Resting-state fMRI is a popular way to study functional networks in the human brain and the abnormalities associated with various neuropsychiatric disorders. However, a lack of quantitative metrics closely associated with underlying neurophysiology characteristics has made its translation to clinical settings difficult^1,2. In this study, we propose two voxel-based metrics derived from resting state fMRI measurements as potential biomarkers for sensitive detection of pathological changes resulting from neurological disorders.The study sample consisted of 27 patients Alzheimer’s disease (AD) (m/f =12/15, age=73±6) and 20 healthy controls (m/f =9/11, age=72±7). The resting-state fMRI data acquisition was conducted using a whole-body 3T clinical MRI scanner (Magnetom Trio, Siemens) equipped with a 32-channel phased-array receiving head coil. The protocol included conventional clinical MRI scans and a session of 10min long resting-state fMRI using a 2D GRE EPI technique. The main acquisition parameters included the following: TE/TR 35/1600 ms, flip angle=90°, 26 slices of 4 mm thick, in-plane resolution of 3x3mm2, iPAT=2, and 400 dynamic timeframes³. The resting-state fMRI datasets underwent a preprocessing pipeline based on AFNI^1,2. The main steps included temporal de-spiking, rigid body image registration for motion correction, spatial normalization to the standard MNI template, nuisance signal removal by voxel-wise regression using 14 regressors based on the motion correction parameters, the average signal of the ventricles, baseline detrending removal up to the third order polynomial, low-pass filtering at 0.08Hz, and local Gaussian smoothing up to FWHM=5mm using an eroded gray matter mask. For each voxel inside the brain, we compute the Pearson’s cross-correlation coefficients of the resting-state fMRI time course with that of every other voxel inside the brain. To conduct quantitative data-drive analysis (QDA) of the resting-state fMRI data we computed the following voxel-wise metrics: (1) Connectivity counter index (CCI) which is defined as the number of voxel pairs involving the current voxel in question with the Pearson’s cross-correlation coefficients above a given threshold. (2) Connectivity strength index (CSI) which is defined as the non-zero mean value of the Pearson’s cross-correlation coefficients above a given threshold for all voxel pairs involving the current voxel in question. To optimize the threshold we systematically computed CCI and CSI by setting the threshold 0.2-0.4 with an incrementing step of 0.01. We performed t-test with derived CCI and CSI images to detect possible differences between AD and healthy controls. The statistical significance of t-test results was assessed using a family-wise error rate at p<0.05.As shown in Figure 1, the derived CCI and CSI metrics depicts high contrast between different brain tissues. The functional connection hubs can be directly identified in the CCI images as the bright brain areas showing higher number of functional connections. The CSI images have similar but less tissue contrast. The brain regions with significantly reduced functional connectivity in AD patients included posterior cingulate, precuneus, inferior parietal lobule and prefrontal cortex as detected by t-test of the CCI and CSI metrics (see Figure 2). This is very consistent with the results reported by previous studies using MRI and other brain imaging modalities³. The optimal threshold value to derive CCI and CSI is about 0.3 as indicated by the maximum detected brain volumes (Figure 3).In addition to the AD data shown here, we have used these metrics to investigate functional connectivity changes in patients with mild trauma brain injuries⁴, migraine^5,6, and psychosis. With a robust pre-processing pipeline for the resting-state fMRI raw data, the derived statistical contrast from CCI and CSI metrics are relatively stable. We have used a consistent data acquisition protocol. It is apparent that we have to investigate further how the metrics are influenced by the data acquisition protocol. We also need to test their robustness in other clinical applications. At higher spatial resolution, the demand for computation capacity can become an issue.The proposed CCI and CSI metrics are relatively stable with good sensitivity to functional connectivity changes related to neuropathology. The metrics can be interpreted straightforward as the local functional connectivity density and strength, respectively. Most importantly, these metrics can be used for quantitative data-driven analysis (e.g., t-test or ANVOA) for direct comparison between subjects groups and over different time periods. Furthermore, they can also be used for studying correlations with other biomarkers and behavior measures.