Fifteen adolescents with nsMDDs were recruited from the outpatient clinic of the Department of Psychiatry at the First Affiliated Hospital of Chongqing Medical University, China, from January 2021 to March 2022. This study was approved by the Human Research and Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (no. 2021-546). Written informed consent was obtained from all adolescents and their caregivers.

All the patients were evaluated by two experienced psychiatrists using the Mini International Neuropsychiatric Interview for Children and Adolescents (MINI-kid). The diagnostic criteria for NSSI were based on the Diagnostic and Statistical Manual of Mental Disorders, the fifth edition (DSM-V). These patients did not take any psychotropic drugs before participating in this study. The primary inclusion criteria were as follows: (i) Han nationality, aged 12–17 years, and right-handed, (ii) first-episode depression with no history of antidepressant treatment or psychotherapy, (iii) reported NSSI events five times or more in the past 12 months and at least once NSSI in the last month according to the Ottawa self-injury inventory (OSI), (iv) exhibited a Hamilton Depression Scale (HAMD-17) score > 17, (v) no MRI scanning contraindication such as metal implants, (vi) no history of brain trauma or other serious physical illnesses, (vii) no substance abuse or dependence, and (viii) no history of any other mental health disorders. The severity of depression was assessed using the 17-item Hamilton Depression Scale (HAMD) (all patients’ score > 17). The patients were once assessed at baseline and then reassessed after 8 weeks of antidepressant medication treatment.

In addition, twenty-two healthy controls (HC) matched for age, gender and education level were recruited from the local school. The inclusion criteria were Han nationality, aged 12–17 years, right-handed, HAMD score < 7, and absence of any severe mental or physical illness.

Sertraline was selected as the main antidepressant drug with an initial dose of 50 mg per day, and the maximum dose was 100–200 mg per day depending on the patients’ condition. The medication continued for 8 weeks.

All MR images were acquired with 3.0-T GE Signa HDxt (General Elecric Healthcare, Chicago, Illinois, United States) scanner equipped with a standard 8-channel head coil. Contiguous sagittal T1-weighted images across the entire brain were acquired with a fast gradient echo (FGRE) sequence: time of echo (TE) = 3.1 msc, time of repetition (TR) = 8 msc, flip angle = 12°, field of view (FOV) = 240 mm, voxel size = 0.938 × 0.938 × 1 mm³, no gap. The rs-fMRI data were obtained using an echo-planar image (EPI) pulse sequence with the following parameters: 33 axial slices, TE = 40 msc, TR = 2000 msc, in-plane resolution = 64 × 64 pixels, flip angle = 90°, FOV = 240 mm, voxel size = 3.75 × 3.75 × 4 mm³, no gap. A total of 240 time points was obtained over 8 min. All participants including both adolescent nsMDDs and HC underwent baseline MR scanning. After 8 weeks of treatment, the adolescent nsMDDs was scanned once again.

Two experienced imaging professionals performed quality control of the original data, and no excessive head movements or abnormal signals were observed in any image. Then, imaging processing was performed with SPM12¹ and RESTplus toolkits² within MATLAB 2013b (MathWorks, Natick, MA, United States).

The procedures were as follows: (i) the original structural and functional image data were firstly converted into NII and NIFTI formats, (ii) the first 10 time points were removed to eliminate nonequilibrium effects of magnetization, (iii) slice timing and head motion correction were then performed to exclude the participants whose head movements exceeded 3 mm translation or 3 degree rotation in any direction, (iv) functional images were spatially normalized to the Montreal Neurological Institute (MNI) EPI template and resampled as 3 × 3 × 3 mm³, (v) a standardization smoothing through the Gaussian filter of 6-mm full width at half maximum (FWHM) was performed to reduce the influence of spatial noise and the differences of brain structure among subjects, and (vi) nuisance regression was performed using the six head motion parameters, white matter, and cerebral spinal fluid BOLD signal as covariates. Linear trends were removed.

ALFF was computed based on the fast Fourier transform (FFT) and the time series of each voxel was transformed to the frequency domain without band-pass filtering. The square root was first calculated at each frequency of the power spectrum, and the mean square root was then obtained across frequency band of 0.01–0.08 Hz for each voxel. Finally, the ALFF of each voxel was divided by the global mean to acquire the mALFF.

