Distance-related functional connectivity alterations in first-episode, drug-naïve major depressive disorder: a resting-state fMRI study with atlas-based neurotransmitter correlates

Abstract

Background

Major depressive disorder (MDD) has been associated with large-scale functional dysconnectivity, yet it remains unclear whether these abnormalities differ as a function of anatomical distance. We investigated distance-related functional connectivity strength (FCS) alterations in first-episode, drug-naïve MDD and examined whether their spatial distribution corresponds to atlas-based neurotransmitter receptors/transporters profiles.

Methods

Resting-state functional MRI data were acquired from 191 patients with first-episode, drug-naïve MDD and 130 healthy controls. Whole-brain global, long-range, and short-range FCS were quantified voxel-wise. Distance-constrained seed-based connectivity analyses were performed to characterize the circuit-level contributions underlying abnormal FCS findings. To provide a molecular context, regional maps of FCS abnormalities were compared with neurotransmitter receptors/transporters distributions derived from a publicly available positron emission tomography atlas.

Results

Compared with healthy controls, patients with MDD showed reduced long-range FCS in the left orbital superior frontal gyrus and left medial superior frontal gyrus, together with increased short-range FCS in the right inferior parietal lobule. Seed-based analyses indicated that reduced long-range FCS was mainly related to weaker long-range connectivity between the left orbital superior frontal gyrus and right inferior temporal gyrus, and between the left medial superior frontal gyrus and right hippocampus. Increased short-range FCS in the right inferior parietal lobule was associated with stronger short-range connectivity with the right inferior frontal operculum, right superior temporal gyrus, and left paracentral lobule. The spatial pattern of FCS abnormalities showed significant correspondence with atlas-based neurotransmitter receptors/transporters distributions; short-range abnormalities spatially covaried with an atlas-derived regional excitation-inhibition balance.

Conclusions

First-episode, drug-naïve MDD is characterized by a distance-related pattern of reduced long-range prefrontal connectivity and increased short-range parietal connectivity. Atlas-based molecular mapping provides a biologically informed, population-level spatial context for these macroscale functional abnormalities.

Introduction

Major depressive disorder (MDD) is among the most prevalent psychiatric disorders globally, characterized by persistent low mood, anhedonia, diminished self-worth, reduced motivation, and prominent somatic symptoms [1, 2]. In severe cases, MDD can involve suicidal ideation and self-harming behavior [3]. The disorder often follows a chronic or recurrent course, imposes substantial functional impairment, and generates considerable socioeconomic burden [4]. Despite extensive research, the pathophysiological mechanisms of MDD remain incompletely understood, reflecting its clinical complexity and phenotypic heterogeneity [5].

Resting-state functional connectivity (FC) has emerged as a pivotal modality for probing large-scale brain networks and the balance between integration and segregation of neural systems [6]. Prior rs-fMRI studies have shown that patients with MDD exhibit increased functional connectivity between the posterior cingulate cortex (PCC) and both the left middle temporal gyrus (MTG) and right middle frontal gyrus (MFG), together with reduced connectivity between the PCC and the left superior frontal gyrus (SFG) and left precuneus [7]. Meta-analytic evidence further indicates that MDD is associated with reduced within-network connectivity of the frontoparietal network (FPN) and weakened coupling between the FPN and parietal regions of the dorsal attention network (DAN). In contrast, within-network coherence in the default network (DN) appears increased, along with abnormally strengthened coupling between FPN control systems and DN regions [8]. Most prior work, however, has been limited in its capacity to capture the full complexity of brain functional organization. The human brain is a spatially embedded, topologically complex network that depends on coordinated interactions across both short- and long-range connections to maintain efficient function [9, 10]. Conventional functional connectivity analyses typically focus on predefined regions or networks, constraining their ability to characterize global patterns of network reorganization [11]. To overcome these limitations, a data-driven approach termed functional connectivity strength (FCS) has been developed. In FCS analysis, each voxel is treated as a seed, and its cumulative connectivity with all other voxels is quantified to generate a whole-brain map of connectivity hubness [12, 13]. This approach removes the need for a priori region or network selection and enables unbiased detection of connectivity abnormalities across the entire brain [14]. Using a related framework, Zou et al. showed that anatomical distance modulates functional connectivity in first-episode, drug-naïve MDD: patients exhibited reduced short-range functional connectivity density in the left superior temporal gyrus (STG), right orbitofrontal cortex (OFC), and bilateral precuneus, as well as decreased long-range connectivity density in occipital regions, including the middle occipital gyrus (MOG), superior occipital gyrus (SOG), and right calcarine fissure [15]. However, previous FCS studies in MDD have not systematically examined how anatomical distance shapes functional reorganization, and it remains unclear whether FCS alterations in MDD are distance dependent.

