RCA 2:

Discover Innovation/ Learning Strategies in Animals
























Research Concentration Area (RCA) 2.i: Animal Learning and Innovation

Animals’ bodies are exquisitely adapted to their habitats.

Embodied intelligence refers to the integrated physical problem solving and innovative capabilities that emerge from the interplay of animals' mechanical and mental processing.

Learning

A fox squirrel is challenged to jump onto a compliant beam after jumping from a stiff control beam. It is able to rapidly learn and accomplish this new task. In the first trial, the fox squirrel nearly fails but by the fifth trial the fox squirrel's ability to jump on the bouncy beam is nearly perfect.

Reference

Hunt, N. Jinn, J., Libby, T., Jacobs, L. F., and Full, R. J. 2015. Learning to launch: targeted leaping from a dynamic obstacle in squirrels. Society of Integrative and Comparative Biology Annual Meeting & Exhibition Final Program and Abstracts. West Palm Beach, FL. January 3-7, 2015.

Innovation

A fox squirrel is challenged in an experiment by longer gap and elevated perch. It unexpectedly parkours off wall, and by doing so generates an intermediate control point, allowing it to easily complete task.

Reference

Hunt, N., Jinn, J., Full, R. J. 2016. Squirrel parkour: wall-jump maneuver adds intermediate control point to ballistic trajectories. Society of Integrative and Comparative Biology Annual Meeting & Exhibition Final Program and Abstracts. Portland, OR. January 3-7, 2016.


RCA2.ii: Physiological Studies in Manipulated Environment

When an animal invents a new solution to a dynamic locomotion problem, it must integrate sensory perception with higher level cognitive representations in order to envisage new locomotor paradigms. To discover high level representations for environmental interaction, and observe how they allow innovation, we propose to examine neurophysiological encoding of spatial navigation during challenging locomotor behaviors in freely behaving animals. We focus on studies of grid cells (Hafting et al, 2005), place cells (O'Keefe, 1976), and other spatial signals in the brain as these animals creatively solve problems in 3D environments. Our studies focus on the grid cells of the Medial Entorhinal Cortex (MEC), egocentric (i.e. "self-based") spatial encoding of external places in the Lateral Entorhinal Cortex (LEC), and place cells of the hippocampus. The MEC is thought to provide a signal of the organism’s current location in an allocentric spatial framework. This signal is primarily derived through a path integration computation, with input from visual landmarks providing a calibration and error-correction signal. New data suggest that the LEC may provide a complementary spatial signal, based on an animal’s egocentric bearing relative to salient, external landmarks or reference points (Wang et al, 2018). Thus, these two systems interact to provide the hippocampus with a self-motion-based signal of the rat’s current location in an allocentric (i.e. "world-centered") coordinate frame along with a signal about the location of salient landmarks and other points external to the organism, based on an egocentric coding representation. Place cells of the hippocampus integrate these two inputs and create context-dependent representations of the animal’s current location, planned trajectories and remembered past sequences of experience. These representations are transmitted to downstream brain regions that make action decision and control motor patterns based in part on the context-dependent memory signals provided by the hippocampus.

Shown above is our proposed augmented reality "dome" apparatus. The dome wall is rendered transparently for illustration, but it is actually opaque. In this system a projector projects visual cues (virtual landmarks or strips) on the inside wall of the dome through a mirror. The rat, attached to a boom arm, can freely walk on the table in one-dimensional circular track. The visual features rotate in real-time proportional to the animal's movement in order to create a virtual reality experience. Based on the speed of rotation of the visual cues, the rat perceives its motion faster or slower than normal and actually adjusts a gain inside the brain known as the path integration gain. A predecessor to this system was used to make new discoveries about the interaction of landmarks and self-motion cues (Jayakumar et al, 2019, Nature 566:533-537)

Novel jumping / ditching apparatus

Stadium-style table for jumping rats, eventually to be placed in our dome, allowing perceptual manipulations during agile behavior

Hippocampal recordings

Hippocampal recordings during novel agile task. Activity of several place cells at multiple locations on the track

Reference

This project uses the system identification approach developed by the Cowan laboratory, summarized in part by this paper, supported by the SLICE MURI.

Madhav, Manu S., and Noah J. Cowan. “The synergy between neuroscience and control theory: the nervous system as inspiration for hard control challenges.” Annual Review of Control, Robotics, and Autonomous Systems , vol. 3, pp. 243-267, 2020.

The schematic figure below was adapted from this paper:

General References:

Shown to the left are "replay" events: encoding future trajectories recorded during sharp wave events. From:

Pfeiffer, Brad E., and David J. Foster. “Hippocampal place-cell sequences depict future paths to remembered goals.” Nature, vol. 497, no. 7447, pp. 74–79, May 2013.

Discovery of place cells was first reported in detail here:

O'Keefe, John. “Place units in the hippocampus of the freely moving rat.” Experimental neurology , vol. 51, no. 1, pp. 78–109, Apr. 1976.

Discovery of grid cells was first reported here:

Hafting, Torkel, Marianne Fyhn, Sturla Molden, May-Britt Moser, and Edvard I. Moser. “Microstructure of a spatial map in the entorhinal cortex.” Nature, vol. 436, no. 7052, pp. 801–806, Aug. 2005.

LEC encodes objects egocentrically:

Wang, Cheng, Xiaojing Chen, Heekyung Lee, Sachin S. Deshmukh, D. Yoganarasimha, Francesco Savelli, and James J. Knierim. “Egocentric coding of external items in the lateral entorhinal cortex.” Science, vol. 362, no. 6417, pp. 945–949, Nov. 2018.