Laparoscopic sleeve gastrectomy induces simultaneous changes in brain functions and structures that are associated with weight loss

Abstract

Introduction

Laparoscopic sleeve gastrectomy (LSG) is one of the most effective interventions to promote weight-loss and induce reversal of brain functional and structural abnormalities. However, it is unclear whether these brain alterations occur simultaneously as well as the relationship with weight-loss.

Methods

We employed magnetic resonance imaging to investigate brain changes in 35 patients with obesity (OB), 36 normal-weight (NW) individuals, and 22 OB who received LSG and were studied at baseline and 1 month after LSG.

Results

Comparisons between OB and NW and between pre- and post-LSG showed (1) Resting-state activity differences in amplitude of low frequency fluctuation (ALFF) in the orbitofrontal cortex (OFC), caudate, hippocampus (HIPP), amygdala, thalamus, anterior cingulate cortex (ACC), and dorsolateral prefrontal cortex. (2) Structural differences in fractional anisotropy (FA) in the middle cerebellar peduncle (MCP), genu corpus collosum (GCC), cingulate (Cin), and superior longitudinal fasciculus. Correlation analysis showed that (1) Reduced BMI/waist-circumference was positively correlated with decreased ALFF in the OFC and negatively associated with increased FA in the GCC following LSG. (2) Increased ALFF in the ACC was positively correlated with increased FA in the GCC and MCP. ALFF in the HIPP was negatively correlated with FA in the Cin at post-LSG.

Conclusions

These findings provide evidence that LSG induces simultaneous changes in brain function and structure which are associated with weight-loss.

Keywords:

• Obesity
• laparoscopic sleeve gastrectomy
• magnetic resonance imaging
• weight-loss

1. Introduction

Neuroimaging studies have reported obese-related functional brain abnormalities in regions involved with homeostasis (i.e. hypothalamus), reward processing (i.e. caudate; putamen), motivation (i.e. orbitofrontal cortex, OFC), emotional reactivity (i.e. amygdala; hippocampus, HIPP), and inhibitory-control (i.e. dorsolateral prefrontal cortex, DLPFC; anterior cingulate cortex, ACC) [1–3]. Obesity-related structural brain abnormalities have also been identified, including decreased white matter (WM) integrity (i.e. corpus collosum, fornix, midbrain, and brainstem tracts), decreased fractional anisotropy (FA) and gray/white matter densities (i.e. bilateral cingulum, Cin; uncinate fasciculus; corpus callosum, CC), and increased radial diffusivity/mean diffusivity (i.e. CC; inferior fronto-occipital tract) [4,5].

Among multiple therapies for obesity interventions, laparoscopic sleeve gastrectomy (LSG) is currently one of the most effective treatments for combating morbid obesity and for the best long-term results regarding weight loss and clinical benefits [6]. Neuroimaging studies revealed that LSG-induced weight loss and improved eating behaviors were associated with brain functional/structural alterations in regions implicated in homeostasis and hedonic processing [7]. Specifically, bariatric surgeryinduced reduction in hypothalamus activation to food-cues was significantly correlated with reduced weight. Resting-state functional connectivity analysis showed that hypothalamus had stronger postsurgical resting state functional connectivity with regions implicated in reward processing and inhibitory-control [8]; increased dopamine (DA) D2 receptor in meso-limbic and meso-striatal pathways were correlated with reduced weight [9]; the increased functional and structural connectivity of DLPFC-ACC were also associated with long-term weight loss and reduced sensitivity to food cues in patients with obesity (OB) after surgery [10]. Structural brain recovery of gray/white matter densities, white matter integrity, and cortical morphometry have also been reported [11–13].

Those previous studies reported the LSG-induced brain functional and structural changes respectively, however, it is unclear whether functional and structural alterations occur simultaneously as well as the relationship with weight loss. Thus, we employed the resting-state functional magnetic resonance imaging (MRI) combined with diffusion tensor imaging to investigate brain functional and structural changes in 35 OB and 36 age- and gender-matched normal weight (NW) individuals, and 22 OB who received LSG and were scanned twice at pre-LSG (PreLSG) and 1-month post-LSG (PostLSG).

