Presented:
INTERNATIONAL UNION OF PHYSIOLOGICAL SCIENCES (IUPS) Meeting 2013
Birmingham, UK, 21-26 July, 2013

Abstract

A transient increase in cortical neuronal activity is followed by a complex sequence of cellular, metabolic, and vascular processes resulting in a transient increase in cerebral blood flow occurring with a slower time-course. This physiological neurovascular coupling is used to infer neuronal activity in vascular-based modern functional brain imaging techniques. Nevertheless, the process of neurovascular coupling itself can become altered with aging and disease and the consequences of such alterations to interpretation of imaging studies remain poorly understood. It thus becomes increasingly important to develop methods to evaluate the process of neurovascular coupling at the whole brain level.
The current methodological approach is to investigate the mean hemodynamic changes in response to repeating somatosensory stimuli that generate identifiable cortical evoked responses that can be recorded non-invasively by electroencephalography (EEG). Here we proposed a novel approach. During general anesthesia, with increasing anesthetic concentrations the EEG becomes discontinuous consisting of whole-brain bursts of activity on a suppressed background referred to as burst-suppression (BS). We hypothesized that neurovascular coupling could be investigated by burst-triggered averaging of the hemodynamic changes during BS.
We investigated the fluctuations in cerebral blood flow during BS induced by chloral hydrate overdose in adult male Wistar rats. The EEG recorded via fronto-occipital epidural electrodes (1-30Hz) was first rectified (rEEG) and then averaged in 200 ms bins. The corresponding binary BS signal, derived by an unsupervised statistical classifier, was then used to identify the bursts onset. The laser Doppler (LD) signal recorded from the frontal cortex was first filtered for cardiac cycle and respiratory fluctuations and then averaged in 200 ms bins. We could then construct a burst-onset triggered average of the change in rEEG and LD from the levels just prior to the burst. We found that following burst onset there was an increase in LD that peaked within 1 second after the rEEG peak, consistent with a neurovascular response. The magnitude of the response was several orders of magnitude smaller than the cardiac LD fluctuation so that averaging of more than 100 bursts was required to construct a reliable response. At a readily controllable burst suppression of 50-75% ratio (the relative time spent in suppression), a 10 minutes recording was typically sufficient.
Our study provides proof of concept that whole-brain neurovascular coupling can be evaluated during spontaneous cerebral activity by inducing a safely reversible anesthetic BS state. This should be accounted for in the design of clinical neurovascular coupling studies combining non-invasive methods such as EEG and near-infrared spectroscopy.