Functional Connectivity Analysis in Neuroimaging Systems

Functional connectivity analysis examines statistical dependencies between spatially distinct brain regions, revealing how neural populations interact to support cognition and behaviour. In both task-based and resting-state paradigms, signals derived from modalities such as functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG) and intracranial electrophysiology are processed to estimate temporal correlations, coherence measures or causal influences. Key steps include signal pre-processing, artefact removal, spatial parcellation into regions of interest and the application of connectivity metrics drawn from network science. Advances in high-resolution acquisition and de-noising approaches have increased sensitivity to genuine neuronal interactions, while methodological standardisation and rigorous comparison of pipelines have improved reproducibility. Functional connectivity has provided insights into the architecture of canonical brain networks—such as the default mode, salience and sensorimotor systems—highlighting frequency-dependent interactions and dynamic reconfiguration in health and disease. Global applications range from biomarker development in neuropsychiatric disorders to guiding neuromodulation therapies.

Research from Nature Portfolio

Recent studies have systematically evaluated hundreds of data-processing pipelines for resting-state fMRI connectomics, assessing the effects of brain parcellation, connectivity definition and global signal regression on motion confounds and test-retest reliability. This work revealed that a majority of pipelines fail to meet rigorous criteria for sensitivity to inter-subject differences and experimental manipulations, while a subset of optimised workflows consistently satisfy reproducibility and sensitivity across datasets collected over minutes, weeks and months. The findings provide a comprehensive performance atlas to inform future best practices in large-scale functional connectome studies.

Functional Connectivity Analysis in Neuroimaging Systems publication trend

The graph below shows the total number of articles in functional connectivity analysis in neuroimaging systems across all publications each year (not limited to Nature Index journals).

Technical terms

Functional connectivity: Statistical dependencies between activity time series of distinct brain regions, used to infer network interactions.

Resting-state fMRI: Acquisition of brain activity while subjects are not performing an explicit task, enabling study of intrinsic connectivity networks.

Global signal regression: A preprocessing step removing the average whole-brain signal to reduce widespread artefacts, at the risk of introducing negative correlations.

Independent component analysis (ICA): A blind source separation technique that decomposes signals into spatially independent maps and time courses, used for artefact removal.

Default mode network (DMN): A set of brain regions that exhibit higher activity during rest and are implicated in self-referential and introspective processes.

References

1. Systematic evaluation of fMRI data-processing pipelines for consistent functional connectomics. Nature Communications (2024).
2. Neuronal dynamics of the default mode network and anterior insular cortex: Intrinsic properties and modulation by salient stimuli. Science Advances (2023).
3. A Tutorial Review of Functional Connectivity Analysis Methods and Their Interpretational Pitfalls. Frontiers in Systems Neuroscience (2016).
4. How reliable are MEG resting-state connectivity metrics?. NeuroImage (2016).