2. Materials and methods

2.1. Participants

Forty-one OB were recruited for LSG at Xijing Hospital affiliated with the Air Force Medical University in Xi’an, China. Patients with psychiatric or neurological diseases, previous intestinal surgery, inflammatory intestinal disease, organ dysfunction or any current medication that could affect the central nervous system were excluded [14]. Individuals who had a waist circumference (WC) greater than the interior diameter of the scanner were also excluded. Given the criteria, six OB were disqualified from the MRI scan (one had a WC greater than the interior diameter of the scanner, three had metal implants, one decided not to have surgery, and one subject’s imaging data were damaged due to a technical problem). Thirty-five remaining OB completed pre-surgical MRI scans and underwent LSG. The identical MRI scans were performed 1-month after surgery. However, 13 OB reported having significant weight loss after LSG via their local clinics, but they were unable to return for follow-up MRI assessment due to long-distance travel. 22 patients remained in the LSG group. The control group consisted of 36 age-, gender-, and education-matched NW subjects. The experimental protocol was approved by the Institutional Review Board of Xijing Hospital and registered in the Chinese Clinical Trial Registry center under number: ChiCTR-OOB-15006346 (http://www.chictr.org.cn). The experiments were conducted in accordance with the Declaration of Helsinki. All participants were informed of the nature of the research and provided written informed consent.

2.2. Experimental design

All participants underwent 12 h fasting overnight. All MRI scans were performed in the morning between 9 and 10 am to ensure consistency across assessment and to minimize circadian influence [15].

2.3. Questionnaires

A designated clinician rated the severity of the subjects’ anxiety using Hamilton Anxiety Rating Scale [16] and depression status using the Hamilton Depression Rating Scale [17]. Subjects were required to complete the Yale Food Addiction Scale (YFAS) evaluation for assessing addictive eating behaviors [18]. All clinical measurements were identically conducted prior to and 1 month after surgery, and the same surgeon performed all surgical procedures.

2.4. MRI acquisition

The experiment was carried out using a 3.0 T Signa Excite HD (GE, Milwaukee, WI, USA) scanner. First, a sagittal three-dimensional T1-weighted fast spoiled gradient recalled echo sequence were acquired with the following parameters: TR = 7.8 ms, TE = 3.0 ms, matrix size = 256 × 256, field of view = 256 × 256 mm², slice thickness = 1 mm and 166 slices). Then, the scan for resting state functional MRI were acquired using a gradient-echo T2*-weighted echo planar imaging sequence with the following parameters: TR = 2000 ms, TE = 30 ms, flip angle = 90 degrees, matrix size = 64 × 64, FOV = 256 × 256 mm², in-plane resolution of 4 mm², slice thickness = 4 mm (with no slice gap) and 32 axial slices. The scan for resting state functional fMRI lasted 360 s. Subjects were instructed to open their eyes and focus on the fixation during the entire scanning procedure [19]. Finally, diffusion-weighted images were acquired using a single-shot echo-planar imaging sequence. The diffusion sensitizing gradients were applied along 60 non-collinear directions (b = 1000 s/mm²) with ten acquisitions without diffusion weighting (b = 0 s/mm²). The imaging parameters were TR = 9400 ms, TE = 84 ms, 75 continuous axial slices and slice thickness = 2 mm, field-of-view = 256 × 256 mm², matrix size = 128 × 128, resulting in 2-mm isotropic voxels.

