Characterizing the Cortical Contributions to Working Memory-Guided Obstacle Locomotion

While walking in complex environments, the ability to acquire information about objects in our surroundings is essential for successful obstacle negotiation. Furthermore, the ease with which most animals can traverse cluttered terrain while grazing, exploring, or hunting is facilitated by the capacity to store obstacle information in working memory (WM). However, the underlying neural substrates supporting such complex behaviours are poorly understood. Therefore, the goal of this thesis is to examine the neural underpinnings of WM-guided obstacle negotiation in the walking cat. Obstacle locomotion was studied in two main paradigms, characterized by whether obstacle presence was detected via vision or touch. In both paradigms, walking was delayed following foreleg obstacle clearance. When walking resumed, elevated hindleg stepping demonstrated that animals successfully remembered the obstacle beneath them. The tactile paradigm was first examined to assess the ability of animals to remember an unexpected obstacle over which the forelegs had tripped. Such tactile input to the forelegs was capable of producing a robust, long-lasting WM of the obstacle, similar to what has been previously described using the visual paradigm. Next, to assess whether regions of the brain associated with spatial representation and movement planning contribute to these behaviours, parietal area 5 was reversibly deactivated as visual or tactile obstacle WM was tested. Such deactivations resulted in substantial WM deficits precluding successful avoidance in both paradigms. To further characterize this cortical contribution, neural activity was then recorded with multi-electrode arrays implanted in area 5. While diverse patterns of task-related modulation were observed, only a small proportion of neurons demonstrated WM-related activity. These neurons exhibited the hallmark property of sustained delay period activity associated with WM maintenance, and were able to reliably discern whether or not the animal had stepped over an obstacle prior to the delay. Therefore, only a specialized subset of area 5 neurons is capable of maintaining stable representations of obstacle information in WM. Altogether, this work extends our understanding of WM-guided obstacle locomotion in the cat. Additionally, these findings provide insight into the neural circuitry within the posterior parietal cortex, which likely supports a variety of WM-guided behaviours