Integrated models of neurovascular coupling and BOLD signals: Responses for varying neural activations

ABSTRACT A state‐of‐the‐art integrated model of neurovascular coupling (NVC) (Dormanns et al., 2015b; Dormanns et al., 2016; Kenny et al., 2018) and the BOLD response (Mathias et al., 2017a; Mathias et al., 2017b) is presented with the ability to simulate the fMRI BOLD responses due to continuous neuronal spiking, bursting and cortical spreading depression (CSD) along with the underlying complex vascular coupling. Simulated BOLD responses are compared to experimental BOLD signals observed in the rat barrel cortex and in the hippocampus under seizure conditions showing good agreement. Bursting phenomena provides relatively clear BOLD signals as long as the time between bursts is not too short. For short burst periods the BOLD signal remains constant even though the neuron is in a predominantly bursting mode. Simulation of CSD exhibits large negative BOLD signals. Visco‐elastic effects of the capillary bed do not seem to have a large effect on the BOLD signal even for relatively high values of oxygen consumption. While the results of the model suggests that potassium ions released during neural activity could act as the main mediator in NVC, it suggests the possibility of other mechanisms that can coexist and increase blood flow such as the arachidonic acid to epoxyeicosatrienoic acid (EET) pathway. The comparison with experimental cerebral blood flow (CBF) data indicates the possible existence of multiple neural pathways influencing the vascular response. Initial negative BOLD signals occur for all simulations due to the rate at which the metabolic oxygen consumption occurs relative to the dilation of the perfusing cerebro‐vasculature. However it is unclear as to whether these are normally seen clinically due to the size of the magnetic field. Experimental comparisons for different animal experiments may very well require variation in the model parameters. The complex integrated model is believed to be the first of its kind to simulate both NVC and the resulting BOLD signal. HIGHLIGHTSA complex neurovascular unit model is integrated with the BOLD response.Simulations are compared to experimental BOLD signals in the cortex and hippocampus.Simulations of cortical spreading depression exhibit large negative BOLD signals.Two neural pathways are hypothesised to simulate experimental blood ow profiles.