Converging evidence suggests that functional connectivity measured by neuroimaging reflects underlying patterns of neuronal activity [16, 17]. Because neuronal communication is largely mediated by neurotransmitters, functional abnormalities in major psychiatric disorders may arise in part from disrupted neurotransmitter signaling, as demonstrated in schizophrenia, obsessive–compulsive disorder, and Parkinson’s disease [18–20]. In MDD, positron emission tomography (PET) studies have shown deficits in both glutamatergic and GABAergic signaling and have implicated serotonin, dopamine D2 receptors, and glutamate in disease pathophysiology [21, 22]. Given that neuromodulatory systems critically regulate the balance between excitation and inhibition within neural circuits [23, 24], and that the blood-oxygen-level–dependent (BOLD) signal predominantly reflects this excitation–inhibition (E/I) balance at the microcircuit level [25, 26], examining the spatial correspondence between FCS alterations and neurotransmitter receptors/transporters distributions may provide a biologically informed context for understanding interregional variability in brain function [27].

The present study aimed to characterize alterations in global and local spontaneous neural activity in first-episode, drug-naïve MDD using rs-fMRI combined with distance-related FCS analysis. We quantified global, short-range, and long-range FCS across the whole brain and examined how these measures relate to clinical features, including symptom severity, illness duration, and age of onset. Building on emerging work that links functional connectivity strength to neurotransmitter distributions, we further tested associations between FCS abnormalities, clinical symptomatology, and neurotransmitter receptors/transporters density profiles, with the goal of providing a molecularly informed context of distance-related functional reorganization in MDD [28, 29].

Materials and methods

Participants

We enrolled 191 treatment-naïve patients experiencing their first major depressive episode from the psychiatric outpatient clinic of the First Affiliated Hospital of Zhengzhou University. Diagnoses were independently confirmed by a chief physician and a trained psychiatrist according to DSM-5 criteria. Symptom severity was assessed using the 17-item Hamilton Depression scale (HAMD-17) [30]. Right-handed, Han Chinese healthy controls (HCs) were recruited from the local community, with no personal or first-degree family history of major psychiatric or neurological disorders. Exclusion criteria for all participants included: comorbid psychiatric or neurological disorders; substance use disorders; prior brain trauma or neurosurgery; major medical illness; and contraindications to MRI. The study protocol was approved by the Medical Ethics Committee of the First Affiliated Hospital of Zhengzhou University and conducted in accordance with institutional and international ethical guidelines. Written informed consent was obtained from all participants.

Scan acquisition

MRI data were acquired on a 3.0 T scanner equipped with an 8-channel phased-array head coil. Foam padding and earplugs were used to minimize head motion and scanner noise. During rs-fMRI, participants were instructed to keep their eyes closed, remain relaxed, and stay awake. Functional images were collected using a gradient-recalled echo-planar imaging (GRE-EPI) sequence with the following parameters: repetition time (TR) = 2000 ms; echo time (TE) = 40 ms; matrix = 64 × 64; flip angle = 90°; field of view (FOV) = 220 × 220 mm²; slice thickness = 4 mm; slice gap = 0.5 mm; voxel size = 3.4 × 3.4 × 4 mm³. Each whole-brain volume consisted of 32 axial slices, and 180 volumes were acquired per run.

