The altered hypothalamic network functional connectivity in chronic insomnia disorder and regulation effect of acupuncture: a randomized controlled neuroimaging study

Abstract

Background

The hypothalamus has been recognized as a core structure in the sleep-wake cycle. However, whether the neuroplasticity of the hypothalamus is involved in the acupuncture treatment of insomnia remains elusive.

Methods

We recruited 42 patients with chronic insomnia disorder (CID) and 23 matched healthy controls (HCs), with CID patients randomly assigned to receive real acupuncture (RA) or sham acupuncture (SA) for four weeks. Insomnia severity was evaluated using the Pittsburgh Sleep Quality Index (PSQI) score, and the resting-state functional connectivity (rsFC) of the hypothalamus was assessed via functional magnetic resonance imaging (fMRI).

Results

In the cross-sectional investigation, CID patients showed increased rsFC between the medial hypothalamus (MH) and left lateral orbital frontal cortex (LOFC), and bilateral medial orbital frontal cortex (MOFC) compared to HCs. In the longitudinal experiment, PSQI scores significantly decreased in the RA group (p = 0.03) but not in the SA group. Interestingly, the increased MH-LOFC connectivity was found to be reduced following RA treatment. In addition, the altered rsFC of MH-LOFC significantly correlated with clinical improvement in the RA group (r = -0.692, p = 0.006).

Conclusion

This randomized neuroimaging study provides preliminary evidence that acupuncture may improve insomnia symptoms by restoring circuits associated with hypothalamic subregions.

Trial registration

This trial has been registered on the Chinese Clinical Trial Registry (www.chictr.org.cn) with the identifier (ChiCTR1800017092). Registered date: 11/07/2018.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12906-024-04703-y.

Introduction

Chronic insomnia disorder (CID) is one of the most prevalent categories of sleep disorders, distinguished by the persistent challenge of initiating and sustaining sleep despite adequate sleep opportunities [1]. Chronically disruptions in sleep not only reduce work productivity and impair the quality of daily life but also precipitate mood disorders such as depression and anxiety [2, 3]. An increasing body of evidence has shown the pivotal role of the central nervous system in the occurrence and maintenance of insomnia [4–6].

Several lines of neuroimaging research have elucidated insomnia-related functional and anatomical alterations in multiple neural networks, including the default mode network, the dorsal attentional network, and the ascending arousal network [7–9]. Among these altered brain regions, the hypothalamus is considered as the central orchestrator of sleep-wake transitions [10, 11], playing a cardinal role in the central pathogenesis of insomnia. Anatomically, the hypothalamic region directly connects to other vital sleep-wake control areas, encompassing the brainstem nuclei of the ascending reticular activating system, the basal forebrain, and the cortex [12, 13]. Indeed, several animal studies have also supported this intricate interaction, demonstrating that hypothalamic neurons engage with these arousal systems to establish a dynamic “sleep-wake switch,” thereby normalizing sleep [14]. Notably, the hypothalamus is a complex brain region with multiple subregions and a variety of functions. Previous research has shown that the hypothalamus can be roughly divided into the lateral hypothalamus (LH) and the medial hypothalamus (MH), which are involved in the regulation of wake and sleep processes, respectively [15, 16]. The LH, well-documented for its role in wakefulness, secretes orexin, which is directly involved in sleep-wake transitions and has been implicated in the pathology of sleep disorders [17]. The role that MH plays in sleep arousal compared to LH is not well defined. However, two crucial nuclei in the MH, the ventromedial hypothalamus (VMH) and dorsomedial hypothalamus (DMH), have been found to be involved in the regulation of both wakefulness and circadian rhythm in recent studies [18, 19]. This suggests that MH in the hypothalamus is also worth further exploration in the pathogenesis of sleep disorders. Considering this context, improving hypothalamic function to alleviate insomnia symptoms has become an essential strategy for the treatment of CID.

Acupuncture, a popular and safe complementary and alternative therapy, boasts a history spanning thousands of years of treating insomnia [20]. Our prior randomized controlled trial [21] demonstrated that real acupuncture, in comparison to sham acupuncture, significantly reduced the severity of insomnia symptoms, a clinical benefit that persisted for at least six weeks. A recent systematic review [22] also confirmed that acupuncture not only improved subjective sleep quality but also exhibited a significant correlation with improvements across various objective sleep parameters. These collective pieces of evidence suggest that acupuncture presents a promising avenue for addressing insomnia. However, whether acupuncture can alleviate insomnia symptoms by modulating the hypothalamus functional network remains an open question.

