: superior occipital gyrus, middle occipital gyrus, calcarine gyrus
Auditory network: middle temporal gyrus
Default‑mode network: superior parietal lobule
Salience / vigilance network: middle frontal gyrus, cerebellum
In short: taVNS "cools down" the over‑activated insula and restores calm to hypersensitive visual‑auditory and vigilance systems.
Figure 3: Between‑group comparison of functional connectivity between taVNS and staVNS groups.
Compared with the control (staVNS) group, after 4‑week taVNS treatment, the treatment group presented significantly decreased resting‑state functional connectivity (rsFC) in left calcarine sulcus (CAL_L), left middle occipital gyrus (MOG_L), left middle temporal gyrus (MTG_L), right superior occipital gyrus (SOG_R), right middle frontal gyrus (MFG_R), left superior parietal gyrus (SPG_L) and left cerebellum (cerebellum_L). Blue areas denote brain regions with decreased rsFC.
taVNS: real transcutaneous auricular vagus nerve‑stimulation treatment; staVNS: sham taVNS; rsFC: resting‑state functional connectivity. R: right; L: left; CAL: calcarine sulcus; MOG: middle occipital gyrus; MTG: middle temporal gyrus; SOG: superior occipital gyrus; MFG: middle frontal gyrus; SPG: superior parietal gyrus.
Table 3: Brain‑region differences in seed‑based FC seeded at the left insula between real taVNS group and StaVNS group among primary‑insomnia patients after 4‑week treatment.
Figure 4: Significant correlations between within‑group changes in left‑insula‑related resting‑state functional connectivity (rsFC) and clinical outcomes.
A: Correlation between changes in left‑cerebellum rsFC (pre‑post‑treatment difference) and improvements in sleep‑disturbance scores. After 4‑week taVNS treatment, decreased rsFC between the left insula (insula_L) and left cerebellum (cerebellum_L) in primary‑insomnia patients showed a significant positive correlation with reductions in sleep‑disturbance scores (r=−0.502, p=0.006, Bonferroni‑corrected).
B: Correlation between changes in right‑superior‑occipital‑gyrus rsFC (pre‑post‑treatment difference) and improvements in sleep‑onset‑latency scores. After 4‑week taVNS treatment, decreased rsFC between the left insula (insula_L) and right superior occipital gyrus (SOG_R) in primary‑insomnia patients showed a significant positive correlation with reductions in sleep‑onset‑latency scores (r=0.391, p=0.040, uncorrected). rsFC: resting‑state functional connectivity; R: right; L: left; SOG: superior occipital gyrus; taVNS: transcutaneous auricular vagus nerve‑stimulation treatment.
Milestone Finding: Baseline Functional Connectivity Predicts Therapeutic Response
The most clinically valuable finding of this study: baseline functional connectivity of the left insula‑visual/auditory circuit can serve as a candidate biomarker for predicting taVNS treatment efficacy.
Prediction correlation coefficient: r=0.534
Statistical validation: p<0.002, stable and reliable results
This means that a single pre‑treatment resting‑state fMRI scan can predict whether taVNS will be effective.
Table 4: Weights of individual brain regions in predicting taVNS therapeutic effects for patients with primary insomnia.
Figure 5: Predictive performance of the method proposed in this study.
A: Line plot showing real versus predicted labels; the model achieved a maximum correlation coefficient of 0.5337 when 25 predictive features were included.
B: Permutation‑test results corresponding to the maximum correlation coefficient (p<0.0002).
Figure 6: Feature‑weight map for predicting taVNS therapeutic response based on the classification model. Red represents positive weights and blue represents negative weights. Major high‑weight brain regions: left middle occipital gyrus, left calcarine sulcus, left middle temporal gyrus, right superior occipital gyrus. R: right; L: left; MOG: middle occipital gyrus; CAL: calcarine sulcus; MTG: middle temporal gyrus; SOG: superior occipital gyrus; taVNS: transcutaneous auricular vagus nerve‑stimulation treatment.
In‑Depth Mechanism Interpretation
1. Core Brain Mechanism of Insomnia
The insula acts as the core of the salience network and functions as the brain’s "master vigilance switch". Abnormally enhanced functional connectivity within the left insula‑visual/auditory network in insomnia patients leads to: brain hypersensitivity toward light, sound and intrusive thoughts, persistently high vigilance, difficulty falling asleep, frequent awakenings and light sleep.
How taVNS Works
3 core pathways:
Calm over‑active visual‑auditory systems
Reduce connectivity between the left insula and occipital‑temporal lobes, dampen amplification of sensory signals, facilitate sleep onset and stabilize nighttime sleep.
Soothe the "restless" default‑mode network of the brain
Reduce insula‑driven activation of the superior parietal lobule, mitigate pre‑sleep rumination and quiet brain activity.
Lower neural vigilance level
Reduce connectivity between the insula and middle‑frontal‑gyrus / cerebellum, relieve sustained hyperarousal and allow the body to enter sleep states.
3. Why this circuit can predict therapeutic response
This study confirmed via support vector regression (SVR) and permutation tests:
Stronger baseline connectivity → more severe insomnia and larger modulatory potential → better taVNS efficacy
More‑normal baseline connectivity → limited modulatory room → relatively limited therapeutic benefit
Therefore, the left insula‑visual/auditory circuit serves both as a pathological marker for insomnia and a predictive marker for taVNS treatment response.
Conclusion
This study lays important scientific foundations for taVNS treatment of insomnia, moving from "effective yet mechanistically unclear" toward "mechanistically understood plus pre‑treatment outcome prediction".
The left insula‑visual/auditory circuit is not only a pathological marker of insomnia but also a predictor for taVNS treatment response. In the future, relying on this biomarker, more insomnia patients can identify optimal personalized treatment options before intervention, escape drug dependence and regain high‑quality sleep.
Disclaimer: This official‑account article only shares cutting‑edge academic findings for non‑commercial purposes. Please contact official‑account backend or send emails promptly for modification or removal if any infringement occurs! E‑mail: 3360739445@qq.com