Imaging data processing

Functional data were preprocessed using the Data Processing & Analysis for Brain Imaging (DPABI, http://rfmri.org/dpabi) toolbox [31, 32]. The preprocessing steps included: (1) conversion from DICOM to NIfTI format; (2) removal of the first 10 volumes to allow for magnetic field stabilization; (3) slice-timing correction; (4) head motion correction, with exclusion of participants showing > 3 mm or rotation > 3°, and calculation of mean framewise displacement (FD) to quantify motion between consecutive volumes [33–36]; (5) spatial normalization to the EPI template and resampling to 3 × 3 × 3 mm³; (6) spatial smoothing with a 6 mm full-width at half-maximum (FWHM) Gaussian kernel and linear trend removal; (7) temporal band-pass filtering (0.01–0.08 Hz); and (8) regression of nuisance covariates, including Friston 24 head motion parameters, white matter, and cerebrospinal fluid signals. Global signal regression was not performed, given evidence that global signal variance and topography may be altered in psychiatric populations [37–39]. Outlier detection and despiking were implemented using cubic spline interpolation based on the median absolute deviation, with outlier time points replaced by interpolated values [40].

Distance-related functional connectivity strength analysis

Whole-brain resting-state FC was assessed after preprocessing. Pearson correlation coefficients were computed between the time series of all gray matter voxels within a gray matter mask, generating individual subject connectivity matrices that were subsequently transformed to z-scores using Fisher’s r-to-z transformation. Voxel-wise FCS was defined as the sum of z-scored connections between each voxel and all other voxels. Of note, our analysis was constrained to correlations surpassing a threshold of r = 0.2, aiming to exclude weak correlations potentially stemming from noise [41]. Connections were classified as short- or long-range based on a Euclidean distance threshold of 75 mm between Montreal Neurological Institute (MNI) coordinates, yielding global, long-range, and short-range FCS maps (gFCS, lrFCS, and sFCS, respectively) [42, 43]. Group differences in gFCS, lrFCS, and sFCS were examined using two-sample t-tests in SPM12 (http://www.fil.ion.ucl.ac.uk/spm), with age, sex, years of education, and mean FD included as covariates. Multiple comparisons were corrected at the cluster level using Gaussian random field (GRF) theory (voxel-level p < 0.005, cluster-level p < 0.05) [44].

Voxel-wise distance-related functional connectivity analysis

To further characterize the regional contributions underlying aberrant short- and long-range FCS, we conducted seed-based FC analyses constrained by distance. Peak coordinates showing significant between-group FCS differences were selected as seeds, each defined as a 6-mm-radius sphere centered on the peak voxel [45]. For each participant, seed-based connectivity maps were generated by calculating Pearson correlation coefficients between the mean time series of each seed and all other brain voxels, followed by Fisher’s r-to-z transformation. Group-level comparisons were conducted using voxel-wise two-sample t-tests. Statistical inference was performed at the cluster level using Gaussian Random Field (GRF) theory, with a voxel-level threshold of p < 0.005 and a cluster-level threshold of p < 0.05. Connectivity analyses were restricted to short- or long-range connections for seeds associated with sFCS and lrFCS alterations, respectively.

Association between FCS abnormalities and clinical features

To relate FCS alterations to clinical measures, spherical regions of interest (ROIs; radius = 6 mm) were centered on peak t-value coordinates within clusters showing significant FCS abnormalities. For each participant, mean gFCS, lrFCS, and sFCS values were extracted from these ROIs wherever group differences were significant. After confirming normality, Pearson correlation analyses were performed between FCS indices and clinical variables, including total HAMD scores, age of onset, and illness duration. Statistical significance was set at p < 0.05 (two-tailed).

Neurotransmitter receptors/transporters associations with FCS abnormalities

To investigate whether FCS alterations in MDD are related to the expression of specific neurotransmitter receptors/transporters, we examined the spatial correspondence between FCS abnormalities and neurotransmitter distributions. Neurotransmitter data were derived from the PET-based atlas compiled by Hansen et al. [27], which provides density maps for serotonin (5-HT1A [46], 5-HT1B [46–52], 5-HT2A [53], 5-HT4 [53], 5-HT6 [54, 55], 5-HTT [53]), acetylcholine (α4β2 [47, 56], M1 [57], VAChT [58, 59]), cannabinoid (CB1 [60–63]), dopamine (D1 [64], D2 [65–68], DAT [69]), GABA (GABAA [70]), histamine (H3 [71]), glutamate (mGluR5 [72, 73], NMDA [74, 75]), opioid (MOR [76]), and norepinephrine (NET [77–80]) systems. PET images from each contributing study were averaged across participants, normalized to the MNI-ICBM 152 nonlinear 2009 template, and parcellated into 268 brain regions. Mean receptor/transporter density values were z-scored across regions [27, 81].