Hence, this study aimed to explore the potential role of the hypothalamus functional network in the neuropathology and recovery mechanisms of CID with a two-session study design (baseline and longitudinal treatment). Firstly, we investigated the differences in hypothalamus resting-state functional connectivity (rsFC) between CID patients and healthy controls (HCs) at baseline. Secondly, we explored how insomnia symptoms and the hypothalamus rsFC changed after different longitudinal interventions (i.e., real, or sham acupuncture). Based on the crucial role of the hypothalamus in regulating sleep and wakefulness, we hypothesized that: (1) patients with CID would exhibit altered rsFC between the hypothalamus and other brain regions compared to HCs, and (2) real acupuncture treatment, but not sham acupuncture, would ameliorate insomnia symptoms and modulate abnormal rsFC in CID patients.

Materials and methods

Study design

This study was comprised of two experiments: a cross-sectional and a longitudinal experiment (Fig. 1A). The cross-sectional experiment examined the neural substrate of CID by comparing the hypothalamus-based rsFC maps between CID patients and HCs. The longitudinal experiment investigated the clinical efficacy of acupuncture and its underlying neurological mechanism by comparing the hypothalamus-based rsFC map before and after treatment. Finally, a spatial overlap map was produced using the rsFC findings of both experiments to evaluate the mechanism of acupuncture therapy in relation to baseline abnormalities. The longitudinal acupuncture experiment was approved by the Institutional Review Board of Chengdu University of Traditional Chinese Medicine (2018KL-041) and submitted to the Chinese Clinical Trial Registry with the identifier (ChiCTR1800017092). All participants provided written informed consent before the initiation of any study procedures. The study adhered to the CONSORT guidelines for reporting.

Fig. 1.

Study design and flowchart

(A) Study design and analytic strategy. (B) Flowchart of the CID patients.

Abbreviations: CID, chronic insomnia disorder; fMRI: functional magnetic resonance imaging; HCs, healthy controls; RA: real acupuncture; SA: sham acupuncture;

Participants

From Augest 2018 to Augest 2020, patients with CID were recruited from the hospital of Chengdu University of Traditional Chinese Medicine (CDUTCM), Chengdu second people hospital, and via online advertisements. The diagnosis of each CID patient was determined by a trained psychiatrist using the International Classification of Sleep Disorders-Third Edition (ICSD-3) [23]. Subsequently, to refine the diagnosis and exclude other neuropsychiatric disorders (such as anxiety and depression), the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) was used as part of our assessment protocol by the same psychiatrist. To be eligible for inclusion in the study, patients must meet the following criteria: (1) be of any gender, between 18 and 65 years of age, and right-handed; (2) meet the diagnostic criteria of ICSD-3 and have experienced difficulty with falling asleep, staying asleep, or early awakening for a minimum of three months; (3) meet the diagnosis of spleen and stomach disharmony in Chinese medicine (see Supplementary Materials); (4) have a score of 7 or above on the Pittsburgh Sleep Quality Index (PSQI); and (5) sign an informed consent form to voluntarily participate in the study. Patients were excluded from the study if they had the following conditions: (1) serious primary diseases affecting the liver, kidney, cardiovascular, cerebrovascular, or hematopoietic systems; (2) a history of other neuropsychiatric disorders, such as major depression or generalized anxiety; (3) systemic illnesses, such as pain, fever, cough, or surgery; (4) insomnia resulting from alcohol and/or psychotropic substance abuse and dependence, including the use of sleeping pills; (5) significant headache, migraine, or history of head trauma; (6) recent use of antibiotics, glucocorticoids, immunosuppressants, or herbal medicines within the past month; (7) pregnancy, preparation for pregnancy, or breastfeeding; (8) typical contraindications to magnetic resonance imaging (MRI) scanning, such as claustrophobia; or (9) participation in another clinical trial within the past month.

Healthy controls were recruited who met the following inclusion criteria: (1) reporting good sleep quality; (2) passing neuropsychological tests; and (3) passing a physical examination without any functional or organic disease or a history of head injury. The exclusion criteria for HCs were the same as for patients with CID.

