The present study examined both anatomical and functional abnormalities in a cohort of drug-naïve MCI patients. Reductions in gray matter volume were observed mainly in the prefrontal, left temporal and posterior cingulate cortex (), and volume reduction in the left inferior frontal gyrus was related to the severity of symptom. The regional functional alteration included decreased regional functional mainly in prefrontal, left parietal regions and right fusiform gyrus, while the increased regional functional in limbic and midbrain (). However, there is no correlation between morphometric and functional deficits, suggesting an independent alteration of morphometry and function at the transitional state between normal aging and dementia.
Consistent with previous studies 
, widespread cortical atrophy was found especially in the bilateral frontal lobe, the left temporal gyrus and posterior cingulate cortex. Furthermore, the decreased GMV in the left middle frontal gyrus survived even after family wised error correction (p
0.047, corrected with FDR). In fact, all these areas are regions typically involved by AD related pathology, as shown by previous MRI 
and PET 
studies, and the gray matter atrophy in the prefrontal and posterior cingulate cortex were even thought to predict the cognitive decline in MCI patients as revealed by longitudinal VBM studies 
. However, we did not find more GMV reduction, especially in the hippocampus which has been the most studied area in MCI cases. In fact, more and more recent studies have shown that structural lesions (eg, neurofibrillary tangles, neuritic plaques) associated with the MCI syndrome are not restricted to hippocampus, but also included the cortical areas 
. Recent study in patients with mild AD even revealed amyloid deposits in the frontal and parietal lobes instead of the medial temporal lobe structures 
. Consistent with these findings, our results confirmed the gray matter deficits in the frontal and posterior cingulate cortex in a cohort of drug-naïve MCI patients, suggesting a more diffused cortical deficits than previously thought in MCI.
Furthermore, the decreased volume in the left inferior frontal gyrus was found to be related to the severity of clinical symptom, as measured by the MMSE. The left inferior frontal gyrus has been implicated as an important part in the pathology of MCI, which was thought to be associated with attention and memory processes, including encoding and retrieval and long- and short-term memory 
. Increased activation in the left inferior frontal gyrus was associated with improved memory processing in subjects with MCI as revealed in a recent double-blind placebo-controlled fMRI study 
. However, previous study only has reported positive correlation between the MMSE and the volume of right inferior frontal gyrus 
. Our results extended previous findings by suggesting that the severity of atrophy in the left inferior frontal gyrus could also be used to evaluate the clinical severity in MCI patients.
Interestingly, resting-state fMRI study revealed both increased and decreased spontaneous activity in MCI group in relative to controls. Abundant previous studies by PET or SPECT have reported hypoactivity in temporal, prefrontal and parietal regions, and decreased glucose metabolism in the temporal and parietal lobes on FDG-PET is recognized as an early imaging marker for the AD pathology 
. Consistent with the previous study 
, decreased ALFF/fALFF was observed in cingulate cortex, medial prefrontal cortex and some parietal regions; while increased ALFF/fALFF was observed in superior temporal gyrus. Furthermore, the changes of ALFF/fALFF in MCI in the slow-5 band were greater than those in the slow-4. However, we also revealed increased ALFF mainly in limbic regions i.e., parahippocampal gyrus, insula and anterior cingulate cortex () and midbrain which had been not been found in the previous study and rarely reported by PET or SPECT studies 
. In fact, task-fMRI studies in MCI have also reported hyperactivation especially in limbic regions during memory encoding 
. One possible explanation for this abnormal hyperactivity is a compensatory neural mechanism, i.e., there is a temporary phase of increased activity in areas like parahippocampus along the course of MCI, which may keep the behavioral performance of MCI patients close to the level of cognitively intact elderly subjects. Around the conversion from MCI to clinical AD, such compensatory ability would be lost, which is then seen as poor task performance. This hypothesis was supported by the finding that greater clinical impairment in MCI subjects as related to smaller hippocampal volumes but greater parahippocampal fMRI activation 
. Another study even found a negative correlation between hippocampal atrophy and left parahippocampal activation only in MCI instead of in the elderly controls or AD patients 
. However, we did not observed GMV reduction in hippocampus in drug naïve MCI, suggesting such compensatory overactivity in limbic regions occur before the detectable hippocampal atrophy by VBM. Another possibility is the relative small sample size which would low the statistical power.
Contradictory to our hypothesis, there is no overlap between the morphometric and functional results, or any correlations between the altered regional GMV and ALFF, though the decreased ALFF in the right middle frontal gyrus and left inferior frontal gyrus were found quite near the decreased GMV in the same gyrus (). By now, only a few studies used anatomical and functional methods in combination in MCI or AD patients. The results from combination of brain volumetrics and PET 
showed that hypometabolism largely exceeds gray matter atrophy in most brain regions of AD patients 
. Another study even observed dissociated effect of atrophy and hypometabolism on episodic memory in MCI patients using the same method 
. Only two recent studies combine VBM and resting state fMRI in MCI patients, and they also found no overlap between regions with gray matter atrophy and regions with altered functional connectivity 
. These results altogether with our findings support that brain volume and function may change independently and act different role in the earliest cognitive symptoms of MCI. The altered regional function may represent effects of early neurodegeneration 
Several methodological issues should be considered when interpreting the present results. First, although DARTEL-VBM was employed in the current study, which minimizes the contamination of brain tissue with non-brain voxels compared to standard VBM 
, the use of a Chinese sample may affect the accuracy of normalization because the generic MNI template differs structurally from the brains of non-Caucasian populations 
. However, as both patients and controls were from a similar ethnic background, the groups differences observed seem unlikely to have been affected by ethnic factors. Second, though we temporally bandpass filtered all fMRI data (0.01–0.08 Hz), and removed components with high correlation to CSF or white matter or with low correlation to gray matter, we cannot completely rule out the influence of physiological noise on our findings due to its variation over time and across subjects. Simultaneous recording of heart rate and respiratory rate and depth during fMRI scanning might help further reduce physiological noise artifacts 
. Nevertheless, the consistency among the resting state connectivity patterns evidenced by the present data and other studies 
does reduce the concern about the magnitude of such potential artifacts. Finally, there is a lack of consensus about the exact physiological nature of ALFF. Though ALFF is thought to reflect spontaneous neural activity 
, its exact basis remains to be fully characterized.
Taken together, current study combines anatomical and functional MRI to examine their relationship in a cohort of drug-naïve MCI patients. Findings from the study document that wide spread brain volume reduction accompanied with decreased and increased regional function in MCI, while the anatomical and functional changes were independently. Therefore, the combination of structural and functional MRI methods would provide complementary information and together advance our understanding of the pathophysiology underlying the symptoms of MCI.