To quantify regional lrFCS and sFCS abnormalities, the mean unthresholded t-statistic was computed for each brain parcel. We then fitted multivariate linear regression models relating these regional FCS abnormalities to the spatial profiles of neurotransmitter receptors/transporters densities. Model significance was assessed using label-wise non-parametric permutation testing with 10,000 iterations, in which regional labels were randomly shuffled to generate empirical null distributions, and multiple comparisons were controlled using Bonferroni correction [82]. Dominance analysis was performed to quantify the relative contribution of each receptor/transporter distribution to overall model fit.

For each region, the excitation–inhibition (E/I) balance was calculated as the ratio of z-scored densities of excitatory to inhibitory receptors [81, 83]. Excitatory receptors included 5-HT2A, 5-HT4, 5-HT6, D1, mGluR5, α4β2, and M1, whereas inhibitory receptors included 5-HT1A, 5-HT1B, CB1, D2, GABAA, H3, and MOR. We then assessed the spatial correspondence between the E/I ratio and maps of lrFCS and sFCS abnormalities separately. Statistical significance of these associations was evaluated using label-wise non-parametric permutation testing (10,000 iterations).

Results

Demographic characteristics

No significant group differences were observed in sex distribution between MDD patients and HCs (p > 0.05). However, individuals in the MDD group were significantly younger (p < 0.001) and had fewer years of education (p < 0.001) than HCs. Detailed demographic and clinical characteristics are summarized in Table 1. Given group differences in age and education, these variables (along with sex and mean FD) were included as covariates in all subsequent imaging analyses.

Table 1.

Demographics and clinical details of the participants

Variables (Mean ± SD)	HC	MDD	p value
N (male/female)	130 (59/71)	191 (96/95)	0.737a
Age (years)	21.05 ± 5.33	18.10 ± 4.50 (14–37)	< 0.001b
Education (years)	13.56 ± 4.49	10.08 ± 2.11	< 0.001b
Age of onset (years)	NA	16.80 ± 4.54 (14–35)	NA
Illness duration (months)	NA	16.10 ± 19.27	NA
HAMD	NA	25.31 ± 9.42	NA

Unless otherwise indicated values shown are mean ± SD

Abbreviations: MDD, major depressive disorder; HC, healthy control; HAMD, the Hamilton Depression Scale

a The P value was obtained by chi-square test

b The P values were obtained by Mann–Whitney U test

Functional connectivity strength analysis

Relative to HCs, patients with MDD showed significantly reduced lrFCS in the left orbital part of the superior frontal gyrus (SFGorb) and the left medial superior frontal gyrus (mSFG) (Fig. 1b; Table 2), and significantly increased sFCS in the right inferior parietal lobule (IPL) (Fig. 1c; Table 2). At the global level, gFCS was significantly decreased in the left medial orbitofrontal cortex (mOFC) and left mSFG, and, conversely, significantly increased in the right IPL (Fig. 1a; Table 2). All effects survived GRF correction for multiple comparisons (voxel-level p < 0.005, cluster-level p < 0.05).

Fig. 1.

Alterations in distance-related FCS and its clinical correlation in MDD. (a-c) Between-group differences in gFCS, lrFCS and sFCS. (d) A weak negative correlation between lrFCS in the left SFGorb and the age of onset in MDD (R² = 0.02, p = 0.039). Abbreviations: MDD, major depressive disorder; gFCS, global functional connectivity strength; lrFCS, long-range functional connectivity strength; sFCS, short-range functional connectivity strength; SFGorb, orbital superior frontal gyrus

Table 2.

