Article Figures & Data
Figures
- Figure 1.
Method details and 2CAP apparatus. A, Wistars and P rats underwent 2 weeks of IAP with 20% alcohol (v/v). Following IAP, Wistars and P rats were matched for consumption and underwent subsequent 2CAP training with 10% alcohol (v/v). Following the successful completion of 2CAP training, the animals underwent surgery and had silicon electrodes chronically implanted. The animals were then reacclimated to the 2CAP protocol as well as habituated to the modified 2CAP chamber. B, After reacclimating, neural recordings were acquired. Each week progressed through different contingencies. We focused the present analysis on congruent sessions and incongruent sessions conducted at the end of the week. Each session type consisted of 48 CS+ trials and 48 CS− trials. C, Wistars and P rats experienced roughly equal numbers of congruent and incongruent sessions. Additionally, neuron yield was relatively matched. D, The 2CAP apparatus and trial structure are detailed. Briefly, during congruent sessions, alcohol access was on the same side as the CS+. During incongruent sessions, alcohol access was on the side opposite the CS+. A 2CAP trial consisted of 4 s of the CS+ being illuminated. One second after the CS+ is turned off, the alcohol-containing sipper descends. The animals then had 8 s of 10% alcohol access prior to the sipper being removed from the chamber.
- Figure 2.
Incongruent sessions disrupt Wistars’ performance. A, Intake in g/kg is shown for each strain and session type. A main effect of strain (*) was detected. B, Sipper occupancy was quantified, and no main effects of either strain or session type were detected. C, The rate of intake is depicted, and a main effect of strain (*) was detected. D, Correct trials were determined per session, and a main effect of strain (*) was detected. E, Incorrect trials were determined per session and a main effect of session type (#) was detected. Post hoc tests indicated that congruent Wistars made fewer mistakes than incongruent Wistars and incongruent P rats. F, Omissions were determined per session, and a main effect of strain (*) was detected. G, CS discrimination was calculated, and a main effect of strain (*) and session type (#) was detected. Additionally, the incongruent Wistars’ ratio was not different from 0.5. H, Latency to approach the correct sipper is shown. A trending main effect of session type was found. I, Latency to approach the incorrect sipper is shown. No main effects of strain or session type were found. A MANOVA detected significant differences between congruent and incongruent sessions in Wistars but not P rats. This suggests an overall difference in the behavioral performance exists in Wistars but not P rats when the session contingency is changed.
- Figure 3.
Wistars make mistakes earlier in a session. A, The probability to make a correct or incorrect approach is shown for each session type for Wistars. A main effect of trial type ($) and session type (#) was detected. Critically, Wistars make more mistakes earlier in the session. B, The same information from A is shown but for P rats. Critically, neither a main effect of trial type nor session type was detected. C, The mean proportions of correct, incorrect, or omissions are shown for the first 15 trials in Wistars. The asterisks (*) denote the trials where the proportions from the incongruent sessions are significantly different than the proportions from the congruent sessions. D, The same information as in C is shown but for P rats.
- Figure 4.
Wistars are faster to approach the correct sipper during congruent sessions and the incorrect sipper during incongruent sessions. A, The distance from the correct sipper was calculated for each condition. Generally, there are no differences in the distances observed between conditions. B, The distance from the incorrect sipper was calculated for each condition. The animals in incongruent sessions are likely to approach the incorrect sipper; however, P rats are faster to correct this behavior. The trial was broken down into discrete epochs. PreCue refers to the time period before the cue. CueOn refers to the time period after the cue onset but before the sipper comes in. SipIn refers to the time period after the sipper comes in and before the sipper is removed. C, The speed at which the correct sipper was approached was calculated for all trial epochs. Critically, it is observed that Wistars in the congruent session move toward the correct sipper during the cue onset period faster than Wistars in the incongruent session. D, The speed at which the incorrect sipper was approached was calculated for all trial epochs. During the cue onset period, Wistars in the congruent session moved slower than Wistars in the incongruent session. E, The speed at which the animals moved toward the correct sipper during the cue onset period was broken down across trials. No conclusive results were observed between strain and session type or between trial blocks. F, The speed at which the animals moved toward the incorrect sipper during the cue onset period was broken down across trials. Again, no differences were observed.