2.5. Functional MRI data preprocessing and regions of interest (ROIs) definition

Functional imaging data were preprocessed and analyzed using Statistical Parametric Mapping 8 (SPM8, https://www.fil.ion.ucl.ac.uk/spm/). Specifically, the first five time points were removed to minimize nonequilibrium effects in the functional MRI signal, then images underwent slice-timing correction for within-scan time differences between slices, and head movement correction [20]. Next, the echo planar images were co-registered to the corresponding T1 anatomical image and then spatially normalized to the template of the Montreal Neurological Institute (MNI) and resampled to a voxel size of 3 × 3 × 3 mm³. Spatial smoothing was applied using a Gaussian kernel of 6 × 6 × 6 mm³ full width at half-maximum to decrease spatial noise. Finally, demeaning/detrending was also performed, and head motion parameters, white matter signals and cerebrospinal fluid signals were regressed out as nuisance covariates [21]. The resting state functional MRI time points that were severely affected by motion were removed using a “scrubbing method” (framewise displacement value > 0.5 mm and ΔBOLD of spatial standard deviation of successive difference images > 0.5%), and < 5% of time points were scrubbed per participant [20].

The amplitude of low frequency fluctuations (ALFF) analysis was carried out using the REST Toolkit (http://resting-fmri.sourceforge.net) to define ROIs. The preprocessed time series was first converted to the frequency domain with a fast Fourier transform, and the power spectrum was obtained. The square root of the power spectrum was computed at each voxel, and the averaged square root was obtained in the bandwidth of 0.01–0.08 Hz at each voxel. For standardization, the ALFF of each voxel was further divided by the whole-brain mean ALFF values. Later, the voxel-wise t-tests (two-sample t-tests, paired t-tests) were employed to compare the differences in ALFF between OB and NW, and between PreLSG and PostLSG, respectively. We performed a regression analysis on the effect of age and gender, and brain regions showing significant ALFF alterations were selected as ROIs (p < 0.05, false discovery rate (FDR) corrected).

2.6. Structural MRI data preprocessing and structural ROI definition

Structural imaging was preprocessed and analyzed by the FMRIB software library (FSL, http://www.fmrib.ox.ac.uk/fsl). First, brain extraction was performed using the FSL Brain Extractor Tool [22] to delete nonbrain tissue. Head motion and eddy-current correction were performed using the FMRIB’s Diffusion Toolbox. DTIFIT in FMRIB’s Diffusion Toolbox was used to fit a diffusion tensor model at each voxel of the preprocessed eddy current corrected diffusion weighted data and to calculate the fractional anisotropy (FA) maps. Second, we used the FA maps as input to investigate group WM differences in a voxel-wise manner by using Tract-Based Spatial Statistics. All FA images were non-linearly registered to an FMRIB58-FA standard space template and aligned to the Montreal Neurological Institute space. The mean FA image was generated and thinned to create a skeletonized mean FA image, which had threshold at the FA value of 0.2. Then, for each subject, the sorted FA data mapped the mean skeleton, and to assess the FA values’ differences between each group, voxel-wise cross-subject statistics were acquired. A permutation nonparametric test (5000 permutations) was employed to assess group-related differences using threshold-free cluster enhancement. Results were corrected for multiple comparisons using family wise error (FWE) corrections at the cluster level correction approach (p_FWE < 0.05) with a cluster defining threshold of p < 0.05. Finally, according to the ICBM-DTI-81 WM label atlas [23], each cluster was disassembled into subregions defined as ROIs, and the values of the mean FA were extracted for ROIs across all of the subjects [12].

2.7. Statistical analysis

While controlling for age and gender, two-sample t-tests were implemented in IBM SPSS (Statistical Package for Social Sciences, Release 22.0, Chicago: SPSS, IL) to compare the group differences in ALFF/FA of ROIs, and behavioral measurements between OB and NW. Paired t-tests were used to assess time effects on ALFF/FA of ROIs, and behavioral measurements between PreLSG and PostLSG. Partial correlations with age and gender as covariates were performed to assess the association between LSG-induced changes in brain function/structure and BMI/WC. Bonferroni correction was applied for multiple comparisons, and level of significance was set at p < 0.0007 (0.05/7*5*2). In addition, the Fisher r-to-z transformation were applied to assess the significance of the differences between two correlation coefficients at PreLSG and PostLSG (p < 0.05) [24].

