Cerebrospinal fluid flow closely tracks behavioral performance during an attention task

2571 

Abstract Submission 

Zinong Yang¹, Stephanie Williams², Stephanie Anakwe³, Emilia Schimmelpfennig¹, Laura Lewis⁴

¹Boston University, bosotn, MA, ²Boston University, Boston, MA, ³Boston University, Bosotn, MA, ⁴Massachusetts Institute of Technology, Cambridge, MA

Zinong Yang  
Boston University
bosotn, MA

Stephanie Williams  
Boston University
Boston, MA

Stephanie Anakwe  
Boston University
Bosotn, MA

Emilia Schimmelpfennig  
Boston University
bosotn, MA

Laura Lewis  
Massachusetts Institute of Technology
Cambridge, MA

Sufficient sleep plays an essential role in maintaining optimal cognitive function and brain health in humans. A single night of lost sleep causes substantial negative cognitive consequences, such as attentional lapses[1]. Recent studies suggest that sleep is the critical stage for cleaning brain neurotoxic waste products that accumulate during wake periods via the glymphatic system, in which metabolites transit through the cerebrospinal fluid (CSF)[2]. A key mechanism driving CSF and clearance is the pulsation of blood vessels in the brain, and sleep enhances this vascular pulsatility and in turn increases CSF flow. However, despite the critical role of sleep and CSF flow in brain health, it remains unclear whether CSF dynamics change after sleep deprivation (SD), and how CSF activity correlates with SD-induced cognitive impairment. Here we characterized how large-scale brain dynamics are altered after acute SD. We first tested the effects of SD on global brain dynamics and CSF activity during wakefulness. We then assessed the correlation between CSF activity and attentional states fluctuation assessed by changes in reaction times, using combined EEG and fast fMRI to simultaneously measure CSF flow.

To study the effects of SD on CSF dynamics and cognitive deficits, we imaged 26 healthy adults. Each subject was scanned in morning MRI sessions, one after typical sleep and one after SD, in counterbalanced order. During the SD session, subjects underwent a 24-hour period of SD under continuous monitoring of study staff, with EEG performed every 3 hours, followed by EEG-fMRI-Eyetracking during wakefulness in the morning. Functional images were acquired using fast BOLD-fMRI (TR=378ms, 2.5 mm isotropic). Accelerated fMRI paradigms contain flow-related enhancement signals in which high-velocity fluids produce bright signals in the outermost image slices[3]. Our acquisition paradigm exploited these signals and has been successfully used previously to measure BOLD dynamics and CSF flow[4,5]. Subjects underwent 4 runs of the psychomotor vigilance task (PVT) and 1 resting-state. No global signal regression nor ventricle regression was used. The fourth ventricle was identified anatomically as a region of interest (ROI) and was extracted to measure CSF flow signals generated by flow-related enhancement.

Consistent with prior studies, spectral power analysis demonstrated that the EEG slow-delta frequency (SWA, 0.5-4 Hz) power increased after SD. We calculated the power spectrum density of the CSF signal at each sleep stage (wake, or stages N1/2/3 of NREM sleep) to test whether SD induces increased CSF flow at ultra-low frequencies (<0.1 Hz) at awake state. We observed a large oscillation in the CSF signal centered around 0.05 Hz during wakefulness after SD, similar to what has been reported during typical rested NREM sleep[4]. We further found a positive correlation between CSF peaks and reaction times during PVT both at baseline and after SD (p<0.05, permutation test). We analyzed the SWA, global cortex BOLD and pupil size relative to the peak of the CSF waves and found that neural, BOLD, pupil size and CSF activity were coupled (1736 peaks identified). Cross-correlation analysis (n=26) indicated that peaks in EEG SWA preceded CSF inflow by 6-8s, suggesting that even during wakefulness a transient local increase in slow wave activity is predictive of CSF inflow.

We found that the increased slow wave activity after total sleep deprivation is linked to large-scale CSF inflow through the fourth ventricle during wakefulness. The coupling between CSF inflow and neuronal activity can be detected in drowsy human subjects during behavioral tasks. Sleep deprivation is detrimental to cognitive ability, and cognitive lapses come at a high cost for organismal function. Our results suggest that SD-induced cognitive impairments reflect moments of increased cerebrospinal fluid flow, due to the need for sleep-like restorative physiology after SD.

BOLD fMRI

EEG

Sleep and Wakefulness ¹

Perception and Attention Other ²

Cerebral Metabolism and Hemodynamics

Cerebro Spinal Fluid (CSF)

Cognition

Computational Neuroscience

Electroencephaolography (EEG)

FUNCTIONAL MRI

MRI

Sleep