Between-group differences in gFCS, lrFCS and sFCS

	Cluster size	Peak MNI			T	Regions	Hemisphere	P_cluster
		X	Y	Z
gFCS	314	-12	36	-24	-4.3647	mOFC	L	0.001
	267	0	66	18	-4.0958	mSFG	L	0.001
	193	51	-42	48	4.2976	IPL	R	0.004
lrFCS	360	-12	36	-24	-4.4328	SFGorb	L	< 0.001
	393	-12	45	3	-4.2879	mSFG	L	< 0.001
sFCS	290	54	-42	48	4.5855	IPL	R	0.002

Abbreviations: gFCS, global functional connectivity strength; lrFCS, long-range functional connectivity strength; sFCS, short-range functional connectivity strength; L, left; R, right; mOFC, medial orbitofrontal cortex; mSFG, medial superior frontal gyrus; IPL, inferior parietal lobule; SFGorb, orbital superior frontal gyrus

Voxel-wise and distance-related functional connectivity analysis

Distance-constrained, seed-based FC analyses were conducted to delineate the networks underlying the observed FCS abnormalities. Because the two clusters showing gFCS differences were primarily driven by lrFCS and sFCS alterations, these analyses focused on the corresponding lrFCS and sFCS changes. Seed regions were defined as 6-mm-radius spheres centered at the peak coordinates of the significant lrFCS/sFCS clusters: left SFGorb (MNI: x = − 12, y = 36, z = − 24), left mSFG (MNI: x = − 12, y = 45, z = 3), and right IPL (MNI: x = 54, y = − 42, z = 48). For each seed, whole-brain FC maps were generated and compared between MDD patients and HCs.

Compared with HCs, MDD patients exhibited significantly reduced long-range functional connectivity (lrFC) between the left SFGorb (x = − 12, y = 36, z = − 24) and the right inferior temporal gyrus (ITG), and significantly reduced lrFC between the left mSFG (x = − 12, y = 45, z = 3) and the right hippocampus (Fig. 2b; Table 3). For the seed region showing sFCS abnormalities (right IPL; x = 54, y = − 42, z = 48), two-sample t-tests revealed significantly increased short-range functional connectivity (sFC) in MDD patients relative to HCs with the right inferior frontal operculum (IFO), right STG, and left paracentral lobule (Fig. 2a; Table 3).

Fig. 2.

Distance-related FC abnormalities in MDD compared with HCs. (a) Distance-related FC abnormalities seeded at peaks of clusters showing short-range FCS abnormalities in MDD. (b) Distance-related FC abnormalities seeded at peaks of clusters showing long-range FCS abnormalities in MDD.Abbreviations: MDD, major depressive disorder; HC, healthy control; FC, functional connectivity; FCS, functional connectivity strength

Table 3.

Between-group differences in lrFCS-seeds and sFCS-seeds

	Seeds	Cluster size	Peak MNI			T	Regions	Hemisphere	P_cluster
			X	Y	Z
lrFCS-seeds	-12, 36, -24	11,519	54	-9	-36	-5.7265	ITG	R	< 0.001
	-12, 45, 3	11,944	18	-24	-9	-5.3199	Hippocampus	R	< 0.001
sFCS-seeds	54, -42, 48	445	51	12	3	3.9049	IFO	R	0.004
		282	66	-39	21	3.1071	STG	R	0.016
		163	0	-36	63	2.9587	Paracentral Lobule	L	0.031

Abbreviations: lrFCS, long-range functional connectivity strength; sFCS, short-range functional connectivity strength; L, left; R, right; ITG, inferior temporal gyrus; IFO, inferior frontal operculum; STG, superior temporal gyrus

Association between FCS alterations and clinical features

We next examined relationships between FCS alterations (lrFCS/sFCS) and clinical variables (HAMD scores, age of onset, and illness duration) within the MDD group. No significant correlations were observed between FCS measures and either HAMD scores or illness duration. In contrast, lrFCS in the left SFGorb showed a weak negative correlation with age of onset (R² = 0.02, p = 0.039) (Fig. 1d). Given the small effect size and exploratory nature of this analysis, this finding should be interpreted with caution.

Associations between neurotransmitter receptors/transporters profiles and FCS abnormalities

We further examined the relationship between neurotransmitter receptors/transporters distributions and FCS abnormalities in MDD. To this end, we constructed multivariate linear models linking the spatial profiles of receptor/transporter densities to regional patterns of altered FCS. These models showed significant fit, with adjusted R² values of 0.3093 for sFCS (F(268, 248) = 7.29, p < 0.01) and 0.336 for lrFCS (F(268, 248) = 8.12, p < 0.01). Dominance analysis indicated that distinct neurotransmitter receptors made different contributions to each FCS pattern: the 5-HT1A receptor emerged as the primary contributor to sFCS abnormalities, whereas the D1 receptor was most influential for lrFCS (Fig. 3a).