Acupuncture interventions

A researcher who was not involved in the process of patient recruitment and treatment, used the SAS 9.2 software (SAS Institute Inc., Cary, NC, USA) to generate random numbers. All CID patients were randomly assigned to one of two groups according to the table of random numbers: real acupuncture (RA) or sham acupuncture (SA). The grouping information was blind for CID patients, recruiting researchers, and statisticians. Due to the particularity of acupuncture operation, acupuncturists were not blinded in grouping information of CID patients.

All acupuncture treatments were performed by two licensed acupuncturists. According to traditional acupuncture practices and several clinical studies [21, 24–26], both groups underwent a total of 20 sessions spanning a four-week treatment duration, with five consecutive daily sessions per week followed by two days off. The selected acupoints for the RA group were Baihui (DU20), Zhongwan (RN12), and Zusanli (ST36) according to a previous similar study [27]. Among them, DU20 is a commonly used acupoint to improve sleep [28, 29], while RN12 and ST36 are commonly used to regulate gastrointestinal function [30, 31]. More importantly, a recent study has found that ST36 can regulate not only GI symptoms but also sleep [32]. The acupuncturist inserted disposable stainless-steel needles vertically into the selected acupoints at a depth of 0.5 to 1.5 cun (a special unit of measurement in acupuncture; 1 cun = 25 mm) and gently twisted, lifted, and pushed the needles with consistent amplitude, force, and speed to induce the Deqi sensation. Deqi refers to the subjective feelings of soreness, numbness, heaviness, and distension experienced by the patient, as well as the sensations of heaviness, astringency, and tightness felt by the acupuncturist. After the patient reported that they had a feeling of Deqi, the needles were left in place for 30 min. Referring to a previous study [33], the SA group adopted the operation method of shallow needle insertion (depth of 2–3 mm) at non-acupoints. The non-acupoints in the SA group were located near the real acupoints (2 cm lateral to DU20, RN12, and ST36), which were not part of any known meridian or traditional acupoint. After the needle was inserted into the skin, no needle manipulation was performed, and the needle was left for 30 min.

Clinical outcome measurement

The primary outcome measurement employed was the Pittsburgh Sleep Quality Index, a well-recognized 7-item inventory to quantify the severity of insomnia symptoms. Additionally, the anxiety levels of CID patients were measured using the Self-Rating Anxiety Scale (SAS), and the depression levels were measured using the Self-Rating Depression Scale (SDS). Clinical assessments were conducted at two-time points: baseline (week 0) and post-treatment (week 4).

MRI data acquisition

The MRI scans of CID patients were conducted both before treatment and after the completion of either real or sham acupuncture treatments. In contrast, the MRI scans for healthy controls were performed only once. All MRI data were collected using a 3.0T MRI scanner (GE Discovery 750, Milwaukee, WI) at the MRI Research Center of the University of Electronic Science and Technology. Structural images were obtained using a high-resolution T1-weighted brain volume MRI sequence with the following parameters: repetition time (TR)/echo time (TE) of 5.964/1.976 ms, slice thickness of 1 mm, 157 slices, flip angle of 9°, field of view (FOV) of 256 × 256 mm [2], and voxel size of 1 × 1 × 1 mm [3]. Functional images were acquired using an echo-planar imaging (EPI) sequence with the following parameters: TR/TE of 2000/30 ms, flip angle of 90°, acquisition matrix of 64 × 64, matrix size of 3.75 × 3.75, FOV of 240 × 240 mm, slice thickness of 4 mm, 35 slices, voxel size of 3.75 × 3.75 × 3.2 mm [3], and total volume of 255. Participants were instructed to close their eyes, remain still and quiet, avoid any head movements during the scan, remain awake, and try not to think about anything.

Demographic and clinical data analyses

The demographic and clinical outcome analyses were carried out using SPSS 22.0 software. In the cross-sectional study, two-sample t-tests and Chi-square tests were utilized to compare the baseline characteristics of participants between CID patients and HCs. In the longitudinal study, the changes in clinical symptoms (PSQI, SAS, and SDS scores) were compared before and after treatment in each group separately using paired t-tests. Subsequently, a 2-way (grouped by time) analysis of variance (ANOVA) was used to examine the differences in clinical symptoms before and after treatment between the real acupuncture and sham acupuncture groups. A p value < 0.05 was considered statistically significant.

