Sleep-dependent modulation of thalamic brain temperature estimates in humans

Abstract

Mammalian thermal physiology is geared towards dissipating heat during sleep. Brain temperature typically decreases during non-rapid-eye-movement (NREM) sleep in all mammalian species measured to date, yet non-invasive confirmation in healthy humans is lacking: the only available human study used invasive recordings in epileptic patients without a wake control condition. Magnetic resonance spectroscopy (MRS) enables brain temperature mapping non-invasively, using the temperature-sensitive chemical shift of water relative to N-Acetyl-Aspartate as an internal reference. First, we validated that with metabolite cycling (MC), a MRS approach that enables the simultaneous acquisition of metabolite and water information without conventional water suppression, can be used to determine estimates of brain temperature in an ex vivo sheep head. Next, we determined that by using a temporal resolution of ~4 minutes, we obtained high-quality spectra without losing excessive temporal resolution. We then re-analyzed an existing set of MRS data (acquired with the MC sequence) of the thalamus in young adults (n=9) who slept in the evening for up to 2 hours in the scanner, to assess sleep-evoked changes in estimates of brain temperature. Of the nine participants, five participants took part in the subsequent wake-only experiment, allowing to control for the potential impact of circadian time and body posture on estimates of brain temperature. Sleep-wake states were quantified with standard polysomnography, confirming that the participants managed to sleep in the scanner. On average, they spent 39.4+/-15.7 min in N2 sleep, 24.8+/-18.1 min in N3 sleep, 10.0+/-4.9 min in N1 sleep, and 37.0+/-19.6 min awake. Estimates of brain temperature were correlated across the sleep and wake-only experiment within the same participant (R^2=0.29, P<0.0001). When participants entered NREM sleep, estimates of brain temperature decreased by 0.2 degrees Celcius. Moreover, the deeper NREM sleep, the more pronounced the decrease in estimates of brain temperature (P=0.014). These data suggest that an evening nap consisting of NREM sleep decreases thalamic temperature. Considering the profound impact temperature can have on brain functioning, these brain temperature changes could be a pathway through which sleep impacts the brain.