Non-stationarity in the "resting brain's" modular architecture

Area of Science:

• Neuroscience
• Medical Imaging
• Gerontology

Background:

• Task-free functional MRI (TF-fMRI) reveals intrinsic brain connectivity but suffers from high variability.
• This variability hinders the development of reliable biomarkers for neurologic illnesses.
• Non-stationarity, or the dynamic nature of connectivity, is a potential source of this variability.

Purpose of the Study:

• To investigate if the non-stationary nature of intrinsic connectivity networks (ICNs) explains variability in TF-fMRI data.
• To determine if altered dwell times in specific brain network configurations are associated with Alzheimer's dementia.
• To explore the role of dynamic connectivity in mild cognitive impairment and at-risk individuals.

Main Methods:

• Constructed a functional brain atlas of 68 regions from a large cohort (n=892) of older adults.
• Created dynamic connectivity graphs using a sliding time window approach.
• Compared default mode network (DMN) sub-network dwell times between Alzheimer's dementia patients and cognitively normal controls.

Main Results:

• Demonstrated the non-stationary, dynamic modular organization of brain connectivity.
• Found that differences in TF-fMRI connectivity in Alzheimer's disease are explained by altered dwell times in DMN sub-networks.
• Identified potential links between DMN dwell time and findings in mild cognitive impairment and at-risk individuals.

Conclusions:

• The non-stationary nature of ICNs is a key feature, not noise, in TF-fMRI.
• Altered dwell times in specific DMN configurations represent a more sensitive marker for Alzheimer's dementia than static connectivity.
• Dynamic connectivity analysis, specifically DMN dwell time, may offer novel biomarkers for early detection and understanding of neurodegenerative diseases.