MRI data preprocessing

Preprocessing of MRI images was performed using the DPARSF (http://rfmri.org) and SPM 12 (www.fil.ion.ucl.ac.uk/spm) toolkits in MATLAB 2014b (MathWorks, Natick, MA). The steps of the preprocessing process including: (1) removal of the first 10 volumes to reduce noise interference; (2) slice timing; (3) realignment of the images to correct for any head movements; (4) co-registration of T1 images with functional images; (5) normalization of T1 images to Montreal Neurological Institute (MNI) space, and segmentation into gray matter, white matter, and cerebrospinal fluid; (6) smoothing of the images using an isotropic Gaussian kernel with a full width at half maximum of 6 mm; (7) temporal filtering using a bandpass filter with a range of 0.01–0.08 Hz, and (8) regression of nuisance signals such as global mean, white matter, cerebrospinal fluid signal, and six motion parameters.

Seed-based functional connectivity analyses

Two subregions of the hypothalamus were selected as the regions of interest (ROIs): the bilateral medial hypothalamus (MH) seed (MNI [x, y, z] = ± 4, -2, -12 mm) and the lateral hypothalamus (LH) seed (MNI [x, y, z] = ± 6, -9, -10 mm) with a 2 mm radius sphere, which had been used in previous studies [34] (Figure S1).

In the first-level analysis, the functional connectivity between the two ROIs, bilateral MH and LH, and the rest of the brain was computed. This was achieved by extracting the blood oxygen level-dependent time course from each ROI, separately, and calculating Pearson’s correlation coefficients between the time course in the LH/MH and every voxel in the whole brain. The correlation coefficients were then transformed to z-scores using Fisher’s z-transformation to improve their normality and enable better analysis using the General Linear Model in the second-level analysis.

For the cross-sectional experiment, the hypothalamus rsFC maps between patients with CID and HCs were compared using a two-sample t-test. For the longitudinal experiment, a paired t-test was used to compare the within-group statistical maps, and a 2-way ANOVA was used to compare the between-group statistical maps. A threshold of p < 0.005 at the voxel level and p < 0.05 at the cluster level, corrected for false discovery rate (FDR), was applied to the group analysis, in accordance with our previous studies [35, 36]. In FDR-corrected clusters, the minimum voxel value was set to 20. If the results did not survive correction for FDR, a small volume correction was applied to the regions associated with sleep and insomnia. The disease-related brain regions were defined as ROIs and generated using the Anatomical Automated Labeling Atlas. To correct for multiple comparisons, Monte Carlo simulations were performed using the 3dFWHMx and 3dClustSim tools from the AFNI software (https://afni.nimh.nih.gov) and applied to the ROIs.

A conjunction analysis was conducted to determine if the altered hypothalamic rsFC in patients with CID would be modulated by acupuncture treatment. The conjunction analysis compared the rsFC alteration maps of CID patients and the effect maps of acupuncture in the RA and SA groups, respectively.

Finally, to explore the association between clinical outcomes and rsFC, we also performed correlation analyses between the significantly altered hypothalamus-based rsFCs (‘post’ minus ‘pre’) and the corresponding PSQI changes in the RA group.

Results

Sample characteristics

A total of 60 patients with CID were enrolled in this study, with five patients withdrawing before undergoing the baseline MRI scan. The baseline MRI scan was completed by 45 patients with CID and 25 age- and gender-matched HCs. After preprocessing the fMRI data, three patients with CID and two HCs had to be excluded from further analysis due to excessive head motion during the scan (> 2 mm or 2°). The baseline analysis was finally conducted on 42 patients with CID and 23 HCs. Two patients did not complete all treatment sessions, and three patients did not participate in the post-treatment MRI scan due to schedule conflicts or unwillingness. The longitudinal analysis finally included 37 patients with CID, 18 in the RA group and 19 in the SA group (Fig. 1B).

Baseline characteristics and therapeutic effects

Table 1 shows the baseline characteristics of the CID patients and the HCs. There were no significant differences between the two groups in terms of age, gender, or educational attainment. The CID group displayed a mean PSQI score of 12.69 ± 2.83. Emotion assessments showed significantly higher SAS and SDS scores in the CID group compared to the HCs group (p < 0.001).

Table 1.

