Detecting rs-fMRI Networks in Disorders of Consciousness: Improving Clinical Interpretability

Abstract

Background

Preserved resting-state functional MRI (rs-fMRI) networks are typically observed in Disorders of Consciousness (DOC). Despite the widespread use of rs-fMRI in DOC, a systematic assessment of networks is needed to improve the interpretability of data in clinical practice. We investigated functional connectivity of the main networks, combining structural MRI to obtain a description of the most observed networks in DOC, their diagnostic ability, and whether they can be related to clinical assessment.

Methods

A group of 109 chronic patients [65 vegetative state/unresponsive wakefulness state (VS/UWS), 34 minimally conscious state (MCS), and 10 emerged from MCS (eMCS)], with different etiologies, and 34 control subjects underwent multimodal assessment. Rs-fMRI data were analyzed with a semi-automatic pipeline to assess residual functional activity in terms of number, type, mean intensity, and structural preservation of networks.

Results

The more networks observed, the better the patient's clinical condition is likely to be. VS/UWS patients display 0–9, MCS 5–9, and eMCS 8–10 networks. Both the presence and intensity of 5 networks (visual networks, temporal, left fronto-parietal and default mode network) are relevant to distinguish VS/UWS from MCS, with AUCs of 0.64–0.69 (95% confidence interval). Etiology and disease duration have an impact on the number and type of preserved networks. High residual functional connectivity observed in VS/UWS patients, as in MCS, is in agreement with neurophysiological and metabolic evaluations.

Conclusions

This systematic assessment of the main rs-fMRI networks in DOC provides basic measures of functional connectivity that can enhance their interpretability in clinical practice.