Baseline functional connectivity of the basal forebrain-cortical circuit predict taVNS treatment response in primary insomnia: a randomized controlled trial and fMRI study

Abstract

Background: Dysfunctional basal forebrain (BF) connectivity contributes to primary insomnia (PI). This study investigated whether transcutaneous auricular vagus nerve stimulation (taVNS) modulates BF functional connectivity (FC) in patients with PI and whether baseline FC predicts taVNS treatment response.

Methods: Seventy patients with PI were randomized to real or sham taVNS for 4 weeks. Clinical assessments-including Pittsburgh Sleep Quality Index (PSQI], Insomnia Severity Index (ISI] and Zung's Self-Rating Anxiety (SAS], and Depression Scale (SDS)-and resting-state fMRI data were collected at baseline and after treatment. FC of the bilateral BF subregions (Ch_123, Ch_4) was analyzed, and pre-to-post intervention changes in FC and clinical scores were compared between groups. Baseline FC was used to predict treatment response using a support vector regression (SVR) model, validated on an independent dataset.

Results: Sixty-seven patients completed the study (33 real taVNS, 34 sham taVNS). Changes in clinical outcomes showed that real taVNS significantly reduce PSQI, ISI, and SAS scores compared to sham. FC analysis revealed reduced connectivity between bilateral BF and areas involved in visual (superior occipital gyrus, SOG; middle occipital gyrus, MOG; fusiform gyrus, FFG), somatosensory (supplementary motor area, SMA) cortex and medial prefrontal cortex (mPFC) after taVNS treatment. Reduced FC between bilateral BF and left MOG correlated positively with ISI improvement (r = 0.490, p = 0.008, Bonferroni correction). The SVR model effectively predicted treatment response based on BF-visual circuit connectivity (r = 0.520, p = 0.0014, 5000 permutation test) and generalized well to an independent dataset (r = 0.443, p = 0.0354, 5000 permutation test).

Conclusions: Our findings suggest that taVNS may alleviate symptoms of primary insomnia through modulation of basal forebrain connectivity with visual, sensorimotor, and medial prefrontal cortical regions. Preliminary investigations indicate that baseline functional connectivity in the BF-visual circuit could represent a candidate biomarker for taVNS response, potentially informing personalized treatment strategies.

Trial registration: The study was registered with the China Clinical Trial Registry (Clinical Trial No. ChiCTR1900022535).

Keywords: Functional connectivity of basal forebrain; Primary insomnia; Transcutaneous auricular vagus nerve stimulation; Treatment efficacy prediction.

Fig. 1

Illustration of the locations of basal forebrain (BF) subregions. The red color represents bilateral Ch_123; the blue color represents bilateral Ch_4

Fig. 2

Flowchart of the study. PI, primary insomnia; taVNS, transcutaneous auricular vagus nerve stimulation treatment; Bi_Bforebrain, bilateral basal forebrain; FC, functional connectivity; ANOVA, analysis of variance; FFG, fusiform gyrus; MOG, middle occipital gyrus; SOG, superior occipital gyrus; PSQI, Pittsburgh Sleep Quality Index; LOOCV, leave-one-out cross-validation; SVR, support vector regression

Fig. 3

Seed-based functional connection analysis result of the bilateral Ch_123 seed. Functional connectivity result of the bilateral Ch_123 seed showed decreased connectivity with the right superior occipital gyrus (SOG), right middle occipital gyrus (MOG), left fusiform gyrus (FFG), left middle occipital gyrus, right medial prefrontal cortex (mPFC), and left supplementary motor area (SMA) in the real taVNS group after treatment compared with the sham taVNS. L, left; R, right

Fig. 4

Partial correlation analysis results. The decreased FC between bilateral Ch-123 and left MOG is positively correlated with ISI score improvement after taVNS treatment (r = 0.490, p = 0.008, Bonferroni correction)

Fig. 5

A Prediction performance of the proposed approach. A Line chart reflecting actual label and predictive label; when including 28 predictive features, the highest correlation coefficient is 0.520. B The result of the permutation test with the highest correlation coefficient (p = 0.0014)

Fig. 6

Regions with the most predictive features. Red means positive weight, and blue means negative weight. The main weight brain areas are located in FFG, MOG, and SOG. R, right; L, left; MOG, middle occipital gyrus; SOG, superior occipital gyrus; FFG, fusiform gyrus