Demographic and clinical traits of CID patients and HCs in the cross-sectional experiment

Characteristic	CID patients (n = 42)	HCs (n = 23)	F/T/X2	P value
Age	33.57(11.16)	32.70(8.80)	0.33	0.75
Gender (M/F)	9/33	9/14	2.32	0.13
Education (years)	15.97(2.68)	16.60(2.88)	0.93	0.36
Duration (months)	47.02(40.61)	NA	NA	NA
PSQI_baseline	12.69(2.83)	3.78(2.21)	13.04	< 0.001
SAS_baseline	39.12(8.55)	27.35(6.24)	5.80	< 0.001
SDS_baseline	40.69(9.15)	29.70(6.91)	4.79	< 0.001

Abbreviations: CID, chronic insomnia disorder; F, female; HCs, healthy controls; M, male; PSQI, Pittsburgh Sleep Quality Index; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale

The within-group analyses showed a significant reduction in PSQI and SDS scores following four weeks of RA treatment in comparison to baseline (t = 7.17, p < 0.001, t = 2.27, p = 0.03), whereas no such changes were observed in the SA group. The between-group analyses showed a significant difference in the change in PSQI scores (“post” minus “pre”) between the RA and SA groups (t = 2.69, p = 0.01), as shown in Table 2.

Table 2.

Demographic and clinical traits of CID patients receiving RA or SA treatment in the longitudinal experiment

Characteristic	RA (n = 18)	SA (n = 19)	F/T/X2	P value
Age	36.81(12.06)	30.33(9.33)	1.67	0.10
Gender (M/F)	5/13	4/15	0.23	0.63
Education (years)	15.29(2.88)	16.67(2.33)	1.44	0.16
Duration (months)	52.90(45.93)	41.14(34.63)	0.65	0.52
PSQI_baseline	12.95(2.60)	12.43(3.09)	1.01	0.32
GSRS_baseline	36.22(14.35)	37.05(12.55)	0.19	0.85
SAS_baseline	40.33(9.51)	37.90(7.51)	0.45	0.66
SDS_baseline	45.00(10.5)	28.35(8.00)	1.01	0.32
PSQI_change	4.61(2.73)	2.16(2.81)	2.69	0.01
GSRS_change	6.50(10.45)	1.58(6.52)	1.73	0.09
SAS_change	4.28(7.99)	3.68(8.32)	0.22	0.82
SDS_change	3.56(7.52)	0.89(7.77)	1.06	0.30

Abbreviations: CID, chronic insomnia disorder; F, female; GSRS, Gastrointestinal Symptom Rating Scale; M, male; PSQI, Pittsburgh Sleep Quality Index; RA: real acupuncture; SA: sham acupuncture; SAS, Self-Rating Anxiety Scale; SDS, Self-Rating Depression Scale

Abnormal hypothalamus functional network in chronic insomnia disorder

To explore the baseline differences in hypothalamus subregion rsFC, a a two-sample t-test was employed to compare all CID patients (n = 42) and HCs (n = 23). Patients with CID, in contrast to HCs, showed significantly higher rsFC between the MH and left lateral orbital frontal cortex (LOFC) and left medial orbital frontal cortex (MOFC), while they had a lower rsFC between the MH and right posterior insula (PI) at baseline (Fig. 2A). No significant differences were observed in the LH rsFC network between the two groups at the threshold set.

Fig. 2.

Alterations of the hypothalamus rsFC network in cross-sectional and longitudinal studies

(A) The abnormal hypothalamus rsFC network in patients with CID compared to HCs. (B) The treatment effect of real and sham acupuncture on the hypothalamus rsFC network in patients with CID. Abbreviations: AI, anterior insular; CID, chronic insomnia disorder; DACC, dorsal anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex; FFA, fusiform area; HCs, healthy controls; LOFC, lateral orbital frontal cortex; MH, medial hypothalamus; MOFC, medial orbital frontal cortex; PI, posterior insula; RA: real acupuncture; rsFC, resting-state functional connectivity; SA: sham acupuncture; SFG, superior frontal gyrus; SMA, supplemental motor area.

