To study Alzheimer’s disease (AD) and mild cognitive impairment (MCI), the medial temporal lobes (MTL), identified as the components of neuronal system for declarative memory [1], have received considerable interest. perirhinal cortices (PRC) and parahippocampal cortices (PHC) are both part of MTL. However, the possible functional connections between PRC or PHC and hippocampus subfields in patients with MCI and AD were unknown. Therefore, this study aimed to explore the different functional connectivity among PHC/PRC and hippocampus subfields in NC, MCI and AD patients with resting-state blood oxygen level–dependent (BOLD) sequence. Based on the research by Laura A. Libby et all[2],we divided the hippocampus into three subfields as head, body and tail.

Subjects: a total of 90 subjects, including 32 AD patients, 31 MCI patients and 27 NC, were recruited from the Department of Neurology of the Affiliated Drum Tower Hospital of Nanjing University Medical School in this study. The characteristics of patients are presented in Table1.

Sequence: each subject underwent a 3DT1W scan and a resting-state BOLD scan, respectively, with a 3 Tesla MR scanner (Achieva 3.0T TX dual-source parallel RF excitation and transmission technology, Philips Medical Systems, The Netherlands). The parameters of the 3DT1W and BOLD scan are shown in table 2.

Data Processing: The pre-processing was taken on the Brainnetome fMRI Toolkit (BRAT 1.0), which is designed by the Brainnetome Center, Institute of Automation, Chinese Academy of Science. All patients passed the quality control of head motion which is 3mm and 3 degree. The right and left PHC masks were taken from AAL(Anatomical Automatic Labeling) Template and the right and left PRC masks were taken from Bordmann Template. Therefore, the region of interest (ROI) included PHC(L/R), PRC(L/R), hippocampus head(L/R), hippocampus body(L/R) and hippocampus tail(L/R). Function connectivity among ROIs was processed on the BRAT 1.0 software with the pre-processed data.

Statistical Analysis: Statistics analyses including One-way analysis of variance (ANOVA) and multiple comparison correction (MC) with a p value of < 0.05. ANOVA was taken on SPSS (V 21.0) and multiple comparison correction was applied with Matlab code based on FDR principle. Finally, Gaussian mixture model was used to cluster the MCI data.

As shown in table 3, the functional connectivity in three pairs of ROIs has statistical significant differences among NC, MCI and AD. The three pairs of ROIs are right PRC connected with right hippocampus tail, left PRC connected with right hippocampus tail, and right PHC connected with right hippocampus head. Figure 1 and table 3 show that in AD group more functional connection decrease between PRC and right hippocampus tail sub-region occurs compared with MCI. In contract, the functional connection between right PHC and the right hippocampus head sub-region significantly increases in MCI (p=0.003) and decreases in AD (p=0.03) compared with NC.

The result reveals that the decrease of functional connectivity between PRC and right hippocampus tail affects the cognitive ability. Compared with NC, there is a significant increase of functional connectivity between right PHC and the right hippocampus head sub-region in MCI subjects. It appears that, in the course of MCI there may be hyper-activation in some functional connections, possibly representing inefficient compensatory mechanism for memory encoding activity.

MCI is a heterogeneous clinical entity with multiple sources of heterogeneity. According to the result of functional connections between PRC / PHC and the three hippocampus subfields, we clustered MCI subjects in this research into two types, shown in figure 2, using Gaussian mixture model (GGM). This result suggests that the different functional connection patterns in MCI patients can be the basis and another criterion of MCI classification.

The limitation is that MCI subjects in this research were not divided into different subtypes with clinical manifestation, therefore it’s hard to find the difference of functional connectivity in MCI subtypes as well as the relationship between subtypes and the classification result we achieved. In the next step, we will continue our research on comparing the functional connectivity in different MCI subtypes and follow the progress of clinical changes in MCI subjects by follow-up visit. It’s promising that our research can help the classification of AD, MCI and NC, as well as benefiting the early diagnosis of AD.

Our result indicates significant differences of functional connectivity between right PRC and right hippocampus tail, between left PRC and right hippocampus tail, and between right PHC and right hippocampus head among AD, MCI and NC subjects. The result may reveal some neuronal alterations with the disease evolvement, and make further efforts on classification of AD, MCI and NC.