Resting-state fMRI (R-fMRI) is a major method for studying human brain networks. It relies on correlations in spontaneous fluctuations of BOLD signal. However BOLD signal fluctuations are poorly understood, leading to ambiguity in interpretation of R-fMRI derived networks. For task fMRI (T-fMRI) we, and others, have shown a direct relationship between BOLD signal changes of glutamatergic neuronal signaling and oxidative demand. The evoked changes in oxidative demand (CMRO2), measured by calibrated T-fMRI combines measurements of BOLD signal, blood flow (CBF), blood volume (CBV), and subject-specific biophysical constants. However a similar level of metabolic understanding for R-fMRI has been elusive. In the past cycle, using calibrated T-fMRI with 1H[13C] MRS and extracellular recordings, we developed and validated transfer functions of BOLD, CBF, CBV, and CMRO2 signals that related them directly to extracellular recordings of multi-unit activity (MUA) and local field potentials (LFP). Inverse of the transfer functions allowed calibrated T-fMRI maps to be converted into neuronal activity maps. However this approach has limitations when applied to calibrated R-fMRI, in particular the small spatial scale of MUA/LFP recordings is not suitable for correlating with large regions involved in correlated network fluctuations and lack of a stimulus trigger in resting-state to allow recordings in and out of scanner to be time synchronized. We will overcome these limitations by simultaneous calibrated R-fMRI and calcium (Ca2+) imaging in Snap25-GCaMP6 mice, which contain genetically encoded fluorescent Ca2+ reporters. Since Ca2+ imaging directly measures neuronal activity in these transgenic mice, we will improve the biomarker potential of R-fMRI derived functional connectivity density (FCD) differences in health and disease. Preliminary data show that the spatiotemporal structure of R-fMRI and resting Ca2+ (R-Ca2+) networks quite similarly, and long-term mitochondrial health (e.g., aging) can also perturb R-fMRI network patterns. Since we, and others, have shown that the resting-state fluctuations depend on the total activity, an independent and absolute measure of total activity (by high-resolution 1H[13C] spectroscopic imaging) is critical, so that we can bridge the gap between BOLD signal and underlying activities of neuronal populations (by electrophysiology). We build on these preliminary results and propose: In Aim 1 we will develop technologies for simultaneous calibrated R-fMRI and R-Ca2+ imaging. In Aim 2 we will measure and validate state-dependent neurovascular and neurometabolic transfer functions. In Aim 3 we will apply calibrated R-fMRI and R-Ca2+ imaging in aging, with and without calorie restriction,to validate that the methods can accurately track how neuronal resting-state fluctuations are altered and sensitivity to an intervention. Given that network changes as revealed by fMRI- derived FCD is a feature of brain disorders (e.g., autism, schizophrenia, healthy aging) calibrating R- fMRI to R-Ca2+ could help interpret the altered network dynamics that underlie disease.

Spontaneous modulations of oxygen supply in relation to oxygen demand are critical for brain's health and these fluctuations are also the basis of contrast generation in fMRI. Resting-state fMRI (R-fMRI), a major method for generating human brain networks, uses these spontaneous fluctuations but whose neuronal basis is misunderstood. Validating R-fMRI on the basis of how metabolic and hemodynamic signals are linked to calcium imaging of wide-scale neuronal activity will improve the biomarker potential of R-fMRI based functional connectivity changes observed in brain health and disease.