Acupuncture effects on the hypothalamus functional network

A direct comparison of before and after RA treatment showed a significant decrease in MH rsFC within the left LOFC and left superior frontal gyrus (SFG) and an increased rsFC within the right fusiform area (FFA) in CID patients (Table 3; Fig. 2B). Interestingly, the decreased MH-LOFC rsFC after RA intervention (“post” vs. “pre”) overlapped with the increased MH-LOFC rsFC at baseline (CID patients vs. HCs), indicating that the effect of acupuncture treatment can normalize the altered MH-LOFC rsFC (Fig. 3A).

Table 3.

Hypothalamus rsFC network results

Brain region	BA	Voxel size	MNI coordinates			Peak Z Score
			x	y	z
CID patients vs. HCs
Left LOFC	11	67	-33	60	-9	4.22
Bilateral MOFC	11	70	-6	63	-15	4.17
Right PI	13	22	48	-33	18	-3.2
Post vs. pre in the RA treatment group
Right FFA	37	24	45	-48	-24	4.59
Left SFG	8	32	-12	36	51	-4.86
Left LOFC	11	32	-39	51	-12	-4.22
Post vs. pre in the SA treatment group
Left DLPFC	9	49	-42	30	39	6.11
Left AI	13	70	-24	18	-12	4.93
Right SMA	6	39	9	9	60	5.04
Left PI	13	28	-42	-36	18	-4.53
Different changes between the two treatment groups
Left DACC	33	57	-21	12	36	3.57
Left SFG	8	69	-9	15	60	3.45
Left DLPFC	9	84	-39	30	39	4.45

Abbreviations: AI, anterior insular; BA, Brodmann’s area; CID, chronic insomnia disorder; DACC, dorsal anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex; FFA, fusiform area; HCs, healthy controls; rsFC, resting-state functional connectivity; LOFC, lateral orbital frontal cortex; MOFC, medial orbital frontal cortex; PI, posterior insula; SFG, superior frontal gyrus; SMA, supplemental motor area

Fig. 3.

The clinical association of overlap region (i.e., LOFC) in patients with CID

A. The real acupuncture treatment effect on LOFC is overlapped with the abnormal hypothalamus rsFC network in patients with CID; B. The MH-LOFC rsFC in the CID and HCs groups, and the alteration between pre-and post-treatment in the RA group. C. The altered MH-LOFC rsFC was significantly associated with the alteration of the PSQI score in the RA treatment group. Abbreviations: CID, chronic insomnia disorder; HCs, healthy controls; RA, real acupuncture; LOFC, lateral orbital frontal cortex; MH, medial hypothalamus; PSQI, Pittsburgh Sleep Quality Index; rsFC, resting-state functional connectivity.

The SA group showed greater rsFC between the MH and left dorsolateral prefrontal cortex (DLPFC), left anterior insula (AI), and right supplemental motor area (SMA), as well as lower rsFC between the MH and left PI at the end of treatment compared to pre-treatment (Table 3; Fig. 2B).

The intergroup comparison between the RA and SA groups showed a significant increase in rsFC between the MH and left DLPFC, left dorsal anterior cingulate cortex (DACC), and left SFG (Table 3; Fig. 2B) using a 2-way ANOVA. There was no significant effects of group and treatment on the hypothalamus rsFC network after correction for multiple comparisons. The uncorrected results are presented in Figure S3.

Conjoint analysis of rsFC results and association with PSQI improvement

As Fig. 3A presents, in the left LOFC, the baseline altered rsFC map and the longitudinal altered rsFC map overlapped. We extracted the average z-values of the overlapping area from CID patients at baseline (n = 42), HCs at baseline (n = 23), CID patients pre-RA treatment (n = 18), and CID patients after RA treatment (n = 18). The two-sample t-test analysis showed that the rsFC between the MH and left LOFC was higher in CID patients than in HCs at baseline (t = 3.68, p < 0.001). After RA treatment, the rsFC between the MH and left LOFC in CID patients decreased significantly to levels close to HCs (t = 3.37, p = 0.002; Fig. 3B).

Additionally, a correlation analysis between the change in rsFC values and the improvement in PSQI scores and GSRS scores was performed only in the RA group, as there was no significant improvement in PSQI scores in the SA group. The results showed a negative correlation between the change in the MH-LOFC rsFC values in the overlapping cluster and the improvement of PSQI scores in the RA group (r = -0.692, p = 0.006; Fig. 3C), no significant association between altered MH-LOFC rsFC and the change in GSRS scores in the RA group (r = -0.069, p = 0.784).

