| Objective(1) To compare the sensitivity and validity of one automated segmentation and volumetric method-IBASPM and two voxel-based morphometry approachs:VBM5standard and VBM8-DARTEL with manual hippocampal volumetry in first episode, treatment-naive young MDD and schizophrenia patients.(2) To detect hippocampal structural changes in MDD and schizophrenia patients by using the four volumetric methods above simultaneously in the same data。MethodHigh-resolution three-dimensional (3D) T1-weighted images of32first episode, treatment-naive young MDD patients and30healthy controls were acquired on a1.5-T whole-body scanner (Avanto, Siemens,Germany) of the Third Xiangya Hospital. In addition, high-resolution three-dimensional (3D) T1-weighted images of32first episode, treatment-naive patients with schizophrenia and27healthy controls were acquired on a3.0-T whole-body scanner (Intera Achieva, Phillips, Holland) of the Second Xiangya Hospital. The measurement of bilateral hippocampal structure were compared respectively in two MRI data using the manual segmentation and volumetric method, IBASPM and the hippocampal-ROI analysis in VBM5standard and VBM8-DARTELResults(1) In the MDD study, the Repeated Measure ANCOVA with manual hippocampal volume standardized by Cendes'method as the dependent variable (gender and age as covariates) revealed a significant group effect (F=5.192, df=58, p=0.026), and significant differences restricted to the left hippocampal region (F=9.977, df=58, p=0.003). For the manual raw hippocampal volume of MDD patients, the Repeated Measure ANCOVA with gender, age and the manual estimated TIV as covariates also revealed a significant group effect (F=10.113, df=57, p=0.002), and significant differences restricted to the left hippocampal region (F=13.943, df=57, p<0.001). For the hippocampal volume measured by IBASPM in MDD patients, the Repeated Measure ANCOVA with gender, age and the individual brain volume as covariates revealed a significant group effect (F=5.337, df=57, p=0.025). When doing hippocampus-ROI analyses with the standard VBM5procedure, ANCOVA showed no significant group effect (multiple comparisons correction:p<0.01, cluster>18). However, when doing hippocampus-ROI analyses with the VBM8-DARTEL procedure, ANCOVA showed significant left hippocampal differences between the MDD patients and healthy controls (-36-31-12:z=2.87, multiple comparisons correction:P<0.01, cluster>16).(2) The Pearson's correlation is used to find the correlations among the TⅣ/brain volume measured by four methods in the all subjects, healthy control group, and MDD group respectively. The highest correlation were found within three samples between the brain volume measured by the standard VBM5and VBM8-DARTEL(0.896(P<0.01),0.947(P<0.01)与0.863(P<0.01)), whereas the lowest correlation were found within three samples between the manual measured TIV and the brain volume measured by IBASPM(0.209(P>0.05),0.444(P<0.05)与0.055(P>0.05)). In addition, the correlation between the he manual measured TIV and the brain volume measured by VBM8within three samples were0.209(P>0.05),0.444(P<0.05)与0.055(P>0.05). For the hippocampal volume measured by IBASPM, the Repeated Measure ANCOVA with gender, age and the individual brain volume measured by VBM8as covariates revealed no significant group effect (F=0.987, df=57, p=0.325). However, the same analysis for the manual raw hippocampal volume still revealed a significant group effect (F=7.690, df=57, p=0.008), and significant differences restricted to the left hippocampal region (F=11.964, df=57, p=0.001).(3) In the schizophrenia study, the Repeated Measure ANCOVA with manual hippocampal volume standardized by Cendes's method (gender and age as covariates) revealed a significant group effect (F=7.500, df=55, p=0.008), but no significant side effect and groupxside interaction. For the manual raw hippocampal volume of MDD patients, the Repeated Measure ANCOVA with gender, age and the manual estimated TIV as covariates also revealed a significant group effect (F=23.116, df=54, p<0.001), but no significant side effect and groupx side interaction. For the hippocampal volume measured by IBASPM, the Repeated Measure ANCOVA with gender, age and the individual brain volume (measured by VBM8) as covariates revealed a significant group effect (F=5.937, df=54, p=0.018), but no significant side effect and groupxside interaction. When doing hippocampus-ROI analyses with the standard VBM5procedure, ANCOVA showed no significant group effect (multiple comparisons correction:p<0.01, cluster>18). However, when doing hippocampus-ROI analyses with the VBM8-DARTEL procedure, ANCOVA showed significant bilateral hippocampal differences between the schizophrenia patients and healthy controls ((-20-15-11:z=2.83;23-369:z=2.72;-14-3310:z=2.75; multiple comparisons correction:P<0.01, cluster>16).Conclution(1) The VBM8-DARTEL methods showed good agreement with manual hippocampal volumetry, thus suggesting the high sensitivity and validilty of this VBM methodology. This technique could be very useful for detecting the localized morphologic alterations in the patients with severe mental illness.(2) The sensitivity and validilty of standard VBM5for the identification of hippocampal morphologic alterations are not as good as manual volumetry method and VBM8-DARTEL, especially for the patients with severe mental illness.(3) Although the automated segmentation and volumetric method-IBASPM could detect the hippocampal differences in the patients with severe mental illness, the results indicate that IBASPM need to be improved. Furthermore, the total brain volume measured by IBASPM was questionable.(4) The left hippocampal volumes of the patients with MDD were significantly smaller than that of the normal controls.(5) The bilateral hippocampal volumes of the patients with schizophrenia were significantly smaller than that of the normal controls.
|
PDF Full Text Request