3. Results

3.1. Clinical and demographic characteristics

There were no significant differences in age and gender between OB and NW groups. At baseline, OB relative to NW group had significantly higher weight (p < 0.001), body mass index (BMI, p < 0.001), WC (p < 0.001), YFAS (p < 0.001), HAMD (p = 0.002), and HAMA (p < 0.001, Table 1). LSG significantly decreased weight (p = 0.032), BMI (p = 0.022), WC (p = 0.022), and YFAS (p = 0.010, Table 1). There were no significant differences in depression and anxiety between pre- and post-surgery sessions (p = 0.818, p = 0.322, Table 1).

Table 1. Demographic and clinical information of OB and NW.

	OB (N = 35) (Mean ± SE)	NW (N = 36) (Mean ± SE)	OB vs. NW	LSG (N = 22)		PreLSG vs. PostLSG
				PreLSG	PostLSG
			p Value	(Mean ± SE)	(Mean ± SE)	p Value
Age	28.54 ± 1.51	26.31 ± 1.19	0.248	26.63 ± 1.83	26.63 ± 1.83	1.000
Gender	17M/18F	14M/22F	0.411	9M/13F	9M/13F	NAN
Duration of obesity	11.73 ± 1.61	0	NAN	12.75 ± 2.19	NAN	NAN
Weight (kg)	111.06 ± 3.16	59.06 ± 1.62	<0.001	111.46 ± 3.85	99.32 ± 3.90	0.032
BMI (kg/m^2)	38.44 ± 0.97	21.27 ± 0.42	<0.001	38.99 ± 1.28	34.65 ± 1.31	0.022
WC (cm)	119.06 ± 2.28	79.34 ± 1.74	<0.001	119.95 ± 2.76	108.89 ± 3.73	0.022
YFAS	4.88 ± 0.55	1.81 ± 0.24	<0.001	5.36 ± 0.64	3.23 ± 0.47	0.010
HAMD	12.20 ± 1.56	6.78 ± 0.76	0.002	12.41 ± 2.17	11.77 ± 1.68	0.818
HAMA	10.37 ± 1.33	4.97 ± 0.57	<0.001	10.86 ± 1.68	8.73 ± 1.32	0.322

OB: patients with obesity; NW: normal weight; LSG: Laparoscopic sleeve gastrectomy; PreLSG: patients with obesity who had MRI scan before surgery; PostLSG: patients with obesity who received LSG and were scanned again 1 month after surgery; SE: standard error; BMI: body mass index; WC: waist circumference; YFAS: Yale Food Addiction Scale; HAMD: Hamilton Depression Rating Scale; HAMA: Hamilton Anxiety Rating Scale.

3.2. Alterations in ALFF

Two sample t-tests between OB and NW groups showed that OB had increased ALFF in the OFC, DLPFC, HIPP, ventral tegmental area, putamen, and amygdala, and decreased ALFF in the ACC, ventral medial prefrontal cortex, thalamus and caudate (p_FDR < 0.05; Table 2; Figure 1(A)). Paired t-tests in LSG group showed that PostLSG had increased ALFF in the ACC, DLPFC, caudate, and thalamus, and decreased ALFF in the OFC, HIPP and amygdala (p_FDR < 0.05, Table 2; Figure 1(B)).

Figure 1. Functional mapping of brain areas demonstrating significant ALFF alterations between OB and NW groups, and between PreLSG and PostLSG groups during resting state (p_FDR < 0.05). (A) Compared to NW, OB had increased ALFF in the OFC, DLPFC, HIPP, VTA, PUTA, and AMY, and decreased ALFF in the ACC, VMPFC, THA, and CAU. (B) LSG showed that PostLSG had increased ALFF in the ACC, DLPFC, CAU, and THA, and decreased ALFF in the OFC, HIPP, and AMY. ALFF: Amplitude of low frequency fluctuations; OB: patients with obesity; NW: normal weight; LSG: Laparoscopic sleeve gastrectomy; PreLSG: patients with obesity who had MRI scan before surgery; PostLSG: patients with obesity who received LSG and were scanned again 1 month after surgery; FDR: the false discovery rate; OFC: orbitofrontal cortex; DLPFC: dorsal lateral prefrontal cortex; HIPP: hippocampus; VTA: ventral tegmental area; PUTA: putamen; AMY: amygdala; ACC: anterior cingulate cortex; VMPFC: ventral lateral prefrontal cortex; THA: thalamus; CAU: caudate.

