On 2015 Jun 01, UFRJ Neurobiology and Reproducibility Journal Club commented:

Although the findings of the study are interesting, we would like to point out that, due to differences in the interpretation of the term “spontaneous recovery”, the comparison between these results and previous results from the literature (Monfils et al., 2009, Clem et al., 2010) performed by the authors is not warranted. Spontaneous recovery is an intrinsic characteristic of the extinction process, and has been defined as a recovery of responding to the conditioned stimulus with the passage of time following the conclusion of extinction. (Bouton et al., 2002). However, one should observe that, since the decay of fear within an extinction session involves different mechanisms than those involved in extinction from one session to the next (Plendl & Wotjak, 2010, Almeida-Corrêa et al., 2015), a difference between the end of an extinction session and the beginning of the next test session is actually to be expected, and does not constitute evidence of spontaneous recovery, as the freezing epochs are not comparable (Plendl & Wotjak, 2010). Thus, the adequate way to evaluate spontaneous recovery is to compare freezing between two test sessions of similar duration, spaced apart and happening after the extinction session (such as in Monfils et al., 2009 and Flavell et al. 2011). Furthermore, it is worth highlighting that the training-test interval in this article and in the ones cited in Table 1 are clearly not the same. In Monfils et al., 2009 and Clem et al., 2010, the return of fear is prevented in the retrieval-extinction group 1 month and 1 week after extinction, respectively, but neither show a return of fear 24h after the extinction session. Chan et al. 2010 also don’t see a return of fear in either experimental group when they are tested in the extinction context, while in Stafford et al. 2013, freezing during the extinction session is not shown. Thus, rather than contradicting the previous findings, as discussed by the authors, the current results actually confirm those of Monfils et al. (2009) and others – namely, that a retrieval session before extinction does not alter freezing in a session 24 h later. It is surprising that the authors discuss many variables that might have influenced the “contradictory” results between this and the abovementioned papers, but don’t consider the different definitions of spontaneous recovery and the different delays between extinction and testing used in the articles. To clear up these misunderstandings in future studies, we propose, as others have done (Plendl & Wotjak, 2010), that the term “spontaneous recovery”, irrespectively of the time interval, should be based on comparisons between test sessions of similar duration happening after extinction, which is not the case when the end of a long extinction session is compared to a short test session. As has been discussed in Riebe et al 2012, the increasing number in studies being published in fear extinction don’t follow harmonic experimental designs, nomenclature or data analysis, therefore making it difficult to compare studies among themselves. The newly explored sub-field of post-retrieval extinction seems to be repeating this pattern, and we would like to reinforce the importance of avoiding such repetition.

On 2015 Jun 01, UFRJ Neurobiology and Reproducibility Journal Club commented:

Although the findings of the study are interesting, we would like to point out that, due to differences in the interpretation of the term “spontaneous recovery”, the comparison between these results and previous results from the literature (Monfils et al., 2009, Clem et al., 2010) performed by the authors is not warranted. Spontaneous recovery is an intrinsic characteristic of the extinction process, and has been defined as a recovery of responding to the conditioned stimulus with the passage of time following the conclusion of extinction. (Bouton et al., 2002). However, one should observe that, since the decay of fear within an extinction session involves different mechanisms than those involved in extinction from one session to the next (Plendl & Wotjak, 2010, Almeida-Corrêa et al., 2015), a difference between the end of an extinction session and the beginning of the next test session is actually to be expected, and does not constitute evidence of spontaneous recovery, as the freezing epochs are not comparable (Plendl & Wotjak, 2010). Thus, the adequate way to evaluate spontaneous recovery is to compare freezing between two test sessions of similar duration, spaced apart and happening after the extinction session (such as in Monfils et al., 2009 and Flavell et al. 2011). Furthermore, it is worth highlighting that the training-test interval in this article and in the ones cited in Table 1 are clearly not the same. In Monfils et al., 2009 and Clem et al., 2010, the return of fear is prevented in the retrieval-extinction group 1 month and 1 week after extinction, respectively, but neither show a return of fear 24h after the extinction session. Chan et al. 2010 also don’t see a return of fear in either experimental group when they are tested in the extinction context, while in Stafford et al. 2013, freezing during the extinction session is not shown. Thus, rather than contradicting the previous findings, as discussed by the authors, the current results actually confirm those of Monfils et al. (2009) and others – namely, that a retrieval session before extinction does not alter freezing in a session 24 h later. It is surprising that the authors discuss many variables that might have influenced the “contradictory” results between this and the abovementioned papers, but don’t consider the different definitions of spontaneous recovery and the different delays between extinction and testing used in the articles. To clear up these misunderstandings in future studies, we propose, as others have done (Plendl & Wotjak, 2010), that the term “spontaneous recovery”, irrespectively of the time interval, should be based on comparisons between test sessions of similar duration happening after extinction, which is not the case when the end of a long extinction session is compared to a short test session. As has been discussed in Riebe et al 2012, the increasing number in studies being published in fear extinction don’t follow harmonic experimental designs, nomenclature or data analysis, therefore making it difficult to compare studies among themselves. The newly explored sub-field of post-retrieval extinction seems to be repeating this pattern, and we would like to reinforce the importance of avoiding such repetition.