Healthy sleep is crucial to physical and mental health, as well as social well-being. In recent years, public awareness of sleep health has been gradually increasing (1, 2). Factors such as the fast pace of modern life, heightened stress, and widespread use of electronic devices have markedly elevated the risk of insomnia. It is estimated that approximately 30–35% of the global population suffers from insomnia (3). Notably, insomnia prevalence varies significantly by gender (higher in females) and age (rising with increasing age) (4). With its prevalence on the rise, insomnia has become an unignorable public health concern.

The primary harms of insomnia include inducing negative emotions (e.g., significantly increasing the risk of depression) (5, 6) and impairing cognitive function (7). Additionally, chronic sleep deprivation is associated with a higher risk of multiple diseases, such as cardiovascular diseases (e.g., coronary heart disease), cerebrovascular diseases, diabetes, and obesity (8–10). This imposes a substantial disease burden on both individuals and society. Given these impacts, early identification and intervention for insomnia are vital to improving patients’ quality of life, reducing the risk of comorbid diseases, and alleviating the social burden.

Currently, the diagnosis of primary insomnia primarily relies on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) (11), while the Insomnia Severity Index (ISI) is a commonly used self-report scale to assess symptom severity (12). Studies have shown that an ISI score of ≥8 indicates clinically significant insomnia (13). However, diagnoses based solely on patients’ subjective symptoms have inherent limitations. Against the backdrop of rising insomnia prevalence, further exploration of its neuropathological mechanisms is urgently needed to improve diagnostic accuracy and optimize treatment efficacy.

Research indicates that the pathophysiological mechanisms of PI involve abnormalities in sleep–wake rhythm regulation and hyperarousal states (14). Immune and metabolic factors (15) as well as electroencephalographic features are also involved (16). Functional magnetic resonance imaging (fMRI)—particularly resting-state fMRI (rs-fMRI)—has provided new methodological support for investigating neuropsychiatric disorders such as insomnia (17). Functional connectivity (FC) analysis has revealed abnormal connectivity in brain regions related to sleep–wake regulation in PI patients (18). Voxel-wise degree centrality (DC) can effectively identify key hub nodes in brain networks at the voxel level, serving as a “detector of functional hubs” (19) and exhibiting high test–retest reliability (20). Thus, DC can reflect the importance of hub nodes in brain networks and their changing characteristics at the voxel level. To date, however, no stable abnormal hub regions have been identified in PI research.

In brain structural studies, voxel-based morphometry (VBM) has detected abnormal gray matter volume (GMV) in PI patients, but results remain inconsistent (21). In contrast, surface-based morphometry (SBM) has advantages in measuring cortical characteristics, including cortical thickness and cortical folding degree (22). Preliminary SBM studies suggest that PI may be associated with cortical thickening (23), and that circadian rhythm-related rapid eye movement (REM) sleep regulation exerts a protective effect against age-related brain structural changes (24). Nevertheless, research on this topic remains limited.

While previous studies have reported abnormal brain functional networks and structural microalterations in PI patients, most of these studies are single-modality studies. Thus, the present study aimed to integrate rs-fMRI and 3D T1WI data, applying voxel-wise DC and SBM approaches to examine PI patients who met clinical diagnostic criteria (ISI score ≥8). It focused on alterations in functional hubs and cortical structures at the whole-brain level, and explores the relationships between these abnormal alterations and sleep quality, cognitive function, and emotional states. The ultimate goal is to comprehensively reveal the brain functional and structural alterations associated with PI, elucidate their potential neurobiological mechanisms.

Our findings demonstrate that the bilateral precuneus/posterior cingulate gyrus is one of the key hubs in the abnormal functional network of PI patients, exhibiting aberrant activation patterns. As a core node of the posterior Default Mode Network (DMN) (30), this brain region plays a pivotal role in cognitive processes including episodic memory retrieval and self-reflection, with a particular specialization in contextual information processing (31–34). Among the pathological mechanisms underlying PI, the hyperarousal hypothesis is the most widely accepted. This framework posits that difficulties in sleep onset and maintenance are primarily driven by elevated physical or cognitive hyperarousal (35, 36). Cortical hyperarousal is thought to arise from the brain‘s excessive, repetitive focus on negative cognitions—encompassing hyperactive problem-solving processes, recurrent self-reflection, and persistent worry (37, 38). Accordingly, the aberrant activation of the bilateral precuneus/posterior cingulate gyrus in PI patients aligns closely with the hyperarousal hypothesis; this phenomenon may reflect amplified rumination on negative events in PI patients, ultimately precipitating a state of sustained hyperarousal. Previous studies have reported that the glucose metabolic rate in the precuneus/posterior cingulate gyrus of PI patients shows minimal discrepancy between sleep and wakefulness (39). This observation suggests that these brain regions maintain sustained functional activity at the molecular level in insomniacs, even across sleep–wake transitions, thereby providing theoretical evidence for aberrant activation of the posterior DMN in this patient population. Moreover, existing research has demonstrated that excessive DMN activation is associated with emotional dysregulation and enhanced retrieval of negative memories (40). The elevated anxiety and depression scores observed in the PI group in the present study further support this association. Taken together, our results lend further support to the hyperarousal hypothesis of insomnia.