Discussion

This study explored alterations in the functional connectivity of hypothalamic subregions (i.e., MH and LH) in patients with CID, as well as the effects of acupuncture modulation on these connections. The results showed that: (1) compared to HCs, CID patients exhibited increased intrinsic rsFC between the MH and the LOFC; (2) RA treatment can reduce the PSQI scores and decrease the rsFC of the MH-LOFC in CID patients; and (3) the rsFC between the MH and the overlapping region (i.e., the LOFC) was significantly correlated with improved clinical symptoms. These findings provide evidence for the involvement of the hypothalamus-OFC circuit in the pathogenesis of CID and partially reveal the neural plasticity mechanisms underlying acupuncture treatment for CID.

As a pivotal region within the central autonomic network, the hypothalamus is involved in both sympathetic and parasympathetic activities and has been implicated in regulating sleep-wake cycles, body temperature, feeding behavior, and memory function [37, 38]. Using cellular, molecular, and macroscopic neuroimaging measurements, the significance of the hypothalamus in sleep-wake regulation has been widely documented in animal studies [39, 40]. Recent advancements in human fMRI research have largely corroborated these findings [10, 41, 42]. For instance, Boes et al. demonstrated human hypothalamus functional connectivity involved in sleep-wake regulation using the resting-state fMRI approach [41]. Jiang et al. further investigated alterations in hypothalamic functional connectivity in brain regions associated with non-rapid eye movement (NREM) sleep, establishing a correspondence between the organization of functional networks and hypothalamic sleep-wake regulation in humans [10]. These findings support the crucial role of the hypothalamic circuit in orchestrating sleep-wake cycles in humans.

Hypothalamus is typically partitioned into two distinct parts: the lateral hypothalamus (LH) and the medial hypothalamus (MH), each carrying out specific functions in sleep-wake regulation. The LH, also known as the “arousal” center, contains orexin and melanin-concentrating hormones [43]. In our study, there were no significant alterations in LH functional connectivity networks in patients with CID. Although, with a more lenient threshold, changes in brain regions such as the middle frontal cortex, inferior temporal gyrus, postcentral lobe, and supplemental motor area could be identified (voxel-level p = 0.05, not correct, Figure S2). The MH accommodates gamma-aminobutyric acid and glutamate interneurons that appear to innervate both sleep- and wake-promoting circuits [44, 45]. Lesions in the MH have been shown to eliminate the sleep-wake circadian rhythm in animal studies [46]. Consistent with these studies, a primary finding of our study was that patients with CID exhibited aberrant MH-OFC connectivity compared to HCs.

Numerous investigations have reported the abnormal volume and metabolism of OFC among insomnia patients [47, 48]. The OFC is comprised of five major subdivisions [49], and the present study found that in CID patients, the abnormal functional connectivity of the OFC is predominantly located in the medial and lateral portions. Previous research has indicated that these portions of the OFC, known as the medial and lateral OFC, are structurally and functionally connected to the hypothalamus, and these connections play a role in emotion and cognitive control [50, 51]. The OFC is mainly involved in reward learning, decision-making, affective regulation, cognitive flexibility, social interactions, and inhibitory control [52, 53]. Studies have demonstrated that activation within the OFC often coincides with efforts to suppress unpleasant or distressing experiences [53]. The strong connections observed between the hypothalamus and OFC in our study suggest that the brain employs a self-regulation strategy to deal with unpleasant experiences related to insomnia, such as recurrent painful hyperarousal states.

Interestingly, the increased MH-LOFC connectivity at baseline was found to be diminish following the RA intervention. The medial and lateral segments of the OFC serve different roles, with the LOFC believed to be involved in suppressing emotional information [54]. Patients with CID are frequently in an emotionally triggered state, and they are vulnerable to negative thoughts, such as anxiety and stress, due to their concern about the recurrence of insomnia [55]. Therefore, our findings might indicate that effective treatment may reverse the negative emotions (e.g., worries related to sleep) associated with a high arousal state, thereby prompting a shift from wakefulness to sleep.

Importantly, further correlation analysis revealed a significant association between the decreased rsFC of MH-LOFC and the improvement of insomnia symptoms, implying that RA may attain therapeutic effects by modulating the hypothalamic-LOFC network. This brain-behavior correlation result reinforces the hypothesis that effective acupuncture treatment can modulate the neural plasticity of the hypothalamus-LOFC circuit to treat insomnia.

