Mapping the subcortical connectivity of the human default mode network

Jian Li, William H. Curley, Bastien Guérin, Darin D. Dougherty, Adrian V. Dalca, Bruce Fischl, Andreas Horn, Brian L. Edlow

TL;DR

This paper maps the subcortical connections of the human Default Mode Network (DMN) using advanced fMRI analysis techniques. This detailed map is crucial for understanding consciousness and identifying precise targets for therapeutic interventions in patients with disorders of consciousness.

What Problem Does This Paper Address?

The Default Mode Network (DMN) is fundamental to self-awareness and introspection, playing a key role in human consciousness. While therapies to restore consciousness in patients with severe brain injuries often target subcortical regions, the precise subcortical connectivity of the DMN remains poorly understood. This lack of detailed mapping hinders the identification of optimal targets for neuromodulation therapies, such as deep brain stimulation (DBS) or focused ultrasound. The challenge lies in the low signal-to-noise ratio (SNR) and spatial resolution of fMRI data in subcortical areas, making it difficult to accurately map functional connectivity.

Methods

The authors developed and employed a novel tensor decomposition method called Nadam-Accelerated SCAlable and Robust (NASCAR). This method allows for the joint analysis of cortical and subcortical fMRI data, creating a more comprehensive DMN map. Key methodological innovations include:

  • Inter-subject Temporal Synchronization (BrainSync): To overcome the asynchronous nature of resting-state fMRI data across subjects, BrainSync was used to align data in a spatiotemporal manner, a crucial assumption for the tensor model.
  • NASCAR Tensor Decomposition: This algorithm models brain networks as low-rank components, enabling the identification of the DMN and its subcortical connections from high-dimensional fMRI data. It uses a "warm start" approach for robustness and scalability.
  • High-Resolution Neuroanatomic Overlay: The resulting functional connectivity maps were overlaid onto a high-resolution (100 µm) ex vivo MRI dataset for precise anatomical localization within brainstem, thalamus, hypothalamus, basal forebrain, and basal ganglia.
  • Region-of-Interest (ROI)-Based Analysis: Standard atlases (FreeSurfer, PTN, AAN, BF/HT) were used to segment specific subcortical nuclei, allowing for quantitative analysis of DMN connectivity within these regions.
  • Comparison with Seed-Based Analysis: The NASCAR results were compared with traditional seed-based correlation analyses to highlight differences in connectivity patterns and interpretation, particularly concerning global signal regression.

Datasets

  • Human Connectome Project (HCP) 7 Tesla (7T) resting-state fMRI data: Data from 168 healthy volunteers were used. The 7T data was chosen for its superior SNR, especially in subcortical regions.
  • 7T 100 µm ex vivo MRI dataset: This high-resolution structural dataset was used for precise neuroanatomic mapping of the functional connectivity results.
  • Human Brainstem Specimen: Used for immunostaining (tyrosine hydroxylase) to validate functional connectivity findings, particularly for the ventral tegmental area (VTA).

Key Insights and Ideas

  • The DMN extends into subcortical regions: The study provides strong evidence that the DMN is not purely a cortical network but has significant subcortical components.
  • Central Lateral Nucleus (CL) of the thalamus and Ventral Tegmental Area (VTA) of the midbrain are key subcortical DMN hubs: These regions showed particularly strong positive correlations with the cortical DMN, supporting their hypothesized role in consciousness modulation.
  • Anti-correlations reveal mesocircuit dynamics: Negative correlations between the DMN and regions like the putamen and globus pallidus provide fMRI evidence for the "mesocircuit hypothesis" of consciousness, suggesting an inhibitory role in modulating cortical activity.
  • NASCAR offers advantages over traditional seed-based methods: The NASCAR approach effectively decouples the "global physiological" network, allowing for clearer visualization of negative correlations and avoiding the ambiguity introduced by global signal regression in seed-based analyses.
  • High-resolution fMRI is critical for subcortical mapping: The study highlights the necessity of high-field (7T) and high-resolution imaging for accurately characterizing subcortical functional connectivity.