Meanwhile, we identified the right parietal lobe (encompassing the inferior parietal lobule, superior parietal lobule, and supramarginal gyrus) as another critical hub in the abnormal functional network of PI patients. This hub region is primarily part of the Central Executive Network (CEN) and is implicated in a suite of high-order cognitive processes, including working memory, decision-making, and the regulation of goal-directed behaviors (41). Within the parietal lobe, subregions affiliated with the DMN—such as the precuneus investigated in the present study—directly subserve memory retrieval functions. In contrast, parietal regions outside the DMN are more involved in post-retrieval memory processes, particularly memory-based decision-making (31). In this study, we observed a significant reduction in DC values of the right parietal lobe among whole-brain functional network nodes in PI patients. Previous research has demonstrated that insomniac patients exhibit diminished functional activity in brain regions linked to executive working memory tasks (e.g., the bilateral inferior parietal lobule) (42), accompanied by significantly decreased amplitude of low-frequency fluctuations (ALFF) in the inferior parietal lobule (43) and right supramarginal gyrus (44). These findings corroborate our results, suggesting that CEN-associated parietal regions may undergo functional hypoactivity during episodic memory-based decision-making in PI patients. This functional suppression may be linked to the regulation of hyperarousal in insomniac populations; however, direct empirical evidence supporting a putative “protective effect” of this suppression remains elusive.

In the seed-based FC analysis, we observed reduced FC between the bilateral posterior cingulate gyrus/precuneus and specific subregions of the bilateral occipital lobe in PI patients. The occipital lobe is classically implicated in visual processing, spanning from basic sensory encoding of visual stimuli to higher-order abstract visual functions (e.g., dream imagery). Previous studies have also documented that aberrant neural activity within the occipital lobe may correlate with disruptions in sleep architecture and circadian rhythm regulation (45). Notably, all participants in the present study were instructed to keep their eyes closed throughout the MRI scanning session, thereby eliminating confounding effects of exogenous simple visual stimuli. Recent research has reported that patients with objectively short sleep duration exhibit significantly reduced FC between the left inferior occipital gyrus and the bilateral occipital, parietal, and temporal lobes (46). Furthermore, sleep deprivation studies have demonstrated diminished brain activation in visual-associated cortical regions, including the occipital cortex, lingual gyrus, and fusiform gyrus (47). Building on our findings, the posterior DMN is likely involved in the hyperarousal pathway of insomniac patients—consistent with the elevated DC values of the bilateral precuneus/posterior cingulate gyrus reported in Section 3.2. The reduced FC between the posterior DMN and occipital cortex probably reflects impaired functional interaction between these two brain regions. One plausible interpretation is that this FC reduction may serve to modulate excessive activation of complex abstract visual functions in PI patients, which aligns well with the hyperarousal hypothesis. However, this interpretation remains merely speculative at present: in the absence of objective sleep monitoring data (e.g., polysomnography-recorded sleep stages and parameters) or intervention-based validation (e.g., FC alterations following effective insomnia treatment), we cannot definitively establish this as a bona fide “protective mechanism.”

Additionally, we observed a significant reduction in seed-based FC between the right parietal lobe and multiple subregions of the right posterior cerebellum in PI patients. Although the cerebellum is classically known for its core role in modulating motor coordination and balance, it is also implicated in the regulation of sleep–wake states and sleep rhythm homeostasis (48). The cerebellum establishes tight bidirectional anatomical and functional connections with the cerebral cortex via the thalamus, pontomesencephalic junction, and inferior olivary nucleus. Oscillatory activity within this circuit is highly sensitive to fluctuations in sleep–wake states (49–51). Animal studies have further demonstrated that lesions targeting the cerebellar vermis and cerebellar hemispheres in cats prolong the average duration of non-rapid eye movement (NREM) sleep and total rapid eye movement (REM) sleep duration, while concurrently reducing the frequency of sleep–wake cycle transitions (52, 53). Collectively, these lines of evidence suggest that the reduced FC between the right posterior cerebellum and right parietal cortex may be linked to dysregulated sleep rhythms in PI patients.

This study demonstrated structural aberrations in the frontoparietal cortex of PI patients, characterized by cortical atrophy in the left frontoparietal region accompanied by increased cortical GI, and reduced cortical FD in partial subregions of the right parietal cortex. Notably, the frontoparietal cortex is a key component of the CEN (54). Cerebral cortical activity serves a pivotal role in the execution of high-order cognitive tasks (55), with the prefrontal cortex (56) and parietal cortex (57, 58) being particularly critical to these processes. Previous studies have indicated that reduced GMV in the left precuneus and bilateral inferior parietal cortical atrophy may act as potential biomarkers for the progression from amnestic mild cognitive impairment (aMCI) to Alzheimer’s disease (AD) (59). Furthermore, a bidirectional interplay has been proposed to exist between insomnia and cognitive decline (60)—chronic insomnia could accelerate cognitive impairment, while incipient cognitive deficits might also disrupt sleep regulatory mechanisms.

