Origins of Spontaneous Hemodynamic Signals in the Awake Brai

Cerebral blood flow is regulated on a sub-millimeter scale by local neural populations. Increased neural activity gives rise to vessel dilations and greater flow. This neurovascular coupling is important for the maintenance of brain health but also serves a practical role in the study of the human brain. Imaging modalities, such as fMRI, can non-invasively detect ongoing brain-wide hemodynamics changes and use them to infer the underlying neural activity patterns. However, the neural mechanisms which give rise to cerebral hemodynamics are not clear. In particular, the relationship between neural activity and cerebral blood in the absence of sensory input or motor output must be elucidated to enable an accurate interpretation of fMRI data.;We tested how well cerebral blood flow can report local neural activity during quiescent behavior by simultaneously measuring neural activity and hemodynamics in a localized sensory region of the cerebral cortex of awake mice. We found that the quantified relationship (ie: the transfer function) between neural activity and hemodynamics during quiescence was similar to that obtained from periods of voluntary movement and sensory stimulation. We used this function to predict individual hemodynamic changes from the corresponding neural activity in order to determine what fraction of the ongoing signal could explained by local brain activity. We found that this fraction was exceedingly low during periods of rest. To determine the cause of this poor correspondence we infused drugs into the local region which silenced various neural signaling mechanisms. We found that hemodynamic signals continued to change even in the absence of localized neural activity indicating that the neuro-vascular relationship was obscured by an additional mechanism which drives blood changes in the brain and is uncorrelated to behavior and local neural activity. These data indicate that individual spontaneous hemodynamic signals are not reflective of the underlying neural processes and must be interpreted with caution in functional imaging data.