Author response:
Reviewer #1 (Public review):
Summary:
The manuscript reports a series of experiments designed to test whether optogenetic activation of infralimbic (IL) neurons facilitates extinction retrieval and whether this depends on animals' prior experience. In Experiment 1, rats underwent fear conditioning followed by either one or two extinction sessions, with IL stimulation given during the second extinction; stimulation facilitated extinction retrieval only in rats with prior extinction experience. Experiments 2 and 3 examined whether backward conditioning (CS presented after the US) could establish inhibitory properties that allowed IL stimulation to enhance extinction, and whether this effect was specific to the same stimulus or generalized to different stimuli. Experiments 5 - 7 extended this approach to appetitive learning: rats received backward or forward appetitive conditioning followed by extinction, and then fear conditioning, to determine whether IL stimulation could enhance extinction in contexts beyond aversive learning and across conditioning sequences. Across studies, the key claim is that IL activation facilitates extinction retrieval only when animals possess a prior inhibitory memory, and that this effect generalizes across aversive and appetitive paradigms.
Strengths:
(1) The design attempts to dissect the role of IL activity as a function of prior learning, which is conceptually valuable.
We thank the Reviewer for their positive assessment.
(2) The experimental design of probing different inhibitory learning approaches to probe how IL activation facilitates extinction learning was creative and innovative.
We thank the Reviewer for their positive assessment.
Weaknesses:
(1) Non-specific manipulation.
ChR2 was expressed in IL without distinction between glutamatergic and GABAergic populations. Without knowing the relative contribution of these cell types or the percentage of neurons affected, the circuit-level interpretation of the results is unclear.
ChR2 was intentionally expressed in the infralimbic cortex (IL) without distinction between local neuronal populations for two reasons. First, this manuscript aimed to uncover some of the features characterizing the encoding of inhibitory memories in the IL, and this encoding likely engages interactions among various neuronal populations within the IL. Second, the hypotheses tested in the manuscript derived from findings that indiscriminately stimulated the IL using the GABAA receptor antagonist picrotoxin, which is best mimicked by the approach taken. We agree that it is also important to determine the respective contributions of distinct IL neuronal populations to inhibitory encoding; however, the global approach implemented in the present experiments represents a necessary initial step. This rationale will be incorporated into the revised manuscript, which will also make reference to the need to identify the relative contributions of the various neuronal populations within the IL.
(2) Extinction retrieval test conflates processes
The retrieval test included 8 tones. Averaging across this many tone presentations conflate extinction retrieval/expression (early tones) with further extinction learning (later tones). A more appropriate analysis would focus on the first 2-4 tones to capture retrieval only. As currently presented, the data do not isolate extinction retrieval.
It is unclear when retrieval of what has been learned across extinction ceases and additional extinction learning occurs. In fact, it is only the first stimulus presentation that unequivocally permits a distinction between retrieval and additional extinction learning, as the conditions for this additional learning have not been fulfilled at that presentation. However, confining evidence for retrieval to the first stimulus presentation introduces concerns that other factors could influence performance. For instance, processing of the stimulus present at the start of the session may differ from that present at the end of the previous session, thereby affecting what is retrieved. Such differences between the stimuli present at the start and end of an extinction session have been long recognized as a potential explanation for spontaneous recovery (Estes, 1955). More importantly, whether the test data presented confound retrieval and additional extinction learning or not, the interpretation remains the same with respect to the effects of a prior history of inhibitory learning on enabling the facilitative effects of IL stimulation. Finally, it is unclear how these facilitative effects could occur in the absence of the subjects retrieving the extinction memory formed under the stimulation. Nevertheless, the revised manuscript will provide the trial-by-trial performance during the post-extinction retrieval tests and discuss this issue.
(3) Under-sampling and poor group matching.
Sample sizes appear small, which may explain why groups are not well matched in several figures (e.g., 2b, 3b, 6b, 6c) and why there are several instances of unexpected interactions (protocol, virus, and period). This baseline mismatch raises concerns about the reliability of group differences.
Efforts were made to match group performance upon completion of each training stage and before IL stimulation. Unfortunately, these efforts were not completely successful due to exclusions following post-mortem analyses. However, we acknowledge that the unexpected interactions deserve further discussion, and this will be incorporated into the revised manuscript (see also comment from Reviewer 2). Although we cannot exclude that sample sizes may have contributed to some of these interactions, we remain confident about the reliability of the main findings reported, especially given their replication across the various protocols. Overall, the manuscript provides evidence that IL stimulation does not facilitate brief extinction in the absence of prior inhibitory experience in five different experiments, replicating previous findings (Lingawi et al., 2018; Lingawi et al., 2017). It also replicates these previous findings by showing that prior experience with either fear or appetitive extinction enables IL stimulation to facilitate subsequent fear extinction. Furthermore, the facilitative effects of such stimulation following fear or appetitive backward conditioning are replicated in the present manuscript.