Table 2. Brain regions of interest (ROIs) showing significant brain alterations in ALFF between OB and NW via two sample t-tests (p_FDR < 0.05) and between PreLSG and PostLSG via paired t-tests (p_FDR < 0.05).

ROIs	OB vs. NW					PreLSG vs. PostLSG
	Number of voxels	Max t value	Coordinate (MNI)			Number of voxels	Max t value	Coordinate (MNI)
			x	y	z			x	y	z
ACC	189	−3.0	4	28	−6	23	−5.14	−7	34	30
OFC	216	4.42	7	35	−24	51	3.94	−8	40	−19
DLPFC	137	3.34	−14	40	45	62	−4.81	−28	28	46
Caudate	189	−3.40	−8	22	0	135	−4.92	15	21	3
Amygdala	1593	4.31	−23	1	−16	108	4.08	−21	−3	−24
HIPP	217	3.42	−32	−2	−25	378	5.1	−24	−27	−15
Thalamus	2188	−3.67	−2	−20	9	189	−4.85	6	−12	6
Insula	2349	3.83	37	−2	3	81	−4.37	−39	−15	9
VMPFC	810	−3.57	−14	64	21
VTA	21	3.38	−8	−17	15
Putamen	1782	3.80	−29	−14	0

OB: patients with obesity; NW: normal weight; LSG: Laparoscopic sleeve gastrectomy; PreLSG: patients with obesity who had MRI scan before surgery; PostLSG: patients with obesity who received LSG and were scanned again 1 month after surgery; FDR: the false discovery rate; MNI: Montreal Neurological Institute; ACC: anterior cingulate cortex; OFC: orbitofrontal cortex; DLPFC: dorsal lateral prefrontal cortex; HIPP: hippocampus; VMPFC: ventral lateral prefrontal cortex; VTA: ventral tegmental area.

3.3. Alterations in WM integrity

Relative to NW, OB had lower FA in the anterior corona radiata, body/genu-corpus collosum (BCC/GCC), cingulate (Cin), fornix, middle cerebellar peduncle (MCP), superior longitudinal fasciculus, and sagittal stratum (Figure 2(A)). Paired t-tests in LSG showed that PostLSG had higher FA than PreLSG in the GCC, Cin, superior longitudinal fasciculus, and MCP (Figure 2(B)).

Figure 2. Differences in FA between OB and NW groups, and between PreLSG and PostLSG groups (p_FWE < 0.05). (A) Relative to NW, OB had lower FA in the ACR, BCC, Cin, fornix, GCC, MCP, SLF, and SS. (B) LSG showed that PostLSG had higher FA than PreLSG in the GCC, Cin, SLF, and MCP. FA: Fractional anisotropy; OB: patients with obesity; NW: normal weight; LSG: Laparoscopic sleeve gastrectomy; PreLSG: patients with obesity who had MRI scan before surgery; PostLSG: patients with obesity who received LSG and were scanned again 1 month after surgery; FWE: the family wise error; ACR: anterior corona radiata; BCC: body corpus collosum; Cin: cingulate; GCC: genu corpus collosum; MCP: middle cerebellar peduncle; SLF: superior longitudinal fasciculus; SS: sagittal stratum.