Differences in the clinical and demographic characteristics between patients and controls were assessed using SPSS (v26.0, IBM, Chicago, NY, United States). The whole brain analysis of mALFF alteration in adolescent nsMDDs compared to HC was conducted through the SPM using the two-sample t-tests. Age, gender and education level was entered into the general linear model as covariates. An initial voxel-wise threshold was set at p < 0.05 with subsequent family-wise error (FWE) corrected value of p < 0.05 at the cluster level. Then, region of interest (ROI) analysis was conducted to compare the mean signal at the altered brain regions at pre and post-treatment.

The other whole brain analysis of mALFF changes in adolescent MDD at post-treatment compared to pre-treatment was conducted through the SPM using the paired t-test. An initial voxel-wise threshold was set at p < 0.05 with subsequent family-wise error (FWE) corrected value of p < 0.05 at the cluster level. Also, ROI analysis was conducted to compare the mean signal changes at the altered brain regions at pre and post-treatment.

Depression severity changes were further assessed through the paired t-test. Also, Pearson correlation was carried out to examine relationship between the mALFF alteration and depression severity changes. The xjView toolbox³ was used for imaging visualization.

Fifteen adolescent nsMDDs (13 female/2 male, 14.60+/−1.35 years old) and twenty-two HC (13 female/9 male, 15.27 +/−1.05 years old) to be included into the analysis. The statistical analysis showed no significant differences of gender, age and education years between the two groups. More than half of adolescent nsMDDs had total HAMD scores higher than 24, corresponding to the severe status. The demographic and clinical characteristics of these individuals are shown in Table 1.

At pre-treatment, significant differences were observed in regional mALFF between adolescent nsMDDs and HC groups. Compared to controls, adolescent nsMDDs showed significantly increased mALFF in the right superior occipital extending to left lingual gyrus (Peak value = 4.60; at [15, −87, 33]; k = 1,565 voxels) at p < 0.05 (corrected) (see Figure 1; Table 2). The hyperactivity at occipital cortex still remained significantly at p < 0.001 (corrected) (see Supplementary Figure S1; Table S1). Meanwhile, the mALFF was found decreased at right medial superior frontal gryus (SFG, Peak value = −3.95; at [9, 51, 6]; k = 607 voxels) extending to left medial orbital SFG at p < 0.062 (corrected) (see Figure 1; Table 2).

Compared to pre-treatment, significantly decreased mALFF was found at left orbital middle frontal gyrus (orbMFG, Peak value = −7.34; at [−18, 36, −15]; k = 1,638 voxels) extending to right triangular inferior frontal gyrus (IFG) in adolescent nsMDDs after treatment at p < 0.05 (corrected). The reduced neuronal activity at the orbMFG remained significantly at p < 0.001 (see Supplementary Figure S1; Supplementary Table S1). Also, the significantly decreased mALFF was found at the left lingual gyrus (Peak value = −4.44; at [−15, −54, 0]; k = 814 voxels) extending to right lingual gyrus at p < 0.05 (corrected) (see Figure 2; Table 3).

The ROI analysis found a trend of decreased mALFF at the hyperactive occipital cortex found in adolescent nsMDDs after sertraline treatment. Whereas, a trend of increased mALFF at the hypoactive medial SFG found in adolescent nsMDDs after treatment. Further, ROI analysis found significantly decreased mALFF at left orbMFG in adolescent nsMDDs compared to pre-treatment after sertraline treatment. Also, the mALFF at lingual gyrus in adolescent nsMDDs showed significantly decreased compared to pre-treatment (see Figure 3).

After sertraline treatment, the HAMD scores were significantly decreased in adolescent nsMDDs (t = 5.86, p < 0.001) (see details in Supplementary Table S2). Further, Pearson correlation analysis showed no significant correlation between the HAMD changes with mALFF alteration in adolescent nsMDDs. But the correlation between the changes of mALFF at occipital cortex and HAMD was achieved nearly significant (r = −0.295, p = 0.287 for superior occipital; r = −0.357, p = 0.192 for lingual gyrus) (see Supplementary Table S3). Finally, the NSSI measured by OSI was 1.47 +/−0.64 before treatment, and after therapy, it showed a decrease (1.27 +/−0.96) but no significant difference was observed (t = 0.716, p = 0.486) (see Supplementary Table S4). Pearson correlation showed no significant correlation of the alteration in changes of mALFF and NSSI after treatment (see Supplementary Table S5).

The occipital cortex is involved in primary visual stimulation and is associated with facial emotional awareness and working memory (22). Previous studies have provided evidences of the decreased concentration of gamma-aminobutyric acid (GABA) at the occipital cortex (23–25). As GABA is an inhibitory neurotransmitter, the deficits might cause the abnormal spontaneous activity found in nsMDDs.