(4) Incomplete presentation of conditioning data.
Figure 3 only shows a single conditioning session despite five days of training. Without the full dataset, it is difficult to evaluate learning dynamics or whether groups were equivalent before testing.
We apologize, as we incorrectly labeled the X axis for the backward conditioning data set in Figures 3B, 4B, 4D and 5B. It should have indicated “Days” instead of “Trials”. This error will be corrected in the revised manuscript.
(5) Interpretation stronger than evidence.
The authors conclude that IL activation facilitates extinction retrieval only when an inhibitory memory has been formed. However, given the caveats above, the data are insufficient to support such a strong mechanistic claim. The results could reflect non-specific facilitation or disruption of behavior by broad prefrontal activation. Moreover, there is compelling evidence that optogenetic activation of IL during fear extinction does facilitate subsequent extinction retrieval without prior extinction training (Do-Monte et al 2015, Chen et al 2021), which the authors do not directly test in this study.
As noted above, the revised manuscript will show that the interpretations of the main findings stand whether ore the test data confounds retrieval with additional extinction learning. The revised manuscript will also clarify the plotting of the data for the backward conditioning stages. We do agree that further discussion of the unexpected interactions is necessary, and this will also be incorporated into the revised manuscript. However, the various replications of the core findings provide strong evidence for their reliability and the interpretations advanced in the original manuscript. The proposal that the results reflect non-specific facilitation or disruption of behavior seems highly unlikely. Indeed, the present experiments and previous findings (Lingawi et al., 2018; Lingawi et al., 2017) provide multiple demonstrations that IL stimulation fails to produce any facilitation in the absence of prior inhibitory experience with the target stimulus. Although these demonstrations appear inconsistent with previous studies (Do-Monte et al., 2015; Chen et al., 2021), this inconsistency is likely explained by the fact that these studies manipulated activity in specific IL neuronal populations. Previous work has already revealed differences between manipulations targeting discrete IL neuronal populations as opposed to general IL activity (Kim et al., 2016). Importantly, as previously noted, the present manuscript aimed to generally explore inhibitory encoding in the IL that, as we will acknowledge, is likely to engage several neuronal populations within the IL. Adequate statements on these matters will be included in the revised manuscript.
Impact:
The role of IL in extinction retrieval remains a central question in the fear learning literature. However, because the test used conflates extinction retrieval with new learning and the manipulations lack cell-type specificity, the evidence presented here does not convincingly support the main claims. The study highlights the need for more precise manipulations and more rigorous behavioral testing to resolve this issue.
As noted in our responses, the interpretations of the data presented remain identical whether the test data conflate extinction retrieval with additional extinction learning or not. Although we agree that it is important to establish the role of specific IL neuronal populations in extinction learning, this was beyond the scope of the manuscript and the findings reported remain valuable to our understanding of inhibitory encoding within the IL.
Reviewer #2 (Public review):
Summary:
In this manuscript, the authors examine the mechanisms by which stimulation of the infralimbic cortex (IL) facilitates the retention and retrieval of inhibitory memories. Previous work has shown that optogenetic stimulation of the IL suppresses freezing during extinction but does not improve extinction recall when extinction memory is probed one day later. When stimulation occurs during a second extinction session (following a prior stimulation-free extinction session), freezing is suppressed during the second extinction as well as during the tone test the following day. The current study was designed to further explore the facilitatory role of the IL in inhibitory learning and memory recall. The authors conducted a series of experiments to determine whether recruitment of IL extends to other forms of inhibitory learning (e.g., backward conditioning) and to inhibitory learning involving appetitive conditioning. Further, they assessed whether their effects could be explained by stimulus familiarity. The results of their experiments show that backward conditioning, another form of inhibitory learning, also enabled IL stimulation to enhance fear extinction. This phenomenon was not specific to aversive learning, as backward appetitive conditioning similarly allowed IL stimulation to facilitate extinction of aversive memories. Finally, the authors ruled out the possibility that IL facilitated extinction merely because of prior experience with the stimulus (e.g., reducing the novelty of the stimulus). These findings significantly advance our understanding of the contribution of IL to inhibitory learning. Namely, they show that the IL is recruited during various forms of inhibitory learning, and its involvement is independent of the motivational value associated with the unconditioned stimulus.