3.4. Correlation analysis

In LSG group, there was a significant positive correlation between BMI and ALFF in the OFC at PreLSG (r = 0.76, p = 0.004). Though this correlation was not significant at PostLSG (r = 0.04, p = 0.840), there was significant difference between PreLSG and PostLSG (z = 2.95, p = 0.003). Reduced BMI positively correlated with increased ALFF in the OFC following surgery (r = 0.62, p = 0.002; Figure 3(A)). There were a significant correlation between BMI and ALFF in the HIPP at PreLSG and PostLSG (PreLSG: r = 0.48, p = 0.020; PostLSG: r = 0.61, p = 0.003), but there was no significant difference (z = -0.57, p = 0.569, Figure 3(A)). In addition, there was a significant positive correlation between BMI and FA in the GCC at PreLSG (r = 0.71, p < 0.001), and there was a significant difference in correlation between PreLSG and PostLSG (z = 3.26, p = 0.001). Reduced WC was negatively correlated with increased FA in the GCC following surgery (r = -0.60, p = 0.003; Figure 3(B)). Cross-modalities correlation analysis showed that increased ALFF in the ACC correlated positively with increased FA in the GCC and MCP at PostLSG respectively (r = 0.63, p = 0.002; r = 0.59, p = 0.003). FA in the Cin correlated negatively with ALFF in the HIPP (PreLSG: r = -0.29, p = 0.190; PostLSG: r = -0.60, p = 0.003), but there was no significant difference between PreLSG and PostLSG (z = 1.22, p = 0.223, Figure 3(C)).

Figure 3. Correlation results between the clinical measurements and functional/structural imaging data in LSG group. (A) BMI was positively correlated with ALFF in the OFC at PreLSG, there was significant difference between PreLSG and PostLSG. Reduced BMI positively correlated with increased ALFF in the OFC at PostLSG. BMI was positively correlated with ALFF in the HIPP at both PreLSG and PostLSG. (B) BMI was positively correlated with FA in the GCC at PreLSG, and there was significant difference in correlation between PreLSG and PostLSG. Reduced WC negatively correlated with increased FA in the GCC at PostLSG. (C) Increased ALFF in the ACC correlated positively with increased FA in the GCC and MCP at PostLSG respectively. FA in the Cin correlated negatively with ALFF in the HIPP at PostLSG. LSG: Laparoscopic sleeve gastrectomy; PreLSG: patients with obesity who had MRI scan before surgery; PostLSG: patients with obesity who received LSG and were scanned again 1 month after surgery; ALFF: amplitude of low frequency fluctuations; BMI: body mass index; OFC: orbitofrontal cortex; HIPP: hippocampus; FA: fractional anisotropy; GCC: genu corpus collosum; WC: waist circumference; ACC: anterior cingulate cortex; MCP: middle cerebellar peduncle; Cin: cingulate.

4. Discussion

In the current study, we adopted resting-state MRI and diffusion tensor imaging to investigate LSG-induced brain functional and structural changes. Results showed that increased and decreased ALFF in brain regions implicated in motivation/reward (OFC and caudate), memory/emotion (HIPP, amygdala, and thalamus), and inhibitory-control (ACC and DLPFC). LSG-induced structural recovery of white matter tracts included the GCC, Cin, MCP, and superior longitudinal fasciculus. In addition, correlation analysis showed that BMI was positively correlated with ALFF in the OFC, and correlated with FA in the GCC at PreLSG. There were significant correlations between BMI and ALFF in the HIPP before and after surgery. Reduced BMI/WC were positively correlated with decreased ALFF in the OFC, and negatively correlated with increased FA in the GCC following LSG respectively. Cross modalities correlation analysis showed that increased ALFF in the ACC were positively correlated with increased FA in the GCC and MCP following surgery, and FA in the Cin negatively correlated with ALFF in the HIPP at PostLSG.

