Subcortical neuromodulation of arousal state transitions in the human brain

2565 

Abstract Submission 

Nicholas Cicero¹, Beverly Setzer², Daniel Gomez³, Ewa Beldzik³, Makaila Banks², Juan Eugenio Iglesias⁴, Giorgio Bonmassar⁴, Brian Edlow⁴, Laura Lewis³

¹Boston University, Allston, MA, ²Boston University, Boston, MA, ³Massacusetts Institute of Technology, Cambridge, MA, ⁴Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, MA

Nicholas Cicero  
Boston University
Allston, MA

Beverly Setzer  
Boston University
Boston, MA

Daniel Gomez, Ph.D.  
Massacusetts Institute of Technology
Cambridge, MA

Ewa Beldzik, Ph.D.  
Massacusetts Institute of Technology
Cambridge, MA

Makaila Banks  
Boston University
Boston, MA

Juan Eugenio Iglesias, Ph.D.  
Athinoula A. Martinos Center for Biomedical Imaging
Charlestown, MA

Giorgio Bonmassar, Ph.D.  
Athinoula A. Martinos Center for Biomedical Imaging
Charlestown, MA

Brian Edlow, M.D.  
Athinoula A. Martinos Center for Biomedical Imaging
Charlestown, MA

Laura Lewis, Ph.D.  
Massacusetts Institute of Technology
Cambridge, MA

During sleep, our brains undergo a transformation in state, marked by profound changes in cognition and behavior (1). During a typical night of sleep, the brain cycles through stereotyped stages of sleep (2-9), including many spontaneous, brief arousals from sleep. Emerging evidence indicates that several subcortical neuromodulatory nuclei, known as the ascending arousal network (AAN), generate and maintain this sleep architecture (2,3,6). Individual nuclei within the AAN have been demonstrated in rodents to initiate arousal state transitions; however, it is unknown how the diverse components of this network operate together to generate and sustain sleep structure, and the small size of these brain regions has historically limited studies of the AAN in humans. To study the underlying neural mechanisms controlling sleep-wake state structure in humans, we performed simultaneous acquisition of electroencephalography (EEG) and ultra-high field functional magnetic resonance imaging (fMRI) to identify subcortical and cortical dynamics across sleep-wake state transitions.

Data were collected from nine healthy volunteers. Participants arrived at the lab around their typical bedtime and underwent simultaneous EEG-fMRI at 7 Tesla. After an MPRAGE scan (0.75mm isotropic), we performed a resting state fMRI scan using high spatial resolution sagittal partial-brain EPI acquisition covering the brainstem and midline cortex (voxel size=1.2mm isotropic; TR=0.98s; SMS=2, GRAPPA=4) (Fig 1A). Participants performed a self-paced button press task to indicate their behavioral state.
Anatomical segmentation of the AAN was performed within each individual to localize nuclei with high spatial precision (Fig 1B). To delineate arousal dynamics within regions of the AAN, we extracted BOLD activity in each region of interest (ROI) at the moment of behavioral arousals, defined as the first button press after at least 20 seconds without a button press (Fig 2A). We next separated behavioral arousals into two types: sustained (at least 5 button presses following arousals) and transient (two or fewer button presses). We also separated behavioral arousals by the duration of the unresponsive period preceding the arousal. Significant differences between each ROI's fMRI signals in different arousal types were identified by computing a paired t-test across 1-s bins and the p-value was Bonferroni corrected for the number of time bins and ROIs tested.

Analysis of fMRI dynamics at the moment of behavioral arousal revealed that all regions of the AAN activated prior to behavioral arousals (Fig 2B). We next tested whether specific regions were linked to subsequent maintenance of the arousal state following a sleep-to-wake transition. We found that select subcortical regions, namely the dorsal raphe, periaqueductal gray, mesencephalic reticular formation, ventral tegmental area, lateral hypothalamus, and thalamus, were more strongly activated following sustained compared to transient arousals (Bonferroni corrected paired t-test, p<0.05) (Fig 2C). The dynamics within specific subcortical regions also reflected the arousal state prior to behavioral arousal, with the locus coeruleus, pedunculopontine tegmentum, and lateral hypothalamus displaying significantly greater activation (Bonferroni corrected paired t-test, p<0.05) when subjects were emerging from a longer unresponsive period.

Using simultaneous EEG-fMRI at 7 Tesla, we identify distinct patterns of subcortical dynamics that reflect the moments before, during, and after arousal state transitions. We find that all regions of the AAN activate at the moment of a behavioral arousal state transition, but specific subsets of the AAN display activity are linked to the preceding depth and subsequent stability of an arousal state transition. This work identifies unified and unique properties of the AAN in humans and demonstrates that distinct dynamics within the AAN reflect different aspects of transitions between brain states.

BOLD fMRI ²

EEG

Sleep and Wakefulness ¹

Brainstem

Electroencephaolography (EEG)

FUNCTIONAL MRI

Sleep

Sub-Cortical