- Figure 5.
PCA methodologies for cue-centered and approach-centered analyses. Cue-centered PCA analyses were accomplished by first smoothing 100 ms binned firing rates with a Gaussian kernel. Then an NxMxT matrix was created, where N refers to the number of neurons, M refers to the number of time points, and T refers to the number of trials. Four seconds prior to the cue onset and 18 s after the cue onset were analyzed [−4 18] resulting in 22 total seconds. At 100 ms bins, this resulted in 221 time bins. The number of trials, T, was the first 15 trials of the session. The mean of trials was calculated such that an NxM matrix remained. The NxM matrix for the congruent and incongruent Wistar sessions was vertically concatenated separately from the P rat sessions. PCA was then performed, and the resulting projections and coefficients were analyzed. In the approach-centered analyses, everything was the same except that the number of time bins was greatly reduced and centered on the approach rather than the cue onset. In the left versus right analyses, up to 15 correct trials were averaged. All correct left trial means were averaged separately from all correct right trial means. Then the left and right trial means were horizontally concatenated, which doubled our M dimension. Congruent Wistar and congruent P rat sessions were then vertically concatenated. The same was done with incongruent Wistar and incongruent P rat sessions. We then ran the PCA to obtain the coefficient and score products. Following this, all PCA products were filtered into the left versus right.
- Figure 6.
Principal components highlight several differing features between strains and session types. A, The mean firing rates for each strain collapsed across session types are depicted. Further analysis of the mean firing rates can be found in Extended Data Figure 6-1. B, The top 5 PCs for Wistars performing the congruent sessions are shown. Additionally, the inset describes the explained variance of each, and the dashed line indicates the broken stick point. Interestingly, several of the PCs encode the onset of the cue. C, The same information as in B except for P rats is shown. Several of the PCs encode the sipper descent and ascent. D, The top 5 PCs for Wistars performing the incongruent sessions are shown. Cue encoding during incongruent sessions is blunted in Wistars. F, The same as in D but for P rats. PC3 shows the enhanced sipper encoding during the descent and ascent.
- Figure 7.
Incongruent Wistar sessions show the differential encoding of cue onset. A, The mean firing rates for the full population of Wistar neurons are displayed for both congruent and incongruent sessions focused on the cue onset. During incongruent sessions, the firing rate drifts downward in later trials suggesting that Wistars are differentially encoding cue information after the contingencies are changed in incongruent sessions. The asterisks below indicate the time points where the trial blocks differed from one another. Trials 1–5 versus Trials 6–10 are indicated by black asterisks. Trials 1–5 versus Trials 11–15 are indicated by blue asterisks. Trials 6–10 versus Trials 11–15 are indicated by red asterisks. B, The mean firing rates for the full population of P rat neurons are displayed for both congruent and incongruent sessions focused on the cue onset. C, PC space representations for each strain and session type are depicted. Critically, the major differences observed between congruent and incongruent Wistar sessions are observed near the cue onset period suggesting that the encoding of this period is altered during incongruent sessions. These differences are not observed in P rats.
- Figure 8.
Wistars show greater spatial encoding. A, B, The mean firing rates for the left and right choices are depicted. C, D, The top 5 PCs for congruent and incongruent sessions are shown. The solid lines depict the left trials, and the dashed lines depict the right trials. E, A description of the analysis performed to determine the encoding differences between the left and right trials. In brief, several permutations in a 3D PC space were identified. In each permutation, 1,000 random left and right trials were compared. The mean of those 1,000 trials was then taken for each permutation. F, The results of the permutation analysis are shown. Briefly, the mean of the bootstrap samples was found for each permutation. The mean was then taken prior to approach to determine the separation of the left versus right choices prior to approach initiation. An overall effect was detected. *Tukey–Kramer post hoc tests identified that the distance between the left and right trials in the Wistar congruent session was higher than that in every other strain and session type condition.