4.1. Alterations in regions associated with motivation/reward

The OFC is involved in the processing of motivation, salience attribution and output of compulsive behaviors. The OFC directly projections to DA neurons and to nucleus accumbens that implicated in conditional responses to food cues, excessively neural activity in the OFC is likely to reflect downstream DA functions in encoding of motivation and salience attribution driving overeating behaviors [25]. This is consistent with a previous study showing that, a positive association between higher YFAS scores and greater OFC activation in response to high-calorie food cues [26], suggesting that hypersensitivity to the reward value may serve as a risk factor for obesity. Our data also showed that OB group had increased activation in the OFC, and it was associated with higher BMI, whereas LSG normalized it at PostLSG, which provides evidence that weight-loss concomitantly alters brain activation, and improved brain function might also contribute to long-term weight-loss. In addition, positive correlation between decreased ALFF in the OFC and reduced BMI indicated that differences in weight before and after surgery were captured by brain imaging measures, which might be used as an indicator to predict the effect of LSG, and also as a potential target for behavioral retraining in the treatment of obesity.

The caudate is involved in reward-related processing. The DA responses play an important role in the rewarding effects of foods, dysregulated dopamine D2 receptor striatal signaling might result in hyperactivity of the mesolimbic system toward reward-predicting cues in OB [1]. With regards to the contribution of obesity to the imbalance between inhibitory-control and reward circuits, repeated exposure to the food reward, the decreased striatal dopamine D2 receptor is likely to reduce the sensitivity to natural rewards [2], this might explain why OB had decreased activity in the caudate. In addition, LSG-induced increased activity in the caudate in the OB group after surgery might reflect improved DA function in response to food consumption in the reward circuitry.

4.2. Alterations associated with memory/emotion

HIPP region traditionally has been associated with learning and memory, wherein higher desire to eat a specific food was associated with activation of the HIPP, which is likely to reflect its involvement in storing and retrieving memories for the desired food [12]. In fact, the HIPP region is also implicated in a specific function of memory inhibition and emotional regulation that could contribute to the suppression of food intake. Specifically, gastric stimulation activates the HIPP region presumably from downstream stimulation from the vague nerve and solitary nucleus, associations between gastric stimulation-induced changes in HIPP metabolites and the scores on emotional eating and uncontrolled eating also support the function [27].

Our data also revealed OB group had increased ALFF in the HIPP, and showed decreased activity at PostLSG compared to PreLSG, which is consistent with previous finding that LSG produced a greater decrease in appetite and brain activity in the HIPP [12]. This finding suggests that LSG might enhance HIPP ability in interfering with the retrieval of food reward memories, and contribute to decreases in food intake and weight loss following surgery. Although activation of the HIPP was significantly correlated with BMI at both PreLSG and PostLSG, LSG significantly decreased ALFF in the HIPP and BMI in OB group following surgery, suggesting that functional improvement in HIPP may develop gradually at PostLSG in tandem with weight-loss. It worth noting that, brain activities in the HIPP were significantly negatively correlated with FA in the Cin at PostLSG but not at PreLSG, replicating a finding that LSG can induce neuroplastic recovery of brain structural abnormalities, which changes are also accompanied by functional alterations [15]. The Cin is a relatively larger white matter tract, and plays a key role in the processing and regulation of emotions. It represents a major association with fiber tract of the limbic-cortical networks involved in obesity and depression [28]. Food ingestion evokes greater activation in the HIPP and ACC regions implicated in memory and executive control. Participants who had stronger predicted reward for food-intake memory, the greater activation in the emotional circuit would drive to procure it, that ultimately influences unhealthy eating behaviors and result in obesity [29].

Increased resting-state activity and activation were reported in the amygdala during resting-state and exposed to visual food-cues in OB, which engaged at an early stage of processing by evaluating novel stimuli to determine their value as reinforcers [3]. The amygdala is also involved with the emotional component of food intake. Previous study showed that overweight women with high chronic stress are more vulnerable to the acute effects of stress on amygdala response to milkshake, and this activation can be dampened during the neutral-relaxing condition [30]. Our data also showed that LSG had increased ALFF in the amygdala, and it was reduced following surgery, suggesting that LSG improved interoceptive awareness of bodily states, and decreased sensitivity of the negative mood states.

The thalamus plays an important role in integrating and relaying the flow of information from the basal ganglia, midbrain, and brainstem to the prefrontal cortex [31]. Neuroimaging studies showed that the thalamus displayed a greater response to the food cues when they predicted immediate receipt of caloric drink compared to predicting the arrival of a tasteless solution [32], suggesting that thalamus not only implicated in sensory and emotional processing, but also represented predictive encoding the subjective pleasantness of food cues.

