Extended Data Fig. 3: Left-right alternating sweeps in grid cells and place cells.
From: Left–right-alternating theta sweeps in entorhinal–hippocampal maps of space
a, Number of detected sweeps scales with number of recorded MEC-parasubiculum cells (left; Pearson correlation: r = 0.69, p = 0.003, n = 16 animals, 1 session per animal), number of grid cells (middle; r = 0.67, p = 0.005, n = 16 animals, 1 session per animal), and number of spikes per theta cycle (right; r = 0.95 ± 0.01 (mean ± s.e.m.), p < 0.01 in all 16 animals, 1 session per animal), suggesting that the reported fraction of theta cycles with sweeps is underestimated. Data from individual animals are plotted as dots (left and middle) or lines (right). b, Stacked firing-rate autocorrelograms (±350 ms) for all theta-modulated grid cells (2,399/3,194 or 75.1% of grid cells were theta-modulated), sorted by tendency to fire on alternating theta cycles. Firing rates are colour-coded. Lags corresponding to repetitions of the ~8 Hz theta rhythm is indicated with red arrowheads. Note that a large fraction of theta-modulated grid cells (1,289/2,399 or 53.7%) display theta skipping, with prominent peaks at lags corresponding to every other theta cycle. c, Left: scatter of directions of all sweeps and tracked head direction at the corresponding times during one example recording session (same session as Fig. 1a). Sweep directions are distributed bimodally around the animal’s head direction (circular correlation between sweep direction and head direction r = 0.55 ± 0.03, p < 0.001 in all 16 animals, absolute mean offset 3.1 ± 2.8 deg, mean ± s.e.m. across 16 animals). Right: histogram of angles between successive sweeps. Note that these angles are clustered around ±60 deg. Sweeps directed to the right or left of the previous sweep are defined as ‘right’ and ‘left’ sweeps, respectively. d, Circular histograms of head-centered sweep directions for the 10 animals with the highest fraction of theta cycles with detected sweeps. Sweeps are rotated to head-centred coordinates (head orientation is vertical) and sorted based on whether the previous sweep was directed to the right (blue) or left (red). Note clustering of sweeps around two principal head-centered directions offset ~30 deg to the left and right of the animal’s head direction. e, Sweeps averaged across theta cycles for all cells of a grid module, as in Fig. 1b, but now showing all modules. Sweeps are plotted with reference to the lowpass-filtered decoded trajectory (origin) and rotated to head-centred coordinates (head orientation is vertical). Sweep lengths are normalized by the grid spacing of each module. Note that sweeps have similar lengths relative to the scale of the grid cells. f, Averaged sweeps decoded from hippocampal ensemble activity (8 rats), plotted as in Fig. 1b. Hippocampal sweeps alternated from side to side in 78.5 ± 3.1% (mean ± s.e.m.) of successive triplets of theta cycles with detected sweeps, significantly more than when sweep directions were shuffled (>99.9th percentile for all animals). g, Left column: forward progression of decoded sweeps (projected onto the rats’ head axis) as a function of time from the beginning of each theta cycle for MEC-parasubiculum (top) and hippocampus (bottom) for all 6 rats with dual HC/MEC implants during open field foraging, plotted as in Fig. 1i. Data from each of the rats are plotted as separate lines (1 session per rat). Note that hippocampal sweeps are delayed relative to MEC sweeps. Right column: top plot shows forward progression of entorhinal (blue) and hippocampal (green) sweeps during a linear track session from an example animal (plotted as in Fig. 1i). Bottom plot shows cross-correlations of decoded positions in MEC and hippocampus for 4 animals with dual HC/MEC implants that were tested on the linear track (1 session per animal), plotted as in Fig. 1k. Hippocampal sweeps are delayed by 16.2 ± 4.0 ms (mean ± s.e.m.) with respect to entorhinal sweeps on the linear track. h, Left: histological section from a rat with an implanted probe in the hippocampus. Probe track through hippocampal subregions is indicated by arrowheads. Green channel: autofluorescence; red channel: mCherry (expressed in hippocampus after a viral injection prior to implantation surgery). Right panel: left-right alternating sweeps from neurons recorded simultaneously in CA1 (top right) and CA3 (bottom right) during foraging in the open field task (data from 1 session from animal shown to the left; sweeps averaged across theta cycles, as in e,f). Mean sweep lengths were 28.8 cm and 22.6 cm; directions were 25.5 and 27.3 degrees to either side of the head-axis. Cells were recorded on different sites from the same probes (left). i, Panels show fraction of epochs with consistent left-right-alternation of internal direction (left) or sweep direction (right), with epoch lengths ranging from 3–20 theta cycles. Solid lines show the median across animals; whiskers show 25th and 75th percentiles (n = 16 animals, 1 session per animal). Shuffles were generated by randomly shuffling head-centered directions across theta cycles. Credit: rat, scidraw.io/Gil Costa.
