Since there is more and more delicious food in our daily life, people cannot resist the attraction to food. Therefore, obesity has become an important issue in modern society. Previous studies used food pictures to stimulate obese patients and used functional MRI to find the brain regions with increased activity [1]. However, few studies mentioned about particular brain structure changes in obese patient. Noninvasive diffusion tensor imaging (DTI) are able to observe the water diffusion in the brain on the microscopic level for the early detection of white matter structural changes. Therefore, we used DTI to find the differences of brain structures between obese patients and healthy controls. The correlation between clinical and the DTI indices were also calculated and discussed. The clinical indices included body mass index (BMI), and measures of anxiety and depression.

Diffusion imaging scans of 20 obese patients (BMI = 37.9 ± 5.2) and 30 healthy controls (BMI = 22.6 ± 3.4) were obtained. All patients underwent a brain MRI examination on a 1.5T MRI system (Ingenia, Phillips, Netherlands). The scanning parameters were as follows: 40 slices; 128 x 128 matrix; 1.75 x 1.75 x 3 mm³ voxel size; 224 x 224 mm² FOV; 3 mm slice thickness; repetition time = 3279 ms; echo time = 110 ms; 67 diffusion orientations; and b-values of 0, 1000, 2000 s/mm². The scan time for each patient was almost 21 minutes.

Each participant’s original image was done Eddy Current Correction using FSL (FMRIB Software Library). Then, the diffusion images were spatially normalized to the Montreal Neurological Institute (MNI) T2W template using parameters determined from the normalization of the diffusion null image to the T2W template using Statistical Parametric Mapping (SPM). For the DTI analysis, DTI reconstruction was performed using DSI Studio, and the fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD) mapping were calculated. For the statistical analysis, a two sample t-test was used to detect the significant differences between the obese patients and the healthy controls on the DTI indices. Moreover, multiple regression was used to detect the correlation between the clinical and the DTI indices for the 50 participants.

The FA value of the corpus callosum, cingulate gyrus and hippocampus was lower in the obese patients compared with the healthy controls. The RD and MD values of the frontal gyrus were also lower in the obese patients compared with the healthy controls (Fig. 1). There was no significant difference in the AD values of the two groups. In contrast to the healthy controls, the patients with an eating disorder did not recruit the frontal gyrus during the body size estimation task [2]. This meant that the patients with an eating disorder might not realize that their body size was so large that they had to change their eating habits. Frontal gyrus activity has been proposed to represent cognitive effort to trigger an avoidance tendency to food [3]. In a previous study [4], FA was reduced in patients with eating disorder in corpus callosum. People have also been found to exhibit lethargy and loss of interest in sports after cingulate lesions [5]. The anterior cingulate, meanwhile, has been described as triggering compensatory adjustments in cognitive control [6]. These compensatory adjustments could make a person more likely to engage in exercises and more accepting of weight control advice [7]. Furthermore, increases in hippocampal volume suggest that exercise may elicit higher levels of brain derived neurotrophic factor (BDNF) [8].

In our study, significant negative correlation between BMI and FA, MD values in the cingulate gyrus was found. There was no significant correlation, however, between BMI and AD or RD values (Fig. 2). Specifically, when the BMI value was higher, the DTI indices in the cingulate gyrus were smaller. It should be noted that a higher BMI value for a person indicates that he/she is more obese. As shown in Fig. 1, we found that the DTI indices of the cingulate gyrus were indeed lower in the obese patients compared with the healthy controls.

A significant negative correlation between anxiety scores and FA values in cingulate gyrus was found. Significant negative correlation between anxiety scores and AD, MD values in the hippocampus were also found, as was a significant negative correlation between anxiety scores and RD values in the frontal gyrus (Fig. 3). A significant negative correlation between depression scores and AD values in the corpus callosum was also found, as was a significant negative correlation between depression scores and RD values in the frontal gyrus (Fig. 4). There was no significant correlation, however, between depression scores and FA or MD values. Patients with an eating disorder exhibit higher incidences of anxiety and depression [9]. As shown in Fig. 3, there was a significant negative correlation between the DTI indices and anxiety scores. Specifically, when the anxiety score was higher, the DTI indices in the frontal gyrus, cingulate gyrus and hippocampus were smaller. As shown in Fig. 4, there was a significant negative correlation between the DTI indices and depression scores. Specifically, when the depression score was higher, the DTI indices in the frontal gyrus and corpus callosum were smaller. It should be noted that a higher anxiety or depression score for a person indicates that he/she is more anxious or depressed. As shown in Fig. 1, we found that the DTI indices of the frontal gyrus, cingulate gyrus, hippocampus and corpus callosum were indeed lower in the obese patients compared with the healthy controls. Therefore, we could conclude that obese patients might also have feelings of anxiety and depression.

In summary, the results of our study indicated that the DTI indices of the frontal gyrus, corpus callosum, cingulate gyrus, and hippocampus were lower in obese patients compared with the healthy controls. The obese patients not only had higher BMI values but were also more likely to have feelings of anxiety and depression. Based on the results of the study, we may have gained a better understanding of obese patients and provide a slight contribution to the clinical treatment of obese patients.