Spatial cognition and neuro-mimetic navigation: a model of hippocampal place cell activity

Abstract

A computational model of hippocampal activity during spatial cognition and navigation tasks is presented. The spatial representation in our model of the rat hippocampus is built on-line during exploration via two processing streams. An allothetic vision-based representation is built by unsupervised Hebbian learning extracting spatio-temporal properties of the environment from visual input. An idiothetic representation is learned based on internal movement-related information provided by path integration. On the level of the hippocampus, allothetic and idiothetic representations are integrated to yield a stable representation of the environment by a population of localized overlapping CA3-CA1 place fields. The hippocampal spatial representation is used as a basis for goal-oriented spatial behavior. We focus on the neural pathway connecting the hippocampus to the nucleus accumbens. Place cells drive a population of locomotor action neurons in the nucleus accumbens. Reward-based learning is applied to map place cell activity into action cell activity. The ensemble action cell activity provides navigational maps to support spatial behavior. We present experimental results obtained with a mobile Khepera robot.

Modelling Path Integrator Recalibration Using Hippocampal Place Cells

“…It is consistent with fundamental electro-physiological properties of place cells [1]. This work extends previous models [7,8,9] by equipping them with a new visual system that can deal with realistic sensory input and an adaptive recalibration mechanism used to combine path integration and visual input. …”

Path planning versus cue responding: a bio-inspired model of switching between navigation strategies

“…During a reward-based goal planning phase, this representation is used to plan and execute goal-directed path. Since extra-maze cues are stable in the experiments that we will simulate, we use a simple model of Place Cells as described later (see Arleo and Gerstner 2000;Sheynikhovich et al 2009 for more detailed models of Place Cells that integrate information from distal cues and path integration). The population of Place Cells in our model is created before the learning is started, and the activity of place cell j is given by…”

“…This is in contrast to standard reinforcement learning algorithms in which exploration is chosen according to a predefined stochastic scheme. For example, Arleo and Gerstner (2000) and Chavarriaga et al (2005) use an -greedy scheme, in which novel actions are tested with small probability on each time step, while Foster et al (2000) use a soft-max selection where actions with high Q-values have a higher probability of being chosen. In robotic experiments (Cuperlier et al 2007;Barrera and Weitzenfeld 2007), the exploration is chosen when the animat cannot associate its location with any existing node in its topological map.…”

Spatial Representation and Navigation in a Bio-inspired Robot

“…The model of Arleo et al [29,30,31] is an earlier version of the model presented in the next section. In this model position and direction information extracted from the visual input were combined with information extracted from the self-motion signals and merged into a single space representation which was then used for goal navigation.…”