Skip to main content
. 2023 Oct 16;12:e83930. doi: 10.7554/eLife.83930

Figure 3. Distance and Time Cells classification.

(a) Cells type classified by the CellType metric, averaged over all animals and trials, for fixed-distance and fixed-time experiments (n=18 experiments, mean ± SEM) p<<0.001 by Pearson's chi-squared test using two categories. Diagonal lines represent individual animals. (b) ROC curve (red) showing that the chosen discriminating threshold of 0 (black point) is optimal. The True Positive Rate (TPR) is the percentage of cells classified as distance cells on the fixed-distance session, while the False Positive Rate (FPR) is the percentage of cells classified as distance cells on the fixed-time sessions. (c) Shuffling distribution of the three metrics: CellType, FIT and P-Value. The type of experiment, either fixed-time or fixed-distance, was randomized 1000 times for each of the sessions. All experiments were truncated to 16 s in order to prevent biases. The vertical axis is the Time-Distance balance index (TDI), defined as (#DistanceCells-#TimeCells)/(#DistanceCells + TimeCells) and is between 0 and 1 if there are more distance cells than time cells and between –1 and 0 if there are more time cells than distance cells. The arrows are indicating the actual results which are significant compared to the shuffle distribution.

Figure 3.

Figure 3—figure supplement 1. Additional metrics.

Figure 3—figure supplement 1.

Cells type classified by the FIT metric (a) and p-Value metric (b), averaged over all animals and trials, for fixed distance and fixed time experiments (n=18 experiments, mean ± SEM) p<<0.001 by Pearson's chi-squared test using two categories. The whiskers show the SEM and the points and the lines connecting the points represent individual animal results. (c) Comparison of onset vs peak analysis cells count in each category (d–f) CellType, Fit and p-Value metrics results for a peak-based analysis.

Figure 3—figure supplement 2. The distribution of Time and Place cells, per animal, at various metrics, for Time and Distance sessions.

Figure 3—figure supplement 2.

Animals BK26 (n=3 experiments) and BK33 (n=1 experiment) participated only in distance sessions, while BK45 (n=3 experiments) participated only in a time session, and BK35 (n=3 experiments), BK41 (n=4 experiments) and BK49 (n=4 experiments) participated in both types of sessions. All distributions show a significant dependency on the session type (χ2(1)>12, p<<0.001 for BK35,BK49,BK45,BK33 and χ2(1)=5.7, p<0.02 for BK26 by Pearson's chi-squared test using two categories.), except for animal BK41 (χ2(1)=2.38, p=0.12). The bars represent the percentage of the respective cell type from the entire population of cells in the respective session. The numbers inside the bars represent the number of cells. For the single session animals, the reference distribution used for the χ2 test was taken from the entire population distribution (55% Distance Cells and 45% Time Cells).

Figure 3—figure supplement 3. Distribution of max and average firing rates for Time and Distance cells.

Figure 3—figure supplement 3.

Total 304 time cells and 374 distance cells evaluated. The time cells and distance cells distributions are similar (p=0.975, Kolmogorov-Smirnov test for the max firing rate, p=0.675 for the average firing rate).