Fig. 4: Illustration of 2-D toroidal representations of grid cell activity during running in 1-D and 2-D environments (left and center) and natural behavior (right).
From: Uncovering 2-D toroidal representations in grid cell ensemble activity during 1-D behavior
While studying single cell activity in simple tasks may reveal specific tuning to behavioral variables (top row), the population activity of all recorded neurons together is often difficult to interpret, in this case only recovering the properties of the task itself (bottom row, the activity covering the geometry of the environment). Focusing instead on distinct neural ensembles, such as grid cell modules, we can uncover the intrinsic structure of the neural computations (middle row, toroidal representation as shown in 1- and 2-D environments in this paper and during sleep4), that is likely to generalize to natural behavior (right).
