Movement invariants in the motor cortex


July 17, 2017 - 12:00pm
Northwest Building, Room 243
About the Speaker
Naama Kadmon Harpaz
Speaker Title: 
PhD Candidate
Speaker Affiliation: 
Weizmann Institute of Science

A goal-directed action can be performed in multiple ways, such as by moving in different paths or by activating different muscle groups. Yet, certain geometrical properties of the movements are remarkably similar across repetitions performed by an individual or different individuals, as well as across spatial and temporal transformations of the movements (e.g. when performing the same movement with different speeds). One possible explanation for these invariant solutions is that they reflect conserved neural patterns that undergo various transformations to generate the entire movement repertoire. I will present evidence supporting this hypothesis from an fMRI study in which we simultaneously recorded subjects’ complex hand trajectories and associated BOLD responses. Our findings show that movements of similar shapes but different amplitudes were associated with similar multi-voxel patterns in the primary motor cortex (M1) and in the anterior intraparietal sulcus (aIPS), implying scale-invariant encoding in these areas. We further studied this hypothesis by asking whether we can directly detect invariant neural patterns in an unsupervised manner, and by examining the temporal decomposition of movements based on such patterns. To this end, we modeled the spiking activity recorded from M1 of macaque monkeys while they performed a random target pursuit task using a hidden Markov model. We found a distinct decomposition of the neural activity that correlates with specific behavioral epochs, namely acceleration and deceleration segments with directional selectivity, invariant to the exact position in the workspace or to the executed extent and speed. These overall findings will be discussed with reference to a unified model of motor control that posits an exploration-exploitation balance between invariance and inherent variability of the motor output, allowing the motor system to be both consistent and flexible at the same time.