PROJECT SUMMARY/ABSTRACT
How do we perceive the three-dimensional (3D) movement of objects in the world when our eyes only sense
two-dimensional (2D) projections like a movie on a screen? Accurate and precise perception of 3D object motion
is essential to intercept objects (e.g., catch a ball) and evade others (e.g., dodge a passing bicyclist). The goal
of this proposal is to elucidate the cortical networks that transform ambiguous 2D retinal signals into high-level
3D object-motion representations. To achieve this goal, we will utilize a synergistic combination of behavioral,
electrophysiological, and causal manipulation approaches with macaque monkeys. In Aim 1, we will distinguish
2D retinal motion selectivity from 3D object-motion selectivity at the single neuron level and evaluate functional
correlations with behavior. We will test the hypothesis that 3D object-motion representations are created within
a cortical network consisting of the middle temporal area (MT), the fundus of the superior temporal sulcus (FST),
and the lateral subdivision of the medial superior temporal sulcus (MSTl). The experiments will combine a 3D
object-motion discrimination task with simultaneous high-density neuronal recordings from all three areas.
Importantly, the stimulus set was rigorously vetted through previous perceptual and computational studies, and
maximally discriminates 2D retinal vs. 3D object-motion representations. This work will be the first to assess
functional correlations between neuronal activity and the behavioral discrimination of 3D object-motion. To
evaluate the cortical network organization of MT, FST, and MSTl, we will compare the areas’ functional properties
and measure the Granger causal influences between them using simultaneously recorded local field potentials.
In Aim 2, we will apply a complementary approach to assess the causal contributions of each area to 3D motion
perception. Specifically, we will use electrical microstimulation (EM) with weak currents to manipulate neuronal
activity while the monkeys perform the 3D object-motion discrimination task. These experiments will be the first
to use EM to causally probe the relationship between neuronal activity and 3D object-motion perception.
Critically, the predicted relationship between neuronal response properties at the site of EM and the induced
behavioral biases depends on whether the stimulated neurons are either: (i) selective for 2D retinal motion (with
outputs that are used by downstream neurons to compute 3D object-motion, otherwise no effect of EM would be
expected) or (ii) selective for 3D object-motion. We will test the predictions locally (i.e., at the level of individual
neurons within each area) to assess area-specific functional heterogeneity and globally (i.e., between areas) to
assess hierarchical differences across the network. The proposed experiments will together explicate differences
in the functional properties of three interconnected cortical areas as well as their causal contributions to 3D
motion perception. By elucidating the cortical networks that transform 2D retinal signals into ecologically relevant
representations of 3D object-motion, insights from this work will facilitate future studies that explore how neuronal
representations of dynamic, object-level information support interactions with the 3D world.