Abstracts

Rationale: Small case series demonstrated the feasibility of task-based EEG and f-MRI to identify patients with cognitive motor dissociation (CMD). This state is characterized by impaired ability to express preserved or recovered command following using EEG or MRI linked command following tasks. The prevalence of patients with CMD and prognostic significance in the ICU context remain uncertain. We aimed at quantifying the prevalence and significance of CMD on a large patient sample. Methods: We prospectively recorded EEG in a large consecutive series of unconscious, acutely brain injured ICU patients while presenting repeated auditory commands, asking them to open and close their hand. Serial Coma Recovery Scale-Revised (CRS-R) scores were used to categorize patients into coma, unresponsive wakefulness, minimally conscious minus and plus, and conscious states. Power spectral density algorithms were applied to determine the accuracy to distinguish between the active (hand opening-closing) and passive (rest) task. Following methodology applied in prior reports, we considered unresponsive patients as being in CMD when significance remained after 500 random permutations (p-value Results: From 119 patients (age 60+/-17 years; 47% female; 34% intracerebral hemorrhage, 29% cardiac arrest, 15% subarachnoid hemorrhage, 12% subdural hematoma/traumatic brain injury, 10% other), we recorded a total of 270 artifact-free task-based EEG sessions. Task-based EEG command following was identified in 31%(N=83) of the recordings; 27%(N=34) of EEG recordings in a comatose, 28%(N=15) unresponsive wakefulness, 36%(N=22) minimally conscious minus, 47%(N=8) minimally conscious plus, and 40%(N=4) of conscious states. Of 98 unresponsive patients, 43 were considered in CMD on at least one recording. CMD was associated with GOS-E at 12 months (F[99,6]=4.1; p-value=0.001). Conclusions: Task-based EEG paradigms classify almost half of unresponsive acutely brain injured patients as being in CMD. It is uncertain if the detected signal truly represents conscious comprehension of the auditory instruction. However, detection of CMD acutely after brain injury using currently applied methodology was associated with good long term functional recovery. It remains to be shown if stricter cut offs for command following may further improve the clinical relevance. Funding: DANA Foundation, CTSA award, McDonnell Foundation, National Library of Medicine/NIH