Hemodynamic signals in the brain are used as surrogates of neural activity, but how these hemodynamic signals depend on arousal state is poorly understood. Here, we monitored neural activity and hemodynamic signals in un-anesthetized, head-fixed mice to understand how sleep and awake states impact cerebral hemodynamics. In parallel with electrophysiological recordings, we used intrinsic optical signal imaging to measure bilateral changes in cerebral hemoglobin ([HbT]), and two-photon laser scanning microscopy (2PLSM) to measure dilations of individual arterioles. We concurrently monitored body motion, whisker movement, muscle EMG, cortical LFP, and hippocampal LFP to classify the arousal state of the mouse into awake, NREM sleep, or REM sleep. We found that mice invariably fell asleep during imaging, and these sleep states were interspersed with periods of awake. During both NREM and REM sleep, mice showed large increases in [HbT] relative to the awake state, showing increase in hemoglobin and arteriole diameter two to five times larger than those seen in response to sensory stimulation. During NREM sleep, the amplitude of bilateral low-frequency oscillations in [HbT] increased markedly, and coherency between neural activity and hemodynamic signals was higher than the awake resting and REM states. Bilateral correlations in neural activity and [HbT] were highest during NREM sleep, and lowest in the awake state. Our results show that hemodynamic signals in the cortex are strongly modulated by arousal state, with hemodynamic changes during sleep being substantially larger than sensory-evoked responses. These results underscore the critical importance of behavioral monitoring during studies of spontaneous activity, as sleep-related hemodynamics dominate measures of neurovascular coupling and functional connectivity.

The AnalysisResults.mat file contails all the data required to reproduce the figures in Neurovascular coupling and bilateral connectivity during NREM and REM sleep by K.L. Turner, K.W. Gheres, E.A. Proctor and P.J. Drew.

To reproduce the figures download and run the Matlab code available from the below GitHub repository:
https://github.com/DrewLab/Turner_Gheres_Proctor_Drew_eLife2020

Add the AnalysisResults.mat file to the MATLAB file path by dragging it into the folder containing the code. Open the MATLAB function MainScript_Manuscript2020.m and run.

Software/System Requirements: Code was tested with MATLAB 2019b. Running MainScript_Manuscript2020.m took < 5 minutes to run on a 2018 Macbook Pro (2.6 Ghz 6-Core Intel i7 with 16 Gb 2400 MHz DDR4 RAM, Radeon Pro 560X GPU).

If you would like to automatically save the MATLAB figures and statistical read-outs, change line 29 of MainScript_Manuscript2020.m to saveFigs = 'y'; This will increase the analysis time to create a new folder /Summary Figures and Structures/MATLAB Analysis Figures/.