The spatiotemporal characteristics of the hemodynamic response function (HRF) to focal brain activity is a topic of great relevance as it dictates the accuracy of functional neuroimaging techniques in mapping activation regions, establishes the ultimately achievable spatial and temporal resolution, and influences the interpretation of the data. We optimized stimulus parameters and measured the ensuing HRF, with the goal of determining the ultimate spatial domain of cerebral blood flow (CBF) control and its associated temporal evolution. This work requires brief stimuli, delivered under well-controlled conditions, to elicit minute, yet measurable vascular events, which serve as the building blocks of the integrative CBF response to more complex stimuli. It also requires better animal models, and we have ported most of our research to marmosets (Callithrix jacchus), a small New World non-human primate that we believe is especially suitable to bridge the translational animal-to-human gap, and one in which transgenic techniques are becoming available to allow us to interrogate neurons and blood vessels in interesting and novel ways. We have trained marmosets to be scanned awake and we are able to routinely obtain functional maps while the animal undergoes somatosensory, auditory, or visual stimulation. The head is restrained with a custom-built helmet that is able to hold the head still without sacrificing comfort. After undergoing training, the marmosets produce robust and reproducible fMRI responses in all senses. From somatosensory cortex, S1, S2, and caudate produced reliable fMRI responses to a single 333 s-long stimulus. The areas of activation in S1 and S2 grew from just a few voxels located in the middle and upper layers of the respective regions in response to a single pulse (333 s) stimulus, to occupying the entire cortical regions with longer stimulus durations of up to 1 s, showing that functional hyperemia is an integrative process that involves the entire functional cortical depth. The times-to-peak were longer than those obtained in anesthetized rats, but shorter than those obtained in the human visual cortex, indicating that the cortical microvascular length in marmosets may be more similar to humans than to rodents. We also examined the laminar profile of the fMRI response in the somatosensory cortex of marmosets. The earliest onset times occur in the middle layer, around 0.8mm below the cortical surface, showing that, in marmosets, the early hemodynamic changes happen in the input layers of the cortex. Like humans, marmosets are a social species that utilizes a vast array of vocalizations to communicate within their groups and with other species. We implemented an auditory stimulus paradigm that alternated presentations of pure tones with random frequencies randomly selected within a specified bandwidth and we detected activation in primary auditory cortex, as well as in the belt and parabelt areas to tones and broadband noise played in the 0.5-22 kHz range. We elicited robust bilateral responses of silence versus sound. When contrasting low frequency to high-frequency tones, tonotopic maps were obtained. To perform visual stimulation we trained marmosets to attend to images being presented in a monitor placed outside the magnet. Eye gaze was monitored and recorded sing an eye-tracking system. We trained animals to direct their gaze to images of faces, bodies and objects, and measured functional responses in occipitotemporal cortex (OT) and thalamus (LGN and Pulvinar) of awake marmosets using both implanted electrocorticography (ECoG) arrays and fMRI. Using ECoG, we found that responses within the high gamma range (50-150 Hz) were selective for stimulus categories, particularly for faces. Strong category-specific fMRI activation was observed in discrete patches throughout OT. Combining ECoG with fMRI mapping, we identified at least six face-selective patches that appear to occupy two parallel pathways within the ventral stream, similar to previous findings in macaques and humans. The preference for structured versus scrambled stimuli increased gradually along a posterior to anterior gradient. The marmoset has a set of face-processing regions that bear similar organization to those previously described in humans and macaques. Complementary to fMRI, two-photon laser scanning microscopy (2PLSM) allows direct observation of neurons, astrocytes, microglia, and microvascular CBF. We developed 2PLSM techniques to allow visualization of neurons and blood vessels, and measured changes in neural activity and hemodynamics in the cortex of awake marmosets. We verified the optical imaging depth and clarity of the cranial window over 7 months, and optimized imaging parameters to achieve high spatial and temporal resolution. We developed algorithms to measure flow speed and diameter of individual blood vessels in real time, and the 3D topological connectivity of the vessels. We observed dense capillaries in the cortical tissues that connect from an arteriole to a venule in 15 branches. In order to label neurons underneath the cranial window, an intracortical injection of the genetically encoded calcium indicators GCaMP5G or GCaMP6s was given and functionality was verified in vivo by measuring the functional changes in fluorescence in response to somatosensory stimulation. The capillaries, along with the arterioles, dilate when an electrical stimulation was delivered to the wrist of the awake marmoset. Seven months after the viral delivery, 29% of GCaMP-expressing neurons are responsive to external stimulation of the somatosensory cortex. There are many advantages to using non-human primates to understand physiological and pathological processes in the highly evolved human brain and it is now possible to insert transgenes and to perform gene editing in primates, and we want to take advantage of these techniques to improve on our usage of marmosets to our specific goals. Our first priority was to generate transgenic marmosets expressing genetically encoded calcium indicators such as GCaMP5G/6s that will constitute a much-improved animal model for studying neurovascular coupling in relevant conditions. The rationale for first making a GCaMP-expressing marmoset comes from our desire to investigate neurovascular coupling at the spatial resolution of individual neurovascular units. To distinguish neurons, genetically encoded calcium indicators from the GCaMP family have been recently synthetized. When expressed in excitable cells of the brain, these molecules fluoresce upon calcium binding, becoming a visible marker of neural activity. We developed GCaMP5g and GCaMP6s expression constructs under the ubiquitous cytomegalovirus (CMV) and the human synapsin promoters. Marmoset embryos were collected either via nonsurgical uterine flushing from donor females paired with fertile males after natural mating (NAT); or via laparotomic follicular aspiration of unfertilized oocytes (OPU) from superovulated non-pregnant donor females, followed by in vitro maturation and in vitro fertilization. Single cell to morula stage embryos were then placed in 0.25M sucrose supplemented M2 medium, and the concentrated lentivirus was injected into the perivitelline space. NAT embryos were usually collected between the 16-cell stage and morula stage and injected with the lentivirus on the same day, while OPU embryos were infected with the lentivirus at the single cell stage. Transgene expression in the OPU embryos was verified prior to transferring to a recipient female. Prior to embryo transfer, the estrus cycle of donor and recipient animals were synchronized and microinjected embryos were non-surgically transferred. Pregnancies were confirmed with an ultrasound and monitored regularly until delivery. To date 6 transgenic marmosets were born.
Showing the most recent 10 out of 44 publications
