Goal: We seek to understand the circuitry of the mammalian cerebral cortex and how it endows us with the ability to see.

Approach: We study the visual cortex of both monkeys and rodents. With monkeys, we take advantage of their ability to perform difficult perceptual tasks that provide insight into specific aspects of their visual experience. This allows us to define the neural correlates of particular visual behaviors and then begin to study their underlying mechanisms, first at a coarse level, by activating or inactivating higher order visual areas. In rodents, we exploit the ability to continuously monitor neural populations while animals experience “binocular rivalry” and other interesting perceptual states. We then use optogenetics to manipulate specific cell types, in order to test hypotheses generated by the behavioral experiments in both species.

Techniques: In monkeys, our primary tools are extracellular electrophysiology—with multi-electrode arrays and laminar probes—and psychophysics. They are complemented by techniques that allow us to perturb visual areas, including inactivation with cortical cooling and activation with microstimulation. In rodents we use behavior, 2-photon Ca++ imaging and optogenetics.

Current Projects:

Cortico-cortical feedback. Most of the projects in the lab are aimed at deciphering this ubiquitous, but poorly understood, aspect of cortical connectivity. We have previously shown that feedback from V2 and V3 has a relatively selective effect on the non-classical surrounds of V1 receptive fields (Nassi et al. 2013; Nassi et al. 2014). These surrounds are critical for vision, because they allow local, feature-selective responses to be modulated by the context in which they occur. This modulation is surprisingly sophisticated, and appears well suited to reduce redundancy and create sparse representations in visual cortex via input-gain control (Trott & Born 2015). In addition, feedback exerts a surprisingly large influence on the local field potential and associated rhythms, such as gamma oscillations, which, in turn can affect the variability of neuronal spiking (Gómez-Laberge et al. 2016). Taken together, our findings have led us to focus our rodent studies on top-down inputs to the layer 1 apical dendrites of pyramidal cells and to interactions with somatostatin-containing interneurons.

1) Feedback and learning. While animals learn to discriminate the orientation of noisy oriented textures, we record in V1 with multi-electrode arrays to test the predictions of a hierarchical Bayesian model of perceptual inference. (Ariana Sherdil, Postdoctoral Fellow; Camille Gómez-Laberge, Postdoctoral Fellow; collaboration with Dr. Ralf Haefner, Univ. of Rochester).

2) Feedback mechanisms. We have developed a rodent preparation that lets us simultaneously monitor the activity of both neuronal cell bodies in V1 and the synapses projecting back to them from higher visual areas. We do this using genetically encoded Ca++ indicators of different colors that are conjugated to proteins that traffic them to different parts of the neuron. This allows us to measure and, ultimately, manipulate the different sources of information to test circuit-level hypotheses about how top-down influences affect sensory inputs. (Abhinav Grama, Postdoctoral Fellow; Susanne Haridi, Master's Student; Peter Kim, Harvard undergraduate)

3) Layer 1 connectome. This project, motivated by our recent studies of the effects of V2 inactivation on the response properties of V1 neurons in awake, behaving monkeys, is still in the planning stages. Stay tuned! (collaboration with Kathy Rockland, Boston University; HMS labs of Gord Fishell, David Ginty and Wei-Chung Allen Lee)