A negative BOLD response is often observed in early visual cortex. Previous measurements indicate that negative BOLD is the consequence of flow decreases. However, negative BOLD is usually evoked by blocked experimental designs that obscure details of their temporal dynamics. To understand the physiology of the negative BOLD, we used a slow event-related paradigm to elicit strong hemodynamic response functions (HRFs) in primary visual cortex as a function of eccentricity. We used a modified version of the Arterial Impulse Model to predict the neurovascular and neurometabolic mechanisms of negative BOLD. Results indicate that negative BOLD physiology is profoundly different than positive BOLD with greater non-linearity.