Located at the occipital cortex, lingual gyrus plays an important role in emotional processing, logical reasoning, and visual information integration (26, 27). Previous studies proposed the activation in the lingual gyrus in MDD patients may be a sign of dysfunction in emotion processing and involved in suicidal behavior (28, 29). Furthermore, Yan et al. (30) demonstrated that nsMDDs have increased ALFF in the lingual gyrus and they speculated that increased resting-state activity in the lingual gyrus was associated with NSSI behavior. In this study, we found that nsMDDs exhibited increased mALFF in the right superior occipital gyrus and left lingual gyrus compared with HC, which was consistent with previous studies. Besides, the left lingual gyrus is involved in linguistic processing (31, 32) and an additional study indicated that the changes in speech characteristics could predict depression severity (33). These findings could explain the language expression problems such as reduced speech rate and decreased volume of tone in MDD patients.

Moreover, we found that nsMDDs exhibited decreased mALFF in the medial SFG. The SFG is the core region of the default mode network (DMN), involved in memory-related cognitive and motor control tasks, as well as self-related processing (34). Previous studies have reported abnormalities in DMN of MDD patients, and this abnormality could be related with suicidal ideation and NSSI behavior (15, 35). The decreased levels of glutamatergic (Glu) metabolites in the medial frontal cortex were linked with the pathophysiology of depression (36). Contrary to GABA, Glu is an excitatory neurotransmitter, and the deficits might cause the decreased neuronal activity found in the study.

The left orbMFG is located in the prefrontal lobe, and belongs to the reward network involved with decision making processes. Sun et al. (37) found that the decreased activation in this region was associated with the efficacy of antidepressant treatment, which was similar to our results. NSSI behavior is associated with decision making, the decreased activation found in this region might indicate the reduction of NSSI behaviors or thoughts. It is inferred that the significantly reduced HAMD score after treatment might be associated with the reduced NSSI behaviors or thoughts.

The IFG plays an important role in emotional and cognitive control. As an important component of the IFG, the right triangular part is suggested taking part in language comprehension (38). Also, this region was involved in unconscious information processing (39). Further, it was also associated with the early-onset depression (EOD) and late -onset depression (LOD). Zhang et al. (40) found the EOD patients exhibited increased activation in the right triangular IFG compared with the LOD patients. They indicated that the specific brain alterations of EOD and LOD patients could be associated with different clinical symptoms. EOD patients are associated with more suicidal behaviors and thoughts, slower treatment response and higher prevalence of co-morbid personality disorders (40–42). According to these reports, we speculated that the decreased activation in the right triangular IFG was linked with the reduced suicidal behaviors and the efficacy of the sertraline therapy.

Also, nsMDDs exhibited decreased mALFF at the bilateral lingual gyrus after sertraline therapy, which could indicate that the abnormal activation in the occipital gyrus of the patient was relieved after treatment. It has been demonstrated that SSRI treatments could increase the GABA concentration at the occipital cortex in depressed patients (43, 44). With the increased GABA concentration, it could further lead to the decreased functional neuronal activity in the occipital cortex. Therefore, this GABA concentration improvement could be a potential mechanism underlying the sertraline therapy in nsMDDs to diminish the hyperactivity at the occipital cortex.

Moreover, the ROI analysis found downregulation of hyperactivity at the superior occipital cortex and upregulation of hypoactivity at the medial SFG in adolescent nsMDDs after treatment. It is widely acknowledged that the breakdown in brain serotonin (5-HT) signaling is closely related to depression symptoms, and antidepressants have been developed according to this basis (45, 46). Sertraline is an SSRI, which plays an important role in treating depression through regulating the abnormal 5-HT. Therefore, the changes of neuronal activity might indicate the efficacy of sertraline therapy.

Over all, our findings of neuronal activity alteration pattern in nsMDDs were consistent with the previous studies. Also, there were few studies that used the rs-fMRI technique to examine the efficacy of sertraline treatment in adolescent nsMDDs. But there were still some limitations in this study. First, our sample size was relatively small and some participants could not endure long-term MRI examinations. The future study with large sample size could be carried out to replicate and validate our findings. Second, nsMDDs exhibited changes of neuronal activity as well as HAMD scores after the therapy, but there was no significant correlation between these changes. While a previous study based on structural imaging has indicated that the alterations in the first week of treatment might be associated with long-term treatment efficacy (47). Thus, a longitudinal observation would be conducted in the future study. Lastly, some participants did not fully understand the OSI report, thus, we could not exactly determine the frequency of NSSI. We would optimize the procedure in the future study.