4.3. Alterations associated with inhibitory control

The ACC is implicated in executive control of internal/external stimuli-related and involving modulation of emotional responses, brain abnormality in this region may contribute to an imbalance between cognitive and emotional processing, and consequentially an increased risk to overeating [33]. Our data revealed that obese-related functional brain abnormalities in ALFF in the ACC, and structural abnormalities in FA in the GCC and MCP, which could be normalized following surgery. This indicated that LSG can induce brain functional and structural recovery simultaneously, and these differences in weight before and after surgery can be captured by brain imaging. It is worth noting that increased ALFF in the ACC were positively correlated with increased FA in the GCC and MCP at PostLSG, suggesting that LSG-induced brain structural alterations might be associated with brain activity changes, and may result in improvements in eating behaviors and weight loss.

Specifically, the corpus callosum is a larger white matter tracts that connect both hemispheres (including forceps minor, forceps major, splenium and genu), it involves in multiple motor/sensory processing, and cognitive functions [34]. The GCC is part of corpus callosum, our current finding aligns with a previous study showing that decreased FA in the GCC linked to impaired executive functions, poorer memory performance, and changes in reward processing [12,35,36]. In addition, the MCP is a paired-structure that connects the cerebellum to the pons, and is composed entirely of centripetal fibers, which arise from the pontine nucleus on the opposite side and end in the cerebellar cortex [37]. LSG-induced weight loss was associated with increased white matter density and integrity in the MCP [13], which is consistent with our current results, supporting the interpretation that improved brain activity in the ACC and inversed structural brain abnormalities in the GCC and MCP might modulate eating behavior and weigh loss.

The DLPFC plays a crucial role in executive function, which also could be modulated by DA system, wherein decreased activity in the DLPFC is likely to poorer central regulation of eating behavior, and high impulsivity and compulsivity to consume desirable food in OB [38]. Our data showed that OB groups had decreased ALFF in the DLPFC than NW groups, and LSG normalized it, suggesting that improved cognitive control of eating behaviors following surgery may contribute to long-term dietary control for weight loss.

4.4. Limitations

Due to the strict exclusion criteria and difficulty in retaining patients after surgery for follow-up scanning, we did not have a larger cohort size for the LSG group. Associations of inflammation and/or peripheral hormones with functional/structural brain abnormalities have not been assessed. With regards to LSG-induced changes in brain function and structure might be different at different time points, multiple time-point assessments are warranted for future investigations. The correlations between clinical measurements and brain function and structure did not reach significance after multiple comparisons, but certain correlative trends may still be meaningful for delineating their relationship.

5. Conclusion

The current study examined LSG-induced brain functional and structural alterations in OB before and after LSG. Results showed increased and decreased ALFF in a number of brain regions implicated in motivation/reward (OFC and caudate), memory/emotion (HIPP, amygdala, and thalamus), and inhibitory-control (ACC and DLPFC). LSG-induced structural recovery of brain white matter tracts included the GCC, Cin, MCP, and superior longitudinal fasciculus. Cross modalities correlation analysis showed that increased ALFF in the ACC were positively correlated with increased FA in the GCC and MCP following surgery, and ALFF in the HIPP was negatively correlated with the FA in the Cin at PostLSG. Our findings provide evidence that LSG induces simultaneous changes in brain functions and structures which are associated with weight loss.

Author contribution

The authors’ responsibilities were as follows: Conceptualization, YZ; Data acquisition, JW, GA, WC, WJ, GL, YH, WZ; Data analysis, WC; Writing-Original Draft, JW; Writing-Review & Editing, YZ. All authors critically reviewed the content and approved the final version for publication.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China under Grant [number 82172023, 82202252]; Natural Science Basic Research Program of Shaanxi under Grant [number 2022JC-44, 2022JQ-622].