The R-fMRI Maps Project

Submitted by YAN Chao-Gan on

Dear colleagues,

We are pleased to announce the launch of the R-fMRI Maps project.

The aim of the R-fMRI Maps project is to build a big data of intrinsic brain activity indices, which has the potential to allow us addressing critical questions about the brain. Here we share a broad array of the R-fMRI indices of open R-fMRI data (through a standard processing pipeline built in DPABI/DPARSF), and encourage researchers share their processed R-fMRI indices to public through the R-fMRI Maps project.

Recently, data-sharing initiatives (e.g., FCP/INDI, openfMRI, fMRIDC from the grassroots and ADNI, HCP, NDAR, PING from coordinated) enabled big data research model in the brain imaging community. However, raw data sharing requires intensive coordinating efforts, huge manpower demand and large-amount data storing/management facilities. Furthermore, sharing raw data is mired with privacy concerns arising from possibility of being able to identify participants from high dimensional raw data. These concerns, together with the demands of data organization and the limit of large data uploading, prevents a wider imaging community to share their valuable brain imaging datasets to public.

The R-fMRI Maps project was designed to address the above concerns by only sharing the final R-fMRI indices, which only needs light data storing/uploading requirements and removed the privacy concerns on raw data. The project provides a convenient data organizer GUI integrated in DPABI/DPARSF to facilitate efficient data organization. Furthermore, by sharing the processed R-fMRI indices, the projects removed the barriers of computational resources as well as analytic knowledge for the users, thus allows a wider scientific community (especially for machine learning experts) to join in the endeavor of understanding the brain.

With the R-fMRI Maps project, we hope to build an unprecedented big data of brain imaging across a wide variety of individuals: including different neurological and psychiatric disease, as well as healthy people with different traits. We hope such a big data could help to address the need for a neuroscience-based classification approach for parsing a large variety psychiatric illness, as called by the NIMH RDoC Project. We hope more exciting neuroscience findings could be fostered by such a project, and future innovative model could be inspired by it.

We sincerely hope you could join us, either downloading and utilizing the shared data, or sharing your data with this project. If you have further questions, please refer to to discuss.



Chao-Gan YAN, Ph.D.

Principal Investigator

Deputy Director, MRI Research Center

Institute of Psychology, Chinese Academy of Sciences

16 Lincui Road, Chaoyang District, Beijing 100101, China





P.S., Please see below for a list of R-fMRI indices shared.

1.     Amplitude of low frequency fluctuations (ALFF)

2.     Fractional ALFF (fALFF)

3.     Regional Homogeneity (ReHo)

4.     Voxel-mirrored homotopic connectivity (VMHC)

5.     Degree Centrality (DC)

6.     Functional Connectivity Matrices of

a.     Automated Anatomical Labeling (AAL) atlas

b.     Harvard-Oxford atlas

c.     Craddock’s clustering 200 ROIs

d.     Zalesky’s random parcelations

e.     Dosenbach’s 160 functional ROIs

f.     Global signal

g.     Power's 264 functional ROIs

In addition, the gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) density and volume are shared as well.

Please click here for the details of the data structure and processing steps.

Hi Xiangzhen,

Please see here:

6.     ROISignals: The ROI time courses of predefined ROIs.

1~116: Automated Anatomical Labeling (AAL) atlas (Tzourio-Mazoyer et al., 2002)

117~212: Harvard-Oxford atlas (Kennedy et al., 1998)– cortical areas

213~228: Harvard-Oxford atlas (Kennedy et al., 1998)– subcortical areas

229~428: Craddock’s clustering 200 ROIs (Craddock et al., 2012)

429~1408: Zalesky’s random parcelations (compact version: 980 ROIs) (Zalesky et al., 2010)

1409~1568: Dosenbach’s 160 functional ROIs (Dosenbach et al., 2010)




Hi Chao-Gan,

Thanks for your reply. 

I got it. A further question. As I understand, the first time course for AAL atlas is for the region with a lable of 1. The first time couse for the Harvard-Oxford atlas-cortical areas (i.e., the 117th time course in ROISignals) is for the region with a label of 1. Are these correct? 

In addition, would you like to share the atlas (in nifti format) used in these processing online? I find that it's difficult to get some atlas, for example the Zalesky’s random parcelations. 



Hi Xiangzhen,

1. You are right.

2. Please find them under DPABI/Templates (e.g., Zalesky’s random parcelations, with his permission).

You can click "Define ROI" in DPARSFA to see where they are.





Tue, 07/09/2024 - 09:01

Dear Respected Recipients,
Thank you for this wonderful dataset. I am new in this field, and I have some questions about the MDD dataset:
-> After downloading the dataset and from the "ROISignals_FunImgARCWF" folder, we can work on the ROI signals, create FC data and use the graph neural network for ML classification based works.
But my question is, is the original rsfMRI data supposed to be 4D (please correct me if I am mistaken)? However, the current one is a signal/time series data. Can you please explain why is that? Any suggestions will be highly appreciated.