Predictions about sensory inputs (e.g. contextual signals from associative cortical areas to sensory areas) arrive primarily in the superficial layers of the cortical sheet, where they are received by the distal apical dendrites of pyramidal neurons. These distal apical dendrites are long tree-like extensions which are electro- tonically segregated from the cell body under passive propagation of membrane potential. At the same time, incoming sensory inputs primarily target the region closest to the cell body. The relative coupling of activity between these neuronal regions (i.e. simultaneous versus independent activity) has played a key role in various theories of dendritic computation, but it remains controversial. Some studies report little or no functional coupling between the somata and apical dendrites, whereas others report high levels of coupling. In our work, we have developed methods to link distal apical dendritic compartments to somatic compartments from experimental recordings, enabling near-simultaneous imaging of activity in both of these compartments within the same cell. This approach involves acquiring high-resolution volumetric z-stacks of the neuronal populations, a notable advancement for in vivo neurophysiological recordings. To exploit these high quality z-stacks, we have developed complementary methods to denoise, align, combine and segment them, to isolate the neuronal structures and match these to the dendrites and somata from our experimental recordings. We used connectivity information from these z-stacks along with advances in multi-plane simultaneous imaging to directly observe somato-dendritic coupling, in order to experimentally test theories of dendritic computation. We demonstrate the utility of this approach by applying it to experimental data from a predictive learning task. We show that it can be used to identify the degree of coupling of activity between compartments of individual neurons in vivo.