Strengths:
(1) Transparency about the inclusion of both sexes and the representation of data from both sexes in figures.
We thank the Reviewer for their positive assessment.
(2) Very clear representation of groups and experimental design for each figure.
We thank the Reviewer for their positive assessment.
(3) The authors were very rigorous in determining the neurobehavioral basis for the effects of IL stimulation on extinction. They considered multiple interpretations and designed experiments to address these possible accounts of their data.
We thank the Reviewer for their positive assessment.
(4) The rationale for and the design of the experiments in this manuscript are clearly based on a wealth of knowledge about learning theory. The authors leveraged this expertise to narrow down how the IL encodes and retrieves inhibitory memories.
We thank the Reviewer for their positive assessment.
Weaknesses:
(1) In Experiment 1, although not statistically significant, it does appear as though the stimulation groups (OFF and ON) differ during Extinction 1. It seems like this may be due to a difference between these groups after the first forward conditioning. Could the authors have prevented this potential group difference in Extinction 1 by re-balancing group assignment after the first forward conditioning session to minimize the differences in fear acquisition (the authors do report a marginally significant effect between the groups that would undergo one vs. two extinction sessions in their freezing during the first conditioning session)?
As noted (see response to Reviewer 1), efforts were made daily to match group performance across the training stages, but these efforts were ultimately hampered by the necessary exclusions following post-mortem analyses. This will be made explicit in the revised manuscript. Regarding freezing during Extinction 1, as noted by the Reviewer, the difference, which was not statistically significant, was absent across trials during the subsequent forward fear conditioning stage. Likewise, the protocol difference observed during the initial forward fear conditioning was absent in subsequent stages. We are therefore confident that these initial differences (significant or not) did not impact the main findings at test. Importantly, these findings replicate previous work using identical protocols in which no differences were present during the training stages. These considerations will be addressed in the revised manuscript.
(2) Across all experiments (except for Experiment 1), the authors state that freezing during the initial conditioning increased across "days". The figures that correspond to this text, however, show that freezing changes across trials. In the methods, the authors report that backward conditioning occurred over 5 days. It would be helpful to understand how these data were analyzed and collated to create the final figures. Was the freezing averaged across the five days for each trial for analyses and figures?
We apologize, as noted above, we incorrectly labeled the X axis for the backward conditioning data sets in Figures 3B, 4B, 4D and 5B. It should have indicated “Days” instead of “Trials”. The data shown in these Figures use the average of all trials on a given day. This will be clarified in the methods section of the revised manuscript. The labeling errors on the Figures will be corrected.
(3) In Experiment 3, the authors report a significant Protocol X Virus interaction. It would be useful if the authors could conduct post-hoc analyses to determine the source of this interaction. Inspection of Figure 4B suggests that freezing during the two different variants of backward conditioning differs between the virus groups. Did the authors expect to see a difference in backward conditioning depending on the stimulus used in the conditioning procedure (light vs. tone)? The authors don't really address this confounding interaction, but I do think a discussion is warranted.
We agree with the Reviewer that further discussion of the Protocol x Virus interaction that emerged during the backward conditioning and forward conditioning stages of Experiment 3 is warranted. This will be provided in the revised manuscript. Briefly, during both stages, follow-up analyses did not reveal any differences (main effects or interactions) between the two groups trained with the light stimulus (Diff-EYFP and Diff-ChR2). By contrast, the ChR2 group trained with the tone (Back-ChR2) froze more overall than the EYFP group (Back-EYFP), but there were no other significant differences between the two groups. Based on these analyses, the Protocol x Virus interaction appears to be driven by greater freezing in the ChR2 group trained with the tone rather than a difference in the backward conditioning performance based on stimulus identity. Consistent with this, the statistical analyses did not reveal a main effect of Protocol during either the backward conditioning stage or the stimulus trials during the forward conditioning stage. Nevertheless, during this latter stage, a main effect of Protocol emerged during baseline performance, but once again, this seems to be driven by the Back-ChR2 group. Critically, it is unclear how greater stimulus freezing in the Back-ChR2 group during forward conditioning would lead to lower freezing during the post-extinction retrieval test.