Key Results

  • The study successfully created a comprehensive map of subcortical DMN connectivity.
  • The Central Lateral Nucleus (CL) of the thalamus showed significantly higher functional connectivity with the cortical DMN than the average thalamic signal, with some voxels exhibiting the strongest correlation to the DMN.
  • The Ventral Tegmental Area (VTA) of the midbrain also demonstrated significantly higher DMN functional connectivity than the average brainstem signal.
  • Subcortical regions showing strong positive correlations with the DMN included the thalamus, caudate, and brainstem, reaching approximately 30% of the cortical DMN signal strength.
  • The putamen and globus pallidus exhibited strong negative correlations (anti-correlations) with the DMN.
  • The NASCAR method showed better spatial alignment with dopamine immunostaining data compared to seed-based analyses.
  • The reproducibility analysis showed a high correlation (0.987) between spatial maps from independent datasets, indicating the robustness of the NASCAR pipeline.

Limitations and Future Work

  • Spatial resolution limitations: Despite using 7T data, inferences at the voxel level, especially near boundaries of subcortical structures, are cautioned against due to residual spatial misalignment and the inherent resolution of fMRI.
  • SNR in subcortical regions: The SNR in subcortical regions is significantly lower than in the cortex, making it challenging to identify functional connectivity. The study's findings may not be directly reproducible with lower SNR datasets (e.g., 3T).
  • Post-hoc inspection of components: Like other data-driven methods, manual inspection of NASCAR components is necessary. The authors suggest exploring other identified components as a future direction.
  • Hypothesis-generating findings: Exploratory analyses identified additional nuclei with DMN connections, which require validation in future studies.
  • Future directions: The authors plan to apply these methods to 3T data and to patients with disorders of consciousness to assess clinical translation. Further research is needed to validate the biomarkers and explore the precise physiological mechanisms of these subcortico-cortical connections.

Summary updated

Abstract

Abstract The default mode network (DMN) mediates self-awareness and introspection, core components of human consciousness. Therapies to restore consciousness in patients with severe brain injuries have historically targeted subcortical sites in the brainstem, thalamus, hypothalamus, basal forebrain, and basal ganglia, with the goal of reactivating cortical DMN nodes. However, the subcortical connectivity of the DMN has not been fully mapped and optimal subcortical targets for therapeutic neuromodulation of consciousness have not been identified. In this work, we created a comprehensive map of DMN subcortical connectivity by combining high-resolution functional and structural datasets with advanced signal processing methods. We analyzed 7 Tesla resting-state functional MRI (rs-fMRI) data from 168 healthy volunteers acquired in the Human Connectome Project. The rs-fMRI blood-oxygen-level-dependent (BOLD) data were temporally synchronized across subjects using the BrainSync algorithm. Cortical and subcortical DMN nodes were jointly analyzed and identified at the group level by applying a novel Nadam-Accelerated SCAlable and Robust (NASCAR) tensor decomposition method to the synchronized dataset. The subcortical connectivity map was then overlaid on a 7 Tesla 100 micron ex vivo MRI dataset for neuroanatomic analysis using automated segmentation of nuclei within the brainstem, thalamus, hypothalamus, basal forebrain, and basal ganglia. We further compared the NASCAR subcortical connectivity map with its counterpart generated from canonical seed-based correlation analyses. The NASCAR method revealed that BOLD signal in the central lateral nucleus of the thalamus and ventral tegmental area of the midbrain is strongly correlated with that of the DMN. In an exploratory analysis, additional subcortical sites in the median and dorsal raphe, lateral hypothalamus, and caudate nuclei were correlated with the cortical DMN. We also found that the putamen and globus pallidus are negatively correlated (i.e., anti-correlated) with the DMN, providing rs-fMRI evidence for the mesocircuit hypothesis of human consciousness, whereby a striatopallidal feedback system modulates anterior forebrain function via disinhibition of the central thalamus. Seed-based analyses yielded similar subcortical DMN connectivity, but the NASCAR result showed stronger contrast and better spatial alignment with dopamine immunostaining data. The DMN subcortical connectivity map identified here advances understanding of the subcortical regions that contribute to human consciousness and can be used to inform the selection of therapeutic targets in clinical trials for patients with disorders of consciousness.
BibTeX
@inproceedings{li2021mapping,
  author = {Jian Li and William H. Curley and Bastien Guérin and Darin D. Dougherty and Adrian V. Dalca and Bruce Fischl and Andreas Horn and Brian L. Edlow},
  title = {{Mapping the subcortical connectivity of the human default mode network}},
  booktitle = {NeuroImage},
  year = {2021},
  url = {https://openreview.net/forum?id=DLusgR7Qps},
}
Figure