- Figure 9.
Wistars and P rats show differential spatial encoding before and after approach. Extended Data Figure 9-1 details the methodology and rationale for selecting PC4. A, B, The mean firing rates for neurons that either positively or negatively load onto PC4 in both the congruent and incongruent sessions for Wistars. Critically, the mean firing rate diverges at or prior to the time when each animal approaches making it a useful index of neural activity. C, D, The same mean firing rates for positive and negative loading neurons for PC4 in the P rat neural population. The same divergent feature at the approach time emerges. E, The difference between the left and right firing rates was calculated and compared across each session type and strain combination before and after the approach for positive coefficients. A main effect of approach (&) and session type (#) was found. F, The same analysis as in E but for the negative coefficients. A main effect of approach (&) and strain (*) was found. Additionally, an interaction between strain and session type was found. Further probing of this effect showed that the congruent Wistar groups were significantly different from the congruent P rat and incongruent P rat groups. The congruent P rat group was significantly different from the incongruent Wistar and incongruent P rat groups, and the incongruent Wistar group was significantly different from the incongruent P rat group.
- Figure 10.
Spatial encoding is meaningfully related to latency to approach the sipper. A, The correlation between latency to approach the sipper and the distance between the left and right trials for each session is shown. In the congruent condition, latency and distance are nonsignificantly and negatively correlated, whereas in the incongruent conditions, latency and distance are significantly positively correlated. Overall, Wistars show a significant interaction between distance encoding and session type. This suggests that it is beneficial to encode the side prior to moving during congruent sessions but may be detrimental during incongruent sessions. B, The same information as seen in A but for P rats. In P rats, in the congruent condition, latency and distance are nonsignificantly and positively correlated, whereas in the incongruent condition, latency and distance are significantly and positively correlated.
Extended Data
Figure 6-1
Proportion and magnitude of cue and sipper modulation indicates cue and sipper are differentially expressed. A. No differences were observed in the proportion of neurons that were either positively or negatively modulated by the cue. B. A main effect of strain (F (1,1317) = 4.54, p = 0.033) and an interaction of strain and session type were observed (F (1,1317) = 11.37, p < 0.001). Multiple comparisons (Tukey's HSD) indicate that neurons during incongruent P rat sessions were more strongly modulated than neurons in both congruent P rat sessions (CI: [-1.95 -0.40]) and incongruent Wistar sessions (CI: [-1.56 -0.07]). C. No differences were observed in the proportion of neurons that were significantly modulated by the sipper presentation. D. A main effect of strain (F(1,1235) = 6.870, p = 0.009) and a main effect of sign (F(1,1235) = 4.018, p = 0.045) were observed in the modulation indices. Download Figure 6-1, TIF file.
Figure 9-1
PC coefficients were utilized to analyze raw left versus right trial firing rates. A. PC coefficients were utilized to analyze raw left versus right trial firing rates. A. PC Coefficients were split between positive (greater than or equal to 0.01) and negative (less than or equal to -0.01) coefficients. Neurons that displayed positive or negative loading to these coefficients were placed into their own respective groups. These neurons were then averaged to find the mean firing rate of each PC loading. B. On the left, the distribution of neurons that load onto each coefficient condition is detailed for congruent Wistar sessions. On the right, the same is shown for P rats. C. Raw firing rates that belonged to each coefficient condition were compared with a repeated measures ANOVA (RANOVA). PCs that had main effects are shown on the left. PCs that showed interactions between side and time are shown on the right. D. The same as B but for incongruent sessions. E. The same as C but for incongruent sessions. Download Figure 9-1, TIF file.
In this issue