(4) In this same experiment, the authors state that freezing decreased during extinction; however, freezing in the Diff-EYFP group at the start of extinction (first bin of trials) doesn't look appreciably different than their freezing at the end of the session. Did this group actually extinguish their fear? Freezing on the tone test day also does not look too different from freezing during the last block of extinction trials.
We confirm that overall, there was a significant decline in freezing across the extinction session shown in Figure 4B. The Reviewer is correct to point out that this decline was modest (if not negligible) in the Diff-EYFP group, which was receiving its first inhibitory training with the target tone stimulus. It is worth noting that across all experiments, most groups that did not receive infralimbic stimulation displayed a modest decline in freezing during the extinction session since it was relatively brief, involving only 6 or 8 tone alone presentations. This was intentional, as we aimed for the brief extinction session to generate minimal inhibitory learning and thereby to detect any facilitatory effect of infralimbic stimulation. This issue will be clarified and explained in the revised version of the manuscript.
(5) The Discussion explored the outcomes of the experiments in detail, but it would be useful for the authors to discuss the implications of their findings for our understanding of circuits in which the IL is embedded that are involved in inhibitory learning and memory. It would also be useful for the authors to acknowledge in the Discussion that although they did not have the statistical power to detect sex differences, future work is needed to explore whether IL functions similarly in both sexes.
In line with the Reviewer’s suggestion (see also Reviewer 3), the revised manuscript will include a discussion of the broader implications of the findings regarding inhibitory brain circuitry and will acknowledge the need to further explore sex differences and IL functions.
Reviewer #3 (Public review):
Summary:
This is a really nice manuscript with different lines of evidence to show that the IL encodes inhibitory memories that can then be manipulated by optogenetic stimulation of these neurons during extinction. The behavioral designs are excellent, with converging evidence using extinction/re-extinction, backwards/forwards aversive conditioning, and backwards appetitive/forwards aversive conditioning. Additional factors, such as nonassociative effects of the CS or US, are also considered, and the authors evaluate the inhibitory properties of the CS with tests of conditioned inhibition.
Strengths:
The experimental designs are very rigorous with an unusual level of behavioral sophistication.
We thank the Reviewer for their positive assessment.
Weaknesses:
(1) More justification for parametric choices (number of days of backwards vs forwards conditioning) could be provided.
All experimental parameters were based on previously published experiments showing the capacity of the backward conditioning protocols to generate inhibitory learning and the forward conditioning protocols to produce excitatory learning. Although this was mentioned in the methods section, we acknowledge that further explanation is required to justify the need for multiple days of backward training. This will be provided in the revised manuscript.
(2) The current discussion could be condensed and could focus on broader implications for the literature.
The revised manuscript will make an effort to condense the discussion and focus on broader implications for the literature.
References
Chen, Y.-H., Wu, J.-L., Hu, N.-Y., Zhuang, J.-P., Li, W.-P., Zhang, S.-R., Li, X.-W., Yang, J.-M., & Gao, T.-M. (2021). Distinct projections from the infralimbic cortex exert opposing effects in modulating anxiety and fear. J Clin Invest, 131(14), e145692. https://doi.org/10.1172/JCI145692
Do-Monte, F. H., Manzano-Nieves, G., Quiñones-Laracuente, K., Ramos-Medina, L., & Quirk, G. J. (2015). Revisiting the role of infralimbic cortex in fear extinction with optogenetics. J Neurosci, 35(8), 3607-3615. https://doi.org/10.1523/JNEUROSCI.3137-14.2015
Estes, W. K. (1955). Statistical theory of spontaneous recovery and regression. Psychol Rev, 62(3), 145-154. https://doi.org/10.1037/h0048509
Kim, H.-S., Cho, H.-Y., Augustine, G. J., & Han, J.-H. (2016). Selective Control of Fear Expression by Optogenetic Manipulation of Infralimbic Cortex after Extinction. Neuropsychopharmacology, 41(5), 1261-1273. https://doi.org/10.1038/npp.2015.276
Lingawi, N. W., Holmes, N. M., Westbrook, R. F., & Laurent, V. (2018). The infralimbic cortex encodes inhibition irrespective of motivational significance. Neurobiol Learn Mem, 150, 64-74. https://doi.org/10.1016/j.nlm.2018.03.001
Lingawi, N. W., Westbrook, R. F., & Laurent, V. (2017). Extinction and Latent Inhibition Involve a Similar Form of Inhibitory Learning that is Stored in and Retrieved from the Infralimbic Cortex. Cereb Cortex, 27(12), 5547-5556. https://doi.org/10.1093/cercor/bhw322