Fig. 3.

Spatial associations between distance-related FCS abnormalities, neurotransmitter receptors/transporters distributions, and E/I ratio. (a) Dominance analysis for determining the relative importance of predictors in each model. (b) A positive spatial association between sFCS and the E/I balance in MDD (R² = 0.04, p = 0.0004).Abbreviations: sFCS, short-range functional connectivity strength; lrFCS, long-range functional connectivity strength; E/I, excitation–inhibition; NT, neurotransmitter; MDD, major depressive disorder

Differential patterns of FCS abnormalities showed significant spatial associations with E/I distribution. sFCS abnormalities showed a significant positive correlation with the whole-brain E/I ratio (r = 0.1981, p = 4 × 10⁻⁴) (Fig. 3b), whereas lrFCS abnormalities were not significantly associated with the E/I ratio (r = − 0.016, p = 0.394). A direct comparison using Steiger’s Z-test confirmed that the association between sFCS and the E/I ratio was significantly stronger than that between lrFCS and the E/I ratio (Z = − 2.4944, p = 0.0126).

Discussion

Using a distance-related FCS framework, we identified distinct short- and long-range FC abnormalities in first-episode, drug-naïve MDD. Relative to HCs, patients showed reduced lrFCS in the left SFGorb and left mSFG, accompanied by increased sFCS in the right IPL. Cross-distance FC analyses further revealed that reduced lrFCS was driven primarily by weakened long-range connectivity between the left SFGorb and the right ITG, and between the left mSFG and the right hippocampus. In contrast, increased sFCS was driven by enhanced short-range connectivity linking the right IPL with the right IFO, right STG, and left paracentral lobule. Importantly, both lrFCS and sFCS abnormalities showed spatial associations with atlas-based neurotransmitter receptors/transporters distributions, particularly with the regional E/I ratio, providing a molecular context for distance-related connectivity changes in MDD.

Patients with MDD exhibit widespread FCS abnormalities across multiple brain regions. However, previous studies have yielded inconsistent findings. For instance, Shi et al. reported decreased gFCS in the bilateral precuneus and frontal lobes (including the left SFG and right IFG) and increased FCS in the right precentral gyrus, left thalamus, and left temporal lobe (mainly the left ITG) in medication-naïve MDD [84]. Another study described decreased lrFCS in the right IPL and reduced sFCS in the right insula and right STG, although these results were obtained using relatively liberal statistical thresholds [45]. Small sample sizes, heterogeneous clinical profiles, and the failure to explicitly distinguish between short- and long-range FCS may have limited the sensitivity and reproducibility of prior work. In the present study, we leveraged a relatively large sample and applied strict exclusion criteria to reduce potential confounds such as comorbidity and medication exposure [85]. We identified aberrant gFCS in the left mOFC, left mSFG, and right IPL. When decomposed into distance-related components, these global abnormalities became more pronounced: patients showed reduced lrFCS in the left SFGorb and left mSFG, together with increased sFCS in the right IPL. Although the effect size was small (R² = 0.02), the negative association between SFGorb lrFCS and age of onset suggests that even within this first-episode, drug-naïve cohort, variation in age of onset may index etiological heterogeneity. This finding should be interpreted cautiously and validated in age-matched and longitudinal samples.