However, in the present study, although the PI group exhibited global cognitive impairment, the exact relationship between this deficit and frontoparietal cortical structural damage remains unestablished. Notably, we further identified that PI patients had elevated GI in the left frontoparietal cortex, and this index was significantly negatively correlated with the PSQI Daytime Functioning subscale score after FDR multiple comparison correction (q < 0.05)—specifically, higher GI values correlated with lower PSQI Daytime Functioning subscale scores, implying alleviated daytime functional impairment. This finding raises the intriguing possibility that increased cortical folding complexity in this region may reflect a form of beneficial neuroplasticity or neural reserve mechanism in PI patients, which could help buffer daytime functional decline induced by sleep disturbances. This perspective deviates from the traditional view that brain structural abnormalities are simply equated with impaired neural function, suggesting that compensatory or adaptive structural reorganization may occur in the brain during the pathological progression of insomnia. Therefore, future studies focusing on subgroup stratification of insomnia patients may help better elucidate the underlying pathological mechanisms linking frontoparietal structural abnormalities to cognitive dysfunction. Moreover, longitudinal studies combining structural imaging and functional assessments are warranted to verify the existence of this adaptive neuroplasticity and clarify its role in the long-term trajectory of daytime functional status in PI patients.

Meanwhile, this study found reduced cortical FD in the occipital lobe (pericalcarine cortex/cuneus/lingual gyrus) and partial limbic lobe subregions (isthmus of the cingulate gyrus/parahippocampal gyrus) of PI patients. Notably, these regions partially overlap with those exhibiting reduced FC between the posterior DMN and occipital lobe, as reported earlier. This convergence may underpin a structural substrate for the observed functional abnormalities, further reinforcing the critical regulatory role of aberrant functional activity in these regions in the hyperarousal pathophysiology of insomnia. Separately, we observed cortical atrophy in the left temporal lobe of PI patients, specifically involving the posterior bank of the superior temporal sulcus and superior temporal gyrus. Previous studies have documented decreased ALFF in the superior temporal gyrus (44) and reduced GMV in the central and lateral temporal lobes of PI patients (61). Conversely, some researchers have proposed that excessive temporal lobe activation may constitute a core factor undermining insomniac patients’ ability to initiate or maintain sleep (62). Although no abnormal functional activity was detected in the temporal lobe of PI patients in our study, cortical thinning in the posterior bank of the left superior temporal sulcus and left superior temporal gyrus may be linked to sleep onset or maintenance difficulties. Further studies are warranted to elucidate this potential association.

This study has several limitations that should be acknowledged. First, the sample size included in this study is relatively modest, which compromises both statistical power and the generalizability of the findings to a broader, more heterogeneous population of insomniac patients—including those with distinct insomnia subtypes (e.g., onset insomnia, maintenance insomnia) and varying disease durations. Second, our study relied solely on subjective sleep assessment tools, which are susceptible to recall bias and subjective reporting discrepancies. Future studies should therefore integrate objective sleep metrics, such as polysomnography (PSG)-derived parameters (e.g., sleep efficiency, arousal index, sleep stage distribution), to conduct a holistic assessment of patients’ sleep physiology. Third, this study is limited by the lack of white matter structural analysis, as we only investigated gray matter cortical parameters. In line with Hidese et al. (63), who reported that poorer sleep quality correlates with reduced white matter integrity in healthy adults. Thus, our findings may not capture the full spectrum of brain structural abnormalities in PI, and future research should integrate gray and white matter assessments to address this gap. Fourth, the cross-sectional study design precludes the establishment of causal relationships between PI, brain structural/functional abnormalities, and cognitive-emotional disturbances. Moving forward, longitudinal cohort studies stratified by insomnia subtypes, disease severity, or cognitive status are warranted to elucidate the underlying biological mechanisms and directional causality of these associations.

PI patients exhibit distinct functional and structural brain abnormalities. Functionally, there is elevated DC in the posterior DMN (bilateral precuneus/posterior cingulate gyrus) and reduced DC in the parietal lobe component of the CEN (right inferior parietal lobule). Structurally, PI patients present with cortical thinning in the left frontoparietal cortex and an increased cortical GI. Among these abnormalities, reduced FC between the right parietal lobe of the CEN and the right posterior cerebellum is associated with sleep–wake disturbances. Meanwhile, diminished FC between the posterior DMN and the occipital lobe might be involved in the regulation of hyperarousal, though its putative protective role remains unconfirmed. Notably, the increased GI in the left frontoparietal cortex of PI patients—which reflects enhanced cortical folding complexity—may act as a form of beneficial neuroplasticity or neural reserve mechanism that buffers the decline in daytime function induced by sleep disorders. Collectively, this study provides valuable multimodal neuroimaging evidence for the neurobiological mechanisms underlying primary insomnia.