The results of our study also revealed that other brain connections, including the MH-SFG and MH-FFA, were significantly modulated by the RA treatment. Previous neuroimaging studies have documented the involvement of the SFG and FFA in the pathophysiology and recovery of insomnia disorder [48, 56, 57]. However, there were no significant links between changes in these connections and improvements in insomnia symptoms. Further longitudinal studies are required to better understand the role of these regions in the treatment of CID with acupuncture.

After sham acupuncture intervention, an increase in the rsFCs between the MH and the DLPFC, AI, and SMA and a decrease in the rsFC between the MH and PI were observed. These regions are known to be involved in cognition control, emotion regulation, attention performance, and sensory processing and have shown abnormalities in their structure and function in patients with insomnia [58, 59]. The altered rsFCs between the MH and these regions potentially signify the modulatory effects and therapeutic benefits of placebo treatment. Additionally, comparisons between the two groups showed that the RA group had higher rsFCs between the MH and the left DLPFC, left DACC, and left SFG, in contrast to the SA group. These distinct hypothalamic networks induced by RA and SA align with previous studies that have shown differences in macroscopic neural activity between real and sham acupuncture [60, 61].

Several limitations merit consideration. Firstly, the relatively small sample size and short-term treatment period (4 weeks) could potentially restrict the generalizability of the results. Hence, larger cohorts and longer treatment durations are imperative in future studies. Secondly, the utilization of higher-resolution fMRI data in the future would enhance the precision of connectivity pattern mapping, especially for small seed regions such as the hypothalamus. Finally, incorporating multi-mode neuroimaging techniques, such as positron emission tomography, diffusion tensor imaging, and magnetoencephalography, may furnish a more comprehensive understanding of the effect of RA on the neural network of the hypothalamus and its related regions and help optimize the treatment of CID.

Conclusions

Compared to healthy individuals, patients with CID exhibit abnormal rsFC between the MH and LOFC. This abnormal connectivity, as well as insomnia symptoms, can be significantly reversed after real acupuncture intervention. These preliminary findings suggest that acupuncture may ameliorate insomnia symptoms by restoring the hypothalamus-related circuit.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AI

Anterior insula

ANOVA

Analysis of variance

CID

Chronic insomnia disorder

DACC

Dorsal anterior cingulate cortex

DLPFC

Dorsolateral prefrontal cortex

EPI

Echo-planar imaging

FDR

False discovery rate

FFA

Fusiform area

fMRI

Functional magnetic resonance imaging

FOV

Field of view

HCs

Healthy controls

ICSD-3

International Classification of Sleep Disorders-Third Edition

LH

Lateral hypothalamus

LOFC

Lateral orbital frontal cortex

MH

Medial hypothalamus

MNI

Montreal Neurological Institute

MOFC

Medial orbital frontal cortex

MRI

Magnetic resonance imaging

PI

Posterior insula

PSQI

Pittsburgh Sleep Quality Index

RA

Real acupuncture

ROIs

Regions of interest

rsFC

Resting-state functional connectivity

SA

Sham acupuncture

SAS

Self-Rating Anxiety Scale

SDS

Self-Rating Depression Scale

SFG

Superior frontal gyrus

SMA

Supplemental motor area

TE

Repetition time

TR

Repetition time

Author contributions

W.P.: Data collection and analysis, as well as manuscript drafting; H.X.: Data collection and analysis; C.Z.: Manuscript revision; Y.P.: Conception and design of the study; S.Y.: Conception and design of the study, provided decisive feedback and manuscript revision. All authors reviewed the manuscript.

Funding

This study was funded by the National Natural Science Foundation of China (82004488 and 82375590), the Sichuan Provincial Science and Technology Department project in China (2021YJ0176, 2018JY0249), the Project of China Postdoctoral Science Foundation (2023M740438), Chongqing postdoctoral research project special support (2023CQBSHTB3142), Health Commission & Science and Technology Bureau of Chongqing Jiangbei District (ZBKW2023zy010) and Bureau of Science & Technology Nanchong City (20SXQT0246).

Data availability

The data in the present study can be available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This longitudinal acupuncture experiment was approved by the Institutional Review Board of Chengdu University of Traditional Chinese Medicine (2018KL-041). All participants in this study has signed informed consent prior to the trial.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.