Distance-related FC analyses further clarified the circuits underlying these abnormalities. Reduced lrFCS in MDD was primarily driven by weakened connectivity between the left SFGorb and the right ITG, and between the left mSFG and the right hippocampus. Attenuation of long-range connectivity among these regions may be linked to reduced gray matter volume or compromised white-matter integrity within prefrontal–limbic pathways, consistent with structural findings in MDD [86–88]. Reduced FC between the left SFGorb and right ITG indicates a functional decoupling between a prefrontal hub for higher-order emotion regulation and a temporal region supporting perceptual integration [89]. This disruption may constitute a common circuit-level substrate for the frequently reported dissociation between emotion and cognition in psychiatric disorders, including major depression and anxiety. Similarly, disrupted FC between the left mSFG and right hippocampus signifies impaired fronto–limbic coupling. Given that the mSFG/mPFC serves as a core hub for self-referential processing in depression, and that hippocampal–medial prefrontal interactions are crucial for episodic memory formation, such disruption may hinder the integration between self-referential processing and episodic memory [90–92]. Beyond a general reduction in long-range connectivity, we also observed increased short-range connectivity within networks subserving sensory integration, language processing, and motor planning. This pattern of long-range attenuation and short-range amplification may reflect compensatory hyperactivity or local disinhibition within lower-level sensorimotor and associative networks in the context of compromised higher-order regulatory systems [93, 94]. Such locally heightened connectivity may be directly related to core symptoms of MDD, including rumination mediated by the IPL–STG–IFO/insula circuit and psychomotor disturbances mediated by the IPL–paracentral lobule circuit [94, 95].

A key contribution of this study is the linkage of macroscale functional abnormalities to atlas-based molecular architecture. Correlations with neurotransmitter distributions revealed that distance-related FCS patterns showed significant spatial associations with the E/I ratio. Steiger’s Z-test showed that the positive spatial correlation between sFCS abnormalities and the E/I ratio was significantly stronger than the weak association observed for lrFCS. This dissociation suggests that network dysfunction in depression may not be spatially aligned with a single, uniform molecular profile across spatial scales [96]. More specifically, sFCS abnormalities showed a robust positive spatial relationship with the E/I distribution (r = 0.1981, permutation p = 0.0004), suggesting that regions with higher E/I ratios may be more likely to overlap spatially with short-range functional connectivity alterations in MDD [97]. Previous studies have suggested that altered excitation–inhibition balance may be relevant to neural synchronization and local circuit function [98, 99]. However, the present findings do not directly measure E/I balance in patients and should therefore be interpreted as atlas-based spatial correspondence rather than evidence of patient-specific E/I imbalance. By contrast, lrFCS abnormalities showed a weaker association with static macroscale E/I receptor profiles and may also reflect additional factors, including white-matter integrity [100], neuromodulatory dynamics (e.g., dopamine, serotonin) [101], and processes related to synaptic plasticity [102].

Dominance analysis suggested that lrFC and sFC abnormalities showed differential spatial associations with neurotransmitter receptor distributions, with dopamine D1 receptors contributing most to the lrFC model and 5-HT1A receptors contributing most to the sFC model. These findings provide a molecular context for understanding distance-related FC abnormalities and suggest that distinct neurotransmitter systems may show differential spatial correspondence with short- and long-range connectivity patterns. The 5-HT1A receptor acts as an inhibitory autoreceptor on raphe 5-HT neurons and as a heteroreceptor in limbic and cortical regions, and is broadly expressed in brain systems involved in emotional behavior, anxiety, and mood regulation [103, 104]. In the present atlas-based analysis, the spatial distribution of 5-HT1A receptors showed the strongest contribution to the regression model of sFC abnormalities, suggesting that regions with higher 5-HT1A receptor density may spatially correspond to the pattern of short-range functional connectivity alterations observed in MDD. Conversely, dopamine D1 receptors are prominently expressed in the prefrontal cortex (PFC), where they have been implicated in excitatory signaling, working memory, and large-scale functional coordination [105–107]. Our findings showed that the spatial distribution of D1 receptors contributed most prominently to the model of lrFC abnormalities. This result suggests that the spatial pattern of reduced long-range connectivity, including prefrontal–limbic connectivity alterations, may be aligned with dopaminergic receptor architecture [108, 109]. However, because the receptor maps were derived from independent PET atlas data rather than from the participants in this study, these findings should be interpreted as atlas-based spatial associations and do not provide direct evidence for altered 5-HT1A or D1 receptor function in MDD.

Limitations

This study has several limitations. First, the cross-sectional design precludes characterization of the temporal dynamics of short- and long-range FCS abnormalities and prevents firm conclusions about whether the observed connectivity changes represent consequences of depressive episodes or predispositional markers of MDD. Second, anatomical distance between voxels was quantified as the Euclidean distance between their MNI coordinates, which provides an approximate measure of physical distance but may not fully capture the true anatomical trajectories or white-matter pathways connecting brain regions. Third, neurotransmitter receptors/transporters distributions were derived from group-level PET-based atlases rather than from PET measurements obtained in the participants of the present study. Therefore, the observed associations between FCS abnormalities and neurotransmitter maps should be interpreted as population-level spatial correspondences rather than evidence of patient-specific neurotransmitter alterations or causal neurochemical mechanisms. Finally, age and education were not perfectly matched between groups. Although these variables were included as covariates in all statistical analyses, future studies should strive for closer demographic matching to further validate and extend these findings.

Conclusion

By combining rs-fMRI with distance-related FCS analysis in a large sample of first-episode, drug-naïve MDD patients, we show that anatomical distance systematically shapes resting-state whole-brain connectivity patterns in depression. Reduced long-range connectivity in the left SFGorb and left mSFG, together with increased short-range connectivity in the right IPL, characterizes a pattern of distance-related functional reorganization in MDD. At the molecular level, these functional alterations showed significant spatial correspondence with densities of multiple neurotransmitter receptors/transporters and regional E/I ratios, providing a molecular contextualization of macroscale connectivity abnormalities. Collectively, our results highlight distance-related functional reorganization as an important feature of first-episode, drug-naïve MDD and provide a basis for future longitudinal and multimodal PET-fMRI studies.

Abbreviations

MDD

Major depressive disorder

HC

Healthy control

HAMD-17

17-item Hamilton Depression Rating Scale

rs-fMRI

Resting-state functional magnetic resonance imaging

FC

Functional connectivity

FCS

Functional connectivity strength

gFCS

Global functional connectivity strength

lrFCS

Long-range functional connectivity strength

sFCS

Short-range functional connectivity strength

lrFC

Long-range functional connectivity

sFC

Short-range functional connectivity

PET

Positron emission tomography

E/I

Excitation–inhibition

SFGorb

Orbital superior frontal gyrus

mSFG

Medial superior frontal gyrus

IPL

Inferior parietal lobule

mOFC

Medial orbitofrontal cortex

ITG

Inferior temporal gyrus

IFO

Inferior frontal operculum

STG

Superior temporal gyrus

PFC

Prefrontal cortex

mPFC

Medial prefrontal cortex

ROI

Region of interest

MNI

Montreal Neurological Institute

FD

Framewise displacement

GRF

Gaussian random field

DPABI

Data Processing & Analysis for Brain Imaging

Author contributions

YD contributed to methodology, original draft preparation, and manuscript revision. YC contributed to conceptualization, validation, and methodology. YT contributed to data curation and visualization. ZL contributed to data curation and investigation. HY contributed to data curation and investigation. KF contributed to visualization and investigation. RZ contributed to visualization and investigation. SL contributed to visualization and formal analysis. BW contributed to visualization and formal analysis. YW contributed to visualization and formal analysis. YZ contributed to supervision, resources, methodology, and funding acquisition. SH contributed to supervision, resources, investigation, funding acquisition, and critical revision of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Natural Science Foundation of China (62476252), the Henan Clinical Medical Scientist Program (S20240045), the Medical Science and Technology Research Project of Henan Province (262102311012), the Natural Science Foundation of Henan Province (252300421113 and 242300420414), the Medical Science and Technology Research Project of Henan Province (LHGJ20230217), the Medical Science and Technology Research Project of Henan Province (SBGJ202302068), the Henan Provincial Science and Technology Research Project (262102311097), the Key Research Projects of Higher Education Institutions in Henan Province (24A320069), the Graduate Education Research Project of Zhengzhou University (YJSJY2025173), the Medical Science and Technology Research Joint Venture Project of Henan Province (LHGJ20230192), and the Medical Science and Technology Research Project of Henan Province (SBGJ202101013).

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to participant privacy and institutional restrictions but are available from the corresponding author on reasonable request and subject to approval by the relevant institutional review board.

Declarations

Ethics approval and consent to participate

All participants received a full explanation of the study procedures prior to participation and provided written informed consent. The study protocol was approved by the Medical Ethics Committee of the First Affiliated Hospital of Zhengzhou University and was conducted in accordance with institutional and international ethical guidelines, including the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.