Layer specific differences in the mouse auditory corticocollicular system

Recent data suggest that there may be distinct processing streams emanating from auditory cortical layer 5 and layer 6 that influence the auditory midbrain. To examine the functional properties of neurons in these two layers that project to the midbrain, we performed three sets of experiments. First, to determine whether these projections have different physiological properties, we injected rhodamine-tagged latex tracer beads into the inferior colliculus (IC) of >30 day old mice to label these corticofugal cells. Whole-cell recordings were performed on 62 labeled cells to determine their basic electrophysiological properties and cells were filled with biocytin to determine their morphological characteristics. We observed that layer 5 auditory corticocollicular cells have prominent Ih-mediated sag and rebound currents, generate calcium-dependent rhythmic bursts, and have relatively sluggish time constants. In contrast, layer 6 auditory corticocollicular cells are non-bursting, do not demonstrate sag or rebound currents, and have short time constants. Quantitative (Sholl) analysis of morphology showed that layer 6 cells are smaller, have a horizontal orientation, and have very long dendrites (> 500μm) that branch profusely both near the soma and distally near the pia. Layer 5 corticocollicular cells are large pyramidal cells with a long apical dendrite with most of the branching near the pial surface. The marked differences in physiological properties and dendritic arborization between neurons in layer 5 and layer 6 make it likely that each type plays a distinct role in controlling auditory information processing in the midbrain. Very little is currently known about the nature of the inputs from the rest of the auditory cortex onto these cells. Therefore, our second set of experiments was designed to investigate these local inputs. To do this, we utilized laser photo-uncaging of glutamate to stimulate the cells that synapse onto the layer 5 and layer 6 corticocollicular cells in brain slices taken from adult mice. Pre-identified cells were recorded in a whole cell patch configuration then stimulated with a larger grid covering the area from the white matter to the pia. To isolate synaptic responses, in this preparation, we used a low calcium artificial cerebral spinal fluid method which uses a physiological method of parsing out synaptic versus direct stimulation. We contrast this method with the more commonly used time window method. In identified layer 5 and layer 6 corticocollicular recordings, cells show spatial differences in their respective input maps. Layer 5 corticocollicular neurons were shown to receive inputs coming from various layers compared to layer 6 corticocollicular neurons which almost exclusively receives input from layer 6. Combined with our first set of studies, which showed that layer 5 and layer 6 corticocollicular neurons have different electrophysiological properties, the current data suggest that neurons in these two layers play different roles in modifying ascending information at the IC. These differences may explain the varied results seen in the inferior colliculus during in vivo stimulation of the auditory cortex. In sensory cortices, layer 4 is generally considered to be the primary thalamorecipient layer. Recent data, however, have shown that other layers receive thalamic input. With many of the direct inputs of these thalamocortical collaterals onto inhibitory interneurons. Therefore, in our third set of studies, to investigate these inputs, we used a laser to stimulate thalamocortical axons which have been labeled with channelrhodopsin. Pre-identified corticocollicular cells were recorded in a whole cell patch configuration then stimulated with a larger grid covering the area from the white matter to the pia. We compared the excitatory and inhibitory input from the thalamus in both layer 5 and layer 6 corticocollicular neurons. We found that similar to previous results, layer 5 corticocollicular neurons received more input in general than layer 6. However, interestingly, there are both layer 5 and layer 6 corticocollicular neurons that receive considerable thalamic inputs compared to those assessed with glutamate uncaging. It is likely that the coordinated stimulation of these afferents cause spatial summation of input compared to the very focal nature inherent to the methods in glutamate uncaging