fMRI investigators, neural scientists, Psychologists, and clinicians.Quasiperiodic pattersn (QPPs) of BOLD fluctuations, first reported in [1,2] are likely contributors to functional connectivity (FC) due to their spatial and temporal structure. FC has been widely used to explore the altered brain organization in patients suffering from psychological disorders like Major Depressive Disorder (MDD). In this project, we examined the contribution of QPPs to FC in both normal subjects and MDD patients, and determined whether the differences in connectivity observed in MDD patients were primarily from components corresponding to QPP or the residue or distributed across both.

Subjects and data preprocessing Resting state fMRI (rs-fMRI) images were acquired from 14 patients diagnosed with MDD and 16 healthy controls (male, right-handed, 20~25 years old on a 3T scanner (SIEMENS MAGNETOM Trio) with the following parameters: GRE-EPI sequence, TR/TE=2000/30ms, 30 slices, slice thickness=4mm, FOV=210mmx210mm, matrix=64x64, 7minutes per subject. The preprocessing pipeline included DICOM to NIFTI conversion, slice timing correction, realignment, normalization and spatial smoothing (Gaussian kernel=4mm). Nuisance covariates (six motion parameters, global mean signal, white matter and cerebrospinal fluid signal) were regressed out, and resultant data was detrended and filtered (0.01Hz~0.08Hz) using DPARSF (Yan & Zang, 2010, http://www.restfmri.net).

QPP extraction algorithm QPP templates and plots of template strength as a function of time were obtained with the algorithm described in [2]. To separate the contribution of the QPPs from the raw BOLD signal, the convolution of the QPP templates and time course of template strength was calculated as I_hat, and then the component of the raw BOLD signal (I_raw) corresponding to the QPP was calculated as the projection of raw BOLD in the direction of I_hat, noted as I_qpp. Finally the QPP-corrected signal was defined as I_residue=I_raw-I_qpp. Thus, the raw BOLD signal was divided into two parts: I_residue and I_qpp. The spatial distribution of QPP contributions was then calculated as the percentage change in standard deviation after QPP regression in each voxel.

Analysis of FC FC was calculated using a seed in the posterior cingulate cortex (PCC), a prominent node of the default mode network (DMN) where the reduction of variance after QPP regression is high. To examine the contribution of QPPs to FC, a post hoc T-test was used to determine whether there was a significant change in FC after QPP regression in the control group or in the MDD group. To determine whether the differences in connectivity observed in MDD patients were primarily from I_qpp or I_residue or distributed across both, two-sample T-tests were performed on the FC maps of I_raw and I_residue between the two groups, with significance level 0.05 (after alpha-sim multiple comparison correction).

The distribution of QPP contributions reflected by the change in standard deviation (STD) after QPP regression are shown in fig.1. QPPs accounted for up to 25-30% of the temporal STD of the raw BOLD signal. Differences in the FC with the PCC seed before and after QPP regression are shown in fig. 2. As expected from the greater contribution of the QPP to temporal variance in the control group, connectivity differences with greater spatial extent and higher significance levels were observed in the control group than in the MDD group. The connectivity differences covered much of the brain in the control group but were limited to the DMN in MDD patients. When looking at differences in FC between controls and MDDs, significant decreases were observed in areas of the DMN in the MDD group; however, after QPP regression these differences disappeared, other than one small positive area in the frontal lobe (shown in fig.3).Although FC studies have provided new insights into the network organization and alterations due to pathology (F. Sambataro et al, 2014; L.L. Zeng et al., 2012; J Choi & B Jeong, 2011;), the neurophysiological basis of the BOLD fluctuations used to map FC remain poorly understood. This project shows that QPPs are a major contributor to FC. In animal models, QPPs have been linked to infraslow electrical activity [3] , so this is a step toward elucidating the neurophysiology underlying the BOLD fluctuations used in FC studies. Meanwhile, as differences between the control group and the MDD group nearly disappeared after the contribution of QPPs was minimized by regression, QPP analysis provided new insights into the altered FC observed in a major psychiatric disorder. Although the neurophysiological basis of the QPPs in humans is still under investigation, it is clear that QPP abnormality can be a contributor to or marker of psychiatric or neurological disorders.