CONN

SOFTWARE DESCRIPTION

CONN is an open-source, SPM-based, cross-platform software for the computation, display and analysis of functional connectivity from brain fMRI sequences (functional Magnetic Resonance Imaging). It supports analyses of both resting state (rsfMRI) as well as task designs. 

CONN's methods and workflows are designed to enhance ease of use, transparency and reproducibility in functional connectivity research. Key features include:

• • Data Import: Automatically integration of fMRI datasets in standardized formats (e.g. BIDS) and direct download from publicly available repositories (e.g. FCP/INDI). Flexible options to import raw data (e.g. DICOM, ANALYZE, NIfTI) or data partially or fully preprocessed with other tools (e.g. SPM, FSL, AFNI, fMRIPrep). 

  • Preprocessing Pipelines: Reproducible pipelines for preprocessing functional and anatomical images, including susceptibility distortion correction, motion correction/realignment, slice-timing correction, outlier identification, coregistration, tissue-class segmentation, MNI-normalization, and smoothing.

  • Denoising: Strict control of residual physiological and subject motion artifacts using a combination of scrubbing, motion and task regression, anatomical CompCor, and band-pass filtering.  

  • Quality Control: Integrated Quality Control procedures and measures to independently evaluate the quality of the BOLD signal before and after denoising, minimizing biases in functional connectivity measures, and enhancing transparency and replicability of functional connectivity research.

  • Connectivity Analyses: A wide array of functional connectivity measures, including Seed-Based Correlations (SBC), ROI-to-ROI connectivity matrices (RRC), graph theoretical analyses, generalized Psycho-Physiological Interaction models (gPPI), Independent Component Analyses (group-ICA), masked ICA, Amplitude of Low-Frequency Fluctuations (ALFF), Intrinsic and Global Connectivity (ICC & GCOR), Local Correlations (LCOR), Inter-Hemispheric Correlations (IHC), functional connectivity Multivariate Pattern Analyses (fc-MVPA), and dynamic connectivity analyses (dyn-ICA, and sliding-window analyses).

  • Group-Level Analyses: Support for group- and population-level inferences and models, including ANOVA, multiple regression, repeated measures and mixed within- and between-subject designs. Family-wise false positive control using a variety of state-of-the-art parametric (e.g. Random Field Theory, Functional Network Connectivity) and non-parametric techniques (e.g. permutation/randomization analyses, Threshold Free Cluster Enhancement).

CONN can be operated through a user-friendly GUI or via batch scripts/commands. In high-performance computing (HPC) environments, it interacts automatically with most workload manager protocols (e.g. Slurm, Grid Engine, PBS/Torque, LSF), enabling the processing of large datasets in parallel with minimal user intervention. 

Developed in Matlab, CONN is distributed both as Matlab source code and as a pre-compiled executable (standalone release), requiring no Matlab installations or licenses.

For a detailed description of all processing and analysis methods available in CONN, please refer to the fMRI methods section.

SOFTWARE LICENSE

CONN software is distributed free of charge under the MIT license, allowing unrestricted use, modification, and distribution. 

If you find CONN beneficial to your research, please consider supporting its development through the following avenues: 

• Training: Enroll in our training courses to deepen your understanding of CONN's methods and workflows. Proceeds from these courses directly fund ongoing development. 

• Contribute: Visit CONN's GitHub repository to assist in development, testing, or documentation efforts.

• Community engagement: Share your expertise by participating in CONN's NITRC forum, answering questions from other CONN users, or simply help us spread the word about CONN within your network!