Wilson, Daniel E.

Individual neurons in the primary visual cortex respond selectively to different
features of visual stimuli, such as spatial orientation or direction of motion. A longstanding
goal in systems neuroscience has been to understand the transformations single
cells perform as they integrate synaptic inputs to generate spiking output. Recent
technological developments have facilitated these lines of investigation by enabling direct
measurement of the functional properties of single synaptic inputs to neurons in the
neocortex. It remains an outstanding question as to whether the tuning of single
neocortical neurons can be predicted by their excitatory synaptic inputs. Here, I show
that excitatory synaptic inputs exhibit significant functional diversity with respect to
orientation and direction selectivity. I show that cells can use at least two strategies to
overcome this functional diversity to achieve selective responses in the face of broadly
tuned excitatory input: enhancing responses to the preferred stimuli and suppressing
responses to the non-preferred stimuli. In the case of orientation selectivity, synaptic inputs cluster according to orientation preference and evoke local dendritic nonlinearities,
thereby enhancing somatic responses to the preferred direction. For direction selectivity,
cells receive excitatory synaptic inputs tuned to the preferred and null directions, but
selectively suppress inputs tuned for the null direction to enhance direction selectivity.
This suppression comes from direction-tuned GABAergic interneurons that make longrange,
intercolumnar projections to enhance direction selectivity.

The visual cortex of higher mammals, including humans, is arranged as to achieve a
continuously varying map of features such as the orientation of contours in the environment.
Previous studies used intrinsic signal and two-photon imaging to examine the functional
composition of these cortical maps, but lacked the functional resolution to resolve the underlying
synaptic architecture. Here, we exploited recent advances in genetically encoded calcium
indicators to perform in vivo two photon imaging of dendrites and dendritic spines in an animal
with a mapped visual cortex. We found sharp orientation and direction tuning when we presented
drifting gratings and imaged synaptic calcium transients from large numbers of dendritic spines
in single neurons, obtaining synaptic maps of orientation preference. In addition, we
implemented a newly developed two-photon microscope that uses acousto-optical deflectors to
rapidly steer a pulsed laser in three dimensions. This technology allowed us to image 320 single
cells in an 800x800x200 micron three-dimensional volume, which yielded a three-dimensional
orientation map with single-cell resolution. In the future, we will perform fast, three-dimensional
imaging of a single cell and its entire dendritic tree to monitor functional properties of a cell’s
inputs and its somatic spiking output. These experiments will yield important insight into
synaptic integration and sensory processing in cortical maps and how such organizing principles
might be disrupted in disease states.