Skip to main content
. 2019 Oct 17;8:e49536. doi: 10.7554/eLife.49536

Figure 10. DLS and DMS neural activity correlates with over-trained task-performance during extended training.

(A) Left: Distribution of Spearman rank correlation coefficients relating firing rate after the lever insertion (0–500 ms) to the first lever press latency on a trial by trial basis. Example of neurons with a negative (middle) or positive correlation (right) between firing rate and latency. (B) Left: Distribution of Spearman rank correlation coefficients relating firing rate during the sequence (from first to last lever press) to the response rate on a trial by trial basis. Example of neurons with a negative (middle) or positive correlation (right) between firing rate and response rate. (C) Left: Distribution of Spearman rank correlation coefficients relating firing rate after the lever retraction (0–500 ms) to the port entry latency on a trial by trial basis. Example of neurons with a negative (middle) or positive correlation (right) between firing rate and port entry latency. Red lines on correlation coefficient histograms indicate the coefficient median of units significantly correlated. (D) Venn diagram illustrating the number of units with significant correlation between firing rate and 1st lever press latency (yellow), response rate (green) and port entry latency (blue). (E) Proportion of neurons in DMS and DLS expressing correlation with the 1st lever press latency (left), the response rate (middle), and the port entry latency (right).

Figure 10.

Figure 10—figure supplement 1. Comparison of correlations relating neural firing rate and behavior in DMS and DLS.

Figure 10—figure supplement 1.

(A) Distribution of Spearman rank correlation coefficients relating firing rate after lever insertion (0–500 ms) to the first lever press latency, in DLS (left) and DMS (right) neurons. (B) Distribution of Spearman rank correlation coefficients relating firing rate during lever presses to the response rate in DLS (left) and DMS (right) neurons. (C) Distribution of Spearman rank correlation coefficients relating firing rate after lever retraction (0–500 ms) to the port entry latency in DLS (left) and DMS (right) neurons. In each graph, the vertical red line indicates the median correlation coefficient of correlated units. (D) Proportion of neurons per class in DMS and DLS in the whole population of recorded neurons (left), and in populations of neurons correlated with the 1st lever press latency (middle left), the response rate (middle right), and the port entry latency (right).
Figure 10—figure supplement 2. Characterization of correlations relating firing rate post-lever insertion and 1st lever press latency.

Figure 10—figure supplement 2.

(A–B) Distribution of Spearman rank correlation coefficients from 1 iteration of the shuffled data (A) or from the real data (B), relating shuffled (A) or non-shuffled (B) firing in response to the lever insertion st–500 ms after lever insertion) to the animal’s latency to press the lever on a trial-by-trial basis. Bars shaded show significant correlations. In B, the red line indicates the median correlation coefficient of the correlated neurons. (C–D). Distribution of mean Spearman correlation coefficients (C) and percentage of significantly correlated units (D) from 1000 shuffled iterations of firing rate post lever insertion and the first lever press latency, on a trial-by-trial basis. On both graphs, vertical lines indicates values from real data in DLS (blue) or DMS (red).
Figure 10—figure supplement 3. Characterization of correlations relating within-sequence firing rate and response rate.

Figure 10—figure supplement 3.

(A–B) Distribution of Spearman rank correlation coefficients from 1 iteration of the shuffled data (A) or from the real data (B), relating shuffled (A) or non-shuffled (B) within-sequence firing rate (from 1st to last lever press) to the animal’s response rate on a trial-by-trial basis. Bars shaded show significant correlations. In B, the red line indicates the median correlation coefficient of the correlated neurons. (C–D). Distribution of mean Spearman correlation coefficients (C) and percentage of significantly correlated units (D) from 1000 shuffled iterations of within-sequence firing rate and response rate, on a trial-by-trial basis. On both graphs, vertical lines indicates values from real data in DLS (blue) or DMS (red).
Figure 10—figure supplement 4. Characterization of correlations relating firing rate post-lever retraction and port entry latency.

Figure 10—figure supplement 4.

(A–B) Distribution of Spearman rank correlation coefficients from 1 iteration of the shuffled data (A) or from the real data (B), relating shuffled (A) or non-shuffled (B) firing in response to the lever retraction (0–500 ms after lever retraction) and the animal’s latency to enter the port on a trial-by-trial basis. Bars shaded show significant correlations. In B, the red line indicates the median correlation coefficient of the correlated neurons. (C–D) Distribution of mean Spearman correlation coefficients (C) and percentage of significantly correlated units (D) from 1000 shuffled iterations of firing rate post lever retraction and the port entry latency, on a trial-by-trial basis. On both graphs, vertical lines indicates values from real